d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
2c8472682c
freebsd-dev/sys/cam/scsi/scsi_da.c

2920 lines
76 KiB
C
Raw Normal View History

Start each of the license/copyright comments with /*-
2005-01-05 22:34:37 +00:00
/*-
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
* Implementation of SCSI Direct Access Peripheral driver for CAM.
*
* Copyright (c) 1997 Justin T. Gibbs.
* 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. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``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 OR CONTRIBUTORS BE LIABLE FOR
* ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
Use __FBSDID().
2003-06-10 18:14:05 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
#include <sys/param.h>
Implement a new camcontrol function, 'camcontrol format'. libcam/Makefile: Add scsi_da.c to libcam for the new scsi_format_unit() function. camcontrol.8: Update the man page for the new format functionality, and take out the examples section describing how to do it with 'camcontrol cmd'. camcontrol.c: New format functionality. Note that unlike the rest of the camcontrol subcommands, this one is interactive by default. Because of the potential destructiveness of the format command, I thought it necessary to get confirmation from the user before spamming a disk. You can disable the interactive behavior, and the status meter with command line arguments. scsi_da.c: Add the new scsi_format_unit() cdb building function and use #ifdef _KERNEL to make this file compile in both the kernel and userland. The format unit function is currently only defined in the non-kernel case, because nothing in the kernel is using it. If that changes, it should be un-ifdefed and compiled in both cases. scsi_da.h: New function declaration, CDB structure and format data structures. Thanks to Nick Hibma for providing some valuable input on these changes.
2000-05-21 23:57:52 +00:00
#ifdef _KERNEL
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
#include <sys/systm.h>
#include <sys/kernel.h>
Separate the struct bio related stuff out of <sys/buf.h> into <sys/bio.h>. <sys/bio.h> is now a prerequisite for <sys/buf.h> but it shall not be made a nested include according to bdes teachings on the subject of nested includes. Diskdrivers and similar stuff below specfs::strategy() should no longer need to include <sys/buf.> unless they need caching of data. Still a few bogus uses of struct buf to track down. Repocopy by: peter
2000-05-05 09:59:14 +00:00
#include <sys/bio.h>
Add SYSCTL ints for default normal I/O timeout && retry counts. This is useful if you want to dynamically move into a Fibre Channel or Multi-initiator environment that happens to be particularly noisy and ugly that requires a lot of retries (with shorter I/O timeouts) for commands destried by LIPs or Bus Resets. Reviewed by: deafening silence on audit && scsi on the retry counts MFC after: 2 weeks
2001-07-09 19:18:00 +00:00
#include <sys/sysctl.h>
Unbreak buildworld. sys/taskqueue.h is a kernel-only include. Pointy Hat to: ken
2003-09-03 12:31:03 +00:00
#include <sys/taskqueue.h>
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
#include <sys/lock.h>
#include <sys/mutex.h>
Introduce the disk mini-layer and devstat_end_transaction_buf() in cam/scsi. Somewhat reviewed by: ken
1999-10-01 09:34:10 +00:00
#include <sys/conf.h>
Remove obsolete headers. Move kernel specific headers under #ifdefs
2007-04-23 18:01:44 +00:00
#include <sys/devicestat.h>
Implement a new generic mechanism for attaching handler functions to events, in order to pave the way for removing a number of the ad-hoc implementations currently in use. Retire the at_shutdown family of functions and replace them with new event handler lists. Rework kern_shutdown.c to take greater advantage of the use of event handlers. Reviewed by: green
1999-08-21 06:24:40 +00:00
#include <sys/eventhandler.h>
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
#include <sys/malloc.h>
Merge the cons.c and cons.h to the best of my ability. alpha may or may not compile, I can't test it.
1999-08-09 10:35:05 +00:00
#include <sys/cons.h>
Plumb device physical path reporting from CAM devices, through GEOM and DEVFS, and make it accessible via the diskinfo utility. Extend GEOM's generic attribute query mechanism into generic disk consumers. sys/geom/geom_disk.c: sys/geom/geom_disk.h: sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Allow disk providers to implement a new method which can override the default BIO_GETATTR response, d_getattr(struct bio *). This function returns -1 if not handled, otherwise it returns 0 or an errno to be passed to g_io_deliver(). sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Don't copy the serial number to dp->d_ident anymore, as the CAM XPT is now responsible for returning this information via d_getattr()->(a)dagetattr()->xpt_getatr(). sys/geom/geom_dev.c: - Implement a new ioctl, DIOCGPHYSPATH, which returns the GEOM attribute "GEOM::physpath", if possible. If the attribute request returns a zero-length string, ENOENT is returned. usr.sbin/diskinfo/diskinfo.c: - If the DIOCGPHYSPATH ioctl is successful, report physical path data when diskinfo is executed with the '-v' option. Submitted by: will Reviewed by: gibbs Sponsored by: Spectra Logic Corporation Add generic attribute change notification support to GEOM. sys/sys/geom/geom.h: Add a new attrchanged method field to both g_class and g_geom. sys/sys/geom/geom.h: sys/geom/geom_event.c: - Provide the g_attr_changed() function that providers can use to advertise attribute changes. - Perform delivery of attribute change notifications from a thread context via the standard GEOM event mechanism. sys/geom/geom_subr.c: Inherit the attrchanged method from class to geom (class instance). sys/geom/geom_disk.c: Provide disk_attr_changed() to provide g_attr_changed() access to consumers of the disk API. sys/cam/scsi/scsi_pass.c: sys/cam/scsi/scsi_da.c: sys/geom/geom_dev.c: sys/geom/geom_disk.c: Use attribute changed events to track updates to physical path information. sys/cam/scsi/scsi_da.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, and the updated buffer type references our physical path attribute, emit a GEOM attribute changed event via the disk_attr_changed() API. sys/cam/scsi/scsi_pass.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, update the physical patch devfs alias for this pass instance. Submitted by: gibbs Sponsored by: Spectra Logic Corporation
2011-06-14 17:10:32 +00:00
#include <geom/geom.h>
Use bioq_flush() to drain a bio queue with a specific error code. Retain the mistake of not updating the devstat API for now. Spell bioq_disksort() consistently with the remaining bioq_*(). #include <geom/geom_disk.h> where this is more appropriate.
2003-04-01 15:06:26 +00:00
#include <geom/geom_disk.h>
Remove obsolete headers. Move kernel specific headers under #ifdefs
2007-04-23 18:01:44 +00:00
#endif /* _KERNEL */
Use bioq_flush() to drain a bio queue with a specific error code. Retain the mistake of not updating the devstat API for now. Spell bioq_disksort() consistently with the remaining bioq_*(). #include <geom/geom_disk.h> where this is more appropriate.
