d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
062dad0f33
freebsd-skq/sys/dev/ata/chipsets/ata-ahci.c

1013 lines
32 KiB
C
Raw Normal View History

This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/*-
* Copyright (c) 1998 - 2008 S<EFBFBD>ren Schmidt <sos@FreeBSD.org>
* 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.
*
* 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.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_ata.h"
#include <sys/param.h>
#include <sys/module.h>
#include <sys/systm.h>
#include <sys/kernel.h>
#include <sys/ata.h>
#include <sys/bus.h>
#include <sys/endian.h>
#include <sys/malloc.h>
#include <sys/lock.h>
#include <sys/mutex.h>
#include <sys/sema.h>
#include <sys/taskqueue.h>
#include <vm/uma.h>
#include <machine/stdarg.h>
#include <machine/resource.h>
#include <machine/bus.h>
#include <sys/rman.h>
#include <dev/pci/pcivar.h>
#include <dev/pci/pcireg.h>
#include <dev/ata/ata-all.h>
#include <dev/ata/ata-pci.h>
#include <ata_if.h>
/* local prototypes */
Change the way in which AHCI+PATA combined controllers, such as JMicron and Marvell handled. Instead of trying to attach two different drivers to single device, wrapping each call, make one of them (atajmicron, atamarvell) attach do device solely, but create child device for AHCI driver, passing it all required resources. It is quite easy, as none of resources are shared, except IRQ. As result, it: - makes drivers operation more independent and straitforward, - allows to use new ahci(4) driver with such devices, adding support for new features, such as PMP and NCQ, same time keeping legacy PATA support, - will allow to just drop old ataahci driver, when it's time come.
2009-11-16 15:38:27 +00:00
static int ata_ahci_ch_attach(device_t dev);
static int ata_ahci_ch_detach(device_t dev);
static int ata_ahci_ch_suspend(device_t dev);
static int ata_ahci_ch_resume(device_t dev);
static int ata_ahci_ctlr_reset(device_t dev);
static void ata_ahci_reset(device_t dev);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
static int ata_ahci_suspend(device_t dev);
static int ata_ahci_status(device_t dev);
static int ata_ahci_begin_transaction(struct ata_request *request);
static int ata_ahci_end_transaction(struct ata_request *request);
static int ata_ahci_pm_read(device_t dev, int port, int reg, u_int32_t *result);
static int ata_ahci_pm_write(device_t dev, int port, int reg, u_int32_t result);
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
static int ata_ahci_hardreset(device_t dev, int port, uint32_t *signature);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
static u_int32_t ata_ahci_softreset(device_t dev, int port);
static void ata_ahci_dmasetprd(void *xsc, bus_dma_segment_t *segs, int nsegs, int error);
static int ata_ahci_setup_fis(struct ata_ahci_cmd_tab *ctp, struct ata_request *equest);
Quite mechanical ch_detach implementations for all atapci subdrivers. Some dmainit call fixes for previous commit.
2009-02-19 00:32:55 +00:00
static void ata_ahci_dmainit(device_t dev);
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
static void ata_ahci_start(device_t dev);
static void ata_ahci_stop(device_t dev);
static void ata_ahci_clo(device_t dev);
static void ata_ahci_start_fr(device_t dev);
static void ata_ahci_stop_fr(device_t dev);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/*
* AHCI v1.x compliant SATA chipset support functions
*/
static int
ata_ahci_probe(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
char buffer[64];
/* is this a possible AHCI candidate ? */
Before modularization commit, atapci driver was attaching only to devices of storage class. This check was lost. It is not important for the most cases, but as it was reported on current@, it does important for sis driver and surely inportant for AHCI driver. So restore it there. Submitted by: Toshikazu ICHINOSEKI, Andrey V. Elsukov Discussed on: current@
2008-12-18 21:13:46 +00:00
if (pci_get_class(dev) != PCIC_STORAGE ||
pci_get_subclass(dev) != PCIS_STORAGE_SATA)
return (ENXIO);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* is this PCI device flagged as an AHCI compliant chip ? */
Use cached progif instead of reading it again.
2009-04-27 21:45:05 +00:00
if (pci_get_progif(dev) != PCIP_STORAGE_SATA_AHCI_1_0)
- Give generic AHCI driver lower priority than device-specific drivers. - Consistently use BUS_PROBE_GENERIC instead of -100.
2009-04-27 21:34:15 +00:00
return (ENXIO);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
if (bootverbose)
sprintf(buffer, "%s (ID=%08x) AHCI controller",
ata_pcivendor2str(dev), pci_get_devid(dev));
else
sprintf(buffer, "%s AHCI controller", ata_pcivendor2str(dev));
device_set_desc_copy(dev, buffer);
ctlr->chipinit = ata_ahci_chipinit;
- Give generic AHCI driver lower priority than device-specific drivers. - Consistently use BUS_PROBE_GENERIC instead of -100.
2009-04-27 21:34:15 +00:00
return (BUS_PROBE_GENERIC);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
}
Change the way in which AHCI+PATA combined controllers, such as JMicron and Marvell handled. Instead of trying to attach two different drivers to single device, wrapping each call, make one of them (atajmicron, atamarvell) attach do device solely, but create child device for AHCI driver, passing it all required resources. It is quite easy, as none of resources are shared, except IRQ. As result, it: - makes drivers operation more independent and straitforward, - allows to use new ahci(4) driver with such devices, adding support for new features, such as PMP and NCQ, same time keeping legacy PATA support, - will allow to just drop old ataahci driver, when it's time come.
2009-11-16 15:38:27 +00:00
static int
ata_ahci_ata_probe(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
if ((intptr_t)device_get_ivars(dev) >= 0)
return (ENXIO);
device_set_desc_copy(dev, "AHCI SATA controller");
ctlr->chipinit = ata_ahci_chipinit;
return (BUS_PROBE_GENERIC);
}
static int
ata_ahci_ata_attach(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
device_t child;
int unit;
/* do chipset specific setups only needed once */
ctlr->legacy = 0;
ctlr->ichannels = -1;
ctlr->ch_attach = ata_pci_ch_attach;
ctlr->ch_detach = ata_pci_ch_detach;
ctlr->dev = dev;
if (ctlr->chipinit(dev))
return ENXIO;
/* attach all channels on this controller */
for (unit = 0; unit < ctlr->channels; unit++) {
if ((ctlr->ichannels & (1 << unit)) == 0)
continue;
child = device_add_child(dev, "ata",
((unit == 0 || unit == 1) && ctlr->legacy) ?
unit : devclass_find_free_unit(ata_devclass, 2));
if (child == NULL)
device_printf(dev, "failed to add ata child device\n");
else
device_set_ivars(child, (void *)(intptr_t)unit);
}
bus_generic_attach(dev);
return 0;
}
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
int
ata_ahci_chipinit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
MFp4. Log supported AHCI controller capabilities.
2009-06-01 21:42:26 +00:00
int error, speed;
u_int32_t caps, version;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* if we have a memory BAR(5) we are likely on an AHCI part */
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)))
return ENXIO;
/* setup interrupt delivery if not done allready by a vendor driver */
if (!ctlr->r_irq) {
Improve error handling at ata_ahci_chipinit(). Submitted by: Andrey V. Elsukov
2008-12-17 19:17:58 +00:00
if (ata_setup_interrupt(dev, ata_generic_intr)) {
bus_release_resource(dev, ctlr->r_type2, ctlr->r_rid2, ctlr->r_res2);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
return ENXIO;
Improve error handling at ata_ahci_chipinit(). Submitted by: Andrey V. Elsukov
2008-12-17 19:17:58 +00:00
}
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
}
else
device_printf(dev, "AHCI called from vendor specific driver\n");
/* reset controller */
Improve error handling at ata_ahci_chipinit(). Submitted by: Andrey V. Elsukov
2008-12-17 19:17:58 +00:00
if ((error = ata_ahci_ctlr_reset(dev)) != 0) {
bus_release_resource(dev, ctlr->r_type2, ctlr->r_rid2, ctlr->r_res2);
return (error);
};
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* get the number of HW channels */
Give atapci knowledge about set of implemented AHCI ports. It is possible to not allocate them after the recent ata channels enumeration changes. It allows to save some resources, not bother user with unexisting hardware and not check unimplemented ports status on every interrupt.
2009-02-16 19:10:07 +00:00
ctlr->ichannels = ATA_INL(ctlr->r_res2, ATA_AHCI_PI);
Change the way in which AHCI+PATA combined controllers, such as JMicron and Marvell handled. Instead of trying to attach two different drivers to single device, wrapping each call, make one of them (atajmicron, atamarvell) attach do device solely, but create child device for AHCI driver, passing it all required resources. It is quite easy, as none of resources are shared, except IRQ. As result, it: - makes drivers operation more independent and straitforward, - allows to use new ahci(4) driver with such devices, adding support for new features, such as PMP and NCQ, same time keeping legacy PATA support, - will allow to just drop old ataahci driver, when it's time come.
2009-11-16 15:38:27 +00:00
ctlr->channels = MAX(flsl(ctlr->ichannels),
Add experimental support for SATA interface power management. Feature is controlled by hint.ata.X.pm_level tunable: 0 - PM disabled, old behaviour, default. 1 - device is allowed to initiate PM state change, host is passive. 2 - host initiates PARTIAL state transition every time port is idle. 3 - host initiates SLUMBER state transition every time port is idle. PARTIAL state has up to 100us (50us for me) wakeup latency, but for my ICH8M saves 0.5W of power per drive. SLUMBER state has up to 10ms (3.5ms for me) wakeup latency, but saves 0.8W of power. Modes 2 and 3 are implemented only for AHCI driver now. Interface power management is incompatible with device presence detection (host receives no signal from drive, so unable to monitor it), so later is disabled when PM is used.
2009-04-29 21:17:18 +00:00
(ATA_INL(ctlr->r_res2, ATA_AHCI_CAP) & ATA_AHCI_CAP_NPMASK) + 1);
Change the way in which AHCI+PATA combined controllers, such as JMicron and Marvell handled. Instead of trying to attach two different drivers to single device, wrapping each call, make one of them (atajmicron, atamarvell) attach do device solely, but create child device for AHCI driver, passing it all required resources. It is quite easy, as none of resources are shared, except IRQ. As result, it: - makes drivers operation more independent and straitforward, - allows to use new ahci(4) driver with such devices, adding support for new features, such as PMP and NCQ, same time keeping legacy PATA support, - will allow to just drop old ataahci driver, when it's time come.
2009-11-16 15:38:27 +00:00
if (pci_get_devid(dev) == ATA_M88SX6111)
ctlr->channels = 1;
else if (pci_get_devid(dev) == ATA_M88SX6121)
ctlr->channels = 2;
else if (pci_get_devid(dev) == ATA_M88SX6141 ||
pci_get_devid(dev) == ATA_M88SX6145)
ctlr->channels = 4;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
ctlr->reset = ata_ahci_reset;
As soon as they called in only same one place (ata_pcichannel_attach()), join allocate() and dmainit() atapci subdriver's channel initialization methods into single ch_attach() method. As opposite to ch_attach() add new ch_detach() method to deallocate/disable channel.
2009-02-18 22:17:48 +00:00
ctlr->ch_attach = ata_ahci_ch_attach;
Quite mechanical ch_detach implementations for all atapci subdrivers. Some dmainit call fixes for previous commit.
2009-02-19 00:32:55 +00:00
ctlr->ch_detach = ata_ahci_ch_detach;
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
ctlr->ch_suspend = ata_ahci_ch_suspend;
ctlr->ch_resume = ata_ahci_ch_resume;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
ctlr->setmode = ata_sata_setmode;
MFp4: Introduce ATA_CAM kernel option, turning ata(4) controller drivers into cam(4) interface modules. When enabled, this options deprecates all ata(4) peripheral drivers (ad, acd, ...) and interfaces and allows cam(4) drivers (ada, cd, ...) and interfaces to be natively used instead. As side effect of this, ata(4) mode setting code was completely rewritten to make controller API more strict and permit above change. While doing this, SATA revision was separated from PATA mode. It allows DMA-incapable SATA devices to operate and makes hw.ata.atapi_dma tunable work again. Also allow ata(4) controller drivers (except some specific or broken ones) to handle larger data transfers. Previous constraint of 64K was artificial and is not really required by PCI ATA BM specification or hardware. Submitted by: nwitehorn (powerpc part)
2009-12-06 00:10:13 +00:00
ctlr->getrev = ata_sata_getrev;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
ctlr->suspend = ata_ahci_suspend;
ctlr->resume = ata_ahci_ctlr_reset;
MFp4. Log supported AHCI controller capabilities.
2009-06-01 21:42:26 +00:00
/* announce we support the HW */
version = ATA_INL(ctlr->r_res2, ATA_AHCI_VS);
caps = ATA_INL(ctlr->r_res2, ATA_AHCI_CAP);
speed = (caps & ATA_AHCI_CAP_ISS) >> ATA_AHCI_CAP_ISS_SHIFT;
device_printf(dev,
"AHCI v%x.%02x controller with %d %sGbps ports, PM %s\n",
((version >> 20) & 0xf0) + ((version >> 16) & 0x0f),
((version >> 4) & 0xf0) + (version & 0x0f),
(caps & ATA_AHCI_CAP_NPMASK) + 1,
((speed == 1) ? "1.5":((speed == 2) ? "3":
((speed == 3) ? "6":"?"))),
(caps & ATA_AHCI_CAP_SPM) ?
"supported" : "not supported");
if (bootverbose) {
device_printf(dev, "Caps:%s%s%s%s%s%s%s%s %sGbps",
(caps & ATA_AHCI_CAP_64BIT) ? " 64bit":"",
(caps & ATA_AHCI_CAP_SNCQ) ? " NCQ":"",
(caps & ATA_AHCI_CAP_SSNTF) ? " SNTF":"",
(caps & ATA_AHCI_CAP_SMPS) ? " MPS":"",
(caps & ATA_AHCI_CAP_SSS) ? " SS":"",
(caps & ATA_AHCI_CAP_SALP) ? " ALP":"",
(caps & ATA_AHCI_CAP_SAL) ? " AL":"",
(caps & ATA_AHCI_CAP_SCLO) ? " CLO":"",
((speed == 1) ? "1.5":((speed == 2) ? "3":
((speed == 3) ? "6":"?"))));
printf("%s%s%s%s%s%s %dcmd%s%s%s %dports\n",
(caps & ATA_AHCI_CAP_SAM) ? " AM":"",
(caps & ATA_AHCI_CAP_SPM) ? " PM":"",
(caps & ATA_AHCI_CAP_FBSS) ? " FBS":"",
(caps & ATA_AHCI_CAP_PMD) ? " PMD":"",
(caps & ATA_AHCI_CAP_SSC) ? " SSC":"",
(caps & ATA_AHCI_CAP_PSC) ? " PSC":"",
((caps & ATA_AHCI_CAP_NCS) >> ATA_AHCI_CAP_NCS_SHIFT) + 1,
(caps & ATA_AHCI_CAP_CCCS) ? " CCC":"",
(caps & ATA_AHCI_CAP_EMS) ? " EM":"",
(caps & ATA_AHCI_CAP_SXS) ? " eSATA":"",
(caps & ATA_AHCI_CAP_NPMASK) + 1);
}
return 0;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
}
Change the way in which AHCI+PATA combined controllers, such as JMicron and Marvell handled. Instead of trying to attach two different drivers to single device, wrapping each call, make one of them (atajmicron, atamarvell) attach do device solely, but create child device for AHCI driver, passing it all required resources. It is quite easy, as none of resources are shared, except IRQ. As result, it: - makes drivers operation more independent and straitforward, - allows to use new ahci(4) driver with such devices, adding support for new features, such as PMP and NCQ, same time keeping legacy PATA support, - will allow to just drop old ataahci driver, when it's time come.
2009-11-16 15:38:27 +00:00
static int
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
ata_ahci_ctlr_reset(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
int timeout;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* enable AHCI mode */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_GHC, ATA_AHCI_GHC_AE);
/* reset AHCI controller */
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
ATA_OUTL(ctlr->r_res2, ATA_AHCI_GHC, ATA_AHCI_GHC_AE|ATA_AHCI_GHC_HR);
for (timeout = 1000; timeout > 0; timeout--) {
DELAY(1000);
if ((ATA_INL(ctlr->r_res2, ATA_AHCI_GHC) & ATA_AHCI_GHC_HR) == 0)
break;
}
if (timeout == 0) {
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
device_printf(dev, "AHCI controller reset failure\n");
return ENXIO;
}
/* reenable AHCI mode */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_GHC, ATA_AHCI_GHC_AE);
/* clear interrupts */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_IS, ATA_INL(ctlr->r_res2, ATA_AHCI_IS));
/* enable AHCI interrupts */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_GHC,
ATA_INL(ctlr->r_res2, ATA_AHCI_GHC) | ATA_AHCI_GHC_IE);
return 0;
}
static int
ata_ahci_suspend(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(dev);
/* disable interupts so the state change(s) doesn't trigger */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_GHC,
ATA_INL(ctlr->r_res2, ATA_AHCI_GHC) & (~ATA_AHCI_GHC_IE));
return 0;
}
Change the way in which AHCI+PATA combined controllers, such as JMicron and Marvell handled. Instead of trying to attach two different drivers to single device, wrapping each call, make one of them (atajmicron, atamarvell) attach do device solely, but create child device for AHCI driver, passing it all required resources. It is quite easy, as none of resources are shared, except IRQ. As result, it: - makes drivers operation more independent and straitforward, - allows to use new ahci(4) driver with such devices, adding support for new features, such as PMP and NCQ, same time keeping legacy PATA support, - will allow to just drop old ataahci driver, when it's time come.
2009-11-16 15:38:27 +00:00
static int
As soon as they called in only same one place (ata_pcichannel_attach()), join allocate() and dmainit() atapci subdriver's channel initialization methods into single ch_attach() method. As opposite to ch_attach() add new ch_detach() method to deallocate/disable channel.
2009-02-18 22:17:48 +00:00
ata_ahci_ch_attach(device_t dev)
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
{
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;
As soon as they called in only same one place (ata_pcichannel_attach()), join allocate() and dmainit() atapci subdriver's channel initialization methods into single ch_attach() method. As opposite to ch_attach() add new ch_detach() method to deallocate/disable channel.
2009-02-18 22:17:48 +00:00
ata_ahci_dmainit(dev);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* 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.status = ata_ahci_status;
ch->hw.begin_transaction = ata_ahci_begin_transaction;
ch->hw.end_transaction = ata_ahci_end_transaction;
ch->hw.command = NULL; /* not used here */
ch->hw.softreset = ata_ahci_softreset;
ch->hw.pm_read = ata_ahci_pm_read;
ch->hw.pm_write = ata_ahci_pm_write;
MFp4: Introduce ATA_CAM kernel option, turning ata(4) controller drivers into cam(4) interface modules. When enabled, this options deprecates all ata(4) peripheral drivers (ad, acd, ...) and interfaces and allows cam(4) drivers (ada, cd, ...) and interfaces to be natively used instead. As side effect of this, ata(4) mode setting code was completely rewritten to make controller API more strict and permit above change. While doing this, SATA revision was separated from PATA mode. It allows DMA-incapable SATA devices to operate and makes hw.ata.atapi_dma tunable work again. Also allow ata(4) controller drivers (except some specific or broken ones) to handle larger data transfers. Previous constraint of 64K was artificial and is not really required by PCI ATA BM specification or hardware. Submitted by: nwitehorn (powerpc part)
2009-12-06 00:10:13 +00:00
ch->flags |= ATA_NO_SLAVE;
ch->flags |= ATA_SATA;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
ata_ahci_ch_resume(dev);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
return 0;
}
Change the way in which AHCI+PATA combined controllers, such as JMicron and Marvell handled. Instead of trying to attach two different drivers to single device, wrapping each call, make one of them (atajmicron, atamarvell) attach do device solely, but create child device for AHCI driver, passing it all required resources. It is quite easy, as none of resources are shared, except IRQ. As result, it: - makes drivers operation more independent and straitforward, - allows to use new ahci(4) driver with such devices, adding support for new features, such as PMP and NCQ, same time keeping legacy PATA support, - will allow to just drop old ataahci driver, when it's time come.
2009-11-16 15:38:27 +00:00
static int
Quite mechanical ch_detach implementations for all atapci subdrivers. Some dmainit call fixes for previous commit.
2009-02-19 00:32:55 +00:00
ata_ahci_ch_detach(device_t dev)
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
{
ata_ahci_ch_suspend(dev);
ata_dmafini(dev);
return (0);
}
Change the way in which AHCI+PATA combined controllers, such as JMicron and Marvell handled. Instead of trying to attach two different drivers to single device, wrapping each call, make one of them (atajmicron, atamarvell) attach do device solely, but create child device for AHCI driver, passing it all required resources. It is quite easy, as none of resources are shared, except IRQ. As result, it: - makes drivers operation more independent and straitforward, - allows to use new ahci(4) driver with such devices, adding support for new features, such as PMP and NCQ, same time keeping legacy PATA support, - will allow to just drop old ataahci driver, when it's time come.
2009-11-16 15:38:27 +00:00
static int
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
ata_ahci_ch_suspend(device_t dev)
Quite mechanical ch_detach implementations for all atapci subdrivers. Some dmainit call fixes for previous commit.
2009-02-19 00:32:55 +00:00
{
Disable port hardware on detach. First switch it to slumber mode to power-down peer transmitter, then disable completely. Side effect of this is saving about 0.5W of power per detached device.
2009-02-20 23:17:52 +00:00
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;
/* Disable port interrupts. */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_IE + offset, 0);
/* Reset command register. */
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
ata_ahci_stop(dev);
ata_ahci_stop_fr(dev);
Disable port hardware on detach. First switch it to slumber mode to power-down peer transmitter, then disable completely. Side effect of this is saving about 0.5W of power per detached device.
2009-02-20 23:17:52 +00:00
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset, 0);
/* Allow everything including partial and slumber modes. */
ATA_IDX_OUTL(ch, ATA_SCONTROL, 0);
/* Request slumber mode transition and give some time to get there. */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset, ATA_AHCI_P_CMD_SLUMBER);
DELAY(100);
/* Disable PHY. */
ATA_IDX_OUTL(ch, ATA_SCONTROL, ATA_SC_DET_DISABLE);
Quite mechanical ch_detach implementations for all atapci subdrivers. Some dmainit call fixes for previous commit.
2009-02-19 00:32:55 +00:00
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
return (0);
}
Change the way in which AHCI+PATA combined controllers, such as JMicron and Marvell handled. Instead of trying to attach two different drivers to single device, wrapping each call, make one of them (atajmicron, atamarvell) attach do device solely, but create child device for AHCI driver, passing it all required resources. It is quite easy, as none of resources are shared, except IRQ. As result, it: - makes drivers operation more independent and straitforward, - allows to use new ahci(4) driver with such devices, adding support for new features, such as PMP and NCQ, same time keeping legacy PATA support, - will allow to just drop old ataahci driver, when it's time come.
2009-11-16 15:38:27 +00:00
static int
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
ata_ahci_ch_resume(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
uint64_t work;
int offset = ch->unit << 7;
/* Disable port interrupts */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_IE + offset, 0);
/* setup work areas */
work = ch->dma.work_bus + ATA_AHCI_CL_OFFSET;
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CLB + offset, work & 0xffffffff);
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CLBU + offset, work >> 32);
work = ch->dma.work_bus + ATA_AHCI_FB_OFFSET;
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_FB + offset, work & 0xffffffff);
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_FBU + offset, work >> 32);
/* activate the channel and power/spin up device */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset,
Add experimental support for SATA interface power management. Feature is controlled by hint.ata.X.pm_level tunable: 0 - PM disabled, old behaviour, default. 1 - device is allowed to initiate PM state change, host is passive. 2 - host initiates PARTIAL state transition every time port is idle. 3 - host initiates SLUMBER state transition every time port is idle. PARTIAL state has up to 100us (50us for me) wakeup latency, but for my ICH8M saves 0.5W of power per drive. SLUMBER state has up to 10ms (3.5ms for me) wakeup latency, but saves 0.8W of power. Modes 2 and 3 are implemented only for AHCI driver now. Interface power management is incompatible with device presence detection (host receives no signal from drive, so unable to monitor it), so later is disabled when PM is used.
2009-04-29 21:17:18 +00:00
(ATA_AHCI_P_CMD_ACTIVE | ATA_AHCI_P_CMD_POD | ATA_AHCI_P_CMD_SUD |
((ch->pm_level > 1) ? ATA_AHCI_P_CMD_ALPE : 0) |
((ch->pm_level > 2) ? ATA_AHCI_P_CMD_ASP : 0 )));
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
ata_ahci_start_fr(dev);
ata_ahci_start(dev);
Quite mechanical ch_detach implementations for all atapci subdrivers. Some dmainit call fixes for previous commit.
2009-02-19 00:32:55 +00:00
return (0);
}
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
static int
ata_ahci_status(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 action = ATA_INL(ctlr->r_res2, ATA_AHCI_IS);
int offset = ch->unit << 7;
#define ATA_AHCI_STATBITS \
(ATA_AHCI_P_IX_IF|ATA_AHCI_P_IX_HBD|ATA_AHCI_P_IX_HBF|ATA_AHCI_P_IX_TFE)
if (action & (1 << ch->unit)) {
u_int32_t istatus = ATA_INL(ctlr->r_res2, ATA_AHCI_P_IS + offset);
u_int32_t cstatus = ATA_INL(ctlr->r_res2, ATA_AHCI_P_CI + offset);
/* clear interrupt(s) */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_IS + offset, istatus);
Remove unneeded AND operation. We have already checked that bit is set.
2009-02-15 21:54:16 +00:00
ATA_OUTL(ctlr->r_res2, ATA_AHCI_IS, 1 << ch->unit);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* do we have any PHY events ? */
if (istatus & (ATA_AHCI_P_IX_PRC | ATA_AHCI_P_IX_PC))
ata_sata_phy_check_events(dev);
/* do we have a potentially hanging engine to take care of? */
/* XXX SOS what todo on NCQ */
if ((istatus & ATA_AHCI_STATBITS) && (cstatus & 1)) {
u_int32_t cmd = ATA_INL(ctlr->r_res2, ATA_AHCI_P_CMD + offset);
int timeout = 0;
/* kill off all activity on this channel */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset,
cmd & ~(ATA_AHCI_P_CMD_FRE | ATA_AHCI_P_CMD_ST));
/* XXX SOS this is not entirely wrong */
do {
DELAY(1000);
if (timeout++ > 1000) {
device_printf(dev, "stopping AHCI engine failed\n");
break;
}
} while (ATA_INL(ctlr->r_res2,
ATA_AHCI_P_CMD + offset) & ATA_AHCI_P_CMD_CR);
/* start operations on this channel */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset,
cmd | (ATA_AHCI_P_CMD_FRE | ATA_AHCI_P_CMD_ST));
return 1;
}
else
/* XXX SOS what todo on NCQ */
return (!(cstatus & 1));
}
return 0;
}
/* must be called with ATA channel locked and state_mtx held */
static int
ata_ahci_begin_transaction(struct ata_request *request)
{
MFp4: - Remove most of direct relations between ATA(4) peripherial and controller levels. It makes logic more transparent and is a mandatory step to wrap ATA(4) controller level into ATA-native CAM SIM. - Tune AHCI and SATA2 SiI drivers memory allocation a bit to allow bigger I/O transaction sizes without additional cost.
2009-10-31 13:24:14 +00:00
struct ata_pci_controller *ctlr=device_get_softc(device_get_parent(request->parent));
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
struct ata_channel *ch = device_get_softc(request->parent);
struct ata_ahci_cmd_tab *ctp;
struct ata_ahci_cmd_list *clp;
int offset = ch->unit << 7;
MFp4: - Remove most of direct relations between ATA(4) peripherial and controller levels. It makes logic more transparent and is a mandatory step to wrap ATA(4) controller level into ATA-native CAM SIM. - Tune AHCI and SATA2 SiI drivers memory allocation a bit to allow bigger I/O transaction sizes without additional cost.
2009-10-31 13:24:14 +00:00
int port = request->unit & 0x0f;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
int entries = 0;
int fis_size;
/* get a piece of the workspace for this request */
ctp = (struct ata_ahci_cmd_tab *)
MFp4: - Remove most of direct relations between ATA(4) peripherial and controller levels. It makes logic more transparent and is a mandatory step to wrap ATA(4) controller level into ATA-native CAM SIM. - Tune AHCI and SATA2 SiI drivers memory allocation a bit to allow bigger I/O transaction sizes without additional cost.
2009-10-31 13:24:14 +00:00
(ch->dma.work + ATA_AHCI_CT_OFFSET);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* setup the FIS for this request */
if (!(fis_size = ata_ahci_setup_fis(ctp, request))) {
MFp4: - Remove most of direct relations between ATA(4) peripherial and controller levels. It makes logic more transparent and is a mandatory step to wrap ATA(4) controller level into ATA-native CAM SIM. - Tune AHCI and SATA2 SiI drivers memory allocation a bit to allow bigger I/O transaction sizes without additional cost.
2009-10-31 13:24:14 +00:00
device_printf(request->parent, "setting up SATA FIS failed\n");
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
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)) {
if (ch->dma.load(request, ctp->prd_tab, &entries)) {
MFp4: - Remove most of direct relations between ATA(4) peripherial and controller levels. It makes logic more transparent and is a mandatory step to wrap ATA(4) controller level into ATA-native CAM SIM. - Tune AHCI and SATA2 SiI drivers memory allocation a bit to allow bigger I/O transaction sizes without additional cost.
2009-10-31 13:24:14 +00:00
device_printf(request->parent, "setting up DMA failed\n");
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
request->result = EIO;
return ATA_OP_FINISHED;
}
}
/* setup the command list entry */
clp = (struct ata_ahci_cmd_list *)
MFp4: - Remove most of direct relations between ATA(4) peripherial and controller levels. It makes logic more transparent and is a mandatory step to wrap ATA(4) controller level into ATA-native CAM SIM. - Tune AHCI and SATA2 SiI drivers memory allocation a bit to allow bigger I/O transaction sizes without additional cost.
2009-10-31 13:24:14 +00:00
(ch->dma.work + ATA_AHCI_CL_OFFSET);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
clp->prd_length = entries;
clp->cmd_flags = (request->flags & ATA_R_WRITE ? ATA_AHCI_CMD_WRITE : 0) |
(request->flags & ATA_R_ATAPI ?
(ATA_AHCI_CMD_ATAPI | ATA_AHCI_CMD_PREFETCH) : 0) |
(fis_size / sizeof(u_int32_t)) |
(port << 12);
clp->bytecount = 0;
MFp4: - Remove most of direct relations between ATA(4) peripherial and controller levels. It makes logic more transparent and is a mandatory step to wrap ATA(4) controller level into ATA-native CAM SIM. - Tune AHCI and SATA2 SiI drivers memory allocation a bit to allow bigger I/O transaction sizes without additional cost.
2009-10-31 13:24:14 +00:00
clp->cmd_table_phys = htole64(ch->dma.work_bus + ATA_AHCI_CT_OFFSET);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* set command type bit */
if (request->flags & ATA_R_ATAPI)
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset,
ATA_INL(ctlr->r_res2, ATA_AHCI_P_CMD + offset) |
ATA_AHCI_P_CMD_ATAPI);
else
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset,
ATA_INL(ctlr->r_res2, ATA_AHCI_P_CMD + offset) &
~ATA_AHCI_P_CMD_ATAPI);
/* issue command to controller */
MFp4: - Remove most of direct relations between ATA(4) peripherial and controller levels. It makes logic more transparent and is a mandatory step to wrap ATA(4) controller level into ATA-native CAM SIM. - Tune AHCI and SATA2 SiI drivers memory allocation a bit to allow bigger I/O transaction sizes without additional cost.
2009-10-31 13:24:14 +00:00
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CI + offset, 1);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
if (!(request->flags & ATA_R_ATAPI)) {
/* device reset doesn't interrupt */
if (request->u.ata.command == ATA_DEVICE_RESET) {
u_int32_t tf_data;
int timeout = 1000000;
do {
DELAY(10);
tf_data = ATA_INL(ctlr->r_res2, ATA_AHCI_P_TFD + (ch->unit<<7));
} while ((tf_data & ATA_S_BUSY) && timeout--);
if (bootverbose)
device_printf(ch->dev, "device_reset timeout=%dus\n",
(1000000-timeout)*10);
request->status = tf_data;
if (request->status & ATA_S_ERROR)
request->error = tf_data >> 8;
return ATA_OP_FINISHED;
}
}
/* 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)
{
MFp4: - Remove most of direct relations between ATA(4) peripherial and controller levels. It makes logic more transparent and is a mandatory step to wrap ATA(4) controller level into ATA-native CAM SIM. - Tune AHCI and SATA2 SiI drivers memory allocation a bit to allow bigger I/O transaction sizes without additional cost.
2009-10-31 13:24:14 +00:00
struct ata_pci_controller *ctlr=device_get_softc(device_get_parent(request->parent));
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
struct ata_channel *ch = device_get_softc(request->parent);
struct ata_ahci_cmd_list *clp;
u_int32_t tf_data;
int offset = ch->unit << 7;
/* kill the timeout */
callout_stop(&request->callout);
/* get status */
tf_data = ATA_INL(ctlr->r_res2, ATA_AHCI_P_TFD + offset);
request->status = tf_data;
/* if error status get details */
if (request->status & ATA_S_ERROR)
request->error = tf_data >> 8;
/* on control commands read back registers to the request struct */
if (request->flags & ATA_R_CONTROL) {
u_int8_t *fis = ch->dma.work + ATA_AHCI_FB_OFFSET + 0x40;
request->u.ata.count = fis[12] | ((u_int16_t)fis[13] << 8);
request->u.ata.lba = fis[4] | ((u_int64_t)fis[5] << 8) |
((u_int64_t)fis[6] << 16);
MFp4: - Remove most of direct relations between ATA(4) peripherial and controller levels. It makes logic more transparent and is a mandatory step to wrap ATA(4) controller level into ATA-native CAM SIM. - Tune AHCI and SATA2 SiI drivers memory allocation a bit to allow bigger I/O transaction sizes without additional cost.
2009-10-31 13:24:14 +00:00
if (request->flags & ATA_R_48BIT)
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
request->u.ata.lba |= ((u_int64_t)fis[8] << 24) |
((u_int64_t)fis[9] << 32) |
((u_int64_t)fis[10] << 40);
else
request->u.ata.lba |= ((u_int64_t)(fis[7] & 0x0f) << 24);
}
/* record how much data we actually moved */
clp = (struct ata_ahci_cmd_list *)
MFp4: - Remove most of direct relations between ATA(4) peripherial and controller levels. It makes logic more transparent and is a mandatory step to wrap ATA(4) controller level into ATA-native CAM SIM. - Tune AHCI and SATA2 SiI drivers memory allocation a bit to allow bigger I/O transaction sizes without additional cost.
2009-10-31 13:24:14 +00:00
(ch->dma.work + ATA_AHCI_CL_OFFSET);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
request->donecount = clp->bytecount;
/* release SG list etc */
ch->dma.unload(request);
return ATA_OP_FINISHED;
}
static int
ata_ahci_issue_cmd(device_t dev, u_int16_t flags, int timeout)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
struct ata_ahci_cmd_list *clp =
(struct ata_ahci_cmd_list *)(ch->dma.work + ATA_AHCI_CL_OFFSET);
struct ata_ahci_cmd_tab *ctp =
(struct ata_ahci_cmd_tab *)(ch->dma.work + ATA_AHCI_CT_OFFSET);
u_int32_t status = 0;
int offset = ch->unit << 7;
int port = (ctp->cfis[1] & 0x0f);
int count;
clp->prd_length = 0;
clp->cmd_flags = (20 / sizeof(u_int32_t)) | flags | (port << 12);
clp->bytecount = 0;
clp->cmd_table_phys = htole64(ch->dma.work_bus + ATA_AHCI_CT_OFFSET);
/* issue command to controller */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CI + offset, 1);
/* poll for command finished */
for (count = 0; count < timeout; count++) {
DELAY(1000);
if (!((status = ATA_INL(ctlr->r_res2, ATA_AHCI_P_CI + offset)) & 1))
break;
}
/* clear interrupts */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_IS + offset,
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
ATA_INL(ctlr->r_res2, ATA_AHCI_P_IS + offset));
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
Tune AHCI verbose messages to make log more readable.
2009-02-22 14:08:20 +00:00
if (timeout && (count >= timeout)) {
if (bootverbose) {
device_printf(dev, "ahci_issue_cmd timeout: %d of %dms, status=%08x\n",
count, timeout, status);
}
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
return EIO;
Tune AHCI verbose messages to make log more readable.
2009-02-22 14:08:20 +00:00
}
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
return 0;
}
static int
ata_ahci_pm_read(device_t dev, int port, int reg, u_int32_t *result)
{
struct ata_channel *ch = device_get_softc(dev);
struct ata_ahci_cmd_tab *ctp =
(struct ata_ahci_cmd_tab *)(ch->dma.work + ATA_AHCI_CT_OFFSET);
u_int8_t *fis = ch->dma.work + ATA_AHCI_FB_OFFSET + 0x40;
bzero(ctp->cfis, 64);
ctp->cfis[0] = 0x27; /* host to device */
ctp->cfis[1] = 0x8f; /* command FIS to PM port */
ctp->cfis[2] = ATA_READ_PM;
ctp->cfis[3] = reg;
ctp->cfis[7] = port | ATA_D_LBA;
ctp->cfis[15] = ATA_A_4BIT;
if (ata_ahci_issue_cmd(dev, 0, 10)) {
device_printf(dev, "error reading PM port\n");
return EIO;
}
*result = fis[12] | (fis[4] << 8) | (fis[5] << 16) | (fis[6] << 24);
return 0;
}
static int
ata_ahci_pm_write(device_t dev, int port, int reg, u_int32_t value)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
struct ata_ahci_cmd_tab *ctp =
(struct ata_ahci_cmd_tab *)(ch->dma.work + ATA_AHCI_CT_OFFSET);
int offset = ch->unit << 7;
bzero(ctp->cfis, 64);
ctp->cfis[0] = 0x27; /* host to device */
ctp->cfis[1] = 0x8f; /* command FIS to PM port */
ctp->cfis[2] = ATA_WRITE_PM;
ctp->cfis[3] = reg;
ctp->cfis[7] = port | ATA_D_LBA;
ctp->cfis[12] = value & 0xff;
ctp->cfis[4] = (value >> 8) & 0xff;;
ctp->cfis[5] = (value >> 16) & 0xff;;
ctp->cfis[6] = (value >> 24) & 0xff;;
ctp->cfis[15] = ATA_A_4BIT;
if (ata_ahci_issue_cmd(dev, 0, 100)) {
device_printf(dev, "error writing PM port\n");
return ATA_E_ABORT;
}
return (ATA_INL(ctlr->r_res2, ATA_AHCI_P_TFD + offset) >> 8) & 0xff;
}
static void
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
ata_ahci_stop(device_t dev)
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
{
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;
int timeout;
/* 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,
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
cmd & ~ATA_AHCI_P_CMD_ST);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* XXX SOS this is not entirely wrong */
timeout = 0;
do {
DELAY(1000);
if (timeout++ > 1000) {
device_printf(dev, "stopping AHCI engine failed\n");
break;
}
}
while (ATA_INL(ctlr->r_res2, ATA_AHCI_P_CMD + offset) & ATA_AHCI_P_CMD_CR);
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
}
static void
ata_ahci_clo(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;
int timeout;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* issue Command List Override if supported */
MFp4. Log supported AHCI controller capabilities.
2009-06-01 21:42:26 +00:00
if (ATA_INL(ctlr->r_res2, ATA_AHCI_CAP) & ATA_AHCI_CAP_SCLO) {
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
cmd = ATA_INL(ctlr->r_res2, ATA_AHCI_P_CMD + offset);
cmd |= ATA_AHCI_P_CMD_CLO;
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset, cmd);
timeout = 0;
do {
DELAY(1000);
if (timeout++ > 1000) {
device_printf(dev, "executing CLO failed\n");
break;
}
}
while (ATA_INL(ctlr->r_res2, ATA_AHCI_P_CMD+offset)&ATA_AHCI_P_CMD_CLO);
}
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
}
static void
ata_ahci_start(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;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* clear SATA error register */
ATA_IDX_OUTL(ch, ATA_SERROR, ATA_IDX_INL(ch, ATA_SERROR));
/* 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 */
cmd = ATA_INL(ctlr->r_res2, ATA_AHCI_P_CMD + offset);
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_CMD + offset,
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
cmd | ATA_AHCI_P_CMD_ST |
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
(ch->devices & ATA_PORTMULTIPLIER ? ATA_AHCI_P_CMD_PMA : 0));
}
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
static void
ata_ahci_stop_fr(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;
int timeout;
/* 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_FRE);
timeout = 0;
do {
DELAY(1000);
if (timeout++ > 1000) {
device_printf(dev, "stopping AHCI FR engine failed\n");
break;
}
}
while (ATA_INL(ctlr->r_res2, ATA_AHCI_P_CMD + offset) & ATA_AHCI_P_CMD_FR);
}
static void
ata_ahci_start_fr(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;
/* start FIS reception 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_FRE);
}
Add workaround for some ATI chips, failing to soft-reset when port multiplicator supported, but absent.
2009-02-15 16:05:48 +00:00
static int
Increase initial busy waiting time. If drive was hot-plugged, it may need much time to spin-up.
2009-02-21 23:42:28 +00:00
ata_ahci_wait_ready(device_t dev, int t)
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
{
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;
int timeout = 0;
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
uint32_t val;
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
while ((val = ATA_INL(ctlr->r_res2, ATA_AHCI_P_TFD + offset)) &
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
(ATA_S_BUSY | ATA_S_DRQ)) {
DELAY(1000);
Increase initial busy waiting time. If drive was hot-plugged, it may need much time to spin-up.
2009-02-21 23:42:28 +00:00
if (timeout++ > t) {
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
device_printf(dev, "port is not ready (timeout %dms) tfd = %08x\n", t, val);
return (EBUSY);
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
}
}
if (bootverbose)
device_printf(dev, "ready wait time=%dms\n", timeout);
Add workaround for some ATI chips, failing to soft-reset when port multiplicator supported, but absent.
2009-02-15 16:05:48 +00:00
return (0);
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
}
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
static int
ata_ahci_hardreset(device_t dev, int port, uint32_t *signature)
{
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;
*signature = 0xffffffff;
ata_ahci_stop(dev);
/* Reset port */
if (!ata_sata_phy_reset(dev, port, 0))
return (ENOENT);
/* Wait for clearing busy status. */
if (ata_ahci_wait_ready(dev, 10000)) {
device_printf(dev, "hardware reset timeout\n");
return (EBUSY);
}
*signature = ATA_INL(ctlr->r_res2, ATA_AHCI_P_SIG + offset);
ata_ahci_start(dev);
return (0);
}
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
static u_int32_t
ata_ahci_softreset(device_t dev, int port)
{
struct ata_channel *ch = device_get_softc(dev);
struct ata_ahci_cmd_tab *ctp =
(struct ata_ahci_cmd_tab *)(ch->dma.work + ATA_AHCI_CT_OFFSET);
On Soft Reset, read device signature from FIS receive area, instead of PxSIG register. It works better for NVidia chipsets. ahci(4) does the same. PR: kern/140472, i386/138668
2009-12-05 10:30:54 +00:00
u_int8_t *fis = ch->dma.work + ATA_AHCI_FB_OFFSET + 0x40;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
Tune AHCI verbose messages to make log more readable.
2009-02-22 14:08:20 +00:00
if (bootverbose)
device_printf(dev, "software reset port %d...\n", port);
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
/* kick controller into sane state */
ata_ahci_stop(dev);
ata_ahci_clo(dev);
ata_ahci_start(dev);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* pull reset active */
bzero(ctp->cfis, 64);
ctp->cfis[0] = 0x27;
ctp->cfis[1] = port & 0x0f;
//ctp->cfis[7] = ATA_D_LBA | ATA_D_IBM;
ctp->cfis[15] = (ATA_A_4BIT | ATA_A_RESET);
Tune AHCI verbose messages to make log more readable.
2009-02-22 14:08:20 +00:00
if (ata_ahci_issue_cmd(dev, ATA_AHCI_CMD_RESET | ATA_AHCI_CMD_CLR_BUSY,100)) {
device_printf(dev, "software reset set timeout\n");
return (-1);
}
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
Reduce soft-reset assertion time from 5ms to 50us. ATA specification declares minimal reset time of 5us. SATA keeps it, but requires devices to handle commands transmitted even one by one without any gap.
2009-02-20 20:29:32 +00:00
ata_udelay(50);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
/* pull reset inactive -> device softreset */
bzero(ctp->cfis, 64);
ctp->cfis[0] = 0x27;
ctp->cfis[1] = port & 0x0f;
//ctp->cfis[7] = ATA_D_LBA | ATA_D_IBM;
ctp->cfis[15] = ATA_A_4BIT;
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
ata_ahci_issue_cmd(dev, 0, 3000);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
if (ata_ahci_wait_ready(dev, 0)) {
Tune AHCI verbose messages to make log more readable.
2009-02-22 14:08:20 +00:00
device_printf(dev, "software reset clear timeout\n");
Add workaround for some ATI chips, failing to soft-reset when port multiplicator supported, but absent.
2009-02-15 16:05:48 +00:00
return (-1);
Tune AHCI verbose messages to make log more readable.
2009-02-22 14:08:20 +00:00
}
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
On Soft Reset, read device signature from FIS receive area, instead of PxSIG register. It works better for NVidia chipsets. ahci(4) does the same. PR: kern/140472, i386/138668
2009-12-05 10:30:54 +00:00
return (((u_int32_t)fis[6] << 24) |
((u_int32_t)fis[5] << 16) |
((u_int32_t)fis[4] << 8) |
(u_int32_t)fis[12]);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
}
Change the way in which AHCI+PATA combined controllers, such as JMicron and Marvell handled. Instead of trying to attach two different drivers to single device, wrapping each call, make one of them (atajmicron, atamarvell) attach do device solely, but create child device for AHCI driver, passing it all required resources. It is quite easy, as none of resources are shared, except IRQ. As result, it: - makes drivers operation more independent and straitforward, - allows to use new ahci(4) driver with such devices, adding support for new features, such as PMP and NCQ, same time keeping legacy PATA support, - will allow to just drop old ataahci driver, when it's time come.
2009-11-16 15:38:27 +00:00
static void
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
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);
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
u_int32_t signature;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
int offset = ch->unit << 7;
Tune AHCI verbose messages to make log more readable.
2009-02-22 14:08:20 +00:00
if (bootverbose)
device_printf(dev, "AHCI reset...\n");
Disable port interrupts before doing PHY reset. PHY reset causes drive connect/disconnect events, unwanted at this moment.
2009-02-20 20:20:00 +00:00
/* Disable port interrupts */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_IE + offset, 0);
Integrate user/mav/ata branch: Add ch_suspend/ch_resume methods for PCI controllers and implement them for AHCI. Refactor AHCI channel initialization according to it. Fix Port Multipliers operation. It is far from perfect yet, but works now. Tested with JMicron JMB363 AHCI + SiI 3726 PMP pair. Previous version was also tested with SiI 4726 PMP. Hardware sponsored by: Vitsch Electronics / VEHosting.nl
2009-03-30 22:18:38 +00:00
if (ata_ahci_hardreset(dev, -1, &signature)) {
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
if (bootverbose)
Tune AHCI verbose messages to make log more readable.
2009-02-22 14:08:20 +00:00
device_printf(dev, "AHCI reset done: phy reset found no device\n");
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
ch->devices = 0;
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
/* enable wanted port interrupts */
ATA_OUTL(ctlr->r_res2, ATA_AHCI_P_IE + offset,
(ATA_AHCI_P_IX_CPD | ATA_AHCI_P_IX_PRC | ATA_AHCI_P_IX_PC));
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
return;
}
Tunes to AHCI reset sequences: - specification claims that 1 second is just a maximum controller reset time; implement controller reset properly to save almost 1 second of boot, and about half second of resume time; - enable channel interrupts only after channel status reset to fix duplicate device creation on resume due to unwanted device connection event; - as described in specification, wait for disk ready status after channel power-up; it is not so important when disk already touched by BIOS, but solves device not ready problems on resume and probably some other cases. - uncomment channel stop/start on soft-reset as it is declared mandatory by specification; it was commented due to some random drive detection problems on VIA and JMicron controllers, but I hope it is fixed by previous point.
2009-02-14 23:02:59 +00:00
/* 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 |
Add experimental support for SATA interface power management. Feature is controlled by hint.ata.X.pm_level tunable: 0 - PM disabled, old behaviour, default. 1 - device is allowed to initiate PM state change, host is passive. 2 - host initiates PARTIAL state transition every time port is idle. 3 - host initiates SLUMBER state transition every time port is idle. PARTIAL state has up to 100us (50us for me) wakeup latency, but for my ICH8M saves 0.5W of power per drive. SLUMBER state has up to 10ms (3.5ms for me) wakeup latency, but saves 0.8W of power. Modes 2 and 3 are implemented only for AHCI driver now. Interface power management is incompatible with device presence detection (host receives no signal from drive, so unable to monitor it), so later is disabled when PM is used.
2009-04-29 21:17:18 +00:00
((ch->pm_level == 0) ? ATA_AHCI_P_IX_PRC | ATA_AHCI_P_IX_PC : 0) |
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));
Disable PMP probing for Marvell AHCI controllers. It is not working for some reason. Linux does the same.
2009-11-14 08:04:38 +00:00
/*
* Only probe for PortMultiplier if HW has support.
* Ignore Marvell, which is not working,
*/
if ((ATA_INL(ctlr->r_res2, ATA_AHCI_CAP) & ATA_AHCI_CAP_SPM) &&
pci_get_vendor(ctlr->dev) != 0x11ab) {
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
signature = ata_ahci_softreset(dev, ATA_PM);
Add workaround for some ATI chips, failing to soft-reset when port multiplicator supported, but absent.
2009-02-15 16:05:48 +00:00
/* Workaround for some ATI chips, failing to soft-reset
* when port multiplicator supported, but absent.
* XXX: We can also check PxIS.IPMS==1 here to be sure. */
if (signature == 0xffffffff)
signature = ata_ahci_softreset(dev, 0);
} else {
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
signature = ata_ahci_softreset(dev, 0);
}
if (bootverbose)
device_printf(dev, "SIGNATURE: %08x\n", signature);
Use only higher half of device signature to identify device type. Some devices return incorrect values in lower part confusing detection, while higher part itself gives enough information for proper detetion.
2009-02-21 23:46:34 +00:00
switch (signature >> 16) {
case 0x0000:
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
ch->devices = ATA_ATA_MASTER;
break;
Use only higher half of device signature to identify device type. Some devices return incorrect values in lower part confusing detection, while higher part itself gives enough information for proper detetion.
2009-02-21 23:46:34 +00:00
case 0x9669:
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
ch->devices = ATA_PORTMULTIPLIER;
ata_pm_identify(dev);
break;
Use only higher half of device signature to identify device type. Some devices return incorrect values in lower part confusing detection, while higher part itself gives enough information for proper detetion.
2009-02-21 23:46:34 +00:00
case 0xeb14:
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
ch->devices = ATA_ATAPI_MASTER;
break;
default: /* SOS XXX */
if (bootverbose)
Fix spelling. Submitted by: gavin
2009-02-23 08:19:30 +00:00
device_printf(dev, "Unknown signature, assuming disk device\n");
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
ch->devices = ATA_ATA_MASTER;
}
if (bootverbose)
Tune AHCI verbose messages to make log more readable.
2009-02-22 14:08:20 +00:00
device_printf(dev, "AHCI reset done: devices=%08x\n", ch->devices);
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
}
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);
}
}
KASSERT(nsegs <= ATA_AHCI_DMA_ENTRIES, ("too many DMA segment entries\n"));
args->nsegs = nsegs;
}
Quite mechanical ch_detach implementations for all atapci subdrivers. Some dmainit call fixes for previous commit.
2009-02-19 00:32:55 +00:00
static void
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
ata_ahci_dmainit(device_t dev)
{
struct ata_pci_controller *ctlr = device_get_softc(device_get_parent(dev));
struct ata_channel *ch = device_get_softc(dev);
ata_dmainit(dev);
/* note start and stop are not used here */
ch->dma.setprd = ata_ahci_dmasetprd;
Fix kernel panic, when ataahci driver is used on system with increased MAXPHYS. Current ataahci driver memory allocation scheme includes only 64 items in DMA S/G table, and so not guarantied to support transactions with more then 252K data. Approved by: re (kensmith) MFC after: 2 weeks
2009-07-08 06:00:21 +00:00
ch->dma.max_iosize = (ATA_AHCI_DMA_ENTRIES - 1) * PAGE_SIZE;
This is the roumored ATA modulerisation works, and it needs a little explanation. If you just config KERNEL as usual there should be no apparent changes, you'll get all chipset support code compiled in. However there is now a way to only compile in code for chipsets needed on a pr vendor basis. ATA now has the following "device" entries: atacore: ATA core functionality, always needed for any ATA setup atacard: CARDBUS support atacbus: PC98 cbus support ataisa: ISA bus support atapci: PCI bus support only generic chipset support. ataahci: AHCI support, also pulled in by some vendor modules. ataacard, ataacerlabs, ataadaptec, ataamd, ataati, atacenatek, atacypress, atacyrix, atahighpoint, ataintel, ataite, atajmicron, atamarvell, atamicron, atanational, atanetcell, atanvidia, atapromise, ataserverworks, atasiliconimage, atasis, atavia; Vendor support, ie atavia for VIA chipsets atadisk: ATA disk driver ataraid: ATA softraid driver atapicd: ATAPI cd/dvd driver atapifd: ATAPI floppy/flashdisk driver atapist: ATAPI tape driver atausb: ATA<>USB bridge atapicam: ATA<>CAM bridge This makes it possible to config a kernel with just VIA chipset support by having the following ATA lines in the kernel config file: device atacore device atapci device atavia And then you need the atadisk, atapicd etc lines in there just as usual. If you use ATA as modules loaded at boot there is few changes except the rename of the "ata" module to "atacore", things looks just as usual. However under atapci you now have a whole bunch of vendor specific drivers, that you can kldload individually depending on you needs. Drivers have the same names as used in the kernel config explained above.
2008-10-09 12:56:57 +00:00
if (ATA_INL(ctlr->r_res2, ATA_AHCI_CAP) & ATA_AHCI_CAP_64BIT)
ch->dma.max_address = BUS_SPACE_MAXADDR;
}
static int
ata_ahci_setup_fis(struct ata_ahci_cmd_tab *ctp, struct ata_request *request)
{
bzero(ctp->cfis, 64);
if (request->flags & ATA_R_ATAPI) {
bzero(ctp->acmd, 32);
bcopy(request->u.atapi.ccb, ctp->acmd, 16);
}
return ata_request2fis_h2d(request, &ctp->cfis[0]);
}
ATA_DECLARE_DRIVER(ata_ahci);
Change the way in which AHCI+PATA combined controllers, such as JMicron and Marvell handled. Instead of trying to attach two different drivers to single device, wrapping each call, make one of them (atajmicron, atamarvell) attach do device solely, but create child device for AHCI driver, passing it all required resources. It is quite easy, as none of resources are shared, except IRQ. As result, it: - makes drivers operation more independent and straitforward, - allows to use new ahci(4) driver with such devices, adding support for new features, such as PMP and NCQ, same time keeping legacy PATA support, - will allow to just drop old ataahci driver, when it's time come.
2009-11-16 15:38:27 +00:00
static device_method_t ata_ahci_ata_methods[] = {
DEVMETHOD(device_probe, ata_ahci_ata_probe),
DEVMETHOD(device_attach, ata_ahci_ata_attach),
DEVMETHOD(device_detach, ata_pci_detach),
DEVMETHOD(device_suspend, ata_pci_suspend),
DEVMETHOD(device_resume, ata_pci_resume),
DEVMETHOD(device_shutdown, bus_generic_shutdown),
DEVMETHOD(bus_read_ivar, ata_pci_read_ivar),
DEVMETHOD(bus_write_ivar, ata_pci_write_ivar),
DEVMETHOD(bus_alloc_resource, ata_pci_alloc_resource),
DEVMETHOD(bus_release_resource, ata_pci_release_resource),
DEVMETHOD(bus_activate_resource, bus_generic_activate_resource),
DEVMETHOD(bus_deactivate_resource, bus_generic_deactivate_resource),
DEVMETHOD(bus_setup_intr, ata_pci_setup_intr),
DEVMETHOD(bus_teardown_intr, ata_pci_teardown_intr),
{ 0, 0 }
};
static driver_t ata_ahci_ata_driver = {
"atapci",
ata_ahci_ata_methods,
sizeof(struct ata_pci_controller)
};
DRIVER_MODULE(ata_ahci_ata, atapci, ata_ahci_ata_driver, ata_pci_devclass, 0, 0);
Reference in New Issue Copy Permalink