d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
4b55f0c7c7
freebsd-dev/sys/dev/ata/ata-chipset.c

4899 lines
151 KiB
C
Raw Normal View History

First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*-
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
* Copyright (c) 1998 - 2005 S<EFBFBD>ren Schmidt <sos@FreeBSD.org>
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer,
* without modification, immediately at the beginning of the file.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
Use __FBSDID(). Also some minor style cleanups.
2003-08-24 17:55:58 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
#include "opt_ata.h"
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/ata.h>
#include <sys/bus.h>
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
#include <sys/endian.h>
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
#include <sys/malloc.h>
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
#include <sys/lock.h>
#include <sys/mutex.h>
Overhaul of the timeout/reinit framework. This should clear up most of the leftovers from the old version that really doesn't work anymore. Add a reset function for host-end of the ATA channel. This is needed for the SiI3112 in order to whack it back to reality if a device locks up the SATA interface (thereby preventing that we can reset the device). The result is that ATA now recovers from the timeouts that happens with the SiI3112A and more or less all disks based on old PATA electronics with a Marvell PATA->SATA converter. This includes lots of the popular SATA dongles and the WDC Raptor disks..
2004-01-11 22:08:34 +00:00
#include <sys/sema.h>
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
#include <sys/taskqueue.h>
Use UMA instead of plain malloc for getting ATA request storage. This gives +10% performance on simple tests, so definitly worth it. A few percent more could be had by not using M_ZERO'd alloc's, but we then need to clear fields all over the place to be safe, and that was deemed not worth the trouble (and it makes life dangerous).
2004-01-14 21:26:35 +00:00
#include <vm/uma.h>
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
#include <machine/stdarg.h>
#include <machine/resource.h>
#include <machine/bus.h>
#include <sys/rman.h>
Prefer new location of pci include files (which have only been in the tree for two or more years now), except in a few places where there's code to be compatible with older versions of FreeBSD.
2003-08-22 05:54:52 +00:00
#include <dev/pci/pcivar.h>
#include <dev/pci/pcireg.h>
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
#include <dev/ata/ata-all.h>
#include <dev/ata/ata-pci.h>
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
#include <ata_if.h>
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/* local prototypes */
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
static int ata_generic_chipinit(device_t dev);
static void ata_generic_intr(void *data);
static void ata_generic_setmode(device_t dev, int mode);
static void ata_sata_setmode(device_t dev, int mode);
static int ata_sata_connect(struct ata_channel *ch);
static void ata_sata_phy_enable(struct ata_channel *ch);
static void ata_sata_phy_event(void *context, int dummy);
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
static int ata_ahci_allocate(device_t dev);
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
static int ata_ahci_setup_fis(u_int8_t *fis, struct ata_request *request);
static int ata_ahci_begin_transaction(struct ata_request *request);
static int ata_ahci_end_transaction(struct ata_request *request);
static void ata_ahci_intr(void *data);
static void ata_ahci_reset(device_t dev);
static void ata_ahci_dmainit(device_t dev);
static int ata_acard_chipinit(device_t dev);
static void ata_acard_intr(void *data);
static void ata_acard_850_setmode(device_t dev, int mode);
static void ata_acard_86X_setmode(device_t dev, int mode);
static int ata_ali_chipinit(device_t dev);
static int ata_ali_allocate(device_t dev);
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
static int ata_ali_sata_allocate(device_t dev);
Apply fix for "pr82261 DMA-support on Sparc64 broken" The Acer chip or wiring that SUN uses has problems that this patch tries to work around. Original patch by Marius Strobl, hacked into shape by me..
2005-08-23 08:53:01 +00:00
static void ata_ali_reset(device_t dev);
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
static void ata_ali_setmode(device_t dev, int mode);
static int ata_amd_chipinit(device_t dev);
Add support for the ATI IXP[234]00 series chipsets. HW donated by: sentex
2005-10-12 20:00:26 +00:00
static int ata_ati_chipinit(device_t dev);
static void ata_ati_setmode(device_t dev, int mode);
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
static int ata_cyrix_chipinit(device_t dev);
static void ata_cyrix_setmode(device_t dev, int mode);
static int ata_cypress_chipinit(device_t dev);
static void ata_cypress_setmode(device_t dev, int mode);
static int ata_highpoint_chipinit(device_t dev);
static void ata_highpoint_intr(void *data);
static void ata_highpoint_setmode(device_t dev, int mode);
static int ata_highpoint_check_80pin(device_t dev, int mode);
static int ata_intel_chipinit(device_t dev);
static int ata_intel_31244_allocate(device_t dev);
static void ata_intel_31244_intr(void *data);
static void ata_intel_31244_reset(device_t dev);
static int ata_intel_31244_command(struct ata_request *request);
static void ata_intel_intr(void *data);
static void ata_intel_reset(device_t dev);
static void ata_intel_old_setmode(device_t dev, int mode);
static void ata_intel_new_setmode(device_t dev, int mode);
static int ata_ite_chipinit(device_t dev);
static void ata_ite_setmode(device_t dev, int mode);
Add initial support for the Marvell 88SX[56]0[48][01] series of SATA chips. Hardware donated by: Matthew Jacob Hardware donated by: Y!
2005-12-27 17:09:52 +00:00
static int ata_marvell_chipinit(device_t dev);
static int ata_marvell_allocate(device_t dev);
static void ata_marvell_intr(void *data);
static void ata_marvell_dmainit(device_t dev);
static int ata_marvell_command(struct ata_request *request);
static void ata_marvell_reset(device_t dev);
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
static int ata_national_chipinit(device_t dev);
static void ata_national_setmode(device_t dev, int mode);
static int ata_nvidia_chipinit(device_t dev);
static int ata_nvidia_allocate(device_t dev);
static void ata_nvidia_intr(void *data);
static void ata_nvidia_reset(device_t dev);
static int ata_promise_chipinit(device_t dev);
static int ata_promise_mio_allocate(device_t dev);
static void ata_promise_mio_intr(void *data);
static void ata_promise_sx4_intr(void *data);
static void ata_promise_mio_dmainit(device_t dev);
static void ata_promise_mio_reset(device_t dev);
static int ata_promise_mio_command(struct ata_request *request);
static int ata_promise_sx4_command(struct ata_request *request);
static int ata_promise_apkt(u_int8_t *bytep, struct ata_request *request);
static void ata_promise_queue_hpkt(struct ata_pci_controller *ctlr, u_int32_t hpkt);
static void ata_promise_next_hpkt(struct ata_pci_controller *ctlr);
static void ata_promise_tx2_intr(void *data);
static void ata_promise_old_intr(void *data);
static int ata_promise_new_dmastart(device_t dev);
static int ata_promise_new_dmastop(device_t dev);
static void ata_promise_new_dmareset(device_t dev);
static void ata_promise_new_dmainit(device_t dev);
static void ata_promise_setmode(device_t dev, int mode);
static int ata_serverworks_chipinit(device_t dev);
static void ata_serverworks_setmode(device_t dev, int mode);
static int ata_sii_chipinit(device_t dev);
static int ata_sii_allocate(device_t dev);
static void ata_sii_intr(void *data);
static void ata_cmd_intr(void *data);
static void ata_cmd_old_intr(void *data);
static void ata_sii_reset(device_t dev);
static void ata_sii_setmode(device_t dev, int mode);
static void ata_cmd_setmode(device_t dev, int mode);
static int ata_sis_chipinit(device_t dev);
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
static int ata_sis_allocate(device_t dev);
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
static void ata_sis_reset(device_t dev);
static void ata_sis_setmode(device_t dev, int mode);
static int ata_via_chipinit(device_t dev);
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
static int ata_via_allocate(device_t dev);
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
static void ata_via_reset(device_t dev);
static void ata_via_southbridge_fixup(device_t dev);
static void ata_via_family_setmode(device_t dev, int mode);
static struct ata_chip_id *ata_find_chip(device_t dev, struct ata_chip_id *index, int slot);
static int ata_setup_interrupt(device_t dev);
static int ata_serialize(device_t dev, int flags);
static void ata_print_cable(device_t dev, u_int8_t *who);
static int ata_atapi(device_t dev);
static int ata_check_80pin(device_t dev, int mode);
static int ata_mode2idx(int mode);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* generic or unknown ATA chipset support functions */
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int
ata_generic_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
device_set_desc(dev, "GENERIC ATA controller");
ctlr->chipinit = ata_generic_chipinit;
return 0;
}
static int
ata_generic_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (ata_setup_interrupt(dev))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
ctlr->setmode = ata_generic_setmode;
return 0;
}
static void
ata_generic_intr(void *data)
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
int unit;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
for (unit = 0; unit < ctlr->channels; unit++) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!(ch = ctlr->interrupt[unit].argument))
continue;
Fix for those lost interrupts on probe on lots of controller types. Note to self, just because an idea is good, it doesn't apply everywhere.
2004-01-28 21:54:40 +00:00
if (ch->dma && (ch->dma->flags & ATA_DMA_ACTIVE)) {
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
int bmstat = ATA_IDX_INB(ch, ATA_BMSTAT_PORT) & ATA_BMSTAT_MASK;
if ((bmstat & (ATA_BMSTAT_ACTIVE | ATA_BMSTAT_INTERRUPT)) !=
ATA_BMSTAT_INTERRUPT)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
continue;
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, bmstat & ~ATA_BMSTAT_ERROR);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
DELAY(1);
}
ctlr->interrupt[unit].function(ch);
}
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_generic_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_device *atadev = device_get_softc(dev);
mode = ata_limit_mode(dev, mode, ATA_UDMA2);
mode = ata_check_80pin(dev, mode);
if (!ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode))
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
atadev->mode = mode;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* SATA support functions */
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_sata_setmode(device_t dev, int mode)
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_device *atadev = device_get_softc(dev);
Fix more problems with the Silicon Image 3112A: Setup decent transfer mode defaults as some BIOS's seem to put in things that it *knows* doesn't work. (Note to BIOS writers: stop doing that nonsense, we will get things working with your crappy HW anyways, and then recommend users to buy someone else's products that "just works", thankyou.. ) Limit the device transfer mode to ATA100/UDMA5 on generic SATA. Since we dont know if the user is using a pure SATA device or an old PATA drive with a SATA converter dongle, we need to limit the speed used here to cover up the problems with Marvell ATA-SATA bridges used in lots of SATA products. This workaround is enabled for all detectable SATA controllers as they seem to have semilar problems here. One notable exception is all the Promise pdc2037x chips which just always work (cudos to Promise!).
2003-12-08 08:27:52 +00:00
/*
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
* if we detect that the device isn't a real SATA device we limit
* the transfer mode to UDMA5/ATA100.
* this works around the problems some devices has with the
Change the order of ata_dmainit/ata_allocate in preparation of supporting new chipsets where this is needed.
2004-08-12 08:20:36 +00:00
* Marvell 88SX8030 SATA->PATA converters and UDMA6/ATA133.
Fix more problems with the Silicon Image 3112A: Setup decent transfer mode defaults as some BIOS's seem to put in things that it *knows* doesn't work. (Note to BIOS writers: stop doing that nonsense, we will get things working with your crappy HW anyways, and then recommend users to buy someone else's products that "just works", thankyou.. ) Limit the device transfer mode to ATA100/UDMA5 on generic SATA. Since we dont know if the user is using a pure SATA device or an old PATA drive with a SATA converter dongle, we need to limit the speed used here to cover up the problems with Marvell ATA-SATA bridges used in lots of SATA products. This workaround is enabled for all detectable SATA controllers as they seem to have semilar problems here. One notable exception is all the Promise pdc2037x chips which just always work (cudos to Promise!).
2003-12-08 08:27:52 +00:00
*/
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (atadev->param.satacapabilities != 0x0000 &&
atadev->param.satacapabilities != 0xffff) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
if (!ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0,
ata_limit_mode(dev, mode, ATA_UDMA6)))
Update the ICH7 support so it deals better with chips without AHCI. Update Intel MatrixRAID support to be able to pick up RAID0+1 (RAID10) and RAID5 arrays without panic'ing. This has the side effect of now also supporting multiple volumes on MatrixRAID's now I have the metadata better understood.. HW sponsored by: Mullet Scandinavia AB
2005-12-02 10:13:53 +00:00
/* XXX SOS we should query SATA STATUS for the speed */
atadev->mode = ATA_SA150;
Cleanup the storing and printing of the device transfermode for SATA.
2004-09-01 12:15:44 +00:00
}
else {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ATA_UDMA5);
if (!ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode))
Cleanup the storing and printing of the device transfermode for SATA.
2004-09-01 12:15:44 +00:00
atadev->mode = mode;
}
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
}
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
static int
ata_sata_connect(struct ata_channel *ch)
{
u_int32_t status;
int timeout;
/* wait up to 1 second for "connect well" */
for (timeout = 0; timeout < 100 ; timeout++) {
status = ATA_IDX_INL(ch, ATA_SSTATUS);
if ((status & ATA_SS_CONWELL_MASK) == ATA_SS_CONWELL_GEN1 ||
(status & ATA_SS_CONWELL_MASK) == ATA_SS_CONWELL_GEN2)
break;
ata_udelay(10000);
}
if (timeout >= 100) {
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
if (bootverbose)
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
device_printf(ch->dev, "SATA connect status=%08x\n", status);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
return 0;
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
}
/* clear SATA error register */
ATA_IDX_OUTL(ch, ATA_SERROR, ATA_IDX_INL(ch, ATA_SERROR));
/* find out what type device we got poll for spec'd 31 seconds */
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
/* XXX SOS 10 secs for now as I have little patience */
ch->devices = 0;
for (timeout = 0; timeout < 1000; timeout++) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
if (ATA_IDX_INB(ch, ATA_STATUS) & ATA_S_BUSY)
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
DELAY(10000);
else
break;
}
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
if (bootverbose)
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
device_printf(ch->dev, "SATA connect ready time=%dms\n", timeout * 10);
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
if (timeout < 1000) {
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
if ((ATA_IDX_INB(ch, ATA_CYL_LSB) == ATAPI_MAGIC_LSB) &&
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
(ATA_IDX_INB(ch, ATA_CYL_MSB) == ATAPI_MAGIC_MSB))
ch->devices = ATA_ATAPI_MASTER;
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
else
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
ch->devices = ATA_ATA_MASTER;
}
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
if (bootverbose)
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
device_printf(ch->dev, "sata_connect devices=0x%b\n",
ch->devices, "\20\3ATAPI_MASTER\1ATA_MASTER");
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
return 1;
}
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
static void
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ata_sata_phy_enable(struct ata_channel *ch)
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
{
int loop, retry;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
if ((ATA_IDX_INL(ch, ATA_SCONTROL) & ATA_SC_DET_MASK) == ATA_SC_DET_IDLE) {
Refine the SATA PHY code so the time consumed during probe is reduced. Add the older nVidia nForce2 SATA chip to supported HW.
2005-04-10 10:20:25 +00:00
ata_sata_connect(ch);
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
return;
Refine the SATA PHY code so the time consumed during probe is reduced. Add the older nVidia nForce2 SATA chip to supported HW.
2005-04-10 10:20:25 +00:00
}
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
for (retry = 0; retry < 10; retry++) {
for (loop = 0; loop < 10; loop++) {
Refine the SATA PHY code so the time consumed during probe is reduced. Add the older nVidia nForce2 SATA chip to supported HW.
2005-04-10 10:20:25 +00:00
ATA_IDX_OUTL(ch, ATA_SCONTROL, ATA_SC_DET_RESET);
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
ata_udelay(100);
Refine the SATA PHY code so the time consumed during probe is reduced. Add the older nVidia nForce2 SATA chip to supported HW.
2005-04-10 10:20:25 +00:00
if ((ATA_IDX_INL(ch, ATA_SCONTROL) &
ATA_SC_DET_MASK) == ATA_SC_DET_RESET)
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
break;
}
ata_udelay(5000);
for (loop = 0; loop < 10; loop++) {
Refine the SATA PHY code so the time consumed during probe is reduced. Add the older nVidia nForce2 SATA chip to supported HW.
2005-04-10 10:20:25 +00:00
ATA_IDX_OUTL(ch, ATA_SCONTROL, ATA_SC_DET_IDLE);
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
ata_udelay(100);
Refine the SATA PHY code so the time consumed during probe is reduced. Add the older nVidia nForce2 SATA chip to supported HW.
2005-04-10 10:20:25 +00:00
if ((ATA_IDX_INL(ch, ATA_SCONTROL) & ATA_SC_DET_MASK) == 0) {
ata_sata_connect(ch);
return;
}
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
}
}
}
static void
ata_sata_phy_event(void *context, int dummy)
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
{
struct ata_connect_task *tp = (struct ata_connect_task *)context;
Harden the hotplug support for SATA devices. This also fixes a few races that was present in the timeout/detach code. Sponsored by: pair.com
2005-09-14 12:45:06 +00:00
struct ata_channel *ch = device_get_softc(tp->dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
device_t *children;
int nchildren, i;
mtx_lock(&Giant); /* newbus suckage it needs Giant */
if (tp->action == ATA_C_ATTACH) {
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
device_printf(tp->dev, "CONNECTED\n");
ata_sata_connect(ch);
Properly hook in devices found by SATA connect events. This broke on the changes done to get atapicam happy earlier.
2005-04-20 12:51:54 +00:00
ata_identify(tp->dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
}
if (tp->action == ATA_C_DETACH) {
if (!device_get_children(tp->dev, &children, &nchildren)) {
for (i = 0; i < nchildren; i++)
if (children[i])
device_delete_child(tp->dev, children[i]);
free(children, M_TEMP);
}
Harden the hotplug support for SATA devices. This also fixes a few races that was present in the timeout/detach code. Sponsored by: pair.com
2005-09-14 12:45:06 +00:00
mtx_lock(&ch->state_mtx);
ch->state = ATA_IDLE;
mtx_unlock(&ch->state_mtx);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
device_printf(tp->dev, "DISCONNECTED\n");
}
mtx_unlock(&Giant); /* suckage code dealt with, release Giant */
free(tp, M_ATA);
}
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
/*
* AHCI v1.0 compliant SATA chipset support functions
*/
struct ata_ahci_dma_prd {
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
u_int64_t dba;
u_int32_t reserved;
u_int32_t dbc; /* 0 based */
#define ATA_AHCI_PRD_MASK 0x003fffff /* max 4MB */
#define ATA_AHCI_PRD_IPC (1<<31)
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
} __packed;
struct ata_ahci_cmd_tab {
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
u_int8_t cfis[64];
u_int8_t acmd[32];
u_int8_t reserved[32];
struct ata_ahci_dma_prd prd_tab[16];
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
} __packed;
struct ata_ahci_cmd_list {
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
u_int16_t cmd_flags;
u_int16_t prd_length; /* PRD entries */
u_int32_t bytecount;
u_int64_t cmd_table_phys; /* 128byte aligned */
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
} __packed;
static int
ata_ahci_allocate(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
int offset = (ch->unit << 7);
/* XXX SOS this is a hack to satisfy various legacy cruft */
ch->r_io[ATA_CYL_LSB].res = ctlr->r_res2;
ch->r_io[ATA_CYL_LSB].offset = ATA_AHCI_P_SIG + 1 + offset;
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
ch->r_io[ATA_CYL_MSB].res = ctlr->r_res2;
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
ch->r_io[ATA_CYL_MSB].offset = ATA_AHCI_P_SIG + 3 + offset;
ch->r_io[ATA_STATUS].res = ctlr->r_res2;
ch->r_io[ATA_STATUS].offset = ATA_AHCI_P_TFD + offset;
ch->r_io[ATA_ALTSTAT].res = ctlr->r_res2;
ch->r_io[ATA_ALTSTAT].offset = ATA_AHCI_P_TFD + offset;
/* set the SATA resources */
ch->r_io[ATA_SSTATUS].res = ctlr->r_res2;
ch->r_io[ATA_SSTATUS].offset = ATA_AHCI_P_SSTS + offset;
ch->r_io[ATA_SERROR].res = ctlr->r_res2;
ch->r_io[ATA_SERROR].offset = ATA_AHCI_P_SERR + offset;
ch->r_io[ATA_SCONTROL].res = ctlr->r_res2;
ch->r_io[ATA_SCONTROL].offset = ATA_AHCI_P_SCTL + offset;
ch->r_io[ATA_SACTIVE].res = ctlr->r_res2;
ch->r_io[ATA_SACTIVE].offset = ATA_AHCI_P_SACT + offset;
ch->hw.begin_transaction = ata_ahci_begin_transaction;
ch->hw.end_transaction = ata_ahci_end_transaction;
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
ch->hw.command = NULL; /* not used here */
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
/* setup the work areas */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CLB + offset,
ch->dma->work_bus + ATA_AHCI_CL_OFFSET);
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CLBU + offset, 0x00000000);
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_FB + offset,
ch->dma->work_bus + ATA_AHCI_FB_OFFSET);
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_FBU + offset, 0x00000000);
/* enable wanted port interrupts */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_IE + offset,
(ATA_AHCI_P_IX_CPD | ATA_AHCI_P_IX_TFE | ATA_AHCI_P_IX_HBF |
ATA_AHCI_P_IX_HBD | ATA_AHCI_P_IX_IF | ATA_AHCI_P_IX_OF |
ATA_AHCI_P_IX_PRC | ATA_AHCI_P_IX_PC | ATA_AHCI_P_IX_DP |
ATA_AHCI_P_IX_UF | ATA_AHCI_P_IX_SDB | ATA_AHCI_P_IX_DS |
ATA_AHCI_P_IX_PS | ATA_AHCI_P_IX_DHR));
/* start operations on this channel */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset,
(ATA_AHCI_P_CMD_ACTIVE | ATA_AHCI_P_CMD_FRE |
ATA_AHCI_P_CMD_POD | ATA_AHCI_P_CMD_SUD | ATA_AHCI_P_CMD_ST));
return 0;
}
static int
ata_ahci_setup_fis(u_int8_t *fis, struct ata_request *request)
{
struct ata_device *atadev = device_get_softc(request->dev);
int idx = 0;
/* XXX SOS add ATAPI commands support later */
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
ata_modify_if_48bit(request);
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
fis[idx++] = 0x27; /* host to device */
fis[idx++] = 0x80; /* command FIS (note PM goes here) */
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
fis[idx++] = request->u.ata.command;
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
fis[idx++] = request->u.ata.feature;
fis[idx++] = request->u.ata.lba;
fis[idx++] = request->u.ata.lba >> 8;
fis[idx++] = request->u.ata.lba >> 16;
Unbreak the ICH6 AHCI support in !48Bit mode.
2005-06-05 18:37:56 +00:00
fis[idx] = ATA_D_LBA | atadev->unit;
if (atadev->flags & ATA_D_48BIT_ACTIVE)
idx++;
else
fis[idx++] |= (request->u.ata.lba >> 24 & 0x0f);
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
fis[idx++] = request->u.ata.lba >> 24;
fis[idx++] = request->u.ata.lba >> 32;
fis[idx++] = request->u.ata.lba >> 40;
fis[idx++] = request->u.ata.feature >> 8;
fis[idx++] = request->u.ata.count;
fis[idx++] = request->u.ata.count >> 8;
fis[idx++] = 0x00;
fis[idx++] = ATA_A_4BIT;
fis[idx++] = 0x00;
fis[idx++] = 0x00;
fis[idx++] = 0x00;
fis[idx++] = 0x00;
return idx;
}
/* must be called with ATA channel locked and state_mtx held */
static int
ata_ahci_begin_transaction(struct ata_request *request)
{
struct ata_pci_controller *ctlr=device_get_softc(GRANDPARENT(request->dev));
struct ata_channel *ch = device_get_softc(device_get_parent(request->dev));
struct ata_ahci_cmd_tab *ctp;
struct ata_ahci_cmd_list *clp;
int fis_size, entries;
int tag = 0;
/* get a piece of the workspace for this request */
ctp = (struct ata_ahci_cmd_tab *)
(ch->dma->work + ATA_AHCI_CT_OFFSET + (ATA_AHCI_CT_SIZE * tag));
/* setup the FIS for this request */ /* XXX SOS ATAPI missing still */
if (!(fis_size = ata_ahci_setup_fis(&ctp->cfis[0], request))) {
device_printf(request->dev, "setting up SATA FIS failed\n");
request->result = EIO;
return ATA_OP_FINISHED;
}
/* if request moves data setup and load SG list */
if (request->flags & (ATA_R_READ | ATA_R_WRITE)) {
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
if (ch->dma->load(ch->dev, request->data, request->bytecount,
request->flags & ATA_R_READ,
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
ctp->prd_tab, &entries)) {
device_printf(request->dev, "setting up DMA failed\n");
request->result = EIO;
return ATA_OP_FINISHED;
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
}
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
}
/* setup the command list entry */
clp = (struct ata_ahci_cmd_list *)
(ch->dma->work + ATA_AHCI_CL_OFFSET + (ATA_AHCI_CL_SIZE * tag));
clp->prd_length = entries;
clp->cmd_flags = (request->flags & ATA_R_WRITE ? (1<<6) : 0) |
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
(request->flags & ATA_R_ATAPI ? (1<<5) : 0) |
(fis_size / sizeof(u_int32_t));
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
clp->bytecount = 0;
clp->cmd_table_phys = htole64(ch->dma->work_bus + ATA_AHCI_CT_OFFSET +
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
(ATA_AHCI_CT_SIZE * tag));
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
/* clear eventual ACTIVE bit */
ATA_IDX_OUTL(ch, ATA_SACTIVE, ATA_IDX_INL(ch, ATA_SACTIVE) & (1 << tag));
/* issue the command */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CI + (ch->unit << 7), (1 << tag));
/* start the timeout */
callout_reset(&request->callout, request->timeout * hz,
(timeout_t*)ata_timeout, request);
return ATA_OP_CONTINUES;
}
/* must be called with ATA channel locked and state_mtx held */
static int
ata_ahci_end_transaction(struct ata_request *request)
{
struct ata_pci_controller *ctlr=device_get_softc(GRANDPARENT(request->dev));
struct ata_channel *ch = device_get_softc(device_get_parent(request->dev));
struct ata_ahci_cmd_list *clp;
u_int32_t tf_data;
int tag = 0;
/* kill the timeout */
callout_stop(&request->callout);
/* get status */
tf_data = ATA_INL(ctlr->r_res2, ATA_AHCI_P_TFD + (ch->unit << 7));
request->status = tf_data;
/* if error status get details */
if (request->status & ATA_S_ERROR)
request->error = tf_data >> 8;
/* record how much data we actually moved */
clp = (struct ata_ahci_cmd_list *)
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
(ch->dma->work + ATA_AHCI_CL_OFFSET + (ATA_AHCI_CL_SIZE * tag));
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
request->donecount = clp->bytecount;
/* release SG list etc */
ch->dma->unload(ch->dev);
return ATA_OP_FINISHED;
}
static void
ata_ahci_intr(void *data)
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
u_int32_t port, status, error, issued;
int unit;
int tag = 0;
port = ATA_INL(ctlr->r_res2, ATA_AHCI_IS);
for (unit = 0; unit < ctlr->channels; unit++) {
if (port & (1 << unit)) {
if ((ch = ctlr->interrupt[unit].argument)) {
struct ata_connect_task *tp;
int offset = (ch->unit << 7);
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
error = ATA_INL(ctlr->r_res2, ATA_AHCI_P_SERR + offset);
status = ATA_INL(ctlr->r_res2, ATA_AHCI_P_IS + offset);
issued = ATA_INL(ctlr->r_res2, ATA_AHCI_P_CI + offset);
Unbreak hotplug support on the ICH6 and ICH7 chipsets.
2005-12-09 14:23:24 +00:00
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_SERR + offset, error);
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_IS + offset, status);
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
/* do we have cold connect surprise */
if (status & ATA_AHCI_P_IX_CPD) {
printf("ata_ahci_intr status=%08x error=%08x issued=%08x\n",
status, error, issued);
}
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
/* check for and handle connect events */
Unbreak hotplug support on the ICH6 and ICH7 chipsets.
2005-12-09 14:23:24 +00:00
if (((status & (ATA_AHCI_P_IX_PRC | ATA_AHCI_P_IX_PC)) ==
ATA_AHCI_P_IX_PC) &&
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
(tp = (struct ata_connect_task *)
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
malloc(sizeof(struct ata_connect_task),
M_ATA, M_NOWAIT | M_ZERO))) {
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
device_printf(ch->dev, "CONNECT requested\n");
tp->action = ATA_C_ATTACH;
tp->dev = ch->dev;
TASK_INIT(&tp->task, 0, ata_sata_phy_event, tp);
taskqueue_enqueue(taskqueue_thread, &tp->task);
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
}
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
/* check for and handle disconnect events */
if (((status & (ATA_AHCI_P_IX_PRC | ATA_AHCI_P_IX_PC)) ==
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
ATA_AHCI_P_IX_PRC) &&
(tp = (struct ata_connect_task *)
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
malloc(sizeof(struct ata_connect_task),
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
M_ATA, M_NOWAIT | M_ZERO))) {
device_printf(ch->dev, "DISCONNECT requested\n");
tp->action = ATA_C_DETACH;
tp->dev = ch->dev;
TASK_INIT(&tp->task, 0, ata_sata_phy_event, tp);
taskqueue_enqueue(taskqueue_thread, &tp->task);
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
}
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
/* any drive action to take care of ? */
if (!(issued & (1<<tag)))
ctlr->interrupt[unit].function(ch);
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
}
}
}
ATA_OUTL(ctlr->r_res2, ATA_AHCI_IS, port);
}
static void
ata_ahci_reset(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
u_int32_t cmd;
int offset = (ch->unit << 7);
/* kill off all activity on this channel */
cmd = ATA_INL(ctlr->r_res2, ATA_AHCI_P_CMD + offset);
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset,
cmd & ~(ATA_AHCI_P_CMD_CR | ATA_AHCI_P_CMD_FR |
ATA_AHCI_P_CMD_FRE | ATA_AHCI_P_CMD_ST));
Change the way ioctls are issue to ATA. The most prominent part is that its now possible to issue ata_requests directly to say acd0, instead of going through the cumbersome /dev/ata device.
2005-05-16 13:07:27 +00:00
DELAY(500000); /* XXX SOS this is not entirely wrong */
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
/* spin up device */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset, ATA_AHCI_P_CMD_SUD);
ata_sata_phy_enable(ch);
/* clear any interrupts pending on this channel */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_IS + offset,
ATA_INL(ctlr->r_res2, ATA_AHCI_P_IS + offset));
/* start operations on this channel */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset,
(ATA_AHCI_P_CMD_ACTIVE | ATA_AHCI_P_CMD_FRE |
ATA_AHCI_P_CMD_POD | ATA_AHCI_P_CMD_SUD | ATA_AHCI_P_CMD_ST));
}
static void
ata_ahci_dmasetprd(void *xsc, bus_dma_segment_t *segs, int nsegs, int error)
{
struct ata_dmasetprd_args *args = xsc;
struct ata_ahci_dma_prd *prd = args->dmatab;
int i;
if (!(args->error = error)) {
for (i = 0; i < nsegs; i++) {
prd[i].dba = htole64(segs[i].ds_addr);
prd[i].dbc = htole32((segs[i].ds_len - 1) & ATA_AHCI_PRD_MASK);
}
}
args->nsegs = nsegs;
}
static void
ata_ahci_dmainit(device_t dev)
{
struct ata_channel *ch = device_get_softc(dev);
ata_dmainit(dev);
if (ch->dma) {
/* note start and stop are not used here */
ch->dma->setprd = ata_ahci_dmasetprd;
ch->dma->max_iosize = 8192 * DEV_BSIZE;
}
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*
* Acard chipset support functions
*/
int
ata_acard_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_ATP850R, 0, ATPOLD, 0x00, ATA_UDMA2, "ATP850" },
{ ATA_ATP860A, 0, 0, 0x00, ATA_UDMA4, "ATP860A" },
{ ATA_ATP860R, 0, 0, 0x00, ATA_UDMA4, "ATP860R" },
{ ATA_ATP865A, 0, 0, 0x00, ATA_UDMA6, "ATP865A" },
{ ATA_ATP865R, 0, 0, 0x00, ATA_UDMA6, "ATP865R" },
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{ 0, 0, 0, 0, 0, 0}};
char buffer[64];
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
if (!(idx = ata_match_chip(dev, ids)))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
sprintf(buffer, "Acard %s %s controller",
idx->text, ata_mode2str(idx->max_dma));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_acard_chipinit;
return 0;
}
static int
ata_acard_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
int rid = ATA_IRQ_RID;
Convert callers to the new bus_alloc_resource_any(9) API. Submitted by: Mark Santcroos <marks@ripe.net> Reviewed by: imp, dfr, bde
2004-03-17 17:50:55 +00:00
if (!(ctlr->r_irq = bus_alloc_resource_any(dev, SYS_RES_IRQ, &rid,
RF_SHAREABLE | RF_ACTIVE))) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_printf(dev, "unable to map interrupt\n");
return ENXIO;
}
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if ((bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
ata_acard_intr, ctlr, &ctlr->handle))) {
device_printf(dev, "unable to setup interrupt\n");
return ENXIO;
}
if (ctlr->chip->cfg1 == ATPOLD) {
ctlr->setmode = ata_acard_850_setmode;
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
ctlr->locking = ata_serialize;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
else
ctlr->setmode = ata_acard_86X_setmode;
return 0;
}
static void
ata_acard_intr(void *data)
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
int unit;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
for (unit = 0; unit < ctlr->channels; unit++) {
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
if (!(ch = ctlr->interrupt[unit].argument))
continue;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (ctlr->chip->cfg1 == ATPOLD &&
Change the ata_* methods to use a channel device instead of a controller device. This helps when there is no controller parent to a channel (PPC port).
2005-03-31 15:05:40 +00:00
ATA_LOCKING(ch->dev, ATA_LF_WHICH) != unit)
Fix the PC98 lockups on boot. The interchannel locking for PC98 needed to be updated to match the rest of the locking in ATA.
2004-10-06 19:46:08 +00:00
continue;
Fix for those lost interrupts on probe on lots of controller types. Note to self, just because an idea is good, it doesn't apply everywhere.
2004-01-28 21:54:40 +00:00
if (ch->dma && (ch->dma->flags & ATA_DMA_ACTIVE)) {
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
int bmstat = ATA_IDX_INB(ch, ATA_BMSTAT_PORT) & ATA_BMSTAT_MASK;
if ((bmstat & (ATA_BMSTAT_ACTIVE | ATA_BMSTAT_INTERRUPT)) !=
ATA_BMSTAT_INTERRUPT)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
continue;
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, bmstat & ~ATA_BMSTAT_ERROR);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
DELAY(1);
Second round of updates to the ATA driver. Clean up the DMA interface too much unneeded stuff crept in with the busdma code back when. Modify the ATA_IN* / ATA_OUT* macros so that resource and offset are gotten from a table. That allows for new chipsets that doesn't nessesarily have things ordered the good old way. This also removes the need for the wierd PC98 resource functions. Tested on: i386, PC98, Alpha, Sparc64
2003-03-29 13:37:09 +00:00
ATA_IDX_OUTB(ch, ATA_BMCMD_PORT,
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
ATA_IDX_INB(ch, ATA_BMCMD_PORT)&~ATA_BMCMD_START_STOP);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
DELAY(1);
}
ctlr->interrupt[unit].function(ch);
}
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_acard_850_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int error;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode,
ata_atapi(dev) ? ATA_PIO_MAX : ctlr->chip->max_dma);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
/* XXX SOS missing WDMA0+1 + PIO modes */
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (mode >= ATA_WDMA2) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting %s on %s chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "FAILURE " : "",
ata_mode2str(mode), ctlr->chip->text);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!error) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
u_int8_t reg54 = pci_read_config(gparent, 0x54, 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
reg54 &= ~(0x03 << (devno << 1));
if (mode >= ATA_UDMA0)
reg54 |= (((mode & ATA_MODE_MASK) + 1) << (devno << 1));
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x54, reg54, 1);
pci_write_config(gparent, 0x4a, 0xa6, 1);
pci_write_config(gparent, 0x40 + (devno << 1), 0x0301, 2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
atadev->mode = mode;
return;
}
}
/* we could set PIO mode timings, but we assume the BIOS did that */
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_acard_86X_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int error;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode,
ata_atapi(dev) ? ATA_PIO_MAX : ctlr->chip->max_dma);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_check_80pin(dev, mode);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
/* XXX SOS missing WDMA0+1 + PIO modes */
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (mode >= ATA_WDMA2) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting %s on %s chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "FAILURE " : "",
ata_mode2str(mode), ctlr->chip->text);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!error) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
u_int16_t reg44 = pci_read_config(gparent, 0x44, 2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
reg44 &= ~(0x000f << (devno << 2));
if (mode >= ATA_UDMA0)
reg44 |= (((mode & ATA_MODE_MASK) + 1) << (devno << 2));
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x44, reg44, 2);
pci_write_config(gparent, 0x4a, 0xa6, 1);
pci_write_config(gparent, 0x40 + devno, 0x31, 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
atadev->mode = mode;
return;
}
}
/* we could set PIO mode timings, but we assume the BIOS did that */
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*
* Acer Labs Inc (ALI) chipset support functions
*/
int
ata_ali_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_ALI_5289, 0x00, 2, ALISATA, ATA_SA150, "M5289" },
{ ATA_ALI_5287, 0x00, 4, ALISATA, ATA_SA150, "M5287" },
{ ATA_ALI_5281, 0x00, 2, ALISATA, ATA_SA150, "M5281" },
{ ATA_ALI_5229, 0xc5, 0, ALINEW, ATA_UDMA6, "M5229" },
{ ATA_ALI_5229, 0xc4, 0, ALINEW, ATA_UDMA5, "M5229" },
{ ATA_ALI_5229, 0xc2, 0, ALINEW, ATA_UDMA4, "M5229" },
{ ATA_ALI_5229, 0x20, 0, ALIOLD, ATA_UDMA2, "M5229" },
{ ATA_ALI_5229, 0x00, 0, ALIOLD, ATA_WDMA2, "M5229" },
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{ 0, 0, 0, 0, 0, 0}};
char buffer[64];
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
if (!(idx = ata_match_chip(dev, ids)))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
sprintf(buffer, "AcerLabs %s %s controller",
idx->text, ata_mode2str(idx->max_dma));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_ali_chipinit;
return 0;
}
static int
ata_ali_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (ata_setup_interrupt(dev))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
switch (ctlr->chip->cfg2) {
case ALISATA:
pci_write_config(dev, PCIR_COMMAND,
pci_read_config(dev, PCIR_COMMAND, 2) & ~0x0400, 2);
ctlr->channels = ctlr->chip->cfg1;
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
ctlr->allocate = ata_ali_sata_allocate;
Only try to allocate and use the SATA resource if they are enabled by the BIOS. It seems some BIOS's doesn't get this right, and that would result in ATA panic'ing.
2005-04-25 07:50:51 +00:00
ctlr->setmode = ata_sata_setmode;
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
break;
case ALINEW:
Apply fix for "pr82261 DMA-support on Sparc64 broken" The Acer chip or wiring that SUN uses has problems that this patch tries to work around. Original patch by Marius Strobl, hacked into shape by me..
2005-08-23 08:53:01 +00:00
/* use device interrupt as byte count end */
pci_write_config(dev, 0x4a, pci_read_config(dev, 0x4a, 1) | 0x20, 1);
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
/* enable cable detection and UDMA support on newer chips */
Fix support for Acer chips. The UDMA enable regs wasn't set properly.
2003-02-25 21:22:27 +00:00
pci_write_config(dev, 0x4b, pci_read_config(dev, 0x4b, 1) | 0x09, 1);
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
Apply fix for "pr82261 DMA-support on Sparc64 broken" The Acer chip or wiring that SUN uses has problems that this patch tries to work around. Original patch by Marius Strobl, hacked into shape by me..
2005-08-23 08:53:01 +00:00
/* enable ATAPI UDMA mode */
pci_write_config(dev, 0x53, pci_read_config(dev, 0x53, 1) | 0x01, 1);
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
/* only chips with revision > 0xc4 can do 48bit DMA */
if (ctlr->chip->chiprev <= 0xc4)
device_printf(dev,
"using PIO transfers above 137GB as workaround for "
"48bit DMA access bug, expect reduced performance\n");
Apply fix for "pr82261 DMA-support on Sparc64 broken" The Acer chip or wiring that SUN uses has problems that this patch tries to work around. Original patch by Marius Strobl, hacked into shape by me..
2005-08-23 08:53:01 +00:00
ctlr->reset = ata_ali_reset;
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
ctlr->allocate = ata_ali_allocate;
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
ctlr->setmode = ata_ali_setmode;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
break;
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
case ALIOLD:
/* deactivate the ATAPI FIFO and enable ATAPI UDMA */
Apply fix for "pr82261 DMA-support on Sparc64 broken" The Acer chip or wiring that SUN uses has problems that this patch tries to work around. Original patch by Marius Strobl, hacked into shape by me..
2005-08-23 08:53:01 +00:00
pci_write_config(dev, 0x53, pci_read_config(dev, 0x53, 1) | 0x03, 1);
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
ctlr->setmode = ata_ali_setmode;
break;
}
return 0;
}
static int
ata_ali_allocate(device_t dev)
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
/* setup the usual register normal pci style */
Properly react to allocation failures. Found by: imp@
2005-10-10 19:12:43 +00:00
if (ata_pci_allocate(dev))
return ENXIO;
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
Properly react to allocation failures. Found by: imp@
2005-10-10 19:12:43 +00:00
/* older chips can't do 48bit DMA transfers */
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
if (ctlr->chip->chiprev <= 0xc4)
ch->flags |= ATA_NO_48BIT_DMA;
return 0;
}
static int
ata_ali_sata_allocate(device_t dev)
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
{
device_t parent = device_get_parent(dev);
struct ata_pci_controller *ctlr = device_get_softc(parent);
struct ata_channel *ch = device_get_softc(dev);
Add support for controllers that doesn't have the usual taskfile layout. No functional changes.
2005-04-06 10:22:56 +00:00
struct resource *io = NULL, *ctlio = NULL;
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
int unit01 = (ch->unit & 1), unit10 = (ch->unit & 2);
int i, rid;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
rid = PCIR_BAR(0) + (unit01 ? 8 : 0);
io = bus_alloc_resource_any(parent, SYS_RES_IOPORT, &rid, RF_ACTIVE);
if (!io)
return ENXIO;
rid = PCIR_BAR(1) + (unit01 ? 8 : 0);
Add support for controllers that doesn't have the usual taskfile layout. No functional changes.
2005-04-06 10:22:56 +00:00
ctlio = bus_alloc_resource_any(parent, SYS_RES_IOPORT, &rid, RF_ACTIVE);
if (!ctlio) {
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
bus_release_resource(dev, SYS_RES_IOPORT, ATA_IOADDR_RID, io);
return ENXIO;
}
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
Add support for controllers that doesn't have the usual taskfile layout. No functional changes.
2005-04-06 10:22:56 +00:00
for (i = ATA_DATA; i <= ATA_COMMAND; i ++) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ch->r_io[i].res = io;
ch->r_io[i].offset = i + (unit10 ? 8 : 0);
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
}
Add support for controllers that doesn't have the usual taskfile layout. No functional changes.
2005-04-06 10:22:56 +00:00
ch->r_io[ATA_CONTROL].res = ctlio;
ch->r_io[ATA_CONTROL].offset = 2 + (unit10 ? 4 : 0);
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
ch->r_io[ATA_IDX_ADDR].res = io;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_default_registers(dev);
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
if (ctlr->r_res1) {
for (i = ATA_BMCMD_PORT; i <= ATA_BMDTP_PORT; i++) {
ch->r_io[i].res = ctlr->r_res1;
ch->r_io[i].offset = (i - ATA_BMCMD_PORT)+(ch->unit * ATA_BMIOSIZE);
}
}
Add support for controllers that doesn't have the usual taskfile layout. No functional changes.
2005-04-06 10:22:56 +00:00
ch->flags |= ATA_NO_SLAVE;
Only try to allocate and use the SATA resource if they are enabled by the BIOS. It seems some BIOS's doesn't get this right, and that would result in ATA panic'ing.
2005-04-25 07:50:51 +00:00
/* XXX SOS PHY handling awkward in ALI chip not supported yet */
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_generic_hw(dev);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return 0;
}
Apply fix for "pr82261 DMA-support on Sparc64 broken" The Acer chip or wiring that SUN uses has problems that this patch tries to work around. Original patch by Marius Strobl, hacked into shape by me..
2005-08-23 08:53:01 +00:00
static void
ata_ali_reset(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
device_t *children;
int nchildren, i;
ata_generic_reset(dev);
/*
* workaround for datacorruption bug found on at least SUN Blade-100
* find the ISA function on the southbridge and disable then enable
* the ATA channel tristate buffer
*/
if (ctlr->chip->chiprev == 0xc3 || ctlr->chip->chiprev == 0xc2) {
if (!device_get_children(GRANDPARENT(dev), &children, &nchildren)) {
for (i = 0; i < nchildren; i++) {
if (pci_get_devid(children[i]) == ATA_ALI_1533) {
pci_write_config(children[i], 0x58,
pci_read_config(children[i], 0x58, 1) &
~(0x04 << ch->unit), 1);
pci_write_config(children[i], 0x58,
pci_read_config(children[i], 0x58, 1) |
(0x04 << ch->unit), 1);
break;
}
}
free(children, M_TEMP);
}
}
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_ali_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int error;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ctlr->chip->max_dma);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Fix support for Acer chips. The UDMA enable regs wasn't set properly.
2003-02-25 21:22:27 +00:00
if (ctlr->chip->cfg2 & ALINEW) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (mode > ATA_UDMA2 &&
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_read_config(gparent, 0x4a, 1) & (1 << ch->unit)) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_print_cable(dev, "controller");
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
mode = ATA_UDMA2;
}
}
else
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_check_80pin(dev, mode);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (ctlr->chip->cfg2 & ALIOLD) {
/* doesn't support ATAPI DMA on write */
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
ch->flags |= ATA_ATAPI_DMA_RO;
if (ch->devices & ATA_ATAPI_MASTER && ch->devices & ATA_ATAPI_SLAVE) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/* doesn't support ATAPI DMA on two ATAPI devices */
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "two atapi devices on this channel, no DMA\n");
mode = ata_limit_mode(dev, mode, ATA_PIO_MAX);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
}
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting %s on %s chip\n",
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
(error) ? "FAILURE " : "",
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
ata_mode2str(mode), ctlr->chip->text);
if (!error) {
if (mode >= ATA_UDMA0) {
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
u_int8_t udma[] = {0x0c, 0x0b, 0x0a, 0x09, 0x08, 0x0f, 0x0d};
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
u_int32_t word54 = pci_read_config(gparent, 0x54, 4);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Fix chipset setup problem on older ALI chips.
2003-02-23 16:36:21 +00:00
word54 &= ~(0x000f000f << (devno << 2));
word54 |= (((udma[mode&ATA_MODE_MASK]<<16)|0x05)<<(devno<<2));
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x54, word54, 4);
pci_write_config(gparent, 0x58 + (ch->unit << 2),
Fix support for Acer chips. The UDMA enable regs wasn't set properly.
2003-02-25 21:22:27 +00:00
0x00310001, 4);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
else {
u_int32_t piotimings[] =
{ 0x006d0003, 0x00580002, 0x00440001, 0x00330001,
0x00310001, 0x00440001, 0x00330001, 0x00310001};
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x54, pci_read_config(gparent, 0x54, 4) &
~(0x0008000f << (devno << 2)), 4);
pci_write_config(gparent, 0x58 + (ch->unit << 2),
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
piotimings[ata_mode2idx(mode)], 4);
}
atadev->mode = mode;
}
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
* American Micro Devices (AMD) chipset support functions
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
*/
int
ata_amd_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_AMD756, 0x00, AMDNVIDIA, 0x00, ATA_UDMA4, "756" },
{ ATA_AMD766, 0x00, AMDNVIDIA, AMDCABLE|AMDBUG, ATA_UDMA5, "766" },
{ ATA_AMD768, 0x00, AMDNVIDIA, AMDCABLE, ATA_UDMA5, "768" },
{ ATA_AMD8111, 0x00, AMDNVIDIA, AMDCABLE, ATA_UDMA6, "8111" },
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{ 0, 0, 0, 0, 0, 0}};
char buffer[64];
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
if (!(idx = ata_match_chip(dev, ids)))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
sprintf(buffer, "AMD %s %s controller",
idx->text, ata_mode2str(idx->max_dma));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_amd_chipinit;
return 0;
}
static int
ata_amd_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (ata_setup_interrupt(dev))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Cleanup the defines used for various chipsets. Add new nVidia controllers.
2004-09-03 07:37:53 +00:00
/* disable/set prefetch, postwrite */
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (ctlr->chip->cfg2 & AMDBUG)
pci_write_config(dev, 0x41, pci_read_config(dev, 0x41, 1) & 0x0f, 1);
else
pci_write_config(dev, 0x41, pci_read_config(dev, 0x41, 1) | 0xf0, 1);
ctlr->setmode = ata_via_family_setmode;
return 0;
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
Add support for the ATI IXP[234]00 series chipsets. HW donated by: sentex
2005-10-12 20:00:26 +00:00
/*
* ATI chipset support functions
*/
int
ata_ati_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_ATI_IXP200, 0x00, 0, 0, ATA_UDMA5, "IXP200" },
{ ATA_ATI_IXP300, 0x00, 0, 0, ATA_UDMA6, "IXP300" },
{ ATA_ATI_IXP400, 0x00, 0, 0, ATA_UDMA6, "IXP400" },
{ ATA_ATI_IXP300_S1, 0x00, SIIMEMIO, 0, ATA_SA150, "IXP300" },
{ ATA_ATI_IXP400_S1, 0x00, SIIMEMIO, 0, ATA_SA150, "IXP400" },
{ ATA_ATI_IXP400_S2, 0x00, SIIMEMIO, 0, ATA_SA150, "IXP400" },
Add support for the ATI IXP[234]00 series chipsets. HW donated by: sentex
2005-10-12 20:00:26 +00:00
{ 0, 0, 0, 0, 0, 0}};
char buffer[64];
if (!(idx = ata_match_chip(dev, ids)))
return ENXIO;
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
sprintf(buffer, "ATI %s %s controller",
idx->text, ata_mode2str(idx->max_dma));
Add support for the ATI IXP[234]00 series chipsets. HW donated by: sentex
2005-10-12 20:00:26 +00:00
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
/* the ATI SATA controller is actually a SiI 3112 controller*/
if (ctlr->chip->cfg1 & SIIMEMIO)
ctlr->chipinit = ata_sii_chipinit;
else
ctlr->chipinit = ata_ati_chipinit;
return 0;
}
static int
ata_ati_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
if (ata_setup_interrupt(dev))
return ENXIO;
ctlr->setmode = ata_ati_setmode;
return 0;
}
static void
ata_ati_setmode(device_t dev, int mode)
{
device_t gparent = GRANDPARENT(dev);
struct ata_pci_controller *ctlr = device_get_softc(gparent);
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
int offset = (devno ^ 0x01) << 3;
int error;
u_int8_t piotimings[] = { 0x5d, 0x47, 0x34, 0x22, 0x20, 0x34, 0x22, 0x20,
0x20, 0x20, 0x20, 0x20, 0x20, 0x20 };
u_int8_t dmatimings[] = { 0x77, 0x21, 0x20 };
mode = ata_limit_mode(dev, mode, ctlr->chip->max_dma);
mode = ata_check_80pin(dev, mode);
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
if (bootverbose)
device_printf(dev, "%ssetting %s on %s chip\n",
(error) ? "FAILURE " : "",
ata_mode2str(mode), ctlr->chip->text);
if (!error) {
if (mode >= ATA_UDMA0) {
pci_write_config(gparent, 0x56,
(pci_read_config(gparent, 0x56, 2) &
~(0xf << (devno << 2))) |
((mode & ATA_MODE_MASK) << (devno << 2)), 2);
pci_write_config(gparent, 0x54,
pci_read_config(gparent, 0x54, 1) |
(0x01 << devno), 1);
pci_write_config(gparent, 0x44,
(pci_read_config(gparent, 0x44, 4) &
~(0xff << offset)) |
(dmatimings[2] << offset), 4);
}
else if (mode >= ATA_WDMA0) {
pci_write_config(gparent, 0x54,
pci_read_config(gparent, 0x54, 1) &
~(0x01 << devno), 1);
pci_write_config(gparent, 0x44,
(pci_read_config(gparent, 0x44, 4) &
~(0xff << offset)) |
(dmatimings[mode & ATA_MODE_MASK] << offset), 4);
}
else
pci_write_config(gparent, 0x54,
pci_read_config(gparent, 0x54, 1) &
~(0x01 << devno), 1);
pci_write_config(gparent, 0x4a,
(pci_read_config(gparent, 0x4a, 2) &
~(0xf << (devno << 2))) |
(((mode - ATA_PIO0) & ATA_MODE_MASK) << (devno<<2)),2);
pci_write_config(gparent, 0x40,
(pci_read_config(gparent, 0x40, 4) &
~(0xff << offset)) |
(piotimings[ata_mode2idx(mode)] << offset), 4);
atadev->mode = mode;
}
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*
* Cyrix chipset support functions
*/
int
ata_cyrix_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
if (pci_get_devid(dev) == ATA_CYRIX_5530) {
device_set_desc(dev, "Cyrix 5530 ATA33 controller");
ctlr->chipinit = ata_cyrix_chipinit;
return 0;
}
return ENXIO;
}
static int
ata_cyrix_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (ata_setup_interrupt(dev))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
if (ctlr->r_res1)
Dont use r_bmio on Cyrix chips if not set from BIOS.
2003-03-10 08:20:14 +00:00
ctlr->setmode = ata_cyrix_setmode;
else
ctlr->setmode = ata_generic_setmode;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return 0;
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_cyrix_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
u_int32_t piotiming[] =
{ 0x00009172, 0x00012171, 0x00020080, 0x00032010, 0x00040010 };
u_int32_t dmatiming[] = { 0x00077771, 0x00012121, 0x00002020 };
u_int32_t udmatiming[] = { 0x00921250, 0x00911140, 0x00911030 };
int error;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
ch->dma->alignment = 16;
ch->dma->max_iosize = 126 * DEV_BSIZE;
Up the alignment requirement of the SC1100 to 16 bytes, it has all the bugs and then some of its Cyrix inheritance. Set the max_iosize for the Cyrix to 63K as the SC1100.
2003-10-21 08:53:29 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ATA_UDMA2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting %s on Cyrix chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "FAILURE " : "", ata_mode2str(mode));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!error) {
if (mode >= ATA_UDMA0) {
Second round of updates to the ATA driver. Clean up the DMA interface too much unneeded stuff crept in with the busdma code back when. Modify the ATA_IN* / ATA_OUT* macros so that resource and offset are gotten from a table. That allows for new chipsets that doesn't nessesarily have things ordered the good old way. This also removes the need for the wierd PC98 resource functions. Tested on: i386, PC98, Alpha, Sparc64
2003-03-29 13:37:09 +00:00
ATA_OUTL(ch->r_io[ATA_BMCMD_PORT].res,
Fix typo (that even got cut/pasted 2 times) Found by: phk Approved by: re@
2003-05-12 16:43:13 +00:00
0x24 + (devno << 3), udmatiming[mode & ATA_MODE_MASK]);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
else if (mode >= ATA_WDMA0) {
Second round of updates to the ATA driver. Clean up the DMA interface too much unneeded stuff crept in with the busdma code back when. Modify the ATA_IN* / ATA_OUT* macros so that resource and offset are gotten from a table. That allows for new chipsets that doesn't nessesarily have things ordered the good old way. This also removes the need for the wierd PC98 resource functions. Tested on: i386, PC98, Alpha, Sparc64
2003-03-29 13:37:09 +00:00
ATA_OUTL(ch->r_io[ATA_BMCMD_PORT].res,
Fix typo (that even got cut/pasted 2 times) Found by: phk Approved by: re@
2003-05-12 16:43:13 +00:00
0x24 + (devno << 3), dmatiming[mode & ATA_MODE_MASK]);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
else {
Second round of updates to the ATA driver. Clean up the DMA interface too much unneeded stuff crept in with the busdma code back when. Modify the ATA_IN* / ATA_OUT* macros so that resource and offset are gotten from a table. That allows for new chipsets that doesn't nessesarily have things ordered the good old way. This also removes the need for the wierd PC98 resource functions. Tested on: i386, PC98, Alpha, Sparc64
2003-03-29 13:37:09 +00:00
ATA_OUTL(ch->r_io[ATA_BMCMD_PORT].res,
Fix typo (that even got cut/pasted 2 times) Found by: phk Approved by: re@
2003-05-12 16:43:13 +00:00
0x20 + (devno << 3), piotiming[mode & ATA_MODE_MASK]);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
atadev->mode = mode;
}
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*
* Cypress chipset support functions
*/
int
ata_cypress_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
/*
* the Cypress chip is a mess, it contains two ATA functions, but
* both channels are visible on the first one.
* simply ignore the second function for now, as the right
* solution (ignoring the second channel on the first function)
* doesn't work with the crappy ATA interrupt setup on the alpha.
*/
if (pci_get_devid(dev) == ATA_CYPRESS_82C693 &&
Only use function 1 not 0 on cypress chips.
2003-02-27 09:42:57 +00:00
pci_get_function(dev) == 1 &&
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
pci_get_subclass(dev) == PCIS_STORAGE_IDE) {
device_set_desc(dev, "Cypress 82C693 ATA controller");
ctlr->chipinit = ata_cypress_chipinit;
return 0;
}
return ENXIO;
}
static int
ata_cypress_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (ata_setup_interrupt(dev))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
ctlr->setmode = ata_cypress_setmode;
return 0;
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_cypress_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int error;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ATA_WDMA2);
Only use function 1 not 0 on cypress chips.
2003-02-27 09:42:57 +00:00
Cleanup comments
2005-05-13 10:25:19 +00:00
/* XXX SOS missing WDMA0+1 + PIO modes */
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (mode == ATA_WDMA2) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting WDMA2 on Cypress chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
error ? "FAILURE " : "");
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!error) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, ch->unit ? 0x4e : 0x4c, 0x2020, 2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
atadev->mode = mode;
return;
}
}
/* we could set PIO mode timings, but we assume the BIOS did that */
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*
* HighPoint chipset support functions
*/
int
ata_highpoint_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_HPT374, 0x07, HPT374, 0x00, ATA_UDMA6, "HPT374" },
{ ATA_HPT372, 0x02, HPT372, 0x00, ATA_UDMA6, "HPT372N" },
{ ATA_HPT372, 0x01, HPT372, 0x00, ATA_UDMA6, "HPT372" },
{ ATA_HPT371, 0x01, HPT372, 0x00, ATA_UDMA6, "HPT371" },
{ ATA_HPT366, 0x05, HPT372, 0x00, ATA_UDMA6, "HPT372" },
{ ATA_HPT366, 0x03, HPT370, 0x00, ATA_UDMA5, "HPT370" },
{ ATA_HPT366, 0x02, HPT366, 0x00, ATA_UDMA4, "HPT368" },
{ ATA_HPT366, 0x00, HPT366, HPTOLD, ATA_UDMA4, "HPT366" },
{ ATA_HPT302, 0x01, HPT372, 0x00, ATA_UDMA6, "HPT302" },
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{ 0, 0, 0, 0, 0, 0}};
char buffer[64];
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
if (!(idx = ata_match_chip(dev, ids)))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
strcpy(buffer, idx->text);
if (idx->cfg1 == HPT374) {
if (pci_get_function(dev) == 0)
strcat(buffer, " (channel 0+1)");
else if (pci_get_function(dev) == 1)
strcat(buffer, " (channel 2+3)");
}
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
sprintf(buffer, "HighPoint %s %s controller",
buffer, ata_mode2str(idx->max_dma));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_highpoint_chipinit;
return 0;
}
static int
ata_highpoint_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
int rid = ATA_IRQ_RID;
Convert callers to the new bus_alloc_resource_any(9) API. Submitted by: Mark Santcroos <marks@ripe.net> Reviewed by: imp, dfr, bde
2004-03-17 17:50:55 +00:00
if (!(ctlr->r_irq = bus_alloc_resource_any(dev, SYS_RES_IRQ, &rid,
RF_SHAREABLE | RF_ACTIVE))) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_printf(dev, "unable to map interrupt\n");
return ENXIO;
}
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if ((bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
ata_highpoint_intr, ctlr, &ctlr->handle))) {
device_printf(dev, "unable to setup interrupt\n");
return ENXIO;
}
if (ctlr->chip->cfg2 == HPTOLD) {
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
/* disable interrupt prediction */
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
pci_write_config(dev, 0x51, (pci_read_config(dev, 0x51, 1) & ~0x80), 1);
}
else {
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
/* disable interrupt prediction */
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
pci_write_config(dev, 0x51, (pci_read_config(dev, 0x51, 1) & ~0x03), 1);
pci_write_config(dev, 0x55, (pci_read_config(dev, 0x55, 1) & ~0x03), 1);
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
/* enable interrupts */
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
pci_write_config(dev, 0x5a, (pci_read_config(dev, 0x5a, 1) & ~0x10), 1);
/* set clocks etc */
if (ctlr->chip->cfg1 < HPT372)
pci_write_config(dev, 0x5b, 0x22, 1);
else
pci_write_config(dev, 0x5b,
(pci_read_config(dev, 0x5b, 1) & 0x01) | 0x20, 1);
}
ctlr->setmode = ata_highpoint_setmode;
return 0;
}
static void
ata_highpoint_intr(void *data)
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
int unit;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
for (unit = 0; unit < ctlr->channels; unit++) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!(ch = ctlr->interrupt[unit].argument))
continue;
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
if (ch->dma) {
int bmstat = ATA_IDX_INB(ch, ATA_BMSTAT_PORT) & ATA_BMSTAT_MASK;
if ((bmstat & (ATA_BMSTAT_ACTIVE | ATA_BMSTAT_INTERRUPT)) !=
ATA_BMSTAT_INTERRUPT)
continue;
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, bmstat & ~ATA_BMSTAT_ERROR);
DELAY(1);
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
ctlr->interrupt[unit].function(ch);
}
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_highpoint_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int error;
u_int32_t timings33[][4] = {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* HPT366 HPT370 HPT372 HPT374 mode */
{ 0x40d0a7aa, 0x06914e57, 0x0d029d5e, 0x0ac1f48a }, /* PIO 0 */
{ 0x40d0a7a3, 0x06914e43, 0x0d029d26, 0x0ac1f465 }, /* PIO 1 */
{ 0x40d0a753, 0x06514e33, 0x0c829ca6, 0x0a81f454 }, /* PIO 2 */
{ 0x40c8a742, 0x06514e22, 0x0c829c84, 0x0a81f443 }, /* PIO 3 */
{ 0x40c8a731, 0x06514e21, 0x0c829c62, 0x0a81f442 }, /* PIO 4 */
{ 0x20c8a797, 0x26514e97, 0x2c82922e, 0x228082ea }, /* MWDMA 0 */
{ 0x20c8a732, 0x26514e33, 0x2c829266, 0x22808254 }, /* MWDMA 1 */
{ 0x20c8a731, 0x26514e21, 0x2c829262, 0x22808242 }, /* MWDMA 2 */
{ 0x10c8a731, 0x16514e31, 0x1c829c62, 0x121882ea }, /* UDMA 0 */
{ 0x10cba731, 0x164d4e31, 0x1c9a9c62, 0x12148254 }, /* UDMA 1 */
{ 0x10caa731, 0x16494e31, 0x1c929c62, 0x120c8242 }, /* UDMA 2 */
{ 0x10cfa731, 0x166d4e31, 0x1c8e9c62, 0x128c8242 }, /* UDMA 3 */
{ 0x10c9a731, 0x16454e31, 0x1c8a9c62, 0x12ac8242 }, /* UDMA 4 */
{ 0, 0x16454e31, 0x1c8a9c62, 0x12848242 }, /* UDMA 5 */
{ 0, 0, 0x1c869c62, 0x12808242 } /* UDMA 6 */
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
};
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ctlr->chip->max_dma);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
if (ctlr->chip->cfg1 == HPT366 && ata_atapi(dev))
mode = ata_limit_mode(dev, mode, ATA_PIO_MAX);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_highpoint_check_80pin(dev, mode);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/*
* most if not all HPT chips cant really handle that the device is
* running at ATA_UDMA6/ATA133 speed, so we cheat at set the device to
* a max of ATA_UDMA5/ATA100 to guard against suboptimal performance
*/
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0,
ata_limit_mode(dev, mode, ATA_UDMA5));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting %s on HighPoint chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "FAILURE " : "", ata_mode2str(mode));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!error)
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x40 + (devno << 2),
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
timings33[ata_mode2idx(mode)][ctlr->chip->cfg1], 4);
atadev->mode = mode;
}
static int
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_highpoint_check_80pin(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
u_int8_t reg, val, res;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (ctlr->chip->cfg1 == HPT374 && pci_get_function(gparent) == 1) {
reg = ch->unit ? 0x57 : 0x53;
val = pci_read_config(gparent, reg, 1);
pci_write_config(gparent, reg, val | 0x80, 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
else {
reg = 0x5b;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
val = pci_read_config(gparent, reg, 1);
pci_write_config(gparent, reg, val & 0xfe, 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
res = pci_read_config(gparent, 0x5a, 1) & (ch->unit ? 0x1:0x2);
pci_write_config(gparent, reg, val, 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (mode > ATA_UDMA2 && res) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_print_cable(dev, "controller");
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
mode = ATA_UDMA2;
}
return mode;
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*
* Intel chipset support functions
*/
int
ata_intel_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_I82371FB, 0, 0, 0x00, ATA_WDMA2, "PIIX" },
{ ATA_I82371SB, 0, 0, 0x00, ATA_WDMA2, "PIIX3" },
{ ATA_I82371AB, 0, 0, 0x00, ATA_UDMA2, "PIIX4" },
{ ATA_I82443MX, 0, 0, 0x00, ATA_UDMA2, "PIIX4" },
{ ATA_I82451NX, 0, 0, 0x00, ATA_UDMA2, "PIIX4" },
{ ATA_I82801AB, 0, 0, 0x00, ATA_UDMA2, "ICH0" },
{ ATA_I82801AA, 0, 0, 0x00, ATA_UDMA4, "ICH" },
{ ATA_I82372FB, 0, 0, 0x00, ATA_UDMA4, "ICH" },
{ ATA_I82801BA, 0, 0, 0x00, ATA_UDMA5, "ICH2" },
{ ATA_I82801BA_1, 0, 0, 0x00, ATA_UDMA5, "ICH2" },
{ ATA_I82801CA, 0, 0, 0x00, ATA_UDMA5, "ICH3" },
{ ATA_I82801CA_1, 0, 0, 0x00, ATA_UDMA5, "ICH3" },
{ ATA_I82801DB, 0, 0, 0x00, ATA_UDMA5, "ICH4" },
{ ATA_I82801DB_1, 0, 0, 0x00, ATA_UDMA5, "ICH4" },
{ ATA_I82801EB, 0, 0, 0x00, ATA_UDMA5, "ICH5" },
{ ATA_I82801EB_S1, 0, 0, 0x00, ATA_SA150, "ICH5" },
{ ATA_I82801EB_R1, 0, 0, 0x00, ATA_SA150, "ICH5" },
{ ATA_I6300ESB, 0, 0, 0x00, ATA_UDMA5, "6300ESB" },
{ ATA_I6300ESB_S1, 0, 0, 0x00, ATA_SA150, "6300ESB" },
{ ATA_I6300ESB_R1, 0, 0, 0x00, ATA_SA150, "6300ESB" },
{ ATA_I82801FB, 0, 0, 0x00, ATA_UDMA5, "ICH6" },
{ ATA_I82801FB_S1, 0, AHCI, 0x00, ATA_SA150, "ICH6" },
{ ATA_I82801FB_R1, 0, AHCI, 0x00, ATA_SA150, "ICH6" },
{ ATA_I82801FB_M, 0, AHCI, 0x00, ATA_SA150, "ICH6" },
{ ATA_I82801GB, 0, 0, 0x00, ATA_UDMA5, "ICH7" },
{ ATA_I82801GB_S1, 0, AHCI, 0x00, ATA_SA300, "ICH7" },
{ ATA_I82801GB_R1, 0, AHCI, 0x00, ATA_SA300, "ICH7" },
{ ATA_I82801GB_M, 0, AHCI, 0x00, ATA_SA300, "ICH7" },
{ ATA_I82801GB_AH, 0, AHCI, 0x00, ATA_SA300, "ICH7" },
{ ATA_I31244, 0, 0, 0x00, ATA_SA150, "31244" },
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{ 0, 0, 0, 0, 0, 0}};
char buffer[64];
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
if (!(idx = ata_match_chip(dev, ids)))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
sprintf(buffer, "Intel %s %s controller",
idx->text, ata_mode2str(idx->max_dma));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_intel_chipinit;
return 0;
}
static int
ata_intel_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
int rid = ATA_IRQ_RID;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
if (!ata_legacy(dev)) {
if (!(ctlr->r_irq = bus_alloc_resource_any(dev, SYS_RES_IRQ, &rid,
RF_SHAREABLE | RF_ACTIVE))) {
device_printf(dev, "unable to map interrupt\n");
return ENXIO;
}
if ((bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
ata_intel_intr, ctlr, &ctlr->handle))) {
device_printf(dev, "unable to setup interrupt\n");
return ENXIO;
}
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
/* good old PIIX needs special treatment (not implemented) */
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
if (ctlr->chip->chipid == ATA_I82371FB) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
ctlr->setmode = ata_intel_old_setmode;
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
}
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
/* the intel 31244 needs special care if in DPA mode */
else if (ctlr->chip->chipid == ATA_I31244) {
if (pci_get_subclass(dev) != PCIS_STORAGE_IDE) {
ctlr->r_type2 = SYS_RES_MEMORY;
ctlr->r_rid2 = PCIR_BAR(0);
if (!(ctlr->r_res2 = bus_alloc_resource_any(dev, ctlr->r_type2,
&ctlr->r_rid2,
RF_ACTIVE)))
return ENXIO;
if (bus_teardown_intr(dev, ctlr->r_irq, ctlr->handle) ||
bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
ata_intel_31244_intr, ctlr, &ctlr->handle)) {
device_printf(dev, "unable to setup interrupt\n");
return ENXIO;
}
ctlr->channels = 4;
ctlr->reset = ata_intel_31244_reset;
ctlr->allocate = ata_intel_31244_allocate;
}
ctlr->setmode = ata_sata_setmode;
}
/* non SATA intel chips goes here */
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
else if (ctlr->chip->max_dma < ATA_SA150) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
ctlr->setmode = ata_intel_new_setmode;
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
}
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
/* SATA parts can be either compat or AHCI */
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
else {
Update the ICH7 support so it deals better with chips without AHCI. Update Intel MatrixRAID support to be able to pick up RAID0+1 (RAID10) and RAID5 arrays without panic'ing. This has the side effect of now also supporting multiple volumes on MatrixRAID's now I have the metadata better understood.. HW sponsored by: Mullet Scandinavia AB
2005-12-02 10:13:53 +00:00
/* force all ports active "the legacy way" */
pci_write_config(dev, 0x92, pci_read_config(dev, 0x92, 2) | 0x0f,2);
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
Update the ICH7 support so it deals better with chips without AHCI. Update Intel MatrixRAID support to be able to pick up RAID0+1 (RAID10) and RAID5 arrays without panic'ing. This has the side effect of now also supporting multiple volumes on MatrixRAID's now I have the metadata better understood.. HW sponsored by: Mullet Scandinavia AB
2005-12-02 10:13:53 +00:00
ctlr->reset = ata_intel_reset;
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
Update the ICH7 support so it deals better with chips without AHCI. Update Intel MatrixRAID support to be able to pick up RAID0+1 (RAID10) and RAID5 arrays without panic'ing. This has the side effect of now also supporting multiple volumes on MatrixRAID's now I have the metadata better understood.. HW sponsored by: Mullet Scandinavia AB
2005-12-02 10:13:53 +00:00
/* if we have AHCI capability and BAR(5) as a memory resource */
if (ctlr->chip->cfg1 == AHCI) {
ctlr->r_type2 = SYS_RES_MEMORY;
ctlr->r_rid2 = PCIR_BAR(5);
if ((ctlr->r_res2 = bus_alloc_resource_any(dev, ctlr->r_type2,
&ctlr->r_rid2,
RF_ACTIVE))) {
/* is AHCI or RAID mode enabled in BIOS ? */
if (pci_read_config(dev, 0x90, 1) & 0xc0) {
if (bus_teardown_intr(dev, ctlr->r_irq, ctlr->handle) ||
bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
ata_ahci_intr, ctlr, &ctlr->handle)) {
device_printf(dev, "unable to setup interrupt\n");
return ENXIO;
}
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
Update the ICH7 support so it deals better with chips without AHCI. Update Intel MatrixRAID support to be able to pick up RAID0+1 (RAID10) and RAID5 arrays without panic'ing. This has the side effect of now also supporting multiple volumes on MatrixRAID's now I have the metadata better understood.. HW sponsored by: Mullet Scandinavia AB
2005-12-02 10:13:53 +00:00
/* enable AHCI mode */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_GHC, ATA_AHCI_GHC_AE);
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
Update the ICH7 support so it deals better with chips without AHCI. Update Intel MatrixRAID support to be able to pick up RAID0+1 (RAID10) and RAID5 arrays without panic'ing. This has the side effect of now also supporting multiple volumes on MatrixRAID's now I have the metadata better understood.. HW sponsored by: Mullet Scandinavia AB
2005-12-02 10:13:53 +00:00
/* get the number of HW channels */
ctlr->channels = (ATA_INL(ctlr->r_res2, ATA_AHCI_CAP) &
ATA_AHCI_NPMASK) + 1;
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
Update the ICH7 support so it deals better with chips without AHCI. Update Intel MatrixRAID support to be able to pick up RAID0+1 (RAID10) and RAID5 arrays without panic'ing. This has the side effect of now also supporting multiple volumes on MatrixRAID's now I have the metadata better understood.. HW sponsored by: Mullet Scandinavia AB
2005-12-02 10:13:53 +00:00
/* enable AHCI interrupts */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_GHC,
ATA_INL(ctlr->r_res2, ATA_AHCI_GHC) |
ATA_AHCI_GHC_IE);
ctlr->reset = ata_ahci_reset;
ctlr->dmainit = ata_ahci_dmainit;
ctlr->allocate = ata_ahci_allocate;
}
}
Add support for AHCI compliant ATA devices. For now just support the Intel ICH6 as that the HW at hand. Sponsored by: pair.com
2005-05-11 16:10:08 +00:00
}
ctlr->setmode = ata_sata_setmode;
Update the ICH7 support so it deals better with chips without AHCI. Update Intel MatrixRAID support to be able to pick up RAID0+1 (RAID10) and RAID5 arrays without panic'ing. This has the side effect of now also supporting multiple volumes on MatrixRAID's now I have the metadata better understood.. HW sponsored by: Mullet Scandinavia AB
2005-12-02 10:13:53 +00:00
/* enable PCI interrupt */
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
pci_write_config(dev, PCIR_COMMAND,
pci_read_config(dev, PCIR_COMMAND, 2) & ~0x0400, 2);
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return 0;
}
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
static int
ata_intel_31244_allocate(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
int i;
int ch_offset;
ch_offset = 0x200 + ch->unit * 0x200;
for (i = ATA_DATA; i < ATA_MAX_RES; i++)
ch->r_io[i].res = ctlr->r_res2;
ch->r_io[ATA_DATA].offset = ch_offset + 0x00;
ch->r_io[ATA_FEATURE].offset = ch_offset + 0x06;
ch->r_io[ATA_COUNT].offset = ch_offset + 0x08;
ch->r_io[ATA_SECTOR].offset = ch_offset + 0x0c;
ch->r_io[ATA_CYL_LSB].offset = ch_offset + 0x10;
ch->r_io[ATA_CYL_MSB].offset = ch_offset + 0x14;
ch->r_io[ATA_DRIVE].offset = ch_offset + 0x18;
ch->r_io[ATA_COMMAND].offset = ch_offset + 0x1d;
ch->r_io[ATA_ERROR].offset = ch_offset + 0x04;
ch->r_io[ATA_STATUS].offset = ch_offset + 0x1c;
ch->r_io[ATA_ALTSTAT].offset = ch_offset + 0x28;
ch->r_io[ATA_CONTROL].offset = ch_offset + 0x29;
ch->r_io[ATA_SSTATUS].offset = ch_offset + 0x100;
ch->r_io[ATA_SERROR].offset = ch_offset + 0x104;
ch->r_io[ATA_SCONTROL].offset = ch_offset + 0x108;
ch->r_io[ATA_BMCMD_PORT].offset = ch_offset + 0x70;
ch->r_io[ATA_BMSTAT_PORT].offset = ch_offset + 0x72;
ch->r_io[ATA_BMDTP_PORT].offset = ch_offset + 0x74;
ch->flags |= ATA_NO_SLAVE;
ata_generic_hw(dev);
ch->hw.command = ata_intel_31244_command;
Enable SATA hotplug support. Submitted by: Olivier Houchard
2005-06-10 07:43:10 +00:00
/* enable PHY state change interrupt */
ATA_OUTL(ctlr->r_res2, 0x4,
ATA_INL(ctlr->r_res2, 0x04) | (0x01 << (ch->unit << 3)));
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
return 0;
}
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
static void
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
ata_intel_31244_intr(void *data)
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
int unit;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
for (unit = 0; unit < ctlr->channels; unit++) {
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
if (!(ch = ctlr->interrupt[unit].argument))
continue;
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
/* check for PHY related interrupts on SATA capable HW */
if (ctlr->chip->max_dma >= ATA_SA150) {
u_int32_t status = ATA_IDX_INL(ch, ATA_SSTATUS);
u_int32_t error = ATA_IDX_INL(ch, ATA_SERROR);
struct ata_connect_task *tp;
if (error) {
/* clear error bits/interrupt */
ATA_IDX_OUTL(ch, ATA_SERROR, error);
/* if we have a connection event deal with it */
if ((error & ATA_SE_PHY_CHANGED) &&
(tp = (struct ata_connect_task *)
malloc(sizeof(struct ata_connect_task),
M_ATA, M_NOWAIT | M_ZERO))) {
if ((status & ATA_SS_CONWELL_MASK) == ATA_SS_CONWELL_GEN1) {
device_printf(ch->dev, "CONNECT requested\n");
tp->action = ATA_C_ATTACH;
}
else {
device_printf(ch->dev, "DISCONNECT requested\n");
tp->action = ATA_C_DETACH;
}
tp->dev = ch->dev;
TASK_INIT(&tp->task, 0, ata_sata_phy_event, tp);
taskqueue_enqueue(taskqueue_thread, &tp->task);
}
}
}
/* any drive action to take care of ? */
if (ch->dma && (ch->dma->flags & ATA_DMA_ACTIVE)) {
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
int bmstat = ATA_IDX_INB(ch, ATA_BMSTAT_PORT) & ATA_BMSTAT_MASK;
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
if (!(bmstat & ATA_BMSTAT_INTERRUPT))
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
continue;
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, bmstat & ~ATA_BMSTAT_ERROR);
DELAY(1);
}
ctlr->interrupt[unit].function(ch);
}
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
static void
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
ata_intel_31244_reset(device_t dev)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
struct ata_channel *ch = device_get_softc(dev);
ata_sata_phy_enable(ch);
}
static int
ata_intel_31244_command(struct ata_request *request)
{
struct ata_channel *ch = device_get_softc(device_get_parent(request->dev));
struct ata_device *atadev = device_get_softc(request->dev);
u_int64_t lba;
if (!(atadev->flags & ATA_D_48BIT_ACTIVE))
return (ata_generic_command(request));
lba = request->u.ata.lba;
ATA_IDX_OUTB(ch, ATA_DRIVE, ATA_D_IBM | ATA_D_LBA | atadev->unit);
/* enable interrupt */
ATA_IDX_OUTB(ch, ATA_CONTROL, ATA_A_4BIT);
ATA_IDX_OUTW(ch, ATA_FEATURE, request->u.ata.feature);
ATA_IDX_OUTW(ch, ATA_COUNT, request->u.ata.count);
ATA_IDX_OUTW(ch, ATA_SECTOR, ((lba >> 16) & 0xff00) | (lba & 0x00ff));
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
ATA_IDX_OUTW(ch, ATA_CYL_LSB, ((lba >> 24) & 0xff00) |
((lba >> 8) & 0x00ff));
ATA_IDX_OUTW(ch, ATA_CYL_MSB, ((lba >> 32) & 0xff00) |
((lba >> 16) & 0x00ff));
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
/* issue command to controller */
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
ATA_IDX_OUTB(ch, ATA_COMMAND, request->u.ata.command);
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
return 0;
}
static void
ata_intel_intr(void *data)
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
int unit;
for (unit = 0; unit < ctlr->channels; unit++) {
if (!(ch = ctlr->interrupt[unit].argument))
continue;
if (ch->dma) {
int bmstat = ATA_IDX_INB(ch, ATA_BMSTAT_PORT) & ATA_BMSTAT_MASK;
if ((bmstat & (ATA_BMSTAT_ACTIVE | ATA_BMSTAT_INTERRUPT)) !=
ATA_BMSTAT_INTERRUPT)
continue;
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, bmstat & ~ATA_BMSTAT_ERROR);
DELAY(1);
}
ctlr->interrupt[unit].function(ch);
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_intel_reset(device_t dev)
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t parent = device_get_parent(dev);
Update the ICH6 support so all 4 SATA channels can be used. We still use "normal" mode, AHCI mode is in the works still. HW donated by: Sentex HW donated by: Yahoo!
2004-12-08 18:00:46 +00:00
struct ata_pci_controller *ctlr = device_get_softc(parent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int mask, timeout;
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
Update the ICH7 support so it deals better with chips without AHCI. Update Intel MatrixRAID support to be able to pick up RAID0+1 (RAID10) and RAID5 arrays without panic'ing. This has the side effect of now also supporting multiple volumes on MatrixRAID's now I have the metadata better understood.. HW sponsored by: Mullet Scandinavia AB
2005-12-02 10:13:53 +00:00
/* ICH6 & ICH7 in compat mode has 4 SATA ports as master/slave on 2 ch's */
if (ctlr->chip->cfg1) {
Update the ICH6 support so all 4 SATA channels can be used. We still use "normal" mode, AHCI mode is in the works still. HW donated by: Sentex HW donated by: Yahoo!
2004-12-08 18:00:46 +00:00
mask = (0x0005 << ch->unit);
}
else {
/* ICH5 in compat mode has SATA ports as master/slave on 1 channel */
if (pci_read_config(parent, 0x90, 1) & 0x04)
mask = 0x0003;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
else {
Update the ICH6 support so all 4 SATA channels can be used. We still use "normal" mode, AHCI mode is in the works still. HW donated by: Sentex HW donated by: Yahoo!
2004-12-08 18:00:46 +00:00
mask = (0x0001 << ch->unit);
Update the ICH7 support so it deals better with chips without AHCI. Update Intel MatrixRAID support to be able to pick up RAID0+1 (RAID10) and RAID5 arrays without panic'ing. This has the side effect of now also supporting multiple volumes on MatrixRAID's now I have the metadata better understood.. HW sponsored by: Mullet Scandinavia AB
2005-12-02 10:13:53 +00:00
/* XXX SOS should be in intel_allocate if we grow it */
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
ch->flags |= ATA_NO_SLAVE;
}
Update the ICH6 support so all 4 SATA channels can be used. We still use "normal" mode, AHCI mode is in the works still. HW donated by: Sentex HW donated by: Yahoo!
2004-12-08 18:00:46 +00:00
}
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
pci_write_config(parent, 0x92, pci_read_config(parent, 0x92, 2) & ~mask, 2);
DELAY(10);
pci_write_config(parent, 0x92, pci_read_config(parent, 0x92, 2) | mask, 2);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* wait up to 1 sec for "connect well" */
for (timeout = 0; timeout < 100 ; timeout++) {
if (((pci_read_config(parent, 0x92, 2) & (mask << 4)) == (mask << 4)) &&
(ATA_IDX_INB(ch, ATA_STATUS) != 0xff))
break;
Introduce ata_udelay() that uses tsleep instead of DELAY if possible. In places where we have long delays that doesn't depend on too accurate timing, use ata_udelay() instead of DELAY() so we dont uselessly spin the CPU if not nessesary;
2004-09-03 12:10:44 +00:00
ata_udelay(10000);
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
}
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_generic_reset(dev);
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
}
Add support for the Intel 31244. Most code by the submitters, hammered upon to get the right ATA fell by me. Submitted by: Olivier Houchard Submitted by: Jia-Shiun Li
2005-06-08 20:02:55 +00:00
static void
ata_intel_old_setmode(device_t dev, int mode)
{
/* NOT YET */
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_intel_new_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
u_int32_t reg40 = pci_read_config(gparent, 0x40, 4);
u_int8_t reg44 = pci_read_config(gparent, 0x44, 1);
u_int8_t reg48 = pci_read_config(gparent, 0x48, 1);
u_int16_t reg4a = pci_read_config(gparent, 0x4a, 2);
u_int16_t reg54 = pci_read_config(gparent, 0x54, 2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
u_int32_t mask40 = 0, new40 = 0;
u_int8_t mask44 = 0, new44 = 0;
int error;
u_int8_t timings[] = { 0x00, 0x00, 0x10, 0x21, 0x23, 0x10, 0x21, 0x23,
0x23, 0x23, 0x23, 0x23, 0x23, 0x23 };
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ctlr->chip->max_dma);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
if ( mode > ATA_UDMA2 && !(reg54 & (0x10 << devno))) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_print_cable(dev, "controller");
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
mode = ATA_UDMA2;
}
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting %s on %s chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "FAILURE " : "",
ata_mode2str(mode), ctlr->chip->text);
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
if (error)
return;
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
if (mode >= ATA_UDMA0) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x48, reg48 | (0x0001 << devno), 2);
Fix the long standing problem with poor transferrates on Intel ICHH type chips. The DMA timing value was set on device 0 for all devices :/ Prodded by: Harald Schmalzbauer
2005-06-09 21:13:44 +00:00
pci_write_config(gparent, 0x4a,
(reg4a & ~(0x3 << (devno << 2))) |
((0x01 + !(mode & 0x01)) << (devno << 2)), 2);
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
}
else {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x48, reg48 & ~(0x0001 << devno), 2);
pci_write_config(gparent, 0x4a, (reg4a & ~(0x3 << (devno << 2))), 2);
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
}
Always set half'n'half mode on ICH* chips.
2004-10-01 09:06:22 +00:00
reg54 |= 0x0400;
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
if (mode >= ATA_UDMA2)
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x54, reg54 | (0x1 << devno), 2);
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
else
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x54, reg54 & ~(0x1 << devno), 2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
if (mode >= ATA_UDMA5)
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x54, reg54 | (0x1000 << devno), 2);
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
else
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x54, reg54 & ~(0x1000 << devno), 2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
reg40 &= ~0x00ff00ff;
reg40 |= 0x40774077;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
if (atadev->unit == ATA_MASTER) {
mask40 = 0x3300;
new40 = timings[ata_mode2idx(mode)] << 8;
}
else {
mask44 = 0x0f;
new44 = ((timings[ata_mode2idx(mode)] & 0x30) >> 2) |
(timings[ata_mode2idx(mode)] & 0x03);
}
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (ch->unit) {
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
mask40 <<= 16;
new40 <<= 16;
mask44 <<= 4;
new44 <<= 4;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x40, (reg40 & ~mask40) | new40, 4);
pci_write_config(gparent, 0x44, (reg44 & ~mask44) | new44, 1);
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
atadev->mode = mode;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
/*
* Integrated Technology Express Inc. (ITE) chipset support functions
*/
int
ata_ite_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
Add support for the ITE IT8211F controller.
2005-08-05 13:14:00 +00:00
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_IT8212F, 0x00, 0x00, 0x00, ATA_UDMA6, "IT8212F" },
{ ATA_IT8211F, 0x00, 0x00, 0x00, ATA_UDMA6, "IT8211F" },
Add support for the ITE IT8211F controller.
2005-08-05 13:14:00 +00:00
{ 0, 0, 0, 0, 0, 0}};
char buffer[64];
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
Add support for the ITE IT8211F controller.
2005-08-05 13:14:00 +00:00
if (!(idx = ata_match_chip(dev, ids)))
return ENXIO;
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
sprintf(buffer, "ITE %s %s controller",
idx->text, ata_mode2str(idx->max_dma));
Add support for the ITE IT8211F controller.
2005-08-05 13:14:00 +00:00
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_ite_chipinit;
return 0;
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
}
static int
ata_ite_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
if (ata_setup_interrupt(dev))
return ENXIO;
ctlr->setmode = ata_ite_setmode;
/* set PCI mode and 66Mhz reference clock */
pci_write_config(dev, 0x50, pci_read_config(dev, 0x50, 1) & ~0x83, 1);
/* set default active & recover timings */
pci_write_config(dev, 0x54, 0x31, 1);
pci_write_config(dev, 0x56, 0x31, 1);
return 0;
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_ite_setmode(device_t dev, int mode)
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
int error;
/* correct the mode for what the HW supports */
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ATA_UDMA6);
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
/* check the CBLID bits for 80 conductor cable detection */
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (mode > ATA_UDMA2 && (pci_read_config(gparent, 0x40, 2) &
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
(ch->unit ? (1<<3) : (1<<2)))) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_print_cable(dev, "controller");
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
mode = ATA_UDMA2;
}
/* set the wanted mode on the device */
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%s setting %s on ITE8212F chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "failed" : "success", ata_mode2str(mode));
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
/* if the device accepted the mode change, setup the HW accordingly */
if (!error) {
if (mode >= ATA_UDMA0) {
u_int8_t udmatiming[] =
{ 0x44, 0x42, 0x31, 0x21, 0x11, 0xa2, 0x91 };
/* enable UDMA mode */
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x50,
pci_read_config(gparent, 0x50, 1) &
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
~(1 << (devno + 3)), 1);
/* set UDMA timing */
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent,
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
0x56 + (ch->unit << 2) + ATA_DEV(atadev->unit),
udmatiming[mode & ATA_MODE_MASK], 1);
}
else {
u_int8_t chtiming[] =
{ 0xaa, 0xa3, 0xa1, 0x33, 0x31, 0x88, 0x32, 0x31 };
/* disable UDMA mode */
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x50,
pci_read_config(gparent, 0x50, 1) |
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
(1 << (devno + 3)), 1);
/* set active and recover timing (shared between master & slave) */
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (pci_read_config(gparent, 0x54 + (ch->unit << 2), 1) <
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
chtiming[ata_mode2idx(mode)])
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x54 + (ch->unit << 2),
chtiming[ata_mode2idx(mode)], 1);
Add support for the ITE IT8212F controller. HW donated by: Yahoo!
2004-12-08 11:17:38 +00:00
}
atadev->mode = mode;
}
}
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
Add initial support for the Marvell 88SX[56]0[48][01] series of SATA chips. Hardware donated by: Matthew Jacob Hardware donated by: Y!
2005-12-27 17:09:52 +00:00
/*
* Marvell chipset support functions
*/
#define ATA_MV_HOST_BASE(ch) \
((ch->unit & 3) * 0x0100) + (ch->unit > 3 ? 0x30000 : 0x20000)
#define ATA_MV_EDMA_BASE(ch) \
((ch->unit & 3) * 0x2000) + (ch->unit > 3 ? 0x30000 : 0x20000)
int
ata_marvell_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
{{ ATA_M88SX5040, 0, 4, MV5XXX, ATA_SA150, "88SX5040" },
{ ATA_M88SX5041, 0, 4, MV5XXX, ATA_SA150, "88SX5041" },
{ ATA_M88SX5080, 0, 8, MV5XXX, ATA_SA150, "88SX5080" },
{ ATA_M88SX5081, 0, 8, MV5XXX, ATA_SA150, "88SX5081" },
{ ATA_M88SX6041, 0, 4, MV6XXX, ATA_SA300, "88SX6041" },
{ ATA_M88SX6081, 0, 8, MV6XXX, ATA_SA300, "88SX6081" },
{ 0, 0, 0, 0, 0, 0}};
char buffer[64];
if (pci_get_class(dev) != PCIC_STORAGE ||
pci_get_vendor(dev) != ATA_MARVELL_ID)
return ENXIO;
if (!(idx = ata_match_chip(dev, ids)))
return ENXIO;
sprintf(buffer, "Marvell %s %s controller",
idx->text, ata_mode2str(idx->max_dma));
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_marvell_chipinit;
return 0;
}
static int
ata_marvell_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
int rid = ATA_IRQ_RID;
if (!(ctlr->r_irq = bus_alloc_resource_any(dev, SYS_RES_IRQ, &rid,
RF_SHAREABLE | RF_ACTIVE))) {
device_printf(dev, "unable to map interrupt\n");
return ENXIO;
}
ctlr->r_type1 = SYS_RES_MEMORY;
ctlr->r_rid1 = PCIR_BAR(0);
if (!(ctlr->r_res1 = bus_alloc_resource_any(dev, ctlr->r_type1,
&ctlr->r_rid1, RF_ACTIVE)))
return ENXIO;
/* mask all host controller interrupts */
ATA_OUTL(ctlr->r_res1, 0x01d64, 0x00000000);
/* mask all PCI interrupts */
ATA_OUTL(ctlr->r_res1, 0x01d5c, 0x00000000);
ctlr->reset = ata_marvell_reset;
ctlr->dmainit = ata_marvell_dmainit;
ctlr->allocate = ata_marvell_allocate;
ctlr->setmode = ata_sata_setmode;
ctlr->channels = ctlr->chip->cfg1;
if ((bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
ata_marvell_intr, ctlr, &ctlr->handle))) {
device_printf(dev, "unable to setup interrupt\n");
return ENXIO;
}
/* clear host controller interrupts */
ATA_OUTL(ctlr->r_res1, 0x20014, 0x00000000);
if (ctlr->chip->cfg1 > 4)
ATA_OUTL(ctlr->r_res1, 0x30014, 0x00000000);
/* clear PCI interrupts */
ATA_OUTL(ctlr->r_res1, 0x01d58, 0x00000000);
/* unmask PCI interrupts we want */
ATA_OUTL(ctlr->r_res1, 0x01d5c, 0x007fffff);
/* unmask host controller interrupts we want */
ATA_OUTL(ctlr->r_res1, 0x01d64, 0x000000ff/*HC0*/ | 0x0001fe00/*HC1*/ |
/*(1<<19) | (1<<20) | (1<<21) |*/(1<<22) | (1<<24) | (0x7f << 25));
pci_write_config(dev, PCIR_COMMAND,
pci_read_config(dev, PCIR_COMMAND, 2) & ~0x0400, 2);
return 0;
}
static int
ata_marvell_allocate(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
bus_addr_t wordp = ch->dma->work_bus;
int i;
/* set legacy ATA resources */
for (i = ATA_DATA; i <= ATA_COMMAND; i++) {
ch->r_io[i].res = ctlr->r_res1;
ch->r_io[i].offset = 0x02100 + (i << 2) + ATA_MV_EDMA_BASE(ch);
}
ch->r_io[ATA_CONTROL].res = ctlr->r_res1;
ch->r_io[ATA_CONTROL].offset = 0x02120 + ATA_MV_EDMA_BASE(ch);
ch->r_io[ATA_IDX_ADDR].res = ctlr->r_res1;
ata_default_registers(dev);
/* set SATA resources */
switch (ctlr->chip->cfg2) {
case MV5XXX:
ch->r_io[ATA_SSTATUS].res = ctlr->r_res1;
ch->r_io[ATA_SSTATUS].offset = 0x00100 + ATA_MV_HOST_BASE(ch);
ch->r_io[ATA_SERROR].res = ctlr->r_res1;
ch->r_io[ATA_SERROR].offset = 0x00104 + ATA_MV_HOST_BASE(ch);
ch->r_io[ATA_SCONTROL].res = ctlr->r_res1;
ch->r_io[ATA_SCONTROL].offset = 0x00108 + ATA_MV_HOST_BASE(ch);
break;
case MV6XXX:
ch->r_io[ATA_SSTATUS].res = ctlr->r_res1;
ch->r_io[ATA_SSTATUS].offset = 0x02300 + ATA_MV_EDMA_BASE(ch);
ch->r_io[ATA_SERROR].res = ctlr->r_res1;
ch->r_io[ATA_SERROR].offset = 0x02304 + ATA_MV_EDMA_BASE(ch);
ch->r_io[ATA_SCONTROL].res = ctlr->r_res1;
ch->r_io[ATA_SCONTROL].offset = 0x02308 + ATA_MV_EDMA_BASE(ch);
ch->r_io[ATA_SACTIVE].res = ctlr->r_res1;
ch->r_io[ATA_SACTIVE].offset = 0x02350 + ATA_MV_EDMA_BASE(ch);
break;
}
ch->flags |= ATA_NO_SLAVE;
ch->flags |= ATA_USE_16BIT; /* XXX SOS needed ? */
ata_generic_hw(dev);
ch->hw.command = ata_marvell_command;
/* disable the EDMA machinery */
ATA_OUTL(ctlr->r_res1, 0x02028 + ATA_MV_EDMA_BASE(ch), 0x00000002);
DELAY(100000); /* SOS should poll for disabled */
/* set configuration to non-queued 128b read transfers stop on error */
ATA_OUTL(ctlr->r_res1, 0x02000 + ATA_MV_EDMA_BASE(ch), (1<<11) | (1<<13));
/* request queue base high */
ATA_OUTL(ctlr->r_res1, 0x02010 + ATA_MV_EDMA_BASE(ch), (wordp >> 16) >> 16);
/* request queue in ptr */
ATA_OUTL(ctlr->r_res1, 0x02014 + ATA_MV_EDMA_BASE(ch), wordp & 0xffffffff);
/* request queue out ptr */
ATA_OUTL(ctlr->r_res1, 0x02018 + ATA_MV_EDMA_BASE(ch), 0x0);
/* response queue base high */
wordp += 1024;
ATA_OUTL(ctlr->r_res1, 0x0201c + ATA_MV_EDMA_BASE(ch), (wordp >> 16) >> 16);
/* response queue in ptr */
ATA_OUTL(ctlr->r_res1, 0x02020 + ATA_MV_EDMA_BASE(ch), 0x0);
/* response queue out ptr */
ATA_OUTL(ctlr->r_res1, 0x02024 + ATA_MV_EDMA_BASE(ch), wordp & 0xffffffff);
/* clear SATA error register */
ATA_IDX_OUTL(ch, ATA_SERROR, ATA_IDX_INL(ch, ATA_SERROR));
/* clear any outstanding error interrupts */
ATA_OUTL(ctlr->r_res1, 0x02008 + ATA_MV_EDMA_BASE(ch), 0x0);
/* unmask all error interrupts */
ATA_OUTL(ctlr->r_res1, 0x0200c + ATA_MV_EDMA_BASE(ch), ~0x0);
/* enable EDMA machinery */
ATA_OUTL(ctlr->r_res1, 0x02028 + ATA_MV_EDMA_BASE(ch), 0x00000001);
return 0;
}
static void
ata_marvell_intr(void *data)
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
struct ata_request *request;
u_int32_t cause, icr0 = 0, icr1 = 0;
int unit;
cause = ATA_INL(ctlr->r_res1, 0x01d60);
if ((icr0 = ATA_INL(ctlr->r_res1, 0x20014)))
ATA_OUTL(ctlr->r_res1, 0x20014, ~icr0);
if (ctlr->channels > 4) {
if ((icr1 = ATA_INL(ctlr->r_res1, 0x30014)))
ATA_OUTL(ctlr->r_res1, 0x30014, ~icr1);
}
for (unit = 0; unit < ctlr->channels; unit++) {
u_int32_t icr;
int shift;
if (!(ch = ctlr->interrupt[unit].argument))
continue;
mtx_lock(&ch->state_mtx);
shift = unit << 1;
if (ch->unit < 4)
icr = icr0;
else {
icr = icr1;
shift++;
}
/* do we have any errors flagged ? */
if (cause & (1 << shift)) {
/* clear SATA error register */
ATA_IDX_OUTL(ch, ATA_SERROR, ATA_IDX_INL(ch, ATA_SERROR));
/* clear any outstanding error interrupts */
ATA_OUTL(ctlr->r_res1, 0x02008 + ATA_MV_EDMA_BASE(ch), 0x0);
}
/* EDMA interrupt */
if ((icr & (0x0001 << (unit & 3))) && (request = ch->running)) {
u_int32_t rsp_in, rsp_out;
int slot;
request->dmastat = ch->dma->stop(request->dev);
ch->dma->unload(ch->dev);
callout_stop(&request->callout);
/* get response ptr's */
rsp_in = ATA_INL(ctlr->r_res1, 0x02020 + ATA_MV_EDMA_BASE(ch));
rsp_out = ATA_INL(ctlr->r_res1, 0x02024 + ATA_MV_EDMA_BASE(ch));
slot = (((rsp_in & ~0xffffff00) >> 3)) & 0x1f;
rsp_out &= 0xffffff00;
rsp_out += (slot << 3);
/* XXX SOS get status and error into request */
request->status = 0; /* XXX SOS */
request->error = 0; /* XXX SOS */
if (!(request->flags & ATA_R_TIMEOUT))
request->donecount = request->bytecount;
/* ack response */
ATA_OUTL(ctlr->r_res1, 0x02024 + ATA_MV_EDMA_BASE(ch), rsp_out);
ch->running = NULL;
if (ch->state == ATA_ACTIVE)
ch->state = ATA_IDLE;
mtx_unlock(&ch->state_mtx);
ATA_LOCKING(ch->dev, ATA_LF_UNLOCK);
ata_finish(request);
}
/* legacy ATA interrupt */
else if (icr & (0x0100 << (unit & 3))) {
mtx_unlock(&ch->state_mtx);
ctlr->interrupt[unit].function(ch);
}
/* no device action to handle */
else {
mtx_unlock(&ch->state_mtx);
}
}
}
struct ata_marvell_dma_prdentry {
u_int32_t addrlo;
u_int32_t count;
u_int32_t addrhi;
u_int32_t reserved;
};
static void
ata_marvell_dmasetprd(void *xsc, bus_dma_segment_t *segs, int nsegs, int error)
{
struct ata_dmasetprd_args *args = xsc;
struct ata_marvell_dma_prdentry *prd = args->dmatab;
int i;
if ((args->error = error))
return;
for (i = 0; i < nsegs; i++) {
prd[i].addrlo = htole32(segs[i].ds_addr);
prd[i].addrhi = 0;
prd[i].count = htole32(segs[i].ds_len);
}
prd[i - 1].count |= htole32(ATA_DMA_EOT);
}
static int
ata_marvell_dmastart(device_t dev)
{
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
ch->flags |= ATA_DMA_ACTIVE;
return 0;
}
static int
ata_marvell_dmastop(device_t dev)
{
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
ch->flags &= ~ATA_DMA_ACTIVE;
/* get status XXX SOS */
return 0;
}
static void
ata_marvell_dmainit(device_t dev)
{
struct ata_channel *ch = device_get_softc(dev);
ata_dmainit(dev);
if (ch->dma) {
ch->dma->start = ata_marvell_dmastart;
ch->dma->stop = ata_marvell_dmastop;
ch->dma->setprd = ata_marvell_dmasetprd;
}
}
static int
ata_marvell_command(struct ata_request *request)
{
struct ata_pci_controller *ctlr=device_get_softc(GRANDPARENT(request->dev));
struct ata_channel *ch = device_get_softc(device_get_parent(request->dev));
u_int32_t req_in, req_out;
u_int8_t *bytep;
u_int16_t *wordp;
u_int32_t *quadp;
int i, tag = 0x07;
int slot;
/* only DMA R/W goes through the EMDA machine */
/* XXX SOS add ATAPI commands support later */
if (request->u.ata.command != ATA_READ_DMA &&
request->u.ata.command != ATA_WRITE_DMA) {
/* disable the EDMA machinery */
if (ATA_INL(ctlr->r_res1, 0x02028 + ATA_MV_EDMA_BASE(ch)) & 0x00000001)
ATA_OUTL(ctlr->r_res1, 0x02028 + ATA_MV_EDMA_BASE(ch), 0x00000002);
return ata_generic_command(request);
}
req_out = ATA_INL(ctlr->r_res1, 0x02018 + ATA_MV_EDMA_BASE(ch));
/* get next free request queue slot */
req_in = ATA_INL(ctlr->r_res1, 0x02014 + ATA_MV_EDMA_BASE(ch));
slot = (((req_in & ~0xfffffc00) >> 5) + 0) & 0x1f;
bytep = (u_int8_t *)(ch->dma->work);
bytep += (slot << 5);
wordp = (u_int16_t *)bytep;
quadp = (u_int32_t *)bytep;
quadp[0] = (long)ch->dma->sg_bus & 0xffffffff;
quadp[1] = (ch->dma->sg_bus & 0xffffffff00000000) >> 32;
wordp[4] = (request->flags & ATA_R_READ ? 0x01 : 0x00) | (tag<<1);
i = 10;
bytep[i++] = (request->u.ata.count >> 8) & 0xff;
bytep[i++] = 0x10 | ATA_COUNT;
bytep[i++] = request->u.ata.count & 0xff;
bytep[i++] = 0x10 | ATA_COUNT;
bytep[i++] = (request->u.ata.lba >> 24) & 0xff;
bytep[i++] = 0x10 | ATA_SECTOR;
bytep[i++] = request->u.ata.lba & 0xff;
bytep[i++] = 0x10 | ATA_SECTOR;
bytep[i++] = (request->u.ata.lba >> 32) & 0xff;
bytep[i++] = 0x10 | ATA_CYL_LSB;
bytep[i++] = (request->u.ata.lba >> 8) & 0xff;
bytep[i++] = 0x10 | ATA_CYL_LSB;
bytep[i++] = (request->u.ata.lba >> 40) & 0xff;
bytep[i++] = 0x10 | ATA_CYL_MSB;
bytep[i++] = (request->u.ata.lba >> 16) & 0xff;
bytep[i++] = 0x10 | ATA_CYL_MSB;
bytep[i++] = ATA_D_LBA | ATA_D_IBM | ((request->u.ata.lba >> 24) & 0xf);
bytep[i++] = 0x10 | ATA_DRIVE;
bytep[i++] = request->u.ata.command;
bytep[i++] = 0x90 | ATA_COMMAND;
/* enable EDMA machinery if needed */
if (!(ATA_INL(ctlr->r_res1, 0x02028 + ATA_MV_EDMA_BASE(ch)) & 0x00000001)) {
ATA_OUTL(ctlr->r_res1, 0x02028 + ATA_MV_EDMA_BASE(ch), 0x00000001);
while (!(ATA_INL(ctlr->r_res1,
0x02028 + ATA_MV_EDMA_BASE(ch)) & 0x00000001))
DELAY(10);
}
slot = (((req_in & ~0xfffffc00) >> 5) + 1) & 0x1f;
req_in &= 0xfffffc00;
req_in += (slot << 5);
/* tell EDMA it has a new request */
ATA_OUTL(ctlr->r_res1, 0x02014 + ATA_MV_EDMA_BASE(ch), req_in);
return 0;
}
static void
ata_marvell_reset(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
/* disable the EDMA machinery */
ATA_OUTL(ctlr->r_res1, 0x02028 + ATA_MV_EDMA_BASE(ch), 0x00000002);
DELAY(100000); /* SOS should poll for disabled */
/* clear SATA error register */
ATA_IDX_OUTL(ch, ATA_SERROR, ATA_IDX_INL(ch, ATA_SERROR));
/* clear any outstanding error interrupts */
ATA_OUTL(ctlr->r_res1, 0x02008 + ATA_MV_EDMA_BASE(ch), 0x0);
/* unmask all error interrupts */
ATA_OUTL(ctlr->r_res1, 0x0200c + ATA_MV_EDMA_BASE(ch), ~0x0);
/* enable channel and test for devices */
ata_sata_phy_enable(ch);
/* enable EDMA machinery */
ATA_OUTL(ctlr->r_res1, 0x02028 + ATA_MV_EDMA_BASE(ch), 0x00000001);
}
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
/*
* National chipset support functions
*/
int
ata_national_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
/* this chip is a clone of the Cyrix chip, bugs and all */
if (pci_get_devid(dev) == ATA_SC1100) {
device_set_desc(dev, "National Geode SC1100 ATA33 controller");
ctlr->chipinit = ata_national_chipinit;
return 0;
}
return ENXIO;
}
static device_t nat_host = NULL;
static int
ata_national_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
device_t *children;
int nchildren, i;
if (ata_setup_interrupt(dev))
return ENXIO;
/* locate the ISA part in the southbridge and enable UDMA33 */
if (!device_get_children(device_get_parent(dev), &children,&nchildren)){
for (i = 0; i < nchildren; i++) {
if (pci_get_devid(children[i]) == 0x0510100b) {
nat_host = children[i];
break;
}
}
free(children, M_TEMP);
}
ctlr->setmode = ata_national_setmode;
return 0;
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_national_setmode(device_t dev, int mode)
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
u_int32_t piotiming[] =
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
{ 0x9172d132, 0x21717121, 0x00803020, 0x20102010, 0x00100010,
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
0x00803020, 0x20102010, 0x00100010,
0x00100010, 0x00100010, 0x00100010 };
u_int32_t dmatiming[] = { 0x80077771, 0x80012121, 0x80002020 };
u_int32_t udmatiming[] = { 0x80921250, 0x80911140, 0x80911030 };
int error;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
ch->dma->alignment = 16;
ch->dma->max_iosize = 126 * DEV_BSIZE;
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ATA_UDMA2);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%s setting %s on National chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "failed" : "success", ata_mode2str(mode));
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (!error) {
if (mode >= ATA_UDMA0) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x44 + (devno << 3),
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
udmatiming[mode & ATA_MODE_MASK], 4);
}
else if (mode >= ATA_WDMA0) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x44 + (devno << 3),
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
dmatiming[mode & ATA_MODE_MASK], 4);
}
else {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x44 + (devno << 3),
pci_read_config(gparent, 0x44 + (devno << 3), 4) |
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
0x80000000, 4);
}
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x40 + (devno << 3),
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
piotiming[ata_mode2idx(mode)], 4);
atadev->mode = mode;
}
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*
* nVidia chipset support functions
*/
int
ata_nvidia_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_NFORCE1, 0, AMDNVIDIA, NVIDIA, ATA_UDMA5, "nForce" },
{ ATA_NFORCE2, 0, AMDNVIDIA, NVIDIA, ATA_UDMA6, "nForce2" },
{ ATA_NFORCE2_PRO, 0, AMDNVIDIA, NVIDIA, ATA_UDMA6, "nForce2 Pro" },
{ ATA_NFORCE2_PRO_S1, 0, 0, 0, ATA_SA150, "nForce2 Pro" },
{ ATA_NFORCE3, 0, AMDNVIDIA, NVIDIA, ATA_UDMA6, "nForce3" },
{ ATA_NFORCE3_PRO, 0, AMDNVIDIA, NVIDIA, ATA_UDMA6, "nForce3 Pro" },
{ ATA_NFORCE3_PRO_S1, 0, 0, 0, ATA_SA150, "nForce3 Pro" },
{ ATA_NFORCE3_PRO_S2, 0, 0, 0, ATA_SA150, "nForce3 Pro" },
{ ATA_NFORCE_MCP04, 0, AMDNVIDIA, NVIDIA, ATA_UDMA6, "nForce MCP" },
{ ATA_NFORCE_MCP04_S1, 0, 0, NV4OFF, ATA_SA150, "nForce MCP" },
{ ATA_NFORCE_MCP04_S2, 0, 0, NV4OFF, ATA_SA150, "nForce MCP" },
{ ATA_NFORCE_CK804, 0, AMDNVIDIA, NVIDIA, ATA_UDMA6, "nForce CK804" },
{ ATA_NFORCE_CK804_S1, 0, 0, NV4OFF, ATA_SA300, "nForce CK804" },
{ ATA_NFORCE_CK804_S2, 0, 0, NV4OFF, ATA_SA300, "nForce CK804" },
{ ATA_NFORCE_MCP51, 0, AMDNVIDIA, NVIDIA, ATA_UDMA6, "nForce MCP51" },
{ ATA_NFORCE_MCP51_S1, 0, 0, NV4OFF, ATA_SA300, "nForce MCP51" },
{ ATA_NFORCE_MCP51_S2, 0, 0, NV4OFF, ATA_SA300, "nForce MCP51" },
{ ATA_NFORCE_MCP55, 0, AMDNVIDIA, NVIDIA, ATA_UDMA6, "nForce MCP55" },
{ ATA_NFORCE_MCP55_S1, 0, 0, NV4OFF, ATA_SA300, "nForce MCP55" },
{ ATA_NFORCE_MCP55_S2, 0, 0, NV4OFF, ATA_SA300, "nForce MCP55" },
{ 0, 0, 0, 0, 0, 0}} ;
char buffer[64] ;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
if (!(idx = ata_match_chip(dev, ids)))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
sprintf(buffer, "nVidia %s %s controller",
idx->text, ata_mode2str(idx->max_dma));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_nvidia_chipinit;
return 0;
}
static int
ata_nvidia_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (ata_setup_interrupt(dev))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
if (ctlr->chip->max_dma >= ATA_SA150) {
Only try to allocate and use the SATA resource if they are enabled by the BIOS. It seems some BIOS's doesn't get this right, and that would result in ATA panic'ing.
2005-04-25 07:50:51 +00:00
if (pci_read_config(dev, PCIR_BAR(5), 1) & 1)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ctlr->r_type2 = SYS_RES_IOPORT;
Only try to allocate and use the SATA resource if they are enabled by the BIOS. It seems some BIOS's doesn't get this right, and that would result in ATA panic'ing.
2005-04-25 07:50:51 +00:00
else
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ctlr->r_type2 = SYS_RES_MEMORY;
ctlr->r_rid2 = PCIR_BAR(5);
if ((ctlr->r_res2 = bus_alloc_resource_any(dev, ctlr->r_type2,
Only try to allocate and use the SATA resource if they are enabled by the BIOS. It seems some BIOS's doesn't get this right, and that would result in ATA panic'ing.
2005-04-25 07:50:51 +00:00
&ctlr->r_rid2, RF_ACTIVE))) {
Make nVidia SATA hotplug support DTRT. HW sponsored by: Yahoo!
2005-06-05 21:18:26 +00:00
int offset = ctlr->chip->cfg2 & NV4OFF ? 0x0440 : 0x0010;
Only try to allocate and use the SATA resource if they are enabled by the BIOS. It seems some BIOS's doesn't get this right, and that would result in ATA panic'ing.
2005-04-25 07:50:51 +00:00
if (bus_teardown_intr(dev, ctlr->r_irq, ctlr->handle) ||
bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_nvidia_intr, ctlr, &ctlr->handle)) {
device_printf(dev, "unable to setup interrupt\n");
return ENXIO;
}
Make nVidia SATA hotplug support DTRT. HW sponsored by: Yahoo!
2005-06-05 21:18:26 +00:00
/* enable control access */
pci_write_config(dev, 0x50, pci_read_config(dev, 0x50, 1) | 0x04,1);
/* clear interrupt status */
ATA_OUTB(ctlr->r_res2, offset, 0xff);
/* enable device and PHY state change interrupts */
ATA_OUTB(ctlr->r_res2, offset + 1, 0xdd);
/* enable PCI interrupt */
Only try to allocate and use the SATA resource if they are enabled by the BIOS. It seems some BIOS's doesn't get this right, and that would result in ATA panic'ing.
2005-04-25 07:50:51 +00:00
pci_write_config(dev, PCIR_COMMAND,
pci_read_config(dev, PCIR_COMMAND, 2) & ~0x0400,2);
Make nVidia SATA hotplug support DTRT. HW sponsored by: Yahoo!
2005-06-05 21:18:26 +00:00
Only try to allocate and use the SATA resource if they are enabled by the BIOS. It seems some BIOS's doesn't get this right, and that would result in ATA panic'ing.
2005-04-25 07:50:51 +00:00
ctlr->allocate = ata_nvidia_allocate;
ctlr->reset = ata_nvidia_reset;
}
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
ctlr->setmode = ata_sata_setmode;
}
else {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* disable prefetch, postwrite */
pci_write_config(dev, 0x51, pci_read_config(dev, 0x51, 1) & 0x0f, 1);
ctlr->setmode = ata_via_family_setmode;
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return 0;
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
static int
ata_nvidia_allocate(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
/* setup the usual register normal pci style */
Properly react to allocation failures. Found by: imp@
2005-10-10 19:12:43 +00:00
if (ata_pci_allocate(dev))
return ENXIO;
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ch->r_io[ATA_SSTATUS].res = ctlr->r_res2;
ch->r_io[ATA_SSTATUS].offset = (ch->unit << 6);
ch->r_io[ATA_SERROR].res = ctlr->r_res2;
ch->r_io[ATA_SERROR].offset = 0x04 + (ch->unit << 6);
ch->r_io[ATA_SCONTROL].res = ctlr->r_res2;
ch->r_io[ATA_SCONTROL].offset = 0x08 + (ch->unit << 6);
ch->flags |= ATA_NO_SLAVE;
return 0;
}
static void
ata_nvidia_intr(void *data)
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
int offset = ctlr->chip->cfg2 & NV4OFF ? 0x0440 : 0x0010;
u_int8_t status;
int unit;
Make nVidia SATA hotplug support DTRT. HW sponsored by: Yahoo!
2005-06-05 21:18:26 +00:00
/* get interrupt status */
status = ATA_INB(ctlr->r_res2, offset);
/* clear interrupt status */
ATA_OUTB(ctlr->r_res2, offset, 0xff);
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
Make nVidia SATA hotplug support DTRT. HW sponsored by: Yahoo!
2005-06-05 21:18:26 +00:00
for (unit = 0; unit < ctlr->channels; unit++) {
if ((ch = ctlr->interrupt[unit].argument)) {
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
struct ata_connect_task *tp;
Make nVidia SATA hotplug support DTRT. HW sponsored by: Yahoo!
2005-06-05 21:18:26 +00:00
int maskshift = ch->unit << 2;
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
Make nVidia SATA hotplug support DTRT. HW sponsored by: Yahoo!
2005-06-05 21:18:26 +00:00
/* clear error status */
ATA_IDX_OUTL(ch, ATA_SERROR, ATA_IDX_INL(ch, ATA_SERROR));
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
/* check for and handle connect events */
Make nVidia SATA hotplug support DTRT. HW sponsored by: Yahoo!
2005-06-05 21:18:26 +00:00
if (((status & (0x0c << maskshift)) == (0x04 << maskshift)) &&
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
(tp = (struct ata_connect_task *)
malloc(sizeof(struct ata_connect_task),
M_ATA, M_NOWAIT | M_ZERO))) {
device_printf(ch->dev, "CONNECT requested\n");
tp->action = ATA_C_ATTACH;
tp->dev = ch->dev;
TASK_INIT(&tp->task, 0, ata_sata_phy_event, tp);
taskqueue_enqueue(taskqueue_thread, &tp->task);
}
/* check for and handle disconnect events */
Make nVidia SATA hotplug support DTRT. HW sponsored by: Yahoo!
2005-06-05 21:18:26 +00:00
if ((status & (0x08 << maskshift)) &&
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
(tp = (struct ata_connect_task *)
malloc(sizeof(struct ata_connect_task),
M_ATA, M_NOWAIT | M_ZERO))) {
device_printf(ch->dev, "DISCONNECT requested\n");
tp->action = ATA_C_DETACH;
tp->dev = ch->dev;
TASK_INIT(&tp->task, 0, ata_sata_phy_event, tp);
taskqueue_enqueue(taskqueue_thread, &tp->task);
}
/* any drive action to take care of ? */
Make nVidia SATA hotplug support DTRT. HW sponsored by: Yahoo!
2005-06-05 21:18:26 +00:00
if (status & (0x01 << maskshift)) {
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
if (ch->dma && (ch->dma->flags & ATA_DMA_ACTIVE)) {
int bmstat =
ATA_IDX_INB(ch, ATA_BMSTAT_PORT) & ATA_BMSTAT_MASK;
if (!(bmstat & ATA_BMSTAT_INTERRUPT))
continue;
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, bmstat&~ATA_BMSTAT_ERROR);
DELAY(1);
}
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ctlr->interrupt[unit].function(ch);
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
}
}
}
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_nvidia_reset(device_t dev)
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(dev);
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ata_sata_phy_enable(ch);
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*
* Promise chipset support functions
*/
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
#define ATA_PDC_APKT_OFFSET 0x00000010
#define ATA_PDC_HPKT_OFFSET 0x00000040
#define ATA_PDC_ASG_OFFSET 0x00000080
#define ATA_PDC_LSG_OFFSET 0x000000c0
#define ATA_PDC_HSG_OFFSET 0x00000100
#define ATA_PDC_CHN_OFFSET 0x00000400
#define ATA_PDC_BUF_BASE 0x00400000
#define ATA_PDC_BUF_OFFSET 0x00100000
#define ATA_PDC_MAX_HPKT 8
#define ATA_PDC_WRITE_REG 0x00
#define ATA_PDC_WRITE_CTL 0x0e
#define ATA_PDC_WRITE_END 0x08
#define ATA_PDC_WAIT_NBUSY 0x10
#define ATA_PDC_WAIT_READY 0x18
#define ATA_PDC_1B 0x20
#define ATA_PDC_2B 0x40
struct host_packet {
u_int32_t addr;
TAILQ_ENTRY(host_packet) chain;
};
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
struct ata_promise_sx4 {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct mtx mtx;
TAILQ_HEAD(, host_packet) queue;
int busy;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
};
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int
ata_promise_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_PDC20246, 0, PROLD, 0x00, ATA_UDMA2, "PDC20246" },
{ ATA_PDC20262, 0, PRNEW, 0x00, ATA_UDMA4, "PDC20262" },
{ ATA_PDC20263, 0, PRNEW, 0x00, ATA_UDMA4, "PDC20263" },
{ ATA_PDC20265, 0, PRNEW, 0x00, ATA_UDMA5, "PDC20265" },
{ ATA_PDC20267, 0, PRNEW, 0x00, ATA_UDMA5, "PDC20267" },
{ ATA_PDC20268, 0, PRTX, PRTX4, ATA_UDMA5, "PDC20268" },
{ ATA_PDC20269, 0, PRTX, 0x00, ATA_UDMA6, "PDC20269" },
{ ATA_PDC20270, 0, PRTX, PRTX4, ATA_UDMA5, "PDC20270" },
{ ATA_PDC20271, 0, PRTX, 0x00, ATA_UDMA6, "PDC20271" },
{ ATA_PDC20275, 0, PRTX, 0x00, ATA_UDMA6, "PDC20275" },
{ ATA_PDC20276, 0, PRTX, PRSX6K, ATA_UDMA6, "PDC20276" },
{ ATA_PDC20277, 0, PRTX, 0x00, ATA_UDMA6, "PDC20277" },
{ ATA_PDC20318, 0, PRMIO, PRSATA, ATA_SA150, "PDC20318" },
{ ATA_PDC20319, 0, PRMIO, PRSATA, ATA_SA150, "PDC20319" },
{ ATA_PDC20371, 0, PRMIO, PRCMBO, ATA_SA150, "PDC20371" },
{ ATA_PDC20375, 0, PRMIO, PRCMBO, ATA_SA150, "PDC20375" },
{ ATA_PDC20376, 0, PRMIO, PRCMBO, ATA_SA150, "PDC20376" },
{ ATA_PDC20377, 0, PRMIO, PRCMBO, ATA_SA150, "PDC20377" },
{ ATA_PDC20378, 0, PRMIO, PRCMBO, ATA_SA150, "PDC20378" },
{ ATA_PDC20379, 0, PRMIO, PRCMBO, ATA_SA150, "PDC20379" },
{ ATA_PDC20571, 0, PRMIO, PRCMBO2, ATA_SA150, "PDC20571" },
{ ATA_PDC20575, 0, PRMIO, PRCMBO2, ATA_SA150, "PDC20575" },
{ ATA_PDC20579, 0, PRMIO, PRCMBO2, ATA_SA150, "PDC20579" },
{ ATA_PDC20580, 0, PRMIO, PRCMBO2, ATA_SA150, "PDC20580" },
{ ATA_PDC20617, 0, PRMIO, PRPATA, ATA_UDMA6, "PDC20617" },
{ ATA_PDC20618, 0, PRMIO, PRPATA, ATA_UDMA6, "PDC20618" },
{ ATA_PDC20619, 0, PRMIO, PRPATA, ATA_UDMA6, "PDC20619" },
{ ATA_PDC20620, 0, PRMIO, PRPATA, ATA_UDMA6, "PDC20620" },
{ ATA_PDC20621, 0, PRMIO, PRSX4X, ATA_UDMA5, "PDC20621" },
{ ATA_PDC20622, 0, PRMIO, PRSX4X, ATA_SA150, "PDC20622" },
{ ATA_PDC40518, 0, PRMIO, PRSATA2, ATA_SA150, "PDC40518" },
{ ATA_PDC40519, 0, PRMIO, PRSATA2, ATA_SA150, "PDC40519" },
{ ATA_PDC40718, 0, PRMIO, PRSATA2, ATA_SA300, "PDC40718" },
{ ATA_PDC40719, 0, PRMIO, PRSATA2, ATA_SA300, "PDC40719" },
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{ 0, 0, 0, 0, 0, 0}};
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
char buffer[64];
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
uintptr_t devid = 0;
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
if (!(idx = ata_match_chip(dev, ids)))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
/* if we are on a SuperTrak SX6000 dont attach */
Minor cleanup of the Promise code.
2003-03-03 11:51:08 +00:00
if ((idx->cfg2 & PRSX6K) && pci_get_class(GRANDPARENT(dev))==PCIC_BRIDGE &&
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
!BUS_READ_IVAR(device_get_parent(GRANDPARENT(dev)),
GRANDPARENT(dev), PCI_IVAR_DEVID, &devid) &&
Minor cleanup of the Promise code.
2003-03-03 11:51:08 +00:00
devid == ATA_I960RM)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
strcpy(buffer, idx->text);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
Minor cleanup of the Promise code.
2003-03-03 11:51:08 +00:00
/* if we are on a FastTrak TX4, adjust the interrupt resource */
if ((idx->cfg2 & PRTX4) && pci_get_class(GRANDPARENT(dev))==PCIC_BRIDGE &&
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
!BUS_READ_IVAR(device_get_parent(GRANDPARENT(dev)),
GRANDPARENT(dev), PCI_IVAR_DEVID, &devid) &&
Fix support for the Promise TX4 on amd64. Somehow the bridge on there shows up with another PCI id than it does on x86, no idea why...
2004-02-02 14:05:57 +00:00
((devid == ATA_DEC_21150) || (devid == ATA_DEC_21150_1))) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
static long start = 0, end = 0;
if (pci_get_slot(dev) == 1) {
bus_get_resource(dev, SYS_RES_IRQ, 0, &start, &end);
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
strcat(buffer, " (channel 0+1)");
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
else if (pci_get_slot(dev) == 2 && start && end) {
bus_set_resource(dev, SYS_RES_IRQ, 0, start, end);
start = end = 0;
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
strcat(buffer, " (channel 2+3)");
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
else {
start = end = 0;
}
}
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
sprintf(buffer, "Promise %s %s controller",
buffer, ata_mode2str(idx->max_dma));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_promise_chipinit;
return 0;
}
static int
ata_promise_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
int rid = ATA_IRQ_RID;
Convert callers to the new bus_alloc_resource_any(9) API. Submitted by: Mark Santcroos <marks@ripe.net> Reviewed by: imp, dfr, bde
2004-03-17 17:50:55 +00:00
if (!(ctlr->r_irq = bus_alloc_resource_any(dev, SYS_RES_IRQ, &rid,
RF_SHAREABLE | RF_ACTIVE))) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_printf(dev, "unable to map interrupt\n");
return ENXIO;
}
Second round of updates to the ATA driver. Clean up the DMA interface too much unneeded stuff crept in with the busdma code back when. Modify the ATA_IN* / ATA_OUT* macros so that resource and offset are gotten from a table. That allows for new chipsets that doesn't nessesarily have things ordered the good old way. This also removes the need for the wierd PC98 resource functions. Tested on: i386, PC98, Alpha, Sparc64
2003-03-29 13:37:09 +00:00
Further update to the Promise code. Support should now cover all old and current products.
2005-01-01 19:24:25 +00:00
if (ctlr->chip->max_dma >= ATA_SA150)
ctlr->setmode = ata_sata_setmode;
else
ctlr->setmode = ata_promise_setmode;
Second round of updates to the ATA driver. Clean up the DMA interface too much unneeded stuff crept in with the busdma code back when. Modify the ATA_IN* / ATA_OUT* macros so that resource and offset are gotten from a table. That allows for new chipsets that doesn't nessesarily have things ordered the good old way. This also removes the need for the wierd PC98 resource functions. Tested on: i386, PC98, Alpha, Sparc64
2003-03-29 13:37:09 +00:00
switch (ctlr->chip->cfg1) {
case PRNEW:
Fix the setup of old Promise controllers, clocks was not setup right causing way too low transfer rates. Enable interrupts on old CMD64[89] chips, apparently some bogus BIOS's doesn't get this right.
2003-06-07 15:19:16 +00:00
/* setup clocks */
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
ATA_OUTB(ctlr->r_res1, 0x11, ATA_INB(ctlr->r_res1, 0x11) | 0x0a);
Fix the setup of old Promise controllers, clocks was not setup right causing way too low transfer rates. Enable interrupts on old CMD64[89] chips, apparently some bogus BIOS's doesn't get this right.
2003-06-07 15:19:16 +00:00
ctlr->dmainit = ata_promise_new_dmainit;
Second round of updates to the ATA driver. Clean up the DMA interface too much unneeded stuff crept in with the busdma code back when. Modify the ATA_IN* / ATA_OUT* macros so that resource and offset are gotten from a table. That allows for new chipsets that doesn't nessesarily have things ordered the good old way. This also removes the need for the wierd PC98 resource functions. Tested on: i386, PC98, Alpha, Sparc64
2003-03-29 13:37:09 +00:00
/* FALLTHROUGH */
case PROLD:
Fix the setup of old Promise controllers, clocks was not setup right causing way too low transfer rates. Enable interrupts on old CMD64[89] chips, apparently some bogus BIOS's doesn't get this right.
2003-06-07 15:19:16 +00:00
/* enable burst mode */
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
ATA_OUTB(ctlr->r_res1, 0x1f, ATA_INB(ctlr->r_res1, 0x1f) | 0x01);
Fix the setup of old Promise controllers, clocks was not setup right causing way too low transfer rates. Enable interrupts on old CMD64[89] chips, apparently some bogus BIOS's doesn't get this right.
2003-06-07 15:19:16 +00:00
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if ((bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
Second round of updates to the ATA driver. Clean up the DMA interface too much unneeded stuff crept in with the busdma code back when. Modify the ATA_IN* / ATA_OUT* macros so that resource and offset are gotten from a table. That allows for new chipsets that doesn't nessesarily have things ordered the good old way. This also removes the need for the wierd PC98 resource functions. Tested on: i386, PC98, Alpha, Sparc64
2003-03-29 13:37:09 +00:00
ata_promise_old_intr, ctlr, &ctlr->handle))) {
device_printf(dev, "unable to setup interrupt\n");
return ENXIO;
}
break;
case PRTX:
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if ((bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
Second round of updates to the ATA driver. Clean up the DMA interface too much unneeded stuff crept in with the busdma code back when. Modify the ATA_IN* / ATA_OUT* macros so that resource and offset are gotten from a table. That allows for new chipsets that doesn't nessesarily have things ordered the good old way. This also removes the need for the wierd PC98 resource functions. Tested on: i386, PC98, Alpha, Sparc64
2003-03-29 13:37:09 +00:00
ata_promise_tx2_intr, ctlr, &ctlr->handle))) {
device_printf(dev, "unable to setup interrupt\n");
return ENXIO;
}
break;
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
case PRMIO:
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
// if (ctlr->r_res1)
// bus_release_resource(dev, ctlr->r_type1, ctlr->r_rid1,ctlr->r_res1);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
ctlr->r_type1 = SYS_RES_MEMORY;
Try to narrow down the race window on HW that does not have ways to poll for which channel actually pulled the irq line.
2004-08-05 21:13:41 +00:00
ctlr->r_rid1 = PCIR_BAR(4);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
if (!(ctlr->r_res1 = bus_alloc_resource_any(dev, ctlr->r_type1,
&ctlr->r_rid1, RF_ACTIVE)))
return ENXIO;
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
ctlr->r_type2 = SYS_RES_MEMORY;
Try to narrow down the race window on HW that does not have ways to poll for which channel actually pulled the irq line.
2004-08-05 21:13:41 +00:00
ctlr->r_rid2 = PCIR_BAR(3);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
if (!(ctlr->r_res2 = bus_alloc_resource_any(dev, ctlr->r_type2,
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
&ctlr->r_rid2, RF_ACTIVE))){
bus_release_resource(dev, ctlr->r_type1, ctlr->r_rid1,ctlr->r_res1);
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
return ENXIO;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
}
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
ctlr->reset = ata_promise_mio_reset;
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
ctlr->dmainit = ata_promise_mio_dmainit;
ctlr->allocate = ata_promise_mio_allocate;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (ctlr->chip->cfg2 == PRSX4X) {
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
struct ata_promise_sx4 *hpkt;
u_int32_t dimm = ATA_INL(ctlr->r_res2, 0x000c0080);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if ((bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
ata_promise_sx4_intr, ctlr, &ctlr->handle))) {
device_printf(dev, "unable to setup interrupt\n");
/* XXX SOS release resources */
return ENXIO;
}
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
/* print info about cache memory */
device_printf(dev, "DIMM size %dMB @ 0x%08x%s\n",
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
(((dimm >> 16) & 0xff)-((dimm >> 24) & 0xff)+1) << 4,
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
((dimm >> 24) & 0xff),
ATA_INL(ctlr->r_res2, 0x000c0088) & (1<<16) ?
" ECC enabled" : "" );
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* adjust cache memory parameters */
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
ATA_OUTL(ctlr->r_res2, 0x000c000c,
(ATA_INL(ctlr->r_res2, 0x000c000c) & 0xffff0000));
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* setup host packet controls */
hpkt = malloc(sizeof(struct ata_promise_sx4),
M_TEMP, M_NOWAIT | M_ZERO);
Use the right ordering of args on mtx_init(). No functional changes since the args in question was all zero's. Found by: Jimmy Olgeni <olgeni@FreeBSD.org>
2004-06-22 11:18:25 +00:00
mtx_init(&hpkt->mtx, "ATA promise HPKT lock", NULL, MTX_DEF);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
TAILQ_INIT(&hpkt->queue);
hpkt->busy = 0; //hpkt->head = hpkt->tail = 0;
device_set_ivars(dev, hpkt);
ctlr->channels = 4;
Further update to the Promise code. Support should now cover all old and current products.
2005-01-01 19:24:25 +00:00
return 0;
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
}
Further update to the Promise code. Support should now cover all old and current products.
2005-01-01 19:24:25 +00:00
if ((bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
ata_promise_mio_intr, ctlr, &ctlr->handle))) {
device_printf(dev, "unable to setup interrupt\n");
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* XXX SOS release resources */
Further update to the Promise code. Support should now cover all old and current products.
2005-01-01 19:24:25 +00:00
return ENXIO;
}
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
switch (ctlr->chip->cfg2) {
case PRPATA:
ctlr->channels = ((ATA_INL(ctlr->r_res2, 0x48) & 0x01) > 0) +
((ATA_INL(ctlr->r_res2, 0x48) & 0x02) > 0) + 2;
break;
case PRCMBO:
ATA_OUTL(ctlr->r_res2, 0x06c, 0x000000ff);
ctlr->channels = ((ATA_INL(ctlr->r_res2, 0x48) & 0x02) > 0) + 3;
break;
case PRSATA:
ATA_OUTL(ctlr->r_res2, 0x06c, 0x000000ff);
ctlr->channels = 4;
break;
case PRCMBO2:
ATA_OUTL(ctlr->r_res2, 0x060, 0x000000ff);
ctlr->channels = 3;
break;
case PRSATA2:
ATA_OUTL(ctlr->r_res2, 0x060, 0x000000ff);
ctlr->channels = 4;
break;
default:
/* XXX SOS release resources */
return ENXIO;
}
default:
/* XXX SOS release resources */
return ENXIO;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
return 0;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
static int
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
ata_promise_mio_allocate(device_t dev)
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
struct ata_channel *ch = device_get_softc(dev);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
int offset = (ctlr->chip->cfg2 & PRSX4X) ? 0x000c0000 : 0;
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
int i;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
Add support for controllers that doesn't have the usual taskfile layout. No functional changes.
2005-04-06 10:22:56 +00:00
for (i = ATA_DATA; i <= ATA_COMMAND; i++) {
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
ch->r_io[i].res = ctlr->r_res2;
ch->r_io[i].offset = offset + 0x0200 + (i << 2) + (ch->unit << 7);
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
}
Add support for controllers that doesn't have the usual taskfile layout. No functional changes.
2005-04-06 10:22:56 +00:00
ch->r_io[ATA_CONTROL].res = ctlr->r_res2;
ch->r_io[ATA_CONTROL].offset = offset + 0x0238 + (ch->unit << 7);
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
ch->r_io[ATA_IDX_ADDR].res = ctlr->r_res2;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_default_registers(dev);
Update the Promise support for combo (SATA+PATA) controllers.
2004-12-24 13:36:04 +00:00
if ((ctlr->chip->cfg2 & (PRSATA | PRSATA2)) ||
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
((ctlr->chip->cfg2 & (PRCMBO | PRCMBO2)) && ch->unit < 2)) {
ch->r_io[ATA_SSTATUS].res = ctlr->r_res2;
ch->r_io[ATA_SSTATUS].offset = 0x400 + (ch->unit << 8);
ch->r_io[ATA_SERROR].res = ctlr->r_res2;
ch->r_io[ATA_SERROR].offset = 0x404 + (ch->unit << 8);
ch->r_io[ATA_SCONTROL].res = ctlr->r_res2;
ch->r_io[ATA_SCONTROL].offset = 0x408 + (ch->unit << 8);
Add first shot on support for the new Promise SATAII chips. HW donated by: pil.dk
2004-12-08 10:02:41 +00:00
ch->flags |= ATA_NO_SLAVE;
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
}
ch->flags |= ATA_USE_16BIT;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_generic_hw(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (offset)
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
ch->hw.command = ata_promise_sx4_command;
else
ch->hw.command = ata_promise_mio_command;
return 0;
}
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
static void
ata_promise_mio_intr(void *data)
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
u_int32_t vector = 0, status = 0;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
int unit;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
switch (ctlr->chip->cfg2) {
case PRSATA:
case PRCMBO:
/* read and acknowledge interrupt(s) */
vector = ATA_INL(ctlr->r_res2, 0x040);
/* read and clear interface status */
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
status = ATA_INL(ctlr->r_res2, 0x06c);
ATA_OUTL(ctlr->r_res2, 0x06c, status & 0x000000ff);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
break;
case PRSATA2:
case PRCMBO2:
/* read and acknowledge interrupt(s) */
vector = ATA_INL(ctlr->r_res2, 0x040);
Add first shot on support for the new Promise SATAII chips. HW donated by: pil.dk
2004-12-08 10:02:41 +00:00
ATA_OUTL(ctlr->r_res2, 0x040, vector & 0x0000ffff);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* read and clear interface status */
Add first shot on support for the new Promise SATAII chips. HW donated by: pil.dk
2004-12-08 10:02:41 +00:00
status = ATA_INL(ctlr->r_res2, 0x060);
ATA_OUTL(ctlr->r_res2, 0x060, status & 0x000000ff);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
break;
Add first shot on support for the new Promise SATAII chips. HW donated by: pil.dk
2004-12-08 10:02:41 +00:00
}
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
for (unit = 0; unit < ctlr->channels; unit++) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if ((ch = ctlr->interrupt[unit].argument)) {
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
struct ata_connect_task *tp;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* check for and handle disconnect events */
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
if ((status & (0x00000001 << unit)) &&
(tp = (struct ata_connect_task *)
malloc(sizeof(struct ata_connect_task),
M_ATA, M_NOWAIT | M_ZERO))) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (bootverbose)
device_printf(ch->dev, "DISCONNECT requested\n");
tp->action = ATA_C_DETACH;
tp->dev = ch->dev;
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
TASK_INIT(&tp->task, 0, ata_sata_phy_event, tp);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
taskqueue_enqueue(taskqueue_thread, &tp->task);
}
/* check for and handle connect events */
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
if ((status & (0x00000010 << unit)) &&
(tp = (struct ata_connect_task *)
malloc(sizeof(struct ata_connect_task),
M_ATA, M_NOWAIT | M_ZERO))) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (bootverbose)
device_printf(ch->dev, "CONNECT requested\n");
tp->action = ATA_C_ATTACH;
tp->dev = ch->dev;
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
TASK_INIT(&tp->task, 0, ata_sata_phy_event, tp);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
taskqueue_enqueue(taskqueue_thread, &tp->task);
}
/* active interrupt(s) need to call the interrupt handler */
if (vector & (1 << (unit + 1)))
if ((ch = ctlr->interrupt[unit].argument))
ctlr->interrupt[unit].function(ch);
}
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
}
}
static void
ata_promise_sx4_intr(void *data)
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
u_int32_t vector = ATA_INL(ctlr->r_res2, 0x000c0480);
int unit;
for (unit = 0; unit < ctlr->channels; unit++) {
if (vector & (1 << (unit + 1)))
if ((ch = ctlr->interrupt[unit].argument))
ctlr->interrupt[unit].function(ch);
if (vector & (1 << (unit + 5)))
if ((ch = ctlr->interrupt[unit].argument))
ata_promise_queue_hpkt(ctlr,
htole32((ch->unit * ATA_PDC_CHN_OFFSET) +
ATA_PDC_HPKT_OFFSET));
if (vector & (1 << (unit + 9))) {
ata_promise_next_hpkt(ctlr);
if ((ch = ctlr->interrupt[unit].argument))
ctlr->interrupt[unit].function(ch);
}
if (vector & (1 << (unit + 13))) {
ata_promise_next_hpkt(ctlr);
if ((ch = ctlr->interrupt[unit].argument))
ATA_OUTL(ctlr->r_res2, 0x000c0240 + (ch->unit << 7),
htole32((ch->unit * ATA_PDC_CHN_OFFSET) +
ATA_PDC_APKT_OFFSET));
}
}
}
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_promise_mio_dmainit(device_t dev)
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(dev);
ata_dmainit(dev);
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
if (ch->dma) {
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
/* note start and stop are not used here */
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
}
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_promise_mio_reset(device_t dev)
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
struct ata_promise_sx4 *hpktp;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
Further update to the Promise code. Support should now cover all old and current products.
2005-01-01 19:24:25 +00:00
switch (ctlr->chip->cfg2) {
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
case PRSX4X:
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
Further update to the Promise code. Support should now cover all old and current products.
2005-01-01 19:24:25 +00:00
/* softreset channel ATA module */
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
hpktp = device_get_ivars(ctlr->dev);
Fix the reset sequence for sx4/sx4000 type controllers, it could hang the chip in an indefinite state.
2004-07-30 13:33:09 +00:00
ATA_OUTL(ctlr->r_res2, 0xc0260 + (ch->unit << 7), ch->unit + 1);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
ata_udelay(1000);
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
ATA_OUTL(ctlr->r_res2, 0xc0260 + (ch->unit << 7),
(ATA_INL(ctlr->r_res2, 0xc0260 + (ch->unit << 7)) &
~0x00003f9f) | (ch->unit + 1));
Cleanup comments
2005-05-13 10:25:19 +00:00
/* softreset HOST module */ /* XXX SOS what about other outstandings */
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
mtx_lock(&hpktp->mtx);
ATA_OUTL(ctlr->r_res2, 0xc012c,
(ATA_INL(ctlr->r_res2, 0xc012c) & ~0x00000f9f) | (1 << 11));
DELAY(10);
ATA_OUTL(ctlr->r_res2, 0xc012c,
(ATA_INL(ctlr->r_res2, 0xc012c) & ~0x00000f9f));
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
hpktp->busy = 0;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
mtx_unlock(&hpktp->mtx);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_generic_reset(dev);
Further update to the Promise code. Support should now cover all old and current products.
2005-01-01 19:24:25 +00:00
break;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
case PRCMBO:
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
case PRSATA:
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if ((ctlr->chip->cfg2 == PRSATA) ||
((ctlr->chip->cfg2 == PRCMBO) && (ch->unit < 2))) {
/* mask plug/unplug intr */
ATA_OUTL(ctlr->r_res2, 0x06c, (0x00110000 << ch->unit));
}
Add first shot on support for the new Promise SATAII chips. HW donated by: pil.dk
2004-12-08 10:02:41 +00:00
/* softreset channels ATA module */
ATA_OUTL(ctlr->r_res2, 0x0260 + (ch->unit << 7), (1 << 11));
ata_udelay(10000);
ATA_OUTL(ctlr->r_res2, 0x0260 + (ch->unit << 7),
(ATA_INL(ctlr->r_res2, 0x0260 + (ch->unit << 7)) &
~0x00003f9f) | (ch->unit + 1));
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if ((ctlr->chip->cfg2 == PRSATA) ||
((ctlr->chip->cfg2 == PRCMBO) && (ch->unit < 2))) {
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ata_sata_phy_enable(ch);
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* reset and enable plug/unplug intr */
ATA_OUTL(ctlr->r_res2, 0x06c, (0x00000011 << ch->unit));
}
Rearrange the way the reset code is called. Prepare for different looking controllers.
2005-04-28 22:08:08 +00:00
else
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_generic_reset(dev);
Further update to the Promise code. Support should now cover all old and current products.
2005-01-01 19:24:25 +00:00
break;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
case PRCMBO2:
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
case PRSATA2:
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if ((ctlr->chip->cfg2 == PRSATA2) ||
((ctlr->chip->cfg2 == PRCMBO2) && (ch->unit < 2))) {
/* set portmultiplier port */
ATA_OUTL(ctlr->r_res2, 0x4e8 + (ch->unit << 8), 0x0f);
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* mask plug/unplug intr */
ATA_OUTL(ctlr->r_res2, 0x060, (0x00110000 << ch->unit));
}
Add first shot on support for the new Promise SATAII chips. HW donated by: pil.dk
2004-12-08 10:02:41 +00:00
/* softreset channels ATA module */
ATA_OUTL(ctlr->r_res2, 0x0260 + (ch->unit << 7), (1 << 11));
ata_udelay(10000);
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
ATA_OUTL(ctlr->r_res2, 0x0260 + (ch->unit << 7),
(ATA_INL(ctlr->r_res2, 0x0260 + (ch->unit << 7)) &
~0x00003f9f) | (ch->unit + 1));
Add first shot on support for the new Promise SATAII chips. HW donated by: pil.dk
2004-12-08 10:02:41 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if ((ctlr->chip->cfg2 == PRSATA2) ||
((ctlr->chip->cfg2 == PRCMBO2) && (ch->unit < 2))) {
Add first shot on support for the new Promise SATAII chips. HW donated by: pil.dk
2004-12-08 10:02:41 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* set PHY mode to "improved" */
ATA_OUTL(ctlr->r_res2, 0x414 + (ch->unit << 8),
(ATA_INL(ctlr->r_res2, 0x414 + (ch->unit << 8)) &
~0x00000003) | 0x00000001);
Add first shot on support for the new Promise SATAII chips. HW donated by: pil.dk
2004-12-08 10:02:41 +00:00
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ata_sata_phy_enable(ch);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* reset and enable plug/unplug intr */
ATA_OUTL(ctlr->r_res2, 0x060, (0x00000011 << ch->unit));
Add first shot on support for the new Promise SATAII chips. HW donated by: pil.dk
2004-12-08 10:02:41 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* set portmultiplier port */
ATA_OUTL(ctlr->r_res2, 0x4e8 + (ch->unit << 8), 0x00);
}
Rearrange the way the reset code is called. Prepare for different looking controllers.
2005-04-28 22:08:08 +00:00
else
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_generic_reset(dev);
Further update to the Promise code. Support should now cover all old and current products.
2005-01-01 19:24:25 +00:00
break;
Rearrange the way the reset code is called. Prepare for different looking controllers.
2005-04-28 22:08:08 +00:00
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
}
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
}
static int
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
ata_promise_mio_command(struct ata_request *request)
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
{
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
struct ata_pci_controller *ctlr=device_get_softc(GRANDPARENT(request->dev));
struct ata_channel *ch = device_get_softc(device_get_parent(request->dev));
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
u_int32_t *wordp = (u_int32_t *)ch->dma->work;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
ATA_OUTL(ctlr->r_res2, (ch->unit + 1) << 2, 0x00000001);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
/* XXX SOS add ATAPI commands support later */
switch (request->u.ata.command) {
Sync promise_mio_command with WIP on the Promise line of SATA controllers.
2004-09-03 08:23:04 +00:00
default:
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
return ata_generic_command(request);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
Sync promise_mio_command with WIP on the Promise line of SATA controllers.
2004-09-03 08:23:04 +00:00
case ATA_READ_DMA:
Fix problem on Promise "mio" controllers and 48bit mode caused by last commit
2005-08-22 11:38:53 +00:00
case ATA_READ_DMA48:
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
wordp[0] = htole32(0x04 | ((ch->unit + 1) << 16) | (0x00 << 24));
Sync promise_mio_command with WIP on the Promise line of SATA controllers.
2004-09-03 08:23:04 +00:00
break;
case ATA_WRITE_DMA:
Fix problem on Promise "mio" controllers and 48bit mode caused by last commit
2005-08-22 11:38:53 +00:00
case ATA_WRITE_DMA48:
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
wordp[0] = htole32(0x00 | ((ch->unit + 1) << 16) | (0x00 << 24));
Sync promise_mio_command with WIP on the Promise line of SATA controllers.
2004-09-03 08:23:04 +00:00
break;
}
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
wordp[1] = htole32(ch->dma->sg_bus);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
wordp[2] = 0;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
ata_promise_apkt((u_int8_t*)wordp, request);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
ATA_OUTL(ctlr->r_res2, 0x0240 + (ch->unit << 7), ch->dma->work_bus);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
return 0;
}
static int
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
ata_promise_sx4_command(struct ata_request *request)
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
{
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
device_t gparent = GRANDPARENT(request->dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(request->dev));
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_dma_prdentry *prd = ch->dma->sg;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
caddr_t window = rman_get_virtual(ctlr->r_res1);
u_int32_t *wordp;
int i, idx, length = 0;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
/* XXX SOS add ATAPI commands support later */
switch (request->u.ata.command) {
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
default:
return -1;
case ATA_ATA_IDENTIFY:
case ATA_READ:
Fix problem on Promise "mio" controllers and 48bit mode caused by last commit
2005-08-22 11:38:53 +00:00
case ATA_READ48:
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
case ATA_READ_MUL:
Fix problem on Promise "mio" controllers and 48bit mode caused by last commit
2005-08-22 11:38:53 +00:00
case ATA_READ_MUL48:
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
case ATA_WRITE:
Fix problem on Promise "mio" controllers and 48bit mode caused by last commit
2005-08-22 11:38:53 +00:00
case ATA_WRITE48:
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
case ATA_WRITE_MUL:
Fix problem on Promise "mio" controllers and 48bit mode caused by last commit
2005-08-22 11:38:53 +00:00
case ATA_WRITE_MUL48:
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
ATA_OUTL(ctlr->r_res2, 0x000c0400 + ((ch->unit + 1) << 2), 0x00000001);
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
return ata_generic_command(request);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
case ATA_SETFEATURES:
case ATA_FLUSHCACHE:
Fix problem on Promise "mio" controllers and 48bit mode caused by last commit
2005-08-22 11:38:53 +00:00
case ATA_FLUSHCACHE48:
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
case ATA_SLEEP:
case ATA_SET_MULTI:
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
wordp = (u_int32_t *)
(window + (ch->unit * ATA_PDC_CHN_OFFSET) + ATA_PDC_APKT_OFFSET);
wordp[0] = htole32(0x08 | ((ch->unit + 1)<<16) | (0x00 << 24));
wordp[1] = 0;
wordp[2] = 0;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
ata_promise_apkt((u_int8_t *)wordp, request);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
ATA_OUTL(ctlr->r_res2, 0x000c0484, 0x00000001);
ATA_OUTL(ctlr->r_res2, 0x000c0400 + ((ch->unit + 1) << 2), 0x00000001);
ATA_OUTL(ctlr->r_res2, 0x000c0240 + (ch->unit << 7),
htole32((ch->unit * ATA_PDC_CHN_OFFSET)+ATA_PDC_APKT_OFFSET));
return 0;
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
case ATA_READ_DMA:
Fix problem on Promise "mio" controllers and 48bit mode caused by last commit
2005-08-22 11:38:53 +00:00
case ATA_READ_DMA48:
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
case ATA_WRITE_DMA:
Fix problem on Promise "mio" controllers and 48bit mode caused by last commit
2005-08-22 11:38:53 +00:00
case ATA_WRITE_DMA48:
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
wordp = (u_int32_t *)
(window + (ch->unit * ATA_PDC_CHN_OFFSET) + ATA_PDC_HSG_OFFSET);
i = idx = 0;
do {
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
wordp[idx++] = prd[i].addr;
wordp[idx++] = prd[i].count;
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
length += (prd[i].count & ~ATA_DMA_EOT);
} while (!(prd[i++].count & ATA_DMA_EOT));
wordp = (u_int32_t *)
(window + (ch->unit * ATA_PDC_CHN_OFFSET) + ATA_PDC_LSG_OFFSET);
wordp[0] = htole32((ch->unit * ATA_PDC_BUF_OFFSET) + ATA_PDC_BUF_BASE);
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
wordp[1] = htole32(request->bytecount | ATA_DMA_EOT);
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
wordp = (u_int32_t *)
(window + (ch->unit * ATA_PDC_CHN_OFFSET) + ATA_PDC_ASG_OFFSET);
wordp[0] = htole32((ch->unit * ATA_PDC_BUF_OFFSET) + ATA_PDC_BUF_BASE);
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
wordp[1] = htole32(request->bytecount | ATA_DMA_EOT);
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
wordp = (u_int32_t *)
(window + (ch->unit * ATA_PDC_CHN_OFFSET) + ATA_PDC_HPKT_OFFSET);
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
if (request->flags & ATA_R_READ)
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
wordp[0] = htole32(0x14 | ((ch->unit+9)<<16) | ((ch->unit+5)<<24));
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
if (request->flags & ATA_R_WRITE)
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
wordp[0] = htole32(0x00 | ((ch->unit+13)<<16) | (0x00<<24));
wordp[1] = htole32((ch->unit * ATA_PDC_CHN_OFFSET)+ATA_PDC_HSG_OFFSET);
wordp[2] = htole32((ch->unit * ATA_PDC_CHN_OFFSET)+ATA_PDC_LSG_OFFSET);
wordp[3] = 0;
wordp = (u_int32_t *)
(window + (ch->unit * ATA_PDC_CHN_OFFSET) + ATA_PDC_APKT_OFFSET);
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
if (request->flags & ATA_R_READ)
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
wordp[0] = htole32(0x04 | ((ch->unit+5)<<16) | (0x00<<24));
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
if (request->flags & ATA_R_WRITE)
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
wordp[0] = htole32(0x10 | ((ch->unit+1)<<16) | ((ch->unit+13)<<24));
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
wordp[1] = htole32((ch->unit * ATA_PDC_CHN_OFFSET)+ATA_PDC_ASG_OFFSET);
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
wordp[2] = 0;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
ata_promise_apkt((u_int8_t *)wordp, request);
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
ATA_OUTL(ctlr->r_res2, 0x000c0484, 0x00000001);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
if (request->flags & ATA_R_READ) {
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
ATA_OUTL(ctlr->r_res2, 0x000c0400 + ((ch->unit+5)<<2), 0x00000001);
ATA_OUTL(ctlr->r_res2, 0x000c0400 + ((ch->unit+9)<<2), 0x00000001);
ATA_OUTL(ctlr->r_res2, 0x000c0240 + (ch->unit << 7),
htole32((ch->unit * ATA_PDC_CHN_OFFSET) + ATA_PDC_APKT_OFFSET));
}
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
if (request->flags & ATA_R_WRITE) {
Fix Promise sx4/sx4000 support that broke on the latest race fixes.
2004-08-16 09:29:46 +00:00
ATA_OUTL(ctlr->r_res2, 0x000c0400 + ((ch->unit+1)<<2), 0x00000001);
ATA_OUTL(ctlr->r_res2, 0x000c0400 + ((ch->unit+13)<<2), 0x00000001);
ata_promise_queue_hpkt(ctlr,
htole32((ch->unit * ATA_PDC_CHN_OFFSET) + ATA_PDC_HPKT_OFFSET));
}
return 0;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
}
}
static int
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
ata_promise_apkt(u_int8_t *bytep, struct ata_request *request)
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
{
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
struct ata_device *atadev = device_get_softc(request->dev);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
int i = 12;
bytep[i++] = ATA_PDC_1B | ATA_PDC_WRITE_REG | ATA_PDC_WAIT_NBUSY|ATA_DRIVE;
bytep[i++] = ATA_D_IBM | ATA_D_LBA | atadev->unit;
bytep[i++] = ATA_PDC_1B | ATA_PDC_WRITE_CTL;
bytep[i++] = ATA_A_4BIT;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
if (atadev->flags & ATA_D_48BIT_ACTIVE) {
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
bytep[i++] = ATA_PDC_2B | ATA_PDC_WRITE_REG | ATA_FEATURE;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
bytep[i++] = request->u.ata.feature >> 8;
bytep[i++] = request->u.ata.feature;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
bytep[i++] = ATA_PDC_2B | ATA_PDC_WRITE_REG | ATA_COUNT;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
bytep[i++] = request->u.ata.count >> 8;
bytep[i++] = request->u.ata.count;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
bytep[i++] = ATA_PDC_2B | ATA_PDC_WRITE_REG | ATA_SECTOR;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
bytep[i++] = request->u.ata.lba >> 24;
bytep[i++] = request->u.ata.lba;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
bytep[i++] = ATA_PDC_2B | ATA_PDC_WRITE_REG | ATA_CYL_LSB;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
bytep[i++] = request->u.ata.lba >> 32;
bytep[i++] = request->u.ata.lba >> 8;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
bytep[i++] = ATA_PDC_2B | ATA_PDC_WRITE_REG | ATA_CYL_MSB;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
bytep[i++] = request->u.ata.lba >> 40;
bytep[i++] = request->u.ata.lba >> 16;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
bytep[i++] = ATA_PDC_1B | ATA_PDC_WRITE_REG | ATA_DRIVE;
bytep[i++] = ATA_D_LBA | atadev->unit;
}
else {
bytep[i++] = ATA_PDC_1B | ATA_PDC_WRITE_REG | ATA_FEATURE;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
bytep[i++] = request->u.ata.feature;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
bytep[i++] = ATA_PDC_1B | ATA_PDC_WRITE_REG | ATA_COUNT;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
bytep[i++] = request->u.ata.count;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
bytep[i++] = ATA_PDC_1B | ATA_PDC_WRITE_REG | ATA_SECTOR;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
bytep[i++] = request->u.ata.lba;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
bytep[i++] = ATA_PDC_1B | ATA_PDC_WRITE_REG | ATA_CYL_LSB;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
bytep[i++] = request->u.ata.lba >> 8;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
bytep[i++] = ATA_PDC_1B | ATA_PDC_WRITE_REG | ATA_CYL_MSB;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
bytep[i++] = request->u.ata.lba >> 16;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
bytep[i++] = ATA_PDC_1B | ATA_PDC_WRITE_REG | ATA_DRIVE;
bytep[i++] = (atadev->flags & ATA_D_USE_CHS ? 0 : ATA_D_LBA) |
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
ATA_D_IBM | atadev->unit | ((request->u.ata.lba >> 24)&0xf);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
}
Add support for controllers that doesn't have the usual taskfile layout. No functional changes.
2005-04-06 10:22:56 +00:00
bytep[i++] = ATA_PDC_1B | ATA_PDC_WRITE_END | ATA_COMMAND;
Add support for working around controllers that cannot do DMA in 48bit mode. The workaround use PIO mode above ~137GB to allow using the disk. Add the Acer chips with rev < 0xc4 as first candidate.
2005-08-17 15:00:33 +00:00
bytep[i++] = request->u.ata.command;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
return i;
}
static void
ata_promise_queue_hpkt(struct ata_pci_controller *ctlr, u_int32_t hpkt)
{
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_promise_sx4 *hpktp = device_get_ivars(ctlr->dev);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
mtx_lock(&hpktp->mtx);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (hpktp->busy) {
struct host_packet *hp =
malloc(sizeof(struct host_packet), M_TEMP, M_NOWAIT | M_ZERO);
hp->addr = hpkt;
TAILQ_INSERT_TAIL(&hpktp->queue, hp, chain);
}
else {
hpktp->busy = 1;
ATA_OUTL(ctlr->r_res2, 0x000c0100, hpkt);
}
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
mtx_unlock(&hpktp->mtx);
}
static void
ata_promise_next_hpkt(struct ata_pci_controller *ctlr)
{
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_promise_sx4 *hpktp = device_get_ivars(ctlr->dev);
struct host_packet *hp;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
mtx_lock(&hpktp->mtx);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if ((hp = TAILQ_FIRST(&hpktp->queue))) {
TAILQ_REMOVE(&hpktp->queue, hp, chain);
ATA_OUTL(ctlr->r_res2, 0x000c0100, hp->addr);
free(hp, M_TEMP);
}
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
else
hpktp->busy = 0;
mtx_unlock(&hpktp->mtx);
}
static void
ata_promise_tx2_intr(void *data)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
struct ata_pci_controller *ctlr = data;
Third round of updates to the ATA driver. More DMA cleanups, including fix for breakage on older Promise controllers. Add more ways of getting to the ATA registers.
2003-04-07 14:12:12 +00:00
struct ata_channel *ch;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int unit;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
for (unit = 0; unit < ctlr->channels; unit++) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!(ch = ctlr->interrupt[unit].argument))
continue;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
ATA_IDX_OUTB(ch, ATA_BMDEVSPEC_0, 0x0b);
if (ATA_IDX_INB(ch, ATA_BMDEVSPEC_1) & 0x20) {
Fix for those lost interrupts on probe on lots of controller types. Note to self, just because an idea is good, it doesn't apply everywhere.
2004-01-28 21:54:40 +00:00
if (ch->dma && (ch->dma->flags & ATA_DMA_ACTIVE)) {
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
int bmstat = ATA_IDX_INB(ch, ATA_BMSTAT_PORT) & ATA_BMSTAT_MASK;
if ((bmstat & (ATA_BMSTAT_ACTIVE | ATA_BMSTAT_INTERRUPT)) !=
ATA_BMSTAT_INTERRUPT)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
continue;
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, bmstat & ~ATA_BMSTAT_ERROR);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
DELAY(1);
}
ctlr->interrupt[unit].function(ch);
}
}
}
static void
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
ata_promise_old_intr(void *data)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
int unit;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
for (unit = 0; unit < ctlr->channels; unit++) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!(ch = ctlr->interrupt[unit].argument))
continue;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
if (ATA_INL(ctlr->r_res1, 0x1c) & (ch->unit ? 0x00004000 : 0x00000400)){
Fix for those lost interrupts on probe on lots of controller types. Note to self, just because an idea is good, it doesn't apply everywhere.
2004-01-28 21:54:40 +00:00
if (ch->dma && (ch->dma->flags & ATA_DMA_ACTIVE)) {
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
int bmstat = ATA_IDX_INB(ch, ATA_BMSTAT_PORT) & ATA_BMSTAT_MASK;
if ((bmstat & (ATA_BMSTAT_ACTIVE | ATA_BMSTAT_INTERRUPT)) !=
ATA_BMSTAT_INTERRUPT)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
continue;
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, bmstat & ~ATA_BMSTAT_ERROR);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
DELAY(1);
}
ctlr->interrupt[unit].function(ch);
}
}
}
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
static int
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_promise_new_dmastart(device_t dev)
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
{
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
struct ata_pci_controller *ctlr = device_get_softc(GRANDPARENT(dev));
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
if (atadev->flags & ATA_D_48BIT_ACTIVE) {
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
ATA_OUTB(ctlr->r_res1, 0x11,
ATA_INB(ctlr->r_res1, 0x11) | (ch->unit ? 0x08 : 0x02));
Fix the register offset for the ATAPI count register thats used to wrestle the older Promise chips to do 48bit addressing. Spotted by: Martin Birgmeier
2005-06-09 19:00:37 +00:00
ATA_OUTL(ctlr->r_res1, ch->unit ? 0x24 : 0x20,
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
((ch->dma->flags & ATA_DMA_READ) ? 0x05000000 : 0x06000000) |
(ch->dma->cur_iosize >> 1));
}
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, (ATA_IDX_INB(ch, ATA_BMSTAT_PORT) |
(ATA_BMSTAT_INTERRUPT | ATA_BMSTAT_ERROR)));
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
ATA_IDX_OUTL(ch, ATA_BMDTP_PORT, ch->dma->sg_bus);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
ATA_IDX_OUTB(ch, ATA_BMCMD_PORT,
((ch->dma->flags & ATA_DMA_READ) ? ATA_BMCMD_WRITE_READ : 0) |
ATA_BMCMD_START_STOP);
Try to narrow down the race window on HW that does not have ways to poll for which channel actually pulled the irq line.
2004-08-05 21:13:41 +00:00
ch->flags |= ATA_DMA_ACTIVE;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
return 0;
}
static int
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_promise_new_dmastop(device_t dev)
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
{
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
struct ata_pci_controller *ctlr = device_get_softc(GRANDPARENT(dev));
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
int error;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
if (atadev->flags & ATA_D_48BIT_ACTIVE) {
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
ATA_OUTB(ctlr->r_res1, 0x11,
ATA_INB(ctlr->r_res1, 0x11) & ~(ch->unit ? 0x08 : 0x02));
Fix the register offset for the ATAPI count register thats used to wrestle the older Promise chips to do 48bit addressing. Spotted by: Martin Birgmeier
2005-06-09 19:00:37 +00:00
ATA_OUTL(ctlr->r_res1, ch->unit ? 0x24 : 0x20, 0);
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
}
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
error = ATA_IDX_INB(ch, ATA_BMSTAT_PORT);
ATA_IDX_OUTB(ch, ATA_BMCMD_PORT,
ATA_IDX_INB(ch, ATA_BMCMD_PORT) & ~ATA_BMCMD_START_STOP);
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, ATA_BMSTAT_INTERRUPT | ATA_BMSTAT_ERROR);
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
ch->flags &= ~ATA_DMA_ACTIVE;
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
return error;
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
}
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
static void
ata_promise_new_dmareset(device_t dev)
{
struct ata_channel *ch = device_get_softc(dev);
ATA_IDX_OUTB(ch, ATA_BMCMD_PORT,
ATA_IDX_INB(ch, ATA_BMCMD_PORT) & ~ATA_BMCMD_START_STOP);
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, ATA_BMSTAT_INTERRUPT | ATA_BMSTAT_ERROR);
ch->flags &= ~ATA_DMA_ACTIVE;
}
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_promise_new_dmainit(device_t dev)
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(dev);
ata_dmainit(dev);
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
if (ch->dma) {
ch->dma->start = ata_promise_new_dmastart;
ch->dma->stop = ata_promise_new_dmastop;
Reshape the dma code to be a bit more flexible so it can cope with new HW that has new and different demands. Fix a few nits in former commit in this cleanup crusade. Sponsored by: pair.com
2005-05-03 07:55:07 +00:00
ch->dma->reset = ata_promise_new_dmareset;
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
}
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_promise_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int error;
u_int32_t timings33[][2] = {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* PROLD PRNEW mode */
{ 0x004ff329, 0x004fff2f }, /* PIO 0 */
{ 0x004fec25, 0x004ff82a }, /* PIO 1 */
{ 0x004fe823, 0x004ff026 }, /* PIO 2 */
{ 0x004fe622, 0x004fec24 }, /* PIO 3 */
{ 0x004fe421, 0x004fe822 }, /* PIO 4 */
{ 0x004567f3, 0x004acef6 }, /* MWDMA 0 */
{ 0x004467f3, 0x0048cef6 }, /* MWDMA 1 */
{ 0x004367f3, 0x0046cef6 }, /* MWDMA 2 */
{ 0x004367f3, 0x0046cef6 }, /* UDMA 0 */
{ 0x004247f3, 0x00448ef6 }, /* UDMA 1 */
{ 0x004127f3, 0x00436ef6 }, /* UDMA 2 */
{ 0, 0x00424ef6 }, /* UDMA 3 */
{ 0, 0x004127f3 }, /* UDMA 4 */
{ 0, 0x004127f3 } /* UDMA 5 */
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
};
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ctlr->chip->max_dma);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Second round of updates to the ATA driver. Clean up the DMA interface too much unneeded stuff crept in with the busdma code back when. Modify the ATA_IN* / ATA_OUT* macros so that resource and offset are gotten from a table. That allows for new chipsets that doesn't nessesarily have things ordered the good old way. This also removes the need for the wierd PC98 resource functions. Tested on: i386, PC98, Alpha, Sparc64
2003-03-29 13:37:09 +00:00
switch (ctlr->chip->cfg1) {
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
case PROLD:
case PRNEW:
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (mode > ATA_UDMA2 && (pci_read_config(gparent, 0x50, 2) &
(ch->unit ? 1 << 11 : 1 << 10))) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_print_cable(dev, "controller");
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
mode = ATA_UDMA2;
}
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
if (ata_atapi(dev) && mode > ATA_PIO_MAX)
mode = ata_limit_mode(dev, mode, ATA_PIO_MAX);
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
break;
Second round of updates to the ATA driver. Clean up the DMA interface too much unneeded stuff crept in with the busdma code back when. Modify the ATA_IN* / ATA_OUT* macros so that resource and offset are gotten from a table. That allows for new chipsets that doesn't nessesarily have things ordered the good old way. This also removes the need for the wierd PC98 resource functions. Tested on: i386, PC98, Alpha, Sparc64
2003-03-29 13:37:09 +00:00
case PRTX:
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
ATA_IDX_OUTB(ch, ATA_BMDEVSPEC_0, 0x0b);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (mode > ATA_UDMA2 &&
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
ATA_IDX_INB(ch, ATA_BMDEVSPEC_1) & 0x04) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_print_cable(dev, "controller");
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
mode = ATA_UDMA2;
}
Second round of updates to the ATA driver. Clean up the DMA interface too much unneeded stuff crept in with the busdma code back when. Modify the ATA_IN* / ATA_OUT* macros so that resource and offset are gotten from a table. That allows for new chipsets that doesn't nessesarily have things ordered the good old way. This also removes the need for the wierd PC98 resource functions. Tested on: i386, PC98, Alpha, Sparc64
2003-03-29 13:37:09 +00:00
break;
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
case PRMIO:
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (mode > ATA_UDMA2 &&
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
(ATA_INL(ctlr->r_res2,
(ctlr->chip->cfg2 & PRSX4X ? 0x000c0260 : 0x0260) +
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(ch->unit << 7)) & 0x01000000)) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_print_cable(dev, "controller");
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
mode = ATA_UDMA2;
}
I'm pleased to announce that Promise is now supporting the FreeBSD project by providing documentation (under NDA) and hardware for testing. This commit is the first result of the cooperation, and adds support for several of their new controllers that we didn't support before (and probably newer would have without this arrangement). Add support for the Promise SATA150 TX2/TX4 and the Promise TX4000 controllers. This also adds support for various motherboard fitted Promise SATA/ATA chips. Note that this code uses memory mapped registers to minimize overhead. I belive FreeBSD has made another first in the Open Source world by being able to release support for this :)
2003-05-01 06:20:50 +00:00
break;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting %s on %s chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "FAILURE " : "",
ata_mode2str(mode), ctlr->chip->text);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!error) {
Minor cleanup of the Promise code.
2003-03-03 11:51:08 +00:00
if (ctlr->chip->cfg1 < PRTX)
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x60 + (devno << 2),
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
timings33[ctlr->chip->cfg1][ata_mode2idx(mode)],4);
atadev->mode = mode;
}
return;
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*
* ServerWorks chipset support functions
*/
int
ata_serverworks_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_ROSB4, 0x00, SWKS33, 0x00, ATA_UDMA2, "ROSB4" },
{ ATA_CSB5, 0x92, SWKS100, 0x00, ATA_UDMA5, "CSB5" },
{ ATA_CSB5, 0x00, SWKS66, 0x00, ATA_UDMA4, "CSB5" },
{ ATA_CSB6, 0x00, SWKS100, 0x00, ATA_UDMA5, "CSB6" },
{ ATA_CSB6_1, 0x00, SWKS66, 0x00, ATA_UDMA4, "CSB6" },
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{ 0, 0, 0, 0, 0, 0}};
char buffer[64];
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
if (!(idx = ata_match_chip(dev, ids)))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
sprintf(buffer, "ServerWorks %s %s controller",
idx->text, ata_mode2str(idx->max_dma));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_serverworks_chipinit;
return 0;
}
static int
ata_serverworks_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (ata_setup_interrupt(dev))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (ctlr->chip->cfg1 == SWKS33) {
device_t *children;
int nchildren, i;
/* locate the ISA part in the southbridge and enable UDMA33 */
if (!device_get_children(device_get_parent(dev), &children,&nchildren)){
for (i = 0; i < nchildren; i++) {
if (pci_get_devid(children[i]) == ATA_ROSB4_ISA) {
pci_write_config(children[i], 0x64,
(pci_read_config(children[i], 0x64, 4) &
~0x00002000) | 0x00004000, 4);
break;
}
}
free(children, M_TEMP);
}
}
else {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
pci_write_config(dev, 0x5a,
(pci_read_config(dev, 0x5a, 1) & ~0x40) |
(ctlr->chip->cfg1 == SWKS100) ? 0x03 : 0x02, 1);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
ctlr->setmode = ata_serverworks_setmode;
return 0;
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_serverworks_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
Fix the serverworks modesetting code, of mask offset was wrong.
2004-10-01 09:04:53 +00:00
int offset = (devno ^ 0x01) << 3;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int error;
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
u_int8_t piotimings[] = { 0x5d, 0x47, 0x34, 0x22, 0x20, 0x34, 0x22, 0x20,
0x20, 0x20, 0x20, 0x20, 0x20, 0x20 };
u_int8_t dmatimings[] = { 0x77, 0x21, 0x20 };
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ctlr->chip->max_dma);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_check_80pin(dev, mode);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting %s on %s chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "FAILURE " : "",
ata_mode2str(mode), ctlr->chip->text);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!error) {
if (mode >= ATA_UDMA0) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x56,
(pci_read_config(gparent, 0x56, 2) &
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
~(0xf << (devno << 2))) |
((mode & ATA_MODE_MASK) << (devno << 2)), 2);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x54,
pci_read_config(gparent, 0x54, 1) |
Fix the serverworks modesetting code, of mask offset was wrong.
2004-10-01 09:04:53 +00:00
(0x01 << devno), 1);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x44,
(pci_read_config(gparent, 0x44, 4) &
Fix the serverworks modesetting code, of mask offset was wrong.
2004-10-01 09:04:53 +00:00
~(0xff << offset)) |
(dmatimings[2] << offset), 4);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
}
else if (mode >= ATA_WDMA0) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x54,
pci_read_config(gparent, 0x54, 1) &
Fix the serverworks modesetting code, of mask offset was wrong.
2004-10-01 09:04:53 +00:00
~(0x01 << devno), 1);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x44,
(pci_read_config(gparent, 0x44, 4) &
Fix the serverworks modesetting code, of mask offset was wrong.
2004-10-01 09:04:53 +00:00
~(0xff << offset)) |
Add support for the ATI IXP[234]00 series chipsets. HW donated by: sentex
2005-10-12 20:00:26 +00:00
(dmatimings[mode & ATA_MODE_MASK] << offset), 4);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
else
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x54,
pci_read_config(gparent, 0x54, 1) &
Fix the serverworks modesetting code, of mask offset was wrong.
2004-10-01 09:04:53 +00:00
~(0x01 << devno), 1);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, 0x40,
(pci_read_config(gparent, 0x40, 4) &
Fix the serverworks modesetting code, of mask offset was wrong.
2004-10-01 09:04:53 +00:00
~(0xff << offset)) |
(piotimings[ata_mode2idx(mode)] << offset), 4);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
atadev->mode = mode;
}
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
* Silicon Image Inc. (SiI) (former CMD) chipset support functions
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
*/
int
ata_sii_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
{{ ATA_SII3114, 0x00, SIIMEMIO, SII4CH, ATA_SA150, "SiI 3114" },
{ ATA_SII3512, 0x02, SIIMEMIO, 0, ATA_SA150, "SiI 3512" },
{ ATA_SII3112, 0x02, SIIMEMIO, 0, ATA_SA150, "SiI 3112" },
{ ATA_SII3112_1, 0x02, SIIMEMIO, 0, ATA_SA150, "SiI 3112" },
{ ATA_SII3512, 0x00, SIIMEMIO, SIIBUG, ATA_SA150, "SiI 3512" },
{ ATA_SII3112, 0x00, SIIMEMIO, SIIBUG, ATA_SA150, "SiI 3112" },
{ ATA_SII3112_1, 0x00, SIIMEMIO, SIIBUG, ATA_SA150, "SiI 3112" },
Add another ID for the SiI3112a SATA chip as used on the Adaptec 1210SA. RAID support is still in the works, so for now just normal ATA ops. Sponsored by: Matt Douhan(www.fruitsalad.org)
2003-09-08 13:55:05 +00:00
{ ATA_SII0680, 0x00, SIIMEMIO, SIISETCLK, ATA_UDMA6, "SiI 0680" },
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
{ ATA_CMD649, 0x00, 0, SIIINTR, ATA_UDMA5, "CMD 649" },
{ ATA_CMD648, 0x00, 0, SIIINTR, ATA_UDMA4, "CMD 648" },
{ ATA_CMD646, 0x07, 0, 0, ATA_UDMA2, "CMD 646U2" },
{ ATA_CMD646, 0x00, 0, 0, ATA_WDMA2, "CMD 646" },
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{ 0, 0, 0, 0, 0, 0}};
char buffer[64];
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
if (!(idx = ata_match_chip(dev, ids)))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
sprintf(buffer, "%s %s controller", idx->text, ata_mode2str(idx->max_dma));
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_sii_chipinit;
return 0;
}
static int
ata_sii_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
int rid = ATA_IRQ_RID;
Convert callers to the new bus_alloc_resource_any(9) API. Submitted by: Mark Santcroos <marks@ripe.net> Reviewed by: imp, dfr, bde
2004-03-17 17:50:55 +00:00
if (!(ctlr->r_irq = bus_alloc_resource_any(dev, SYS_RES_IRQ, &rid,
RF_SHAREABLE | RF_ACTIVE))) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_printf(dev, "unable to map interrupt\n");
return ENXIO;
}
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
if (ctlr->chip->cfg1 == SIIMEMIO) {
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if ((bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
ata_sii_intr, ctlr, &ctlr->handle))) {
device_printf(dev, "unable to setup interrupt\n");
return ENXIO;
}
Overhaul of the timeout/reinit framework. This should clear up most of the leftovers from the old version that really doesn't work anymore. Add a reset function for host-end of the ATA channel. This is needed for the SiI3112 in order to whack it back to reality if a device locks up the SATA interface (thereby preventing that we can reset the device). The result is that ATA now recovers from the timeouts that happens with the SiI3112A and more or less all disks based on old PATA electronics with a Marvell PATA->SATA converter. This includes lots of the popular SATA dongles and the WDC Raptor disks..
2004-01-11 22:08:34 +00:00
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
ctlr->r_type2 = SYS_RES_MEMORY;
Try to narrow down the race window on HW that does not have ways to poll for which channel actually pulled the irq line.
2004-08-05 21:13:41 +00:00
ctlr->r_rid2 = PCIR_BAR(5);
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
if (!(ctlr->r_res2 = bus_alloc_resource_any(dev, ctlr->r_type2,
&ctlr->r_rid2, RF_ACTIVE)))
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
return ENXIO;
if (ctlr->chip->cfg2 & SIISETCLK) {
if ((pci_read_config(dev, 0x8a, 1) & 0x30) != 0x10)
pci_write_config(dev, 0x8a,
Add another ID for the SiI3112a SATA chip as used on the Adaptec 1210SA. RAID support is still in the works, so for now just normal ATA ops. Sponsored by: Matt Douhan(www.fruitsalad.org)
2003-09-08 13:55:05 +00:00
(pci_read_config(dev, 0x8a, 1) & 0xcf)|0x10,1);
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
if ((pci_read_config(dev, 0x8a, 1) & 0x30) != 0x10)
device_printf(dev, "%s could not set ATA133 clock\n",
ctlr->chip->text);
}
Add another ID for the SiI3112a SATA chip as used on the Adaptec 1210SA. RAID support is still in the works, so for now just normal ATA ops. Sponsored by: Matt Douhan(www.fruitsalad.org)
2003-09-08 13:55:05 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* if we have 4 channels enable the second set */
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
if (ctlr->chip->cfg2 & SII4CH) {
ATA_OUTL(ctlr->r_res2, 0x0200, 0x00000002);
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
ctlr->channels = 4;
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
}
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* enable PCI interrupt as BIOS might not */
pci_write_config(dev, 0x8a, (pci_read_config(dev, 0x8a, 1) & 0x3f), 1);
/* dont block interrupts from any channel */
pci_write_config(dev, 0x48,
(pci_read_config(dev, 0x48, 4) & ~0x03c00000), 4);
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
ctlr->allocate = ata_sii_allocate;
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
if (ctlr->chip->max_dma >= ATA_SA150) {
ctlr->reset = ata_sii_reset;
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
ctlr->setmode = ata_sata_setmode;
Increase robustness of SATA handling.
2004-06-15 11:02:09 +00:00
}
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
else
ctlr->setmode = ata_sii_setmode;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
else {
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if ((bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
ctlr->chip->cfg2 & SIIINTR ?
Fix support for the cmd646 chip. Spotted by: tmm reviewed by: re@
2003-11-21 22:58:56 +00:00
ata_cmd_intr : ata_cmd_old_intr,
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
ctlr, &ctlr->handle))) {
device_printf(dev, "unable to setup interrupt\n");
return ENXIO;
}
if ((pci_read_config(dev, 0x51, 1) & 0x08) != 0x08) {
device_printf(dev, "HW has secondary channel disabled\n");
ctlr->channels = 1;
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
/* enable interrupt as BIOS might not */
Fix support for the old CMD 646 chips, interrupt setup was wrong.
2003-02-25 12:56:37 +00:00
pci_write_config(dev, 0x71, 0x01, 1);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
ctlr->setmode = ata_cmd_setmode;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
return 0;
}
Prevent probing the secondary channel on CMD646 controllers if the channel has been disabled by BIOS. This prevents a bus timeout machine check on B&W G3 PowerMacs, which have a primary-only CMD646 on the motherboard. Approved by: sos Obtained from: NetBSD
2003-06-27 10:11:54 +00:00
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
static int
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
ata_sii_allocate(device_t dev)
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
Add firstshot support for Acer Labs Inc SATA parts Update ALi PATA support to handle ATA133 as well
2005-04-05 14:51:43 +00:00
struct ata_channel *ch = device_get_softc(dev);
Add support for the Silicon Image 3114 4 channel SATA controller.
2004-01-17 23:34:13 +00:00
int unit01 = (ch->unit & 1), unit10 = (ch->unit & 2);
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
int i;
Prevent probing the secondary channel on CMD646 controllers if the channel has been disabled by BIOS. This prevents a bus timeout machine check on B&W G3 PowerMacs, which have a primary-only CMD646 on the motherboard. Approved by: sos Obtained from: NetBSD
2003-06-27 10:11:54 +00:00
Add support for controllers that doesn't have the usual taskfile layout. No functional changes.
2005-04-06 10:22:56 +00:00
for (i = ATA_DATA; i <= ATA_COMMAND; i++) {
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
ch->r_io[i].res = ctlr->r_res2;
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
ch->r_io[i].offset = 0x80 + i + (unit01 << 6) + (unit10 << 8);
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
}
Add support for controllers that doesn't have the usual taskfile layout. No functional changes.
2005-04-06 10:22:56 +00:00
ch->r_io[ATA_CONTROL].res = ctlr->r_res2;
ch->r_io[ATA_CONTROL].offset = 0x8a + (unit01 << 6) + (unit10 << 8);
ch->r_io[ATA_IDX_ADDR].res = ctlr->r_res2;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_default_registers(dev);
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
ch->r_io[ATA_BMCMD_PORT].res = ctlr->r_res2;
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
ch->r_io[ATA_BMCMD_PORT].offset = 0x00 + (unit01 << 3) + (unit10 << 8);
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
ch->r_io[ATA_BMSTAT_PORT].res = ctlr->r_res2;
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
ch->r_io[ATA_BMSTAT_PORT].offset = 0x02 + (unit01 << 3) + (unit10 << 8);
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
ch->r_io[ATA_BMDTP_PORT].res = ctlr->r_res2;
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
ch->r_io[ATA_BMDTP_PORT].offset = 0x04 + (unit01 << 3) + (unit10 << 8);
Add support for detaching PCI controllers. This adds support for cardbus ATA/SATA controllers. I get roughly the same transfer speeds as on true PCI controllers. Nice to be able to add a couble of "real" disks to a laptop :)
2004-03-15 12:03:48 +00:00
ch->r_io[ATA_BMDEVSPEC_0].res = ctlr->r_res2;
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
ch->r_io[ATA_BMDEVSPEC_0].offset = 0xa1 + (unit01 << 6) + (unit10 << 8);
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (ctlr->chip->max_dma >= ATA_SA150) {
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
ch->r_io[ATA_SSTATUS].res = ctlr->r_res2;
ch->r_io[ATA_SSTATUS].offset = 0x104 + (unit01 << 7) + (unit10 << 8);
ch->r_io[ATA_SERROR].res = ctlr->r_res2;
ch->r_io[ATA_SERROR].offset = 0x108 + (unit01 << 7) + (unit10 << 8);
ch->r_io[ATA_SCONTROL].res = ctlr->r_res2;
ch->r_io[ATA_SCONTROL].offset = 0x100 + (unit01 << 7) + (unit10 << 8);
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
ch->flags |= ATA_NO_SLAVE;
Overhaul of the timeout/reinit framework. This should clear up most of the leftovers from the old version that really doesn't work anymore. Add a reset function for host-end of the ATA channel. This is needed for the SiI3112 in order to whack it back to reality if a device locks up the SATA interface (thereby preventing that we can reset the device). The result is that ATA now recovers from the timeouts that happens with the SiI3112A and more or less all disks based on old PATA electronics with a Marvell PATA->SATA converter. This includes lots of the popular SATA dongles and the WDC Raptor disks..
2004-01-11 22:08:34 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* enable PHY state change interrupt */
ATA_OUTL(ctlr->r_res2, 0x148 + (unit01 << 7) + (unit10 << 8),(1 << 16));
}
Reduce the amount of memory reported to busdma. This made the requirements for bouncebuffers too big with PAE. Cleanup the way size defines for transfers are implemented.
2004-09-10 10:31:37 +00:00
if ((ctlr->chip->cfg2 & SIIBUG) && ch->dma) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* work around errata in early chips */
Reduce the amount of memory reported to busdma. This made the requirements for bouncebuffers too big with PAE. Cleanup the way size defines for transfers are implemented.
2004-09-10 10:31:37 +00:00
ch->dma->boundary = 16 * DEV_BSIZE;
Add support for setting the SG list segment size. Use this for the SiI3112 workaround to get rid of the "oversized DMA" errors. MFC to 6.0 candidate.
2005-10-06 15:44:07 +00:00
ch->dma->segsize = 15 * DEV_BSIZE;
Reduce the amount of memory reported to busdma. This made the requirements for bouncebuffers too big with PAE. Cleanup the way size defines for transfers are implemented.
2004-09-10 10:31:37 +00:00
}
Add support for the Promise command sequencer present on all modern Promise controllers (PDC203** PDC206**). This also adds preliminary support for the Promise SX4/SX4000 but *only* as a "normal" Promise ATA controller (ATA RAID's are supported though but only RAID0, RAID1 and RAID0+1). This cuts off yet another 5-8% of the command overhead on promise controllers, making them the fastest we have ever had support for. Work is now continuing to add support for this in ATA RAID, to accellerate ATA RAID quite a bit on these controllers, and especially the SX4/SX4000 series as they have quite a few tricks in there.. This commit also adds a few fixes to the SATA code needed for proper support.
2004-04-13 09:44:20 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_generic_hw(dev);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return 0;
}
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
static void
ata_sii_intr(void *data)
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
int unit;
Fix Sii3114 support.
2004-05-10 20:23:25 +00:00
for (unit = 0; unit < ctlr->channels; unit++) {
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
if (!(ch = ctlr->interrupt[unit].argument))
continue;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* check for PHY related interrupts on SATA capable HW */
if (ctlr->chip->max_dma >= ATA_SA150) {
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
u_int32_t status = ATA_IDX_INL(ch, ATA_SSTATUS);
u_int32_t error = ATA_IDX_INL(ch, ATA_SERROR);
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
struct ata_connect_task *tp;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (error) {
/* clear error bits/interrupt */
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
ATA_IDX_OUTL(ch, ATA_SERROR, error);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
/* if we have a connection event deal with it */
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
if ((error & ATA_SE_PHY_CHANGED) &&
(tp = (struct ata_connect_task *)
malloc(sizeof(struct ata_connect_task),
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
M_ATA, M_NOWAIT | M_ZERO))) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
if ((status & ATA_SS_CONWELL_MASK) == ATA_SS_CONWELL_GEN1) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
device_printf(ch->dev, "CONNECT requested\n");
tp->action = ATA_C_ATTACH;
}
else {
device_printf(ch->dev, "DISCONNECT requested\n");
tp->action = ATA_C_DETACH;
}
tp->dev = ch->dev;
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
TASK_INIT(&tp->task, 0, ata_sata_phy_event, tp);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
taskqueue_enqueue(taskqueue_thread, &tp->task);
}
}
}
/* any drive action to take care of ? */
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
if (ATA_IDX_INB(ch, ATA_BMDEVSPEC_0) & 0x08) {
Fix for those lost interrupts on probe on lots of controller types. Note to self, just because an idea is good, it doesn't apply everywhere.
2004-01-28 21:54:40 +00:00
if (ch->dma && (ch->dma->flags & ATA_DMA_ACTIVE)) {
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
int bmstat = ATA_IDX_INB(ch, ATA_BMSTAT_PORT) & ATA_BMSTAT_MASK;
Overhaul of the timeout/reinit framework. This should clear up most of the leftovers from the old version that really doesn't work anymore. Add a reset function for host-end of the ATA channel. This is needed for the SiI3112 in order to whack it back to reality if a device locks up the SATA interface (thereby preventing that we can reset the device). The result is that ATA now recovers from the timeouts that happens with the SiI3112A and more or less all disks based on old PATA electronics with a Marvell PATA->SATA converter. This includes lots of the popular SATA dongles and the WDC Raptor disks..
2004-01-11 22:08:34 +00:00
if (!(bmstat & ATA_BMSTAT_INTERRUPT))
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
continue;
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, bmstat & ~ATA_BMSTAT_ERROR);
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
DELAY(1);
}
ctlr->interrupt[unit].function(ch);
}
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
}
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
static void
ata_cmd_intr(void *data)
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
u_int8_t reg71;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int unit;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
for (unit = 0; unit < ctlr->channels; unit++) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!(ch = ctlr->interrupt[unit].argument))
continue;
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
if (((reg71 = pci_read_config(device_get_parent(ch->dev), 0x71, 1)) &
(ch->unit ? 0x08 : 0x04))) {
pci_write_config(device_get_parent(ch->dev), 0x71,
reg71 & ~(ch->unit ? 0x04 : 0x08), 1);
Fix for those lost interrupts on probe on lots of controller types. Note to self, just because an idea is good, it doesn't apply everywhere.
2004-01-28 21:54:40 +00:00
if (ch->dma && (ch->dma->flags & ATA_DMA_ACTIVE)) {
Cleanup the interrupt code that deals with the busmaster bits.
2003-10-28 21:08:14 +00:00
int bmstat = ATA_IDX_INB(ch, ATA_BMSTAT_PORT) & ATA_BMSTAT_MASK;
if ((bmstat & (ATA_BMSTAT_ACTIVE | ATA_BMSTAT_INTERRUPT)) !=
ATA_BMSTAT_INTERRUPT)
continue;
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, bmstat & ~ATA_BMSTAT_ERROR);
DELAY(1);
}
ctlr->interrupt[unit].function(ch);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
}
}
Fix support for the cmd646 chip. Spotted by: tmm reviewed by: re@
2003-11-21 22:58:56 +00:00
static void
ata_cmd_old_intr(void *data)
{
struct ata_pci_controller *ctlr = data;
struct ata_channel *ch;
int unit;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
for (unit = 0; unit < ctlr->channels; unit++) {
Fix support for the cmd646 chip. Spotted by: tmm reviewed by: re@
2003-11-21 22:58:56 +00:00
if (!(ch = ctlr->interrupt[unit].argument))
continue;
if (ch->dma && (ch->dma->flags & ATA_DMA_ACTIVE)) {
int bmstat = ATA_IDX_INB(ch, ATA_BMSTAT_PORT) & ATA_BMSTAT_MASK;
if ((bmstat & (ATA_BMSTAT_ACTIVE | ATA_BMSTAT_INTERRUPT)) !=
ATA_BMSTAT_INTERRUPT)
continue;
ATA_IDX_OUTB(ch, ATA_BMSTAT_PORT, bmstat & ~ATA_BMSTAT_ERROR);
DELAY(1);
}
ctlr->interrupt[unit].function(ch);
}
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_sii_reset(device_t dev)
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
int offset = ((ch->unit & 1) << 7) + ((ch->unit & 2) << 8);
/* disable PHY state change interrupt */
ATA_OUTL(ctlr->r_res2, 0x148 + offset, ~(1 << 16));
Add support for setting the SG list segment size. Use this for the SiI3112 workaround to get rid of the "oversized DMA" errors. MFC to 6.0 candidate.
2005-10-06 15:44:07 +00:00
/* reset controller part for this channel */
ATA_OUTL(ctlr->r_res2, 0x48,
ATA_INL(ctlr->r_res2, 0x48) | (0xc0 >> ch->unit));
DELAY(1000);
ATA_OUTL(ctlr->r_res2, 0x48,
ATA_INL(ctlr->r_res2, 0x48) & ~(0xc0 >> ch->unit));
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ata_sata_phy_enable(ch);
/* enable PHY state change interrupt */
ATA_OUTL(ctlr->r_res2, 0x148 + offset, (1 << 16));
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_sii_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int rego = (ch->unit << 4) + (ATA_DEV(atadev->unit) << 1);
int mreg = ch->unit ? 0x84 : 0x80;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int mask = 0x03 << (ATA_DEV(atadev->unit) << 2);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int mval = pci_read_config(gparent, mreg, 1) & ~mask;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int error;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ctlr->chip->max_dma);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Fix potential cable detection problem on older CMD chips.
2004-01-14 14:24:36 +00:00
if (ctlr->chip->cfg2 & SIISETCLK) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (mode > ATA_UDMA2 && (pci_read_config(gparent, 0x79, 1) &
(ch->unit ? 0x02 : 0x01))) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_print_cable(dev, "controller");
Fix potential cable detection problem on older CMD chips.
2004-01-14 14:24:36 +00:00
mode = ATA_UDMA2;
}
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
}
Fix potential cable detection problem on older CMD chips.
2004-01-14 14:24:36 +00:00
else
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_check_80pin(dev, mode);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting %s on %s chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "FAILURE " : "",
ata_mode2str(mode), ctlr->chip->text);
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
if (error)
return;
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
if (mode >= ATA_UDMA0) {
u_int8_t udmatimings[] = { 0xf, 0xb, 0x7, 0x5, 0x3, 0x2, 0x1 };
u_int8_t ureg = 0xac + rego;
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, mreg,
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
mval | (0x03 << (ATA_DEV(atadev->unit) << 2)), 1);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, ureg,
(pci_read_config(gparent, ureg, 1) & ~0x3f) |
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
udmatimings[mode & ATA_MODE_MASK], 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
}
else if (mode >= ATA_WDMA0) {
u_int8_t dreg = 0xa8 + rego;
u_int16_t dmatimings[] = { 0x2208, 0x10c2, 0x10c1 };
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, mreg,
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
mval | (0x02 << (ATA_DEV(atadev->unit) << 2)), 1);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, dreg, dmatimings[mode & ATA_MODE_MASK], 2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
}
else {
u_int8_t preg = 0xa4 + rego;
u_int16_t piotimings[] = { 0x328a, 0x2283, 0x1104, 0x10c3, 0x10c1 };
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, mreg,
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
mval | (0x01 << (ATA_DEV(atadev->unit) << 2)), 1);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, preg, piotimings[mode & ATA_MODE_MASK], 2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
Update the SATA support code to work more correctly with real SATA disks now that I can test it. Add support for the SiI 3112 SATA chip using memory mapped I/O. Update the support for the SiI 0680 to use the memio interface as well. Sponsored by: David Leimbach <leimy2k@mac.com> (3112 based controller) Sponsored by: FreeBSD Systems (www.FreeBSDsystems.com) (SATA disks)
2003-07-02 10:50:44 +00:00
atadev->mode = mode;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_cmd_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int error;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ctlr->chip->max_dma);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_check_80pin(dev, mode);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting %s on %s chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "FAILURE " : "",
ata_mode2str(mode), ctlr->chip->text);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!error) {
Fix support for the cmd646 chip. Spotted by: tmm reviewed by: re@
2003-11-21 22:58:56 +00:00
int treg = 0x54 + ((devno < 3) ? (devno << 1) : 7);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int ureg = ch->unit ? 0x7b : 0x73;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (mode >= ATA_UDMA0) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int udmatimings[][2] = { { 0x31, 0xc2 }, { 0x21, 0x82 },
{ 0x11, 0x42 }, { 0x25, 0x8a },
{ 0x15, 0x4a }, { 0x05, 0x0a } };
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
u_int8_t umode = pci_read_config(gparent, ureg, 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
umode &= ~(atadev->unit == ATA_MASTER ? 0x35 : 0xca);
Fix UDMA modes setup on CMD chips.
2003-02-27 07:11:42 +00:00
umode |= udmatimings[mode & ATA_MODE_MASK][ATA_DEV(atadev->unit)];
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, ureg, umode, 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
else if (mode >= ATA_WDMA0) {
int dmatimings[] = { 0x87, 0x32, 0x3f };
Correct the PCI ID for the SiS 965, and add support for the SATA part that was left our by accident. Prodded by: Patrik Backlund
2005-03-31 06:25:14 +00:00
pci_write_config(gparent, treg, dmatimings[mode & ATA_MODE_MASK],1);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, ureg,
pci_read_config(gparent, ureg, 1) &
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
~(atadev->unit == ATA_MASTER ? 0x35 : 0xca), 1);
}
else {
int piotimings[] = { 0xa9, 0x57, 0x44, 0x32, 0x3f };
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, treg,
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
piotimings[(mode & ATA_MODE_MASK) - ATA_PIO0], 1);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, ureg,
pci_read_config(gparent, ureg, 1) &
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
~(atadev->unit == ATA_MASTER ? 0x35 : 0xca), 1);
}
atadev->mode = mode;
}
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/*
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
* Silicon Integrated Systems Corp. (SiS) chipset support functions
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
*/
int
ata_sis_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_SIS182, 0x00, SISSATA, 0, ATA_SA150, "182" }, /* south */
{ ATA_SIS181, 0x00, SISSATA, 0, ATA_SA150, "181" }, /* south */
{ ATA_SIS180, 0x00, SISSATA, 0, ATA_SA150, "180" }, /* south */
{ ATA_SIS965, 0x00, SIS133NEW, 0, ATA_UDMA6, "965" }, /* south */
{ ATA_SIS964, 0x00, SIS133NEW, 0, ATA_UDMA6, "964" }, /* south */
{ ATA_SIS963, 0x00, SIS133NEW, 0, ATA_UDMA6, "963" }, /* south */
{ ATA_SIS962, 0x00, SIS133NEW, 0, ATA_UDMA6, "962" }, /* south */
{ ATA_SIS745, 0x00, SIS100NEW, 0, ATA_UDMA5, "745" }, /* 1chip */
{ ATA_SIS735, 0x00, SIS100NEW, 0, ATA_UDMA5, "735" }, /* 1chip */
{ ATA_SIS733, 0x00, SIS100NEW, 0, ATA_UDMA5, "733" }, /* 1chip */
{ ATA_SIS730, 0x00, SIS100OLD, 0, ATA_UDMA5, "730" }, /* 1chip */
{ ATA_SIS635, 0x00, SIS100NEW, 0, ATA_UDMA5, "635" }, /* 1chip */
{ ATA_SIS633, 0x00, SIS100NEW, 0, ATA_UDMA5, "633" }, /* unknown */
{ ATA_SIS630, 0x30, SIS100OLD, 0, ATA_UDMA5, "630S"}, /* 1chip */
{ ATA_SIS630, 0x00, SIS66, 0, ATA_UDMA4, "630" }, /* 1chip */
{ ATA_SIS620, 0x00, SIS66, 0, ATA_UDMA4, "620" }, /* 1chip */
{ ATA_SIS550, 0x00, SIS66, 0, ATA_UDMA5, "550" },
{ ATA_SIS540, 0x00, SIS66, 0, ATA_UDMA4, "540" },
{ ATA_SIS530, 0x00, SIS66, 0, ATA_UDMA4, "530" },
{ ATA_SIS5513, 0xc2, SIS33, 1, ATA_UDMA2, "5513" },
{ ATA_SIS5513, 0x00, SIS33, 1, ATA_WDMA2, "5513" },
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{ 0, 0, 0, 0, 0, 0 }};
char buffer[64];
Update the SiS support to distinguish older southbridges better.
2004-03-16 16:23:28 +00:00
int found = 0;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Add support for the SiS964 ATA/SATA southbridge. This could not have been done without the support from kuriyama. Approved by: re@
2003-11-18 15:27:28 +00:00
if (!(idx = ata_find_chip(dev, ids, -pci_get_slot(dev))))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Update the SiS support to distinguish older southbridges better.
2004-03-16 16:23:28 +00:00
if (idx->cfg2 && !found) {
u_int8_t reg57 = pci_read_config(dev, 0x57, 1);
pci_write_config(dev, 0x57, (reg57 & 0x7f), 1);
if (pci_read_config(dev, PCIR_DEVVENDOR, 4) == ATA_SIS5518) {
found = 1;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
idx->cfg1 = SIS133NEW;
Try to simplify the SiS probe. If SiS stays with the current schema for new nothbridges we should even find the southbridge and be able to use it (if its one we know that is).
2003-12-25 19:37:11 +00:00
idx->max_dma = ATA_UDMA6;
sprintf(buffer, "SiS 962/963 %s controller",
ata_mode2str(idx->max_dma));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
Update the SiS support to distinguish older southbridges better.
2004-03-16 16:23:28 +00:00
pci_write_config(dev, 0x57, reg57, 1);
}
if (idx->cfg2 && !found) {
u_int8_t reg4a = pci_read_config(dev, 0x4a, 1);
pci_write_config(dev, 0x4a, (reg4a | 0x10), 1);
if (pci_read_config(dev, PCIR_DEVVENDOR, 4) == ATA_SIS5517) {
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
struct ata_chip_id id[] =
Update the SiS support to distinguish older southbridges better.
2004-03-16 16:23:28 +00:00
{{ ATA_SISSOUTH, 0x10, 0, 0, 0, "" }, { 0, 0, 0, 0, 0, 0 }};
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
Update the SiS support to distinguish older southbridges better.
2004-03-16 16:23:28 +00:00
found = 1;
Try to simplify the SiS probe. If SiS stays with the current schema for new nothbridges we should even find the southbridge and be able to use it (if its one we know that is).
2003-12-25 19:37:11 +00:00
if (ata_find_chip(dev, id, pci_get_slot(dev))) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
idx->cfg1 = SIS133OLD;
Try to simplify the SiS probe. If SiS stays with the current schema for new nothbridges we should even find the southbridge and be able to use it (if its one we know that is).
2003-12-25 19:37:11 +00:00
idx->max_dma = ATA_UDMA6;
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
else {
idx->cfg1 = SIS100NEW;
Try to simplify the SiS probe. If SiS stays with the current schema for new nothbridges we should even find the southbridge and be able to use it (if its one we know that is).
2003-12-25 19:37:11 +00:00
idx->max_dma = ATA_UDMA5;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
sprintf(buffer, "SiS 961 %s controller",ata_mode2str(idx->max_dma));
}
Update the SiS support to distinguish older southbridges better.
2004-03-16 16:23:28 +00:00
pci_write_config(dev, 0x4a, reg4a, 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
Update the SiS support to distinguish older southbridges better.
2004-03-16 16:23:28 +00:00
if (!found)
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
sprintf(buffer,"SiS %s %s controller",
idx->text, ata_mode2str(idx->max_dma));
Update the SiS support to distinguish older southbridges better.
2004-03-16 16:23:28 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_sis_chipinit;
return 0;
}
static int
ata_sis_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (ata_setup_interrupt(dev))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
switch (ctlr->chip->cfg1) {
Add a few more older SiS ID's.
2003-04-10 12:24:39 +00:00
case SIS33:
break;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
case SIS66:
case SIS100OLD:
Change setup for SiS device to allow PCI native mode.
2004-04-24 15:54:20 +00:00
pci_write_config(dev, 0x52, pci_read_config(dev, 0x52, 1) & ~0x04, 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
break;
case SIS100NEW:
case SIS133OLD:
Change setup for SiS device to allow PCI native mode.
2004-04-24 15:54:20 +00:00
pci_write_config(dev, 0x49, pci_read_config(dev, 0x49, 1) & ~0x01, 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
break;
case SIS133NEW:
Change setup for SiS device to allow PCI native mode.
2004-04-24 15:54:20 +00:00
pci_write_config(dev, 0x50, pci_read_config(dev, 0x50, 2) | 0x0008, 2);
pci_write_config(dev, 0x52, pci_read_config(dev, 0x52, 2) | 0x0008, 2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
break;
Add support for the SiS964 ATA/SATA southbridge. This could not have been done without the support from kuriyama. Approved by: re@
2003-11-18 15:27:28 +00:00
case SISSATA:
Only try to allocate and use the SATA resource if they are enabled by the BIOS. It seems some BIOS's doesn't get this right, and that would result in ATA panic'ing.
2005-04-25 07:50:51 +00:00
ctlr->r_type2 = SYS_RES_IOPORT;
ctlr->r_rid2 = PCIR_BAR(5);
if ((ctlr->r_res2 = bus_alloc_resource_any(dev, ctlr->r_type2,
&ctlr->r_rid2, RF_ACTIVE))) {
pci_write_config(dev, PCIR_COMMAND,
pci_read_config(dev, PCIR_COMMAND, 2) & ~0x0400,2);
ctlr->allocate = ata_sis_allocate;
ctlr->reset = ata_sis_reset;
}
Add support for the SiS964 ATA/SATA southbridge. This could not have been done without the support from kuriyama. Approved by: re@
2003-11-18 15:27:28 +00:00
ctlr->setmode = ata_sata_setmode;
return 0;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
default:
return ENXIO;
}
ctlr->setmode = ata_sis_setmode;
return 0;
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
static int
ata_sis_allocate(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
Fix SiS SATA support, the SATA registers was off. Add support for SiS metadata. HW donated by: obrien
2005-11-28 23:08:37 +00:00
int offset = ch->unit << ((ctlr->chip->chipid == ATA_SIS182) ? 5 : 6);
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
/* setup the usual register normal pci style */
Properly react to allocation failures. Found by: imp@
2005-10-10 19:12:43 +00:00
if (ata_pci_allocate(dev))
return ENXIO;
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ch->r_io[ATA_SSTATUS].res = ctlr->r_res2;
Fix SiS SATA support, the SATA registers was off. Add support for SiS metadata. HW donated by: obrien
2005-11-28 23:08:37 +00:00
ch->r_io[ATA_SSTATUS].offset = 0x00 + offset;
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ch->r_io[ATA_SERROR].res = ctlr->r_res2;
Fix SiS SATA support, the SATA registers was off. Add support for SiS metadata. HW donated by: obrien
2005-11-28 23:08:37 +00:00
ch->r_io[ATA_SERROR].offset = 0x04 + offset;
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ch->r_io[ATA_SCONTROL].res = ctlr->r_res2;
Fix SiS SATA support, the SATA registers was off. Add support for SiS metadata. HW donated by: obrien
2005-11-28 23:08:37 +00:00
ch->r_io[ATA_SCONTROL].offset = 0x08 + offset;
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ch->flags |= ATA_NO_SLAVE;
/* XXX SOS PHY hotplug handling missing in SiS chip ?? */
/* XXX SOS unknown how to enable PHY state change interrupt */
return 0;
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_sis_reset(device_t dev)
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(dev);
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ata_sata_phy_enable(ch);
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_sis_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int error;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ctlr->chip->max_dma);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (ctlr->chip->cfg1 == SIS133NEW) {
if (mode > ATA_UDMA2 &&
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_read_config(gparent, ch->unit ? 0x52 : 0x50,2) & 0x8000) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_print_cable(dev, "controller");
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
mode = ATA_UDMA2;
}
}
else {
if (mode > ATA_UDMA2 &&
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_read_config(gparent, 0x48, 1)&(ch->unit ? 0x20 : 0x10)) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_print_cable(dev, "controller");
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
mode = ATA_UDMA2;
}
}
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting %s on %s chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "FAILURE " : "",
ata_mode2str(mode), ctlr->chip->text);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!error) {
switch (ctlr->chip->cfg1) {
case SIS133NEW: {
u_int32_t timings[] =
{ 0x28269008, 0x0c266008, 0x04263008, 0x0c0a3008, 0x05093008,
Dont use r_bmio on Cyrix chips if not set from BIOS.
2003-03-10 08:20:14 +00:00
0x22196008, 0x0c0a3008, 0x05093008, 0x050939fc, 0x050936ac,
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
0x0509347c, 0x0509325c, 0x0509323c, 0x0509322c, 0x0509321c};
u_int32_t reg;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
reg = (pci_read_config(gparent, 0x57, 1)&0x40?0x70:0x40)+(devno<<2);
pci_write_config(gparent, reg, timings[ata_mode2idx(mode)], 4);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
break;
}
case SIS133OLD: {
u_int16_t timings[] =
{ 0x00cb, 0x0067, 0x0044, 0x0033, 0x0031, 0x0044, 0x0033, 0x0031,
0x8f31, 0x8a31, 0x8731, 0x8531, 0x8331, 0x8231, 0x8131 };
u_int16_t reg = 0x40 + (devno << 1);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, reg, timings[ata_mode2idx(mode)], 2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
break;
}
case SIS100NEW: {
u_int16_t timings[] =
{ 0x00cb, 0x0067, 0x0044, 0x0033, 0x0031, 0x0044, 0x0033,
0x0031, 0x8b31, 0x8731, 0x8531, 0x8431, 0x8231, 0x8131 };
u_int16_t reg = 0x40 + (devno << 1);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, reg, timings[ata_mode2idx(mode)], 2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
break;
}
case SIS100OLD:
Add a few more older SiS ID's.
2003-04-10 12:24:39 +00:00
case SIS66:
case SIS33: {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
u_int16_t timings[] =
{ 0x0c0b, 0x0607, 0x0404, 0x0303, 0x0301, 0x0404, 0x0303,
0x0301, 0xf301, 0xd301, 0xb301, 0xa301, 0x9301, 0x8301 };
u_int16_t reg = 0x40 + (devno << 1);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, reg, timings[ata_mode2idx(mode)], 2);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
break;
}
}
atadev->mode = mode;
}
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
/* VIA Technologies Inc. chipset support functions */
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int
ata_via_ident(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
struct ata_chip_id *idx;
static struct ata_chip_id ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_VIA82C586, 0x02, VIA33, 0x00, ATA_UDMA2, "82C586B" },
{ ATA_VIA82C586, 0x00, VIA33, 0x00, ATA_WDMA2, "82C586" },
{ ATA_VIA82C596, 0x12, VIA66, VIACLK, ATA_UDMA4, "82C596B" },
{ ATA_VIA82C596, 0x00, VIA33, 0x00, ATA_UDMA2, "82C596" },
{ ATA_VIA82C686, 0x40, VIA100, VIABUG, ATA_UDMA5, "82C686B"},
{ ATA_VIA82C686, 0x10, VIA66, VIACLK, ATA_UDMA4, "82C686A" },
{ ATA_VIA82C686, 0x00, VIA33, 0x00, ATA_UDMA2, "82C686" },
{ ATA_VIA8231, 0x00, VIA100, VIABUG, ATA_UDMA5, "8231" },
{ ATA_VIA8233, 0x00, VIA100, 0x00, ATA_UDMA5, "8233" },
{ ATA_VIA8233C, 0x00, VIA100, 0x00, ATA_UDMA5, "8233C" },
{ ATA_VIA8233A, 0x00, VIA133, 0x00, ATA_UDMA6, "8233A" },
{ ATA_VIA8235, 0x00, VIA133, 0x00, ATA_UDMA6, "8235" },
{ ATA_VIA8237, 0x00, VIA133, 0x00, ATA_UDMA6, "8237" },
{ ATA_VIA8251, 0x00, VIA133, 0x00, ATA_UDMA6, "8251" },
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
{ 0, 0, 0, 0, 0, 0 }};
static struct ata_chip_id new_ids[] =
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
{{ ATA_VIA6410, 0x00, 0, 0x00, ATA_UDMA6, "6410" },
{ ATA_VIA6420, 0x00, 7, 0x00, ATA_SA150, "6420" },
{ ATA_VIA6421, 0x00, 6, VIABAR, ATA_SA150, "6421" },
{ ATA_VIA8251, 0x00, 0, VIAAHCI, ATA_SA150, "8251" },
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{ 0, 0, 0, 0, 0, 0 }};
char buffer[64];
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
if (pci_get_devid(dev) == ATA_VIA82C571) {
Update the support for the VIA 8237 southbridge.
2004-01-20 16:51:02 +00:00
if (!(idx = ata_find_chip(dev, ids, -99)))
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
return ENXIO;
}
else {
if (!(idx = ata_match_chip(dev, new_ids)))
return ENXIO;
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Cleanup the strings for printing chipset names, saves quite some space. Add a couble new nVidia chips now we are here.
2005-12-16 08:12:13 +00:00
sprintf(buffer, "VIA %s %s controller",
idx->text, ata_mode2str(idx->max_dma));
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_set_desc_copy(dev, buffer);
ctlr->chip = idx;
ctlr->chipinit = ata_via_chipinit;
return 0;
}
static int
ata_via_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if (ata_setup_interrupt(dev))
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
return ENXIO;
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
if (ctlr->chip->max_dma >= ATA_SA150) {
Only try to allocate and use the SATA resource if they are enabled by the BIOS. It seems some BIOS's doesn't get this right, and that would result in ATA panic'ing.
2005-04-25 07:50:51 +00:00
ctlr->r_type2 = SYS_RES_IOPORT;
ctlr->r_rid2 = PCIR_BAR(5);
if ((ctlr->r_res2 = bus_alloc_resource_any(dev, ctlr->r_type2,
&ctlr->r_rid2, RF_ACTIVE))) {
pci_write_config(dev, PCIR_COMMAND,
pci_read_config(dev, PCIR_COMMAND, 2) & ~0x0400,2);
ctlr->allocate = ata_via_allocate;
ctlr->reset = ata_via_reset;
}
Add support for the VIA 8237 (both PATA and SATA part). Cleanup the SATA support a bit now we are here anyways.
2003-10-30 13:16:21 +00:00
ctlr->setmode = ata_sata_setmode;
return 0;
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/* prepare for ATA-66 on the 82C686a and 82C596b */
if (ctlr->chip->cfg2 & VIACLK)
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(dev, 0x50, 0x030b030b, 4);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/* the southbridge might need the data corruption fix */
if (ctlr->chip->cfg2 & VIABUG)
ata_via_southbridge_fixup(dev);
/* set fifo configuration half'n'half */
pci_write_config(dev, 0x43,
(pci_read_config(dev, 0x43, 1) & 0x90) | 0x2a, 1);
/* set status register read retry */
pci_write_config(dev, 0x44, pci_read_config(dev, 0x44, 1) | 0x08, 1);
/* set DMA read & end-of-sector fifo flush */
pci_write_config(dev, 0x46,
(pci_read_config(dev, 0x46, 1) & 0x0c) | 0xf0, 1);
/* set sector size */
pci_write_config(dev, 0x60, DEV_BSIZE, 2);
pci_write_config(dev, 0x68, DEV_BSIZE, 2);
ctlr->setmode = ata_via_family_setmode;
return 0;
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
static int
ata_via_allocate(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
Fix support for the VIA 6421 controller.
2005-08-29 09:01:57 +00:00
/* newer SATA chips has resources in one BAR for each channel */
if (ctlr->chip->cfg2 & VIABAR) {
struct resource *r_io;
int i, rid;
rid = PCIR_BAR(ch->unit);
if (!(r_io = bus_alloc_resource_any(device_get_parent(dev),
SYS_RES_IOPORT,
&rid, RF_ACTIVE)))
return ENXIO;
for (i = ATA_DATA; i <= ATA_COMMAND; i ++) {
ch->r_io[i].res = r_io;
ch->r_io[i].offset = i;
}
ch->r_io[ATA_CONTROL].res = r_io;
ch->r_io[ATA_CONTROL].offset = 2 + ATA_IOSIZE;
ch->r_io[ATA_IDX_ADDR].res = r_io;
ata_default_registers(dev);
for (i = ATA_BMCMD_PORT; i <= ATA_BMDTP_PORT; i++) {
ch->r_io[i].res = ctlr->r_res1;
ch->r_io[i].offset = i - ATA_BMCMD_PORT;
}
ata_generic_hw(dev);
}
Properly react to allocation failures. Found by: imp@
2005-10-10 19:12:43 +00:00
else {
if (ata_pci_allocate(dev))
return ENXIO;
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ch->r_io[ATA_SSTATUS].res = ctlr->r_res2;
ch->r_io[ATA_SSTATUS].offset = (ch->unit << ctlr->chip->cfg1);
ch->r_io[ATA_SERROR].res = ctlr->r_res2;
ch->r_io[ATA_SERROR].offset = 0x04 + (ch->unit << ctlr->chip->cfg1);
ch->r_io[ATA_SCONTROL].res = ctlr->r_res2;
ch->r_io[ATA_SCONTROL].offset = 0x08 + (ch->unit << ctlr->chip->cfg1);
ch->flags |= ATA_NO_SLAVE;
/* XXX SOS PHY hotplug handling missing in VIA chip ?? */
/* XXX SOS unknown how to enable PHY state change interrupt */
return 0;
}
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_via_reset(device_t dev)
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(dev);
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
ata_sata_phy_enable(ch);
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
static void
ata_via_southbridge_fixup(device_t dev)
{
device_t *children;
int nchildren, i;
if (device_get_children(device_get_parent(dev), &children, &nchildren))
return;
for (i = 0; i < nchildren; i++) {
if (pci_get_devid(children[i]) == ATA_VIA8363 ||
pci_get_devid(children[i]) == ATA_VIA8371 ||
pci_get_devid(children[i]) == ATA_VIA8662 ||
pci_get_devid(children[i]) == ATA_VIA8361) {
u_int8_t reg76 = pci_read_config(children[i], 0x76, 1);
if ((reg76 & 0xf0) != 0xd0) {
device_printf(dev,
"Correcting VIA config for southbridge data corruption bug\n");
pci_write_config(children[i], 0x75, 0x80, 1);
pci_write_config(children[i], 0x76, (reg76 & 0x0f) | 0xd0, 1);
}
break;
}
}
free(children, M_TEMP);
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/* common code for VIA, AMD & nVidia */
static void
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_via_family_setmode(device_t dev, int mode)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_t gparent = GRANDPARENT(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_pci_controller *ctlr = device_get_softc(gparent);
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
Fix the register timings for AMD/VIA/nVidia chipsets.
2003-12-18 17:36:41 +00:00
u_int8_t timings[] = { 0xa8, 0x65, 0x42, 0x22, 0x20, 0x42, 0x22, 0x20,
0x20, 0x20, 0x20, 0x20, 0x20, 0x20 };
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int modes[][7] = {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
{ 0xc2, 0xc1, 0xc0, 0x00, 0x00, 0x00, 0x00 }, /* VIA ATA33 */
{ 0xee, 0xec, 0xea, 0xe9, 0xe8, 0x00, 0x00 }, /* VIA ATA66 */
{ 0xf7, 0xf6, 0xf4, 0xf2, 0xf1, 0xf0, 0x00 }, /* VIA ATA100 */
{ 0xf7, 0xf7, 0xf6, 0xf4, 0xf2, 0xf1, 0xf0 }, /* VIA ATA133 */
{ 0xc2, 0xc1, 0xc0, 0xc4, 0xc5, 0xc6, 0xc7 }}; /* AMD/nVIDIA */
int devno = (ch->unit << 1) + ATA_DEV(atadev->unit);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
int reg = 0x53 - devno;
int error;
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_limit_mode(dev, mode, ctlr->chip->max_dma);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (ctlr->chip->cfg2 & AMDCABLE) {
Fix cable detection on AMD chips.
2003-11-01 09:30:15 +00:00
if (mode > ATA_UDMA2 &&
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
!(pci_read_config(gparent, 0x42, 1) & (1 << devno))) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_print_cable(dev, "controller");
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
mode = ATA_UDMA2;
}
}
else
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
mode = ata_check_80pin(dev, mode);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (ctlr->chip->cfg2 & NVIDIA)
reg += 0x10;
Dont set "address setup timing" on newer VIA chips. closes PR: 69180
2004-07-17 14:48:46 +00:00
if (ctlr->chip->cfg1 != VIA133)
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, reg - 0x08, timings[ata_mode2idx(mode)], 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
error = ata_controlcmd(dev, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (bootverbose)
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
device_printf(dev, "%ssetting %s on %s chip\n",
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
(error) ? "FAILURE " : "", ata_mode2str(mode),
ctlr->chip->text);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (!error) {
if (mode >= ATA_UDMA0)
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, reg,
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
modes[ctlr->chip->cfg1][mode & ATA_MODE_MASK], 1);
else
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
pci_write_config(gparent, reg, 0x8b, 1);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
atadev->mode = mode;
}
}
Add support for the PHY on nVidia, SiS and VIA chips This code is done blindfolded as I dont have such HW here, so reports of success/failure are most welcome. More SATA PHY cleanups.
2005-04-08 15:33:04 +00:00
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
/* misc functions */
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct ata_chip_id *
ata_match_chip(device_t dev, struct ata_chip_id *index)
{
while (index->chipid != 0) {
if (pci_get_devid(dev) == index->chipid &&
pci_get_revid(dev) >= index->chiprev)
return index;
index++;
}
return NULL;
}
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
static struct ata_chip_id *
ata_find_chip(device_t dev, struct ata_chip_id *index, int slot)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
device_t *children;
Fix the cable detection and properly find the host bridge on older SiS chips.
2003-03-16 16:26:51 +00:00
int nchildren, i;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
if (device_get_children(device_get_parent(dev), &children, &nchildren))
return 0;
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
while (index->chipid != 0) {
for (i = 0; i < nchildren; i++) {
Add support for the SiS964 ATA/SATA southbridge. This could not have been done without the support from kuriyama. Approved by: re@
2003-11-18 15:27:28 +00:00
if (((slot >= 0 && pci_get_slot(children[i]) == slot) ||
(slot < 0 && pci_get_slot(children[i]) <= -slot)) &&
Unbork the chip locating code. Approved by: re@
2003-05-19 13:42:23 +00:00
pci_get_devid(children[i]) == index->chipid &&
pci_get_revid(children[i]) >= index->chiprev) {
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
free(children, M_TEMP);
return index;
}
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
Unbork the chip locating code. Approved by: re@
2003-05-19 13:42:23 +00:00
index++;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
free(children, M_TEMP);
Support the ICH5 SATA part. Fix HPT374 UDMA133 timing. Fix Promise ID. Cosmetics on probe print for Promise & HPT. Approved by: re
2003-05-18 16:45:48 +00:00
return NULL;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
static int
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
ata_setup_interrupt(device_t dev)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
int rid = ATA_IRQ_RID;
Make the test for ATA PCI legacy addressing mode more robust. Add code (currently ifdef'd out) to allow ATA PCI native addressing. Fix the altio offset for ATA PCI devices.
2004-04-21 20:03:26 +00:00
if (!ata_legacy(dev)) {
Convert callers to the new bus_alloc_resource_any(9) API. Submitted by: Mark Santcroos <marks@ripe.net> Reviewed by: imp, dfr, bde
2004-03-17 17:50:55 +00:00
if (!(ctlr->r_irq = bus_alloc_resource_any(dev, SYS_RES_IRQ, &rid,
RF_SHAREABLE | RF_ACTIVE))) {
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
device_printf(dev, "unable to map interrupt\n");
return ENXIO;
}
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
if ((bus_setup_intr(dev, ctlr->r_irq, ATA_INTR_FLAGS,
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
ata_generic_intr, ctlr, &ctlr->handle))) {
device_printf(dev, "unable to setup interrupt\n");
return ENXIO;
}
}
return 0;
}
Fix the PC98 lockups on boot. The interchannel locking for PC98 needed to be updated to match the rest of the locking in ATA.
2004-10-06 19:46:08 +00:00
struct ata_serialize {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
struct mtx locked_mtx;
int locked_ch;
int restart_ch;
Fix the PC98 lockups on boot. The interchannel locking for PC98 needed to be updated to match the rest of the locking in ATA.
2004-10-06 19:46:08 +00:00
};
static int
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_serialize(device_t dev, int flags)
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
Fix the PC98 lockups on boot. The interchannel locking for PC98 needed to be updated to match the rest of the locking in ATA.
2004-10-06 19:46:08 +00:00
struct ata_serialize *serial;
static int inited = 0;
int res;
if (!inited) {
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
serial = malloc(sizeof(struct ata_serialize),
Fix the PC98 lockups on boot. The interchannel locking for PC98 needed to be updated to match the rest of the locking in ATA.
2004-10-06 19:46:08 +00:00
M_TEMP, M_NOWAIT | M_ZERO);
mtx_init(&serial->locked_mtx, "ATA serialize lock", NULL, MTX_DEF);
serial->locked_ch = -1;
serial->restart_ch = -1;
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
device_set_ivars(ctlr->dev, serial);
Fix the PC98 lockups on boot. The interchannel locking for PC98 needed to be updated to match the rest of the locking in ATA.
2004-10-06 19:46:08 +00:00
inited = 1;
}
else
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
serial = device_get_ivars(ctlr->dev);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
Fix the PC98 lockups on boot. The interchannel locking for PC98 needed to be updated to match the rest of the locking in ATA.
2004-10-06 19:46:08 +00:00
mtx_lock(&serial->locked_mtx);
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
switch (flags) {
case ATA_LF_LOCK:
Fix the PC98 lockups on boot. The interchannel locking for PC98 needed to be updated to match the rest of the locking in ATA.
2004-10-06 19:46:08 +00:00
if (serial->locked_ch == -1)
serial->locked_ch = ch->unit;
if (serial->locked_ch != ch->unit)
serial->restart_ch = ch->unit;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
break;
case ATA_LF_UNLOCK:
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (serial->locked_ch == ch->unit) {
serial->locked_ch = -1;
if (serial->restart_ch != -1) {
if ((ch = ctlr->interrupt[serial->restart_ch].argument)) {
serial->restart_ch = -1;
mtx_unlock(&serial->locked_mtx);
ata_start(ch->dev);
return -1;
}
}
}
Fix the PC98 lockups on boot. The interchannel locking for PC98 needed to be updated to match the rest of the locking in ATA.
2004-10-06 19:46:08 +00:00
break;
case ATA_LF_WHICH:
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
break;
}
Fix the PC98 lockups on boot. The interchannel locking for PC98 needed to be updated to match the rest of the locking in ATA.
2004-10-06 19:46:08 +00:00
res = serial->locked_ch;
mtx_unlock(&serial->locked_mtx);
return res;
First round off updates/fixes to the ATA driver. This moves all chipset specific code to a new file 'ata-chipset.c'. Extensive use of tables and pointers to avoid having the same switch on chipset type in several places, and to allow substituting various functions for different HW arch needs. Added PIO mode setup and all DMA modes. Support for all known SiS chipsets. Thanks to Christoph Kukulies for sponsoring a nice ASUS P4S8X SiS648 based board for this work! Tested on: i386, PC98, alpha and sparc64
2003-02-20 20:02:32 +00:00
}
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
static void
ata_print_cable(device_t dev, u_int8_t *who)
{
device_printf(dev,
"DMA limited to UDMA33, %s found non-ATA66 cable\n", who);
}
static int
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_atapi(device_t dev)
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_channel *ch = device_get_softc(device_get_parent(dev));
struct ata_device *atadev = device_get_softc(dev);
Generalise the SATA PHY handling code so it wont be duplicated for each SATA chip. Promise and Silicon Image are the current candidates for this.
2005-04-08 09:37:47 +00:00
return ((atadev->unit == ATA_MASTER && ch->devices & ATA_ATAPI_MASTER) ||
(atadev->unit == ATA_SLAVE && ch->devices & ATA_ATAPI_SLAVE));
}
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
static int
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_check_80pin(device_t dev, int mode)
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
{
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
struct ata_device *atadev = device_get_softc(dev);
This is the much rumoured ATA mkIII update that I've been working on. o ATA is now fully newbus'd and split into modules. This means that on a modern system you just load "atapci and ata" to get the base support, and then one or more of the device subdrivers "atadisk atapicd atapifd atapist ataraid". All can be loaded/unloaded anytime, but for obvious reasons you dont want to unload atadisk when you have mounted filesystems. o The device identify part of the probe has been rewritten to fix the problems with odd devices the old had, and to try to remove so of the long delays some HW could provoke. Also probing is done without the need for interrupts, making earlier probing possible. o SATA devices can be hot inserted/removed and devices will be created/ removed in /dev accordingly. NOTE: only supported on controllers that has this feature: Promise and Silicon Image for now. On other controllers the usual atacontrol detach/attach dance is still needed. o Support for "atomic" composite ATA requests used for RAID. o ATA RAID support has been rewritten and and now supports these metadata formats: "Adaptec HostRAID" "Highpoint V2 RocketRAID" "Highpoint V3 RocketRAID" "Intel MatrixRAID" "Integrated Technology Express" "LSILogic V2 MegaRAID" "LSILogic V3 MegaRAID" "Promise FastTrak" "Silicon Image Medley" "FreeBSD PseudoRAID" o Update the ioctl API to match new RAID levels etc. o Update atacontrol to know about the new RAID levels etc NOTE: you need to recompile atacontrol with the new sys/ata.h, make world will take care of that. NOTE2: that rebuild is done differently from the old system as the rebuild is now done piggybacked on read requests to the array, so atacontrol simply starts a background "dd" to rebuild the array. o The reinit code has been worked over to be much more robust. o The timeout code has been overhauled for races. o Support of new chipsets. o Lots of fixes for bugs found while doing the modulerization and reviewing the old code. Missing or changed features from current ATA: o atapi-cd no longer has support for ATAPI changers. Todays its much cheaper and alot faster to copy those CD images to disk and serve them from there. Besides they dont seem to be made anymore, maybe for that exact reason. o ATA RAID can only read metadata from all the above metadata formats, not write all of them (Promise and Highpoint V2 so far). This means that arrays can be picked up from the BIOS, but they cannot be created from FreeBSD. There is more to it than just the missing write metadata support, those formats are not unique to a given controller like Promise and Highpoint formats, instead they exist for several types, and even worse, some controllers can have different formats and its impossible to tell which one. The outcome is that we cannot reliably create the metadata of those formats and be sure the controller BIOS will understand it. However write support is needed to update/fail/rebuild the arrays properly so it sits fairly high on the TODO list. o So far atapicam is not supported with these changes. When/if this will change is up to the maintainer of atapi-cam so go there for questions. HW donated by: Webveveriet AS HW donated by: Frode Nordahl HW donated by: Yahoo! HW donated by: Sentex Patience by: Vife and my boys (and even the cats)
2005-03-30 12:03:40 +00:00
if (mode > ATA_UDMA2 && !(atadev->param.hwres & ATA_CABLE_ID)) {
Take newbusification one step further, ie use the device_t more consequently all way through the code down the layers, instead of the mix'n'match that resulted from the conversion done earlier. Sponsored by: pair.com
2005-04-30 16:22:07 +00:00
ata_print_cable(dev, "device");
This is a major rework of the ATA driver (ATAng) Restructure the way ATA/ATAPI commands are processed, use a common ata_request structure for both. This centralises the way requests are handled so locking is much easier to handle. The driver is now layered much more cleanly to seperate the lowlevel HW access so it can be tailored to specific controllers without touching the upper layers. This is needed to support some of the newer semi-intelligent ATA controllers showing up. The top level drivers (disk, ATAPI devices) are more or less still the same with just corrections to use the new interface. Pull ATA out from under Gaint now that locking can be done in a sane way. Add support for a the National Geode SC1100. Thanks to Soekris engineering for sponsoring a Soekris 4801 to make this support. Fixed alot of small bugs in the chipset code for various chips now we are around in that corner anyways.
2003-08-24 09:22:26 +00:00
mode = ATA_UDMA2;
}
return mode;
}
static int
ata_mode2idx(int mode)
{
if ((mode & ATA_DMA_MASK) == ATA_UDMA0)
return (mode & ATA_MODE_MASK) + 8;
if ((mode & ATA_DMA_MASK) == ATA_WDMA0)
return (mode & ATA_MODE_MASK) + 5;
return (mode & ATA_MODE_MASK) - ATA_PIO0;
}
Reference in New Issue Copy Permalink