2003-04-01 15:06:26 +00:00
Implement a new camcontrol function, 'camcontrol format'. libcam/Makefile: Add scsi_da.c to libcam for the new scsi_format_unit() function. camcontrol.8: Update the man page for the new format functionality, and take out the examples section describing how to do it with 'camcontrol cmd'. camcontrol.c: New format functionality. Note that unlike the rest of the camcontrol subcommands, this one is interactive by default. Because of the potential destructiveness of the format command, I thought it necessary to get confirmation from the user before spamming a disk. You can disable the interactive behavior, and the status meter with command line arguments. scsi_da.c: Add the new scsi_format_unit() cdb building function and use #ifdef _KERNEL to make this file compile in both the kernel and userland. The format unit function is currently only defined in the non-kernel case, because nothing in the kernel is using it. If that changes, it should be un-ifdefed and compiled in both cases. scsi_da.h: New function declaration, CDB structure and format data structures. Thanks to Nick Hibma for providing some valuable input on these changes.
2000-05-21 23:57:52 +00:00
#ifndef _KERNEL
#include <stdio.h>
#include <string.h>
#endif /* _KERNEL */
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
#include <cam/cam.h>
#include <cam/cam_ccb.h>
#include <cam/cam_periph.h>
#include <cam/cam_xpt_periph.h>
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
#include <cam/cam_sim.h>
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
#include <cam/scsi/scsi_message.h>
Implement a new camcontrol function, 'camcontrol format'. libcam/Makefile: Add scsi_da.c to libcam for the new scsi_format_unit() function. camcontrol.8: Update the man page for the new format functionality, and take out the examples section describing how to do it with 'camcontrol cmd'. camcontrol.c: New format functionality. Note that unlike the rest of the camcontrol subcommands, this one is interactive by default. Because of the potential destructiveness of the format command, I thought it necessary to get confirmation from the user before spamming a disk. You can disable the interactive behavior, and the status meter with command line arguments. scsi_da.c: Add the new scsi_format_unit() cdb building function and use #ifdef _KERNEL to make this file compile in both the kernel and userland. The format unit function is currently only defined in the non-kernel case, because nothing in the kernel is using it. If that changes, it should be un-ifdefed and compiled in both cases. scsi_da.h: New function declaration, CDB structure and format data structures. Thanks to Nick Hibma for providing some valuable input on these changes.
2000-05-21 23:57:52 +00:00
#ifndef _KERNEL
#include <cam/scsi/scsi_da.h>
#endif /* !_KERNEL */
#ifdef _KERNEL
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
typedef enum {
DA_STATE_PROBE,
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
DA_STATE_PROBE2,
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
DA_STATE_NORMAL
} da_state;
typedef enum {
DA_FLAG_PACK_INVALID = 0x001,
DA_FLAG_NEW_PACK = 0x002,
DA_FLAG_PACK_LOCKED = 0x004,
DA_FLAG_PACK_REMOVABLE = 0x008,
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
DA_FLAG_SAW_MEDIA = 0x010,
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
DA_FLAG_NEED_OTAG = 0x020,
DA_FLAG_WENT_IDLE = 0x040,
DA_FLAG_RETRY_UA = 0x080,
(dacleanup, dasysctlinit): Defend against calling sysctl_ctx_free on an uninitialized sysctl_ctx, using flag DA_FLAG_SCTX_INIT. This prevents a panic encoutered with some umass units that probe correctly but fail to attach. Same problem, and same fix, as scsi_cd.c rev. 1.86. Reviewed by: njl, ken
2003-10-08 07:12:30 +00:00
DA_FLAG_OPEN = 0x100,
Use READ CAPACITY(16) to get information about device physical sectors. As soon as not all devices support READ CAPACITY(16), automatically fall back to READ CAPACITY(10) if CAM_REQ_INVALID or SSD_KEY_ILLEGAL_REQUEST status returned. It also provides first bits of information about Logical Block Provisioning (aka UNMAP/TRIM) support by the device.
2011-12-23 19:12:02 +00:00
DA_FLAG_SCTX_INIT = 0x200,
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
DA_FLAG_CAN_RC16 = 0x400,
DA_FLAG_PROBED = 0x800
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
} da_flags;
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
typedef enum {
DA_Q_NONE = 0x00,
"Fix" a problem with the Quantum Viking. It appears that this drive does not like the 6-byte read and write commands! It returns illegal request, with the field pointer pointing to byte 9 of a 6 byte CDB. In any case, the work around is to put in a quirk mechanism that makes sure that we don't send 6-byte reads or writes to this device. It's rather sad that this is necessary. You'd think that they would be able to get something that basic to work right in their firmware... Reviewed by: gibbs Reported by: Adam McDougall <bsdx@spawnet.com>
1998-12-02 17:35:28 +00:00
DA_Q_NO_SYNC_CACHE = 0x01,
Add the DA_Q_NO_PREVENT quirk which keeps da(4) from sending PREVENT/ALLOW commands. Add a quirk for the Creative Nomad MuVo USB device that uses it as well as NO_SYNCHRONIZE_CACHE. PR: kern/53094 Submitted by: Richard Nyberg <rnyberg@it.su.se> MFC after: 3 days
2003-08-22 16:35:53 +00:00
DA_Q_NO_6_BYTE = 0x02,
Merge to da driver quirks hinting 4K physical sector sizes for SATA disks connected via SAS or USB. Unluckily I've found that SAS (mps) and USB-SATA I have translate models in different ways, requiring twice more quirks. Unluckily for Hitachi, their model names are trimmed on SAS, making impossible to identify 4K sector drives that way.
2011-12-23 00:10:17 +00:00
DA_Q_NO_PREVENT = 0x04,
DA_Q_4K = 0x08
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
} da_quirks;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
typedef enum {
DA_CCB_PROBE = 0x01,
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
DA_CCB_PROBE2 = 0x02,
DA_CCB_BUFFER_IO = 0x03,
DA_CCB_WAITING = 0x04,
DA_CCB_DUMP = 0x05,
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
DA_CCB_DELETE = 0x06,
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
DA_CCB_TUR = 0x07,
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
DA_CCB_TYPE_MASK = 0x0F,
DA_CCB_RETRY_UA = 0x10
} da_ccb_state;
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
typedef enum {
DA_DELETE_NONE,
DA_DELETE_DISABLE,
DA_DELETE_ZERO,
DA_DELETE_WS10,
DA_DELETE_WS16,
DA_DELETE_UNMAP,
DA_DELETE_MAX = DA_DELETE_UNMAP
} da_delete_methods;
static const char *da_delete_method_names[] =
{ "NONE", "DISABLE", "ZERO", "WS10", "WS16", "UNMAP" };
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/* Offsets into our private area for storing information */
#define ccb_state ppriv_field0
#define ccb_bp ppriv_ptr1
struct disk_params {
u_int8_t heads;
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
u_int32_t cylinders;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
u_int8_t secs_per_track;
u_int32_t secsize; /* Number of bytes/sector */
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
u_int64_t sectors; /* total number sectors */
Merge to da driver quirks hinting 4K physical sector sizes for SATA disks connected via SAS or USB. Unluckily I've found that SAS (mps) and USB-SATA I have translate models in different ways, requiring twice more quirks. Unluckily for Hitachi, their model names are trimmed on SAS, making impossible to identify 4K sector drives that way.
2011-12-23 00:10:17 +00:00
u_int stripesize;
u_int stripeoffset;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
};
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
#define UNMAP_MAX_RANGES 512
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
struct da_softc {
Complete the bio/buf divorce for all code below devfs::strategy Exceptions: Vinum untouched. This means that it cannot be compiled. Greg Lehey is on the case. CCD not converted yet, casts to struct buf (still safe) atapi-cd casts to struct buf to examine B_PHYS
2000-04-15 05:54:02 +00:00
struct bio_queue_head bio_queue;
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
struct bio_queue_head delete_queue;
struct bio_queue_head delete_run_queue;
Back out the previous change to the queue(3) interface. It was not discussed and should probably not happen. Requested by: msmith and others
2000-05-26 02:09:24 +00:00
SLIST_ENTRY(da_softc) links;
LIST_HEAD(, ccb_hdr) pending_ccbs;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
da_state state;
da_flags flags;
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
da_quirks quirks;
"Fix" a problem with the Quantum Viking. It appears that this drive does not like the 6-byte read and write commands! It returns illegal request, with the field pointer pointing to byte 9 of a 6 byte CDB. In any case, the work around is to put in a quirk mechanism that makes sure that we don't send 6-byte reads or writes to this device. It's rather sad that this is necessary. You'd think that they would be able to get something that basic to work right in their firmware... Reviewed by: gibbs Reported by: Adam McDougall <bsdx@spawnet.com>
1998-12-02 17:35:28 +00:00
int minimum_cmd_size;
Fix a race condition in CAM peripheral free handling, locking in the CAM XPT bus traversal code, and a number of other periph level issues. cam_periph.h, cam_periph.c: Modify cam_periph_acquire() to test the CAM_PERIPH_INVALID flag prior to allowing a reference count to be gained on a peripheral. Callers of this function will receive CAM_REQ_CMP_ERR status in the situation of attempting to reference an invalidated periph. This guarantees that a peripheral scheduled for a deferred free will not be accessed during its wait for destruction. Panic during attempts to drop a reference count on a peripheral that already has a zero reference count. In cam_periph_list(), use a local sbuf with SBUF_FIXEDLEN set so that mallocs do not occur while the xpt topology lock is held, regardless of the allocation policy of the passed in sbuf. Add a new routine, cam_periph_release_locked_buses(), that can be called when the caller already holds the CAM topology lock. Add some extra debugging for duplicate peripheral allocations in cam_periph_alloc(). Treat CAM_DEV_NOT_THERE much the same as a selection timeout (AC_LOST_DEVICE is emitted), but forgo retries. cam_xpt.c: Revamp the way the EDT traversal code does locking and reference counting. This was broken, since it assumed that the EDT would not change during traversal, but that assumption is no longer valid. So, to prevent devices from going away while we traverse the EDT, make sure we properly lock everything and hold references on devices that we are using. The two peripheral driver traversal routines should be examined. xptpdperiphtraverse() holds the topology lock for the entire time it runs. xptperiphtraverse() is now locked properly, but only holds the topology lock while it is traversing the list, and not while the traversal function is running. The bus locking code in xptbustraverse() should also be revisited at a later time, since it is complex and should probably be simplified. scsi_da.c: Pay attention to the return value from cam_periph_acquire(). Return 0 always from daclose() even if the disk is now gone. Add some rudimentary error injection support. scsi_sg.c: Fix reference counting in the sg(4) driver. The sg driver was calling cam_periph_release() on close, but never called cam_periph_acquire() (which increments the reference count) on open. The periph code correctly complained that the sg(4) driver was trying to decrement the refcount when it was already 0. Sponsored by: Spectra Logic MFC after: 2 weeks
2012-01-12 00:41:48 +00:00
int error_inject;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
int ordered_tag_count;
Don't use the devstat->busy_count for state decisions in the device drivers. Doing so imposes atomicity and locking constraints on the devstat API. By: ken
2003-03-15 11:00:56 +00:00
int outstanding_cmds;
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
int unmap_max_ranges;
int unmap_max_lba;
int delete_running;
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
int tur;
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
da_delete_methods delete_method;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
struct disk_params params;
Change the disk(9) API in order to make device removal more robust. Previously the "struct disk" were owned by the device driver and this gave us problems when the device disappared and the users of that device were not immediately disappearing. Now the struct disk is allocate with a new call, disk_alloc() and owned by geom_disk and just abandonned by the device driver when disk_create() is called. Unfortunately, this results in a ton of "s/\./->/" changes to device drivers. Since I'm doing the sweep anyway, a couple of other API improvements have been carried out at the same time: The Giant awareness flag has been flipped from DISKFLAG_NOGIANT to DISKFLAG_NEEDSGIANT A version number have been added to disk_create() so that we can detect, report and ignore binary drivers with old ABI in the future. Manual page update to follow shortly.
2004-02-18 21:36:53 +00:00
struct disk *disk;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
union ccb saved_ccb;
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
struct task sysctl_task;
Fix ATAPI/USB/Firewire CDROM drive handling in cd(4) and hopefully fix a number of related problems along the way. - Automatically detect CDROM drives that can't handle 6 byte mode sense and mode select, and adjust our command size accordingly. We have to handle this in the cd(4) driver (where the buffers are allocated), since the parameter list length is different for the 6 and 10 byte mode sense commands. - Remove MODE_SENSE and MODE_SELECT translation removed in ATAPICAM and in the umass(4) driver, since there's no way for that to work properly. - Add a quirk entry for CDROM drives that just hang when they get a 6 byte mode sense or mode select. The reason for the quirk must be documented in a PR, and all quirks must be approved by ken@FreeBSD.org. This is to make sure that we fully understand why each quirk is needed. Once the CAM_NEW_TRAN_CODE is finished, we should be able to remove any such quirks, since we'll know what protocol the drive speaks (SCSI, ATAPI, etc.) and therefore whether we should use 6 or 10 byte mode sense/select commands. - Change the way the da(4) handles the no_6_byte sysctl. There is now a per-drive sysctl to set the minimum command size for that particular disk. (Since you could have multiple disks with multiple requirements in one system.) - Loader tunable support for all the sysctls in the da(4) and cd(4) drivers. - Add a CDIOCCLOSE ioctl for cd(4) (bde pointed this out a long time ago). - Add a media validation routine (cdcheckmedia()) to the cd(4) driver, to fix some problems bde pointed out a long time ago. We now allow open() to succeed no matter what, but if we don't detect valid media, the user can only issue CDIOCCLOSE or CDIOCEJECT ioctls. - The media validation routine also reads the table of contents off the drive. We use the table of contents to implement the CDIOCPLAYTRACKS ioctl using the PLAY AUDIO MSF command. The PLAY AUDIO TRACK INDEX command that we previously used was deprecated after SCSI-2. It works in every SCSI CDROM I've tried, but doesn't seem to work on ATAPI CDROM drives. We still use the play audio track index command if we don't have a valid TOC, but I suppose it'll fail anyway in that case. - Add _len() versions of scsi_mode_sense() and scsi_mode_select() so that we can specify the minimum command length. - Fix a couple of formatting problems in the sense printing code. MFC after: 4 weeks
2003-02-21 06:19:38 +00:00
struct sysctl_ctx_list sysctl_ctx;
struct sysctl_oid *sysctl_tree;
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
struct callout sendordered_c;
Do a minor amount of stylifying. Also, get a Fibre Channel WWPN if one exists for a da unit and create a sysctl OID for it.
2010-06-07 17:41:34 +00:00
uint64_t wwpn;
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
uint8_t unmap_buf[UNMAP_MAX_RANGES * 16 + 8];
Add CAM infrastructure to allow reporting when a drive's long read capacity data changes. cam_ccb.h: Add a new advanced information type, CDAI_TYPE_RCAPLONG, for long read capacity data. cam_xpt_internal.h: Add a read capacity data pointer and length to struct cam_ed. cam_xpt.c: Free the read capacity buffer when a device goes away. While we're here, make sure we don't leak memory for other malloced fields in struct cam_ed. scsi_all.c: Update the scsi_read_capacity_16() to take a uint8_t * and a length instead of just a pointer to the parameter data structure. This will hopefully make this function somewhat immune to future changes in the parameter data. scsi_all.h: Add some extra bit definitions to struct scsi_read_capacity_data_long, and bump up the structure size to the full size specified by SBC-3. Change the prototype for scsi_read_capacity_16(). scsi_da.c: Register changes in read capacity data with the transport layer. This allows the transport layer to send out an async notification to interested parties. Update the dasetgeom() API. Use scsi_extract_sense_len() instead of scsi_extract_sense(). scsi_xpt.c: Add support for the new CDAI_TYPE_RCAPLONG advanced information type. Make sure we set the physpath pointer to NULL after freeing it. This allows blindly freeing it in the struct cam_ed destructor. sys/param.h: Bump __FreeBSD_version from 1000005 to 1000006 to make it easier for third party drivers to determine that the read capacity data async notification is available. camcontrol.c, mptutil/mpt_cam.c: Update these for the new scsi_read_capacity_16() argument structure. Sponsored by: Spectra Logic
2012-01-26 18:09:28 +00:00
struct scsi_read_capacity_data_long rcaplong;
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
struct callout mediapoll_c;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
};
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
struct da_quirk_entry {
struct scsi_inquiry_pattern inq_pat;
da_quirks quirks;
};
Add quirk entries from Andre Albsmeier to disable the sync cache command for the Quantum "MAVERICK 540S" and "LPS525S". Also, add common string variables, since we seem to have a few Quantum and Micropolis drives in here. Fix the 'quantum' variable usage in scsi_all.c that likely got broken when someone staticized things in cam_xpt.c. (That particular problem would cause Quantum Fireball ST drives to not get spun up if they were not already spinning.) Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>
2000-08-29 22:09:23 +00:00
static const char quantum[] = "QUANTUM";
static const char microp[] = "MICROP";
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
static struct da_quirk_entry da_quirk_table[] =
{
Sort quirks into sections.
2003-08-25 18:14:43 +00:00
/* SPI, FC devices */
Add quirk for Fujitsu M2513A MO drives. These drives hang at various operations due to the synchronize cache command. PR: kern/21674 Submitted by: W.Scholten <whs@xs4all.nl> MFC after: 1 week
2001-11-17 14:46:22 +00:00
{
/*
* Fujitsu M2513A MO drives.
* Tested devices: M2513A2 firmware versions 1200 & 1300.
* (dip switch selects whether T_DIRECT or T_OPTICAL device)
* Reported by: W.Scholten <whs@xs4all.nl>
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "FUJITSU", "M2513A", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
{
/* See above. */
{T_OPTICAL, SIP_MEDIA_REMOVABLE, "FUJITSU", "M2513A", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
{
Add quirk entries to disable the synchronize cache command for Micropolis 2217's (reported by Matthew Jacob in NetBSD PR kern/6027) and Fujitsu M2954's (reported by Tom Jackson). Some of the Fujitsus at least hang when they get a cache sync command. (Others just return illegal request.) Also, make error printing in dashutdown() a little more selective. Don't print any error when the sense key is illegal request. Drives that don't support the synchronize cache command usually return illegal request. Also, make sure the scsi status is check condition before going into scsi_sense_print(). Reviewed by: gibbs
1998-10-12 17:16:47 +00:00
/*
* This particular Fujitsu drive doesn't like the
* synchronize cache command.
* Reported by: Tom Jackson <toj@gorilla.net>
*/
{T_DIRECT, SIP_MEDIA_FIXED, "FUJITSU", "M2954*", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
{
/*
* This drive doesn't like the synchronize cache command
* either. Reported by: Matthew Jacob <mjacob@feral.com>
* in NetBSD PR kern/6027, August 24, 1998.
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
*/
Add quirk entries from Andre Albsmeier to disable the sync cache command for the Quantum "MAVERICK 540S" and "LPS525S". Also, add common string variables, since we seem to have a few Quantum and Micropolis drives in here. Fix the 'quantum' variable usage in scsi_all.c that likely got broken when someone staticized things in cam_xpt.c. (That particular problem would cause Quantum Fireball ST drives to not get spun up if they were not already spinning.) Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>
2000-08-29 22:09:23 +00:00
{T_DIRECT, SIP_MEDIA_FIXED, microp, "2217*", "*"},
Add quirk entries to disable the synchronize cache command for Micropolis 2217's (reported by Matthew Jacob in NetBSD PR kern/6027) and Fujitsu M2954's (reported by Tom Jackson). Some of the Fujitsus at least hang when they get a cache sync command. (Others just return illegal request.) Also, make error printing in dashutdown() a little more selective. Don't print any error when the sense key is illegal request. Drives that don't support the synchronize cache command usually return illegal request. Also, make sure the scsi status is check condition before going into scsi_sense_print(). Reviewed by: gibbs
1998-10-12 17:16:47 +00:00
/*quirks*/ DA_Q_NO_SYNC_CACHE
Disable cache syncs for a broken NEC drive. Reviewed by: gibbs Submitted by: Blaz Zupan <blaz@gold.amis.net>
1998-10-13 23:34:54 +00:00
},
Add a quirk that disables SYNCHRONIZE CACHE PR: 8882 Obtained from: Hellmuth Michaelis hm@kts.org
1999-01-05 20:43:41 +00:00
{
/*
* This drive doesn't like the synchronize cache command
* either. Reported by: Hellmuth Michaelis (hm@kts.org)
* (PR 8882).
*/
Add quirk entries from Andre Albsmeier to disable the sync cache command for the Quantum "MAVERICK 540S" and "LPS525S". Also, add common string variables, since we seem to have a few Quantum and Micropolis drives in here. Fix the 'quantum' variable usage in scsi_all.c that likely got broken when someone staticized things in cam_xpt.c. (That particular problem would cause Quantum Fireball ST drives to not get spun up if they were not already spinning.) Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>
2000-08-29 22:09:23 +00:00
{T_DIRECT, SIP_MEDIA_FIXED, microp, "2112*", "*"},
Add a quirk that disables SYNCHRONIZE CACHE PR: 8882 Obtained from: Hellmuth Michaelis hm@kts.org
1999-01-05 20:43:41 +00:00
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
Disable cache syncs for a broken NEC drive. Reviewed by: gibbs Submitted by: Blaz Zupan <blaz@gold.amis.net>
1998-10-13 23:34:54 +00:00
{
/*
* Doesn't like the synchronize cache command.
* Reported by: Blaz Zupan <blaz@gold.amis.net>
*/
{T_DIRECT, SIP_MEDIA_FIXED, "NEC", "D3847*", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
"Fix" a problem with the Quantum Viking. It appears that this drive does not like the 6-byte read and write commands! It returns illegal request, with the field pointer pointing to byte 9 of a 6 byte CDB. In any case, the work around is to put in a quirk mechanism that makes sure that we don't send 6-byte reads or writes to this device. It's rather sad that this is necessary. You'd think that they would be able to get something that basic to work right in their firmware... Reviewed by: gibbs Reported by: Adam McDougall <bsdx@spawnet.com>
1998-12-02 17:35:28 +00:00
},
Add quirk entries from Andre Albsmeier to disable the sync cache command for the Quantum "MAVERICK 540S" and "LPS525S". Also, add common string variables, since we seem to have a few Quantum and Micropolis drives in here. Fix the 'quantum' variable usage in scsi_all.c that likely got broken when someone staticized things in cam_xpt.c. (That particular problem would cause Quantum Fireball ST drives to not get spun up if they were not already spinning.) Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>
2000-08-29 22:09:23 +00:00
{
/*
* Doesn't like the synchronize cache command.
Take constructive advice from njl && reformat previously added quirks slightly.
2005-04-14 04:46:46 +00:00
* Reported by: Blaz Zupan <blaz@gold.amis.net>
Add quirk entries from Andre Albsmeier to disable the sync cache command for the Quantum "MAVERICK 540S" and "LPS525S". Also, add common string variables, since we seem to have a few Quantum and Micropolis drives in here. Fix the 'quantum' variable usage in scsi_all.c that likely got broken when someone staticized things in cam_xpt.c. (That particular problem would cause Quantum Fireball ST drives to not get spun up if they were not already spinning.) Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>
2000-08-29 22:09:23 +00:00
*/
{T_DIRECT, SIP_MEDIA_FIXED, quantum, "MAVERICK 540S", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
{
/*
* Doesn't like the synchronize cache command.
*/
{T_DIRECT, SIP_MEDIA_FIXED, quantum, "LPS525S", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
Applied conservative version of suggested quirk. PR: 57468 Submitted by: walter@pelissero.de MFC after: 1 week
2005-04-14 04:05:00 +00:00
{
/*
* Doesn't like the synchronize cache command.
Take constructive advice from njl && reformat previously added quirks slightly.
2005-04-14 04:46:46 +00:00
* Reported by: walter@pelissero.de
Applied conservative version of suggested quirk. PR: 57468 Submitted by: walter@pelissero.de MFC after: 1 week
2005-04-14 04:05:00 +00:00
*/
{T_DIRECT, SIP_MEDIA_FIXED, quantum, "LPS540S", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
"Fix" a problem with the Quantum Viking. It appears that this drive does not like the 6-byte read and write commands! It returns illegal request, with the field pointer pointing to byte 9 of a 6 byte CDB. In any case, the work around is to put in a quirk mechanism that makes sure that we don't send 6-byte reads or writes to this device. It's rather sad that this is necessary. You'd think that they would be able to get something that basic to work right in their firmware... Reviewed by: gibbs Reported by: Adam McDougall <bsdx@spawnet.com>
1998-12-02 17:35:28 +00:00
{
/*
* Doesn't work correctly with 6 byte reads/writes.
* Returns illegal request, and points to byte 9 of the
* 6-byte CDB.
* Reported by: Adam McDougall <bsdx@spawnet.com>
*/
Add quirk entries from Andre Albsmeier to disable the sync cache command for the Quantum "MAVERICK 540S" and "LPS525S". Also, add common string variables, since we seem to have a few Quantum and Micropolis drives in here. Fix the 'quantum' variable usage in scsi_all.c that likely got broken when someone staticized things in cam_xpt.c. (That particular problem would cause Quantum Fireball ST drives to not get spun up if they were not already spinning.) Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>
2000-08-29 22:09:23 +00:00
{T_DIRECT, SIP_MEDIA_FIXED, quantum, "VIKING 4*", "*"},
"Fix" a problem with the Quantum Viking. It appears that this drive does not like the 6-byte read and write commands! It returns illegal request, with the field pointer pointing to byte 9 of a 6 byte CDB. In any case, the work around is to put in a quirk mechanism that makes sure that we don't send 6-byte reads or writes to this device. It's rather sad that this is necessary. You'd think that they would be able to get something that basic to work right in their firmware... Reviewed by: gibbs Reported by: Adam McDougall <bsdx@spawnet.com>
1998-12-02 17:35:28 +00:00
/*quirks*/ DA_Q_NO_6_BYTE
},
{
Add the DA_Q_NO_PREVENT quirk which keeps da(4) from sending PREVENT/ALLOW commands. Add a quirk for the Creative Nomad MuVo USB device that uses it as well as NO_SYNCHRONIZE_CACHE. PR: kern/53094 Submitted by: Richard Nyberg <rnyberg@it.su.se> MFC after: 3 days
2003-08-22 16:35:53 +00:00
/* See above. */
Add quirk entries from Andre Albsmeier to disable the sync cache command for the Quantum "MAVERICK 540S" and "LPS525S". Also, add common string variables, since we seem to have a few Quantum and Micropolis drives in here. Fix the 'quantum' variable usage in scsi_all.c that likely got broken when someone staticized things in cam_xpt.c. (That particular problem would cause Quantum Fireball ST drives to not get spun up if they were not already spinning.) Submitted by: Andre Albsmeier <andre.albsmeier@mchp.siemens.de>
2000-08-29 22:09:23 +00:00
{T_DIRECT, SIP_MEDIA_FIXED, quantum, "VIKING 2*", "*"},
"Fix" a problem with the Quantum Viking. It appears that this drive does not like the 6-byte read and write commands! It returns illegal request, with the field pointer pointing to byte 9 of a 6 byte CDB. In any case, the work around is to put in a quirk mechanism that makes sure that we don't send 6-byte reads or writes to this device. It's rather sad that this is necessary. You'd think that they would be able to get something that basic to work right in their firmware... Reviewed by: gibbs Reported by: Adam McDougall <bsdx@spawnet.com>
1998-12-02 17:35:28 +00:00
/*quirks*/ DA_Q_NO_6_BYTE
Add a quirk entry for Y-E Data USB floppy drive. Driver follows in the next few days.
2000-03-15 22:44:03 +00:00
},
Apply quirk. PR: 57469 Submitted by: walter@pelissero.de MFC after: 1 week
2005-04-14 04:12:45 +00:00
{
Take constructive advice from njl && reformat previously added quirks slightly.
2005-04-14 04:46:46 +00:00
/*
* Doesn't like the synchronize cache command.
* Reported by: walter@pelissero.de
*/
Apply quirk. PR: 57469 Submitted by: walter@pelissero.de MFC after: 1 week
2005-04-14 04:12:45 +00:00
{T_DIRECT, SIP_MEDIA_FIXED, "CONNER", "CP3500*", "*"},
o Style(9): trim eol w/spaces, w/spaces to tabs, kill an empty line.
2006-11-26 17:15:54 +00:00
/*quirks*/ DA_Q_NO_SYNC_CACHE
Apply quirk. PR: 57469 Submitted by: walter@pelissero.de MFC after: 1 week
2005-04-14 04:12:45 +00:00
},
Move the ciss quirk to the right section, also update the comment stating that the controllers do not support SYNC CACHE since ciss only supports a small subset of the scsi spec.
2003-12-08 06:29:38 +00:00
{
/*
* The CISS RAID controllers do not support SYNC_CACHE
*/
{T_DIRECT, SIP_MEDIA_FIXED, "COMPAQ", "RAID*", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
Sort quirks into sections.
2003-08-25 18:14:43 +00:00
/* USB mass storage devices supported by umass(4) */
Add quirks for the EXATELECOM i-Bead mp3 player. PR: kern/51675 Submitted by: Nicolas Jombart <ecu@ipv42.net> MFC after: 3 days
2003-08-22 05:43:30 +00:00
{
/*
* EXATELECOM (Sigmatel) i-Bead 100/105 USB Flash MP3 Player
* PR: kern/51675
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "EXATEL", "i-BEAD10*", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
Quirk for Jungsoft NEXDISK USB flash key. Fails to mount without NO_SYNCHRONIZE_CACHE. PR: kern/54737 Submitted by: David Thiel <lx@redundancy.redundancy.org> MFC after: 3 days
2003-08-22 05:58:23 +00:00
{
/*
Add miscellaneous USB device quirks. PR: kern/53067 PR: kern/54737 PR: kern/54786 PR: kern/57046 MFC after: 1 day
2004-04-19 04:14:09 +00:00
* Power Quotient Int. (PQI) USB flash key
* PR: kern/53067
Quirk for Jungsoft NEXDISK USB flash key. Fails to mount without NO_SYNCHRONIZE_CACHE. PR: kern/54737 Submitted by: David Thiel <lx@redundancy.redundancy.org> MFC after: 3 days
2003-08-22 05:58:23 +00:00
*/
Add miscellaneous USB device quirks. PR: kern/53067 PR: kern/54737 PR: kern/54786 PR: kern/57046 MFC after: 1 day
2004-04-19 04:14:09 +00:00
{T_DIRECT, SIP_MEDIA_REMOVABLE, "Generic*", "USB Flash Disk*",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
o Style(9): trim eol w/spaces, w/spaces to tabs, kill an empty line.
2006-11-26 17:15:54 +00:00
},
Add the DA_Q_NO_PREVENT quirk which keeps da(4) from sending PREVENT/ALLOW commands. Add a quirk for the Creative Nomad MuVo USB device that uses it as well as NO_SYNCHRONIZE_CACHE. PR: kern/53094 Submitted by: Richard Nyberg <rnyberg@it.su.se> MFC after: 3 days
2003-08-22 16:35:53 +00:00
{
/*
* Creative Nomad MUVO mp3 player (USB)
* PR: kern/53094
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "CREATIVE", "NOMAD_MUVO", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE|DA_Q_NO_PREVENT
},
Add miscellaneous USB device quirks. PR: kern/53067 PR: kern/54737 PR: kern/54786 PR: kern/57046 MFC after: 1 day
2004-04-19 04:14:09 +00:00
{
/*
* Jungsoft NEXDISK USB flash key
* PR: kern/54737
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "JUNGSOFT", "NEXDISK*", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
{
/*
* FreeDik USB Mini Data Drive
* PR: kern/54786
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "FreeDik*", "Mini Data Drive",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
{
/*
* Sigmatel USB Flash MP3 Player
* PR: kern/57046
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "SigmaTel", "MSCN", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE|DA_Q_NO_PREVENT
},
Quirk for Neuros USB audio device. PR: kern/63645 Submitted by: Aron Stansvik <elvstone@osdever.net>
2004-05-02 19:24:23 +00:00
{
/*
* Neuros USB Digital Audio Computer
* PR: kern/63645
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "NEUROS", "dig. audio comp.",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Quirk for SEGRAND NP-900 USB MP3Player PR: kern/64563 Submitted by: Kunitada Kokubun <unix_grandy@yahoo.co.jp> MFC after: 1 week
2004-07-18 05:39:14 +00:00
{
/*
* SEAGRAND NP-900 MP3 Player
* PR: kern/64563
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "SEAGRAND", "NP-900*", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE|DA_Q_NO_PREVENT
},
Add support iRiver iFP MP3 player PR: kern/54881, i386/63941, kern/66124 Submitted by: Magnus <bsdhead.at.spray.dot.se@FreeBSD.org>, Dmitry Dyomin <old@old.com.ua>, Dmitry Sivachenko <mitya@demos.su> MFC after: 1 week
2004-08-08 09:08:37 +00:00
{
/*
* iRiver iFP MP3 player (with UMS Firmware)
* PR: kern/54881, i386/63941, kern/66124
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "iRiver", "iFP*", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
add support Frontier Labs NEX IA+ Digital Audio Player with USB CF card reader/writer PR: kern/70158 Submitted by: Bernd Strau. <no_bs@web.de> MFC after: 1 week
2004-08-12 23:17:09 +00:00
},
{
/*
* Frontier Labs NEX IA+ Digital Audio Player, rev 1.10/0.01
* PR: kern/70158
*/
Apply quirk suggested by submitter. PR: 75486 Submitted by: no_bs@web.de MFC after: 1 week
2005-04-14 03:59:48 +00:00
{T_DIRECT, SIP_MEDIA_REMOVABLE, "FL" , "Nex*", "*"},
add support Frontier Labs NEX IA+ Digital Audio Player with USB CF card reader/writer PR: kern/70158 Submitted by: Bernd Strau. <no_bs@web.de> MFC after: 1 week
2004-08-12 23:17:09 +00:00
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
Quirk for ZICPlay USB MP3 Player. PR: kern/75057 Submitted by: Aurelien Nephtali <aurelien.nephtali wanadoo.fr>
2005-04-14 04:31:48 +00:00
{
/*
* ZICPlay USB MP3 Player with FM
* PR: kern/75057
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "ACTIONS*" , "USB DISK*", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add quirk for TEAC USB floppy drives.
2005-05-05 18:48:41 +00:00
{
/*
* TEAC USB floppy mechanisms
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "TEAC" , "FD-05*", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add a quirk for my pen-drive.
2005-06-09 17:35:04 +00:00
{
/*
* Kingston DataTraveler II+ USB Pen-Drive.
* Reported by: Pawel Jakub Dawidek <pjd@FreeBSD.org>
*/
Style nit.
2005-12-19 03:43:48 +00:00
{T_DIRECT, SIP_MEDIA_REMOVABLE, "Kingston" , "DataTraveler II+",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
Add a quirk for my pen-drive.
2005-06-09 17:35:04 +00:00
},
Add support for the USB DISK Pro PMAP. This patch has sit for 6 years in the PR database. PR: usb/96381 Submitted by: jhs Reviewed by: mav Approved by: cperciva (implicit) MFC after: 3 days
2012-10-15 12:33:53 +00:00
{
/*
* USB DISK Pro PMAP
* Reported by: jhs
* PR: usb/96381
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, " ", "USB DISK Pro", "PMAP"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
Allow to use TransFlash drive, which can be found in Motorola E398 Mobile Phone. PR: usb/89889 Submitted by: Wojciech A. Koszek <dunstan@freebsd.czest.pl> MFC after: 1 week
2005-12-19 03:27:28 +00:00
{
/*
* Motorola E398 Mobile Phone (TransFlash memory card).
* Reported by: Wojciech A. Koszek <dunstan@FreeBSD.czest.pl>
* PR: usb/89889
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "Motorola" , "Motorola Phone",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Allow to use Time DPA20B 1GB MP3 Player [1] Allow to use Qware BeatZkey! Pro [2] PR: usb/81846 [1], usb/79164 [2] Submitted by: Ruben de Groot <rdg@bzerk.org> [1], Radek Kozlowski <radek@raadradd.com> [2] Approved by: njl MFC: 1 week
2006-01-13 11:33:40 +00:00
{
/*
* Qware BeatZkey! Pro
* PR: usb/79164
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "GENERIC", "USB DISK DEVICE",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
{
/*
* Time DPA20B 1GB MP3 Player
* PR: usb/81846
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "USB2.0*", "(FS) FLASH DISK*",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add Product IDs : - Sandisk Cruzer Micro 128MB [5] - DANE-ELEC zMate 512MB USB flash drive [7] - Attache 256MB USB 2.0 Flash Drive [8] - Sandisk Cruzer Micro 256MB [9] Add scsi_da.c quirks : - Samsung USB key 128Mb [1] - Kingston DataTraveler 2.0 [2] - Creative MuVo Slim [3] - United MP 5512 Portable MP3 Player [4] - Sandisk Cruzer Micro 128MB [5], [9] - PNY USB Flash keys [6], [7], [8] Add umass.c quirks : - Sandisk Cruzer Micro 128MB [5] - DANE-ELEC zMate 512MB USB flash drive [7] - Attache 256MB USB 2.0 Flash Drive [8] - Sandisk Cruzer Micro 256MB [9] PR: usb/90081 [1], usb/89196 [2], kern/86131 [3], usb/80487 [4], usb/75970 [5], usb/75578 [6], usb/72344 [7], usb/65436 [8], usb/70835 [9] Submitted by: Henri-Pierre CHARLES <hpc@prism.uvsq.fr> [1], Andrey Simonenko <simon@comsys.ntu-kpi.kiev.ua> [2], Erich Rickheit KSC <rickheit-fbp@numachi.com> [3], tnu@chania.di.uoa.gr [4], Bram Abbekerk <bram@abbekerk.demon.nl> [5], Thomas Pornin <pornin@bolet.org> [6], parv <parv@pair.com> [7], Peter D. Quilty <pdquilty@adelphia.net> [8], Raymundo M. Vega <rvega@ms-smtp-03-eri0.socal.rr.com> [9] Approved by: iedowse MFC after: 3 days
2006-01-26 00:35:54 +00:00
{
/*
* Samsung USB key 128Mb
* PR: usb/90081
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "USB-DISK", "FreeDik-FlashUsb",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
{
/*
* Kingston DataTraveler 2.0 USB Flash memory.
* PR: usb/89196
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "Kingston", "DataTraveler 2.0",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
{
/*
* Creative MUVO Slim mp3 player (USB)
* PR: usb/86131
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "CREATIVE", "MuVo Slim",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE|DA_Q_NO_PREVENT
},
{
/*
* United MP5512 Portable MP3 Player (2-in-1 USB DISK/MP3)
* PR: usb/80487
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "Generic*", "MUSIC DISK",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
{
/*
* SanDisk Micro Cruzer 128MB
* PR: usb/75970
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "SanDisk" , "Micro Cruzer",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add the TOSHIBA TransMemory USB sticks to the list of devices that hate "Synchronize cache" commands. MFC after: 1 day
2006-03-13 22:26:33 +00:00
{
/*
* TOSHIBA TransMemory USB sticks
Add reference to PR to TOSHIBA TransMemory quirk entry.
2006-03-18 21:13:14 +00:00
* PR: kern/94660
Add the TOSHIBA TransMemory USB sticks to the list of devices that hate "Synchronize cache" commands. MFC after: 1 day
2006-03-13 22:26:33 +00:00
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "TOSHIBA", "TransMemory",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add Product IDs : - Sandisk Cruzer Micro 128MB [5] - DANE-ELEC zMate 512MB USB flash drive [7] - Attache 256MB USB 2.0 Flash Drive [8] - Sandisk Cruzer Micro 256MB [9] Add scsi_da.c quirks : - Samsung USB key 128Mb [1] - Kingston DataTraveler 2.0 [2] - Creative MuVo Slim [3] - United MP 5512 Portable MP3 Player [4] - Sandisk Cruzer Micro 128MB [5], [9] - PNY USB Flash keys [6], [7], [8] Add umass.c quirks : - Sandisk Cruzer Micro 128MB [5] - DANE-ELEC zMate 512MB USB flash drive [7] - Attache 256MB USB 2.0 Flash Drive [8] - Sandisk Cruzer Micro 256MB [9] PR: usb/90081 [1], usb/89196 [2], kern/86131 [3], usb/80487 [4], usb/75970 [5], usb/75578 [6], usb/72344 [7], usb/65436 [8], usb/70835 [9] Submitted by: Henri-Pierre CHARLES <hpc@prism.uvsq.fr> [1], Andrey Simonenko <simon@comsys.ntu-kpi.kiev.ua> [2], Erich Rickheit KSC <rickheit-fbp@numachi.com> [3], tnu@chania.di.uoa.gr [4], Bram Abbekerk <bram@abbekerk.demon.nl> [5], Thomas Pornin <pornin@bolet.org> [6], parv <parv@pair.com> [7], Peter D. Quilty <pdquilty@adelphia.net> [8], Raymundo M. Vega <rvega@ms-smtp-03-eri0.socal.rr.com> [9] Approved by: iedowse MFC after: 3 days
2006-01-26 00:35:54 +00:00
{
/*
* PNY USB Flash keys
* PR: usb/75578, usb/72344, usb/65436
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "*" , "USB DISK*",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
- Add a scsi_da.c and a umass.c quirk for Genesys 6-in-1 Card Reader. Reported by: anders on freebsd-usb@ Tested by: anders Approved by: ssouhlal MFC after: 3 days
2006-01-30 20:27:44 +00:00
{
/*
* Genesys 6-in-1 Card Reader
- Add a PR number for future reference. Approved by: njl
2006-03-18 18:28:13 +00:00
* PR: usb/94647
- Add a scsi_da.c and a umass.c quirk for Genesys 6-in-1 Card Reader. Reported by: anders on freebsd-usb@ Tested by: anders Approved by: ssouhlal MFC after: 3 days
2006-01-30 20:27:44 +00:00
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "Generic*", "STORAGE DEVICE*",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add a SCSI da(4) quirk for Rekam USB Camera. PR: usb/98713 Submitted by: Alex <goo@t72.ru> MFC after: 3 days
2006-10-07 19:04:23 +00:00
{
/*
* Rekam Digital CAMERA
* PR: usb/98713
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "CAMERA*", "4MP-9J6*",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add SCSI da(4) quirk for the iRiver H10. PR: usb/102547 Submitted by: Sven Esbjerg <freebsd-pr@xbsd.net> MFC after: 3 days
2006-10-07 18:28:55 +00:00
{
/*
* iRiver H10 MP3 player
* PR: usb/102547
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "iriver", "H10*",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
o Style(9): trim eol w/spaces, w/spaces to tabs, kill an empty line.
2006-11-26 17:15:54 +00:00
},
Support for iRiver U10 USB connection PR: 92306 Submitted by: Soren Dossing Approved by: re (blanket)
2007-06-28 06:12:55 +00:00
{
/*
* iRiver U10 MP3 player
* PR: usb/92306
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "iriver", "U10*",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
o Add a quirk for X-Micro Flash Disk. PR: usb/96901 Submitted by: Szalai Andras Reviewed by: nate (a half of year ago) MFC after: 3 weeks
2006-11-26 17:13:24 +00:00
{
/*
* X-Micro Flash Disk
* PR: usb/96901
*/
o Style: remove a w/space before a comma.
2006-11-26 17:18:31 +00:00
{T_DIRECT, SIP_MEDIA_REMOVABLE, "X-Micro", "Flash Disk",
o Add a quirk for X-Micro Flash Disk. PR: usb/96901 Submitted by: Szalai Andras Reviewed by: nate (a half of year ago) MFC after: 3 weeks
2006-11-26 17:13:24 +00:00
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add quirk for EasyMP3 EM732X usb 2.0 flash mp3 player. (It appears that the quirk proceedures link has disappeared and that this PR complied with it, if there's a problem, please contact me). PR: usb/96546
2007-01-22 04:34:03 +00:00
{
/*
* EasyMP3 EM732X USB 2.0 Flash MP3 Player
* PR: usb/96546
*/
Use the more specific 'EM732X' designation rather than * to disable sync cache commands, per request from njl@.
2007-01-23 17:29:31 +00:00
{T_DIRECT, SIP_MEDIA_REMOVABLE, "EM732X", "MP3 Player*",
o Correct EasyMP3 EM732X usb 2.0 flash mp3 player revision: it's 1.00 not 1.0. PR: usb/96546 Submitted by: Eugene Grosbein MFC after: 1 month
2008-01-12 19:24:55 +00:00
"1.00"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
Add quirk for EasyMP3 EM732X usb 2.0 flash mp3 player. (It appears that the quirk proceedures link has disappeared and that this PR complied with it, if there's a problem, please contact me). PR: usb/96546
2007-01-22 04:34:03 +00:00
},
Quirk for Denver MP3 player usb. Submitted by: Ed Schouten PR: 107101 Approved by: re (blanket for device supprot)
2007-06-28 04:51:19 +00:00
{
/*
* Denver MP3 player
* PR: usb/107101
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "DENVER", "MP3 PLAYER",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Philips USB Key Audio KEY013 Submitted by: Arnoud Engelfriet PR: 68412 Approved by: re (blanket)
2007-06-30 04:34:45 +00:00
{
/*
* Philips USB Key Audio KEY013
* PR: usb/68412
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "PHILIPS", "Key*", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE | DA_Q_NO_PREVENT
},
Add support for JNC MP3 Player Submitted by: User & <dhenin@cypries.cyclopes.org> PR: 94439 Approved by: re (blanket)
2007-06-30 05:03:03 +00:00
{
/*
* JNC MP3 Player
* PR: usb/94439
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "JNC*" , "MP3 Player*",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Samsung MP0402H quirk Submitted by: Gipsy<wint3r@securimail.com> PR: 108427 Approved by: re@ (blanket)
2007-06-30 05:21:18 +00:00
{
/*
* SAMSUNG MP0402H
* PR: usb/108427
*/
{T_DIRECT, SIP_MEDIA_FIXED, "SAMSUNG", "MP0402H", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
quirk for I/O Magic USB flash drive "Giga Bank" Submitted by: Jeff Anton PR: 108810 Approved by: re@ (blanket)
2007-06-30 05:24:25 +00:00
{
/*
* I/O Magic USB flash - Giga Bank
* PR: usb/108810
*/
{T_DIRECT, SIP_MEDIA_FIXED, "GS-Magic", "stor*", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add quirk for JoyFly 128mb USB Flash Drive submitted by: Toni Schmidbauer PR: 96133 Approved by: re@ (blanket)
2007-06-30 14:41:30 +00:00
{
/*
* JoyFly 128mb USB Flash Drive
* PR: 96133
*/
After talking with njl@, change this wildcard entry to be a bit less wild. Approved by: re (blanket)
2007-07-01 17:42:07 +00:00
{T_DIRECT, SIP_MEDIA_REMOVABLE, "USB 2.0", "Flash Disk*",
Add quirk for JoyFly 128mb USB Flash Drive submitted by: Toni Schmidbauer PR: 96133 Approved by: re@ (blanket)
2007-06-30 14:41:30 +00:00
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Quirk for ChipsBnk usb stick Submitted by: Mark Andrews PR: 103702 Approved by: re@ (blanket)
2007-06-30 14:44:09 +00:00
{
/*
* ChipsBnk usb stick
* PR: 103702
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "ChipsBnk", "USB*",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add quirk for the Storcase InfoStation 12bay SATA to FC SAN. PR: 129858 Submitted by: Nick Triantos <nick-freebsd at triantos dot com> MFC after: 1 week
2008-12-23 09:11:05 +00:00
{
/*
* Storcase (Kingston) InfoStation IFS FC2/SATA-R 201A
* PR: 129858
*/
{T_DIRECT, SIP_MEDIA_FIXED, "IFS", "FC2/SATA-R*",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add quirk for the Samsung YP-U3 PR: 125398 Submitted by: Tino Engel <goaengel at gmx dot net> Approved by: imp (mentor, implicit) MFC after: 1 week
2008-07-09 17:00:14 +00:00
{
/*
* Samsung YP-U3 mp3-player
* PR: 125398
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "Samsung", "YP-U3",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add quirks for Netac Onlydisk 2000 USB disk.
2007-12-11 06:41:38 +00:00
{
{T_DIRECT, SIP_MEDIA_REMOVABLE, "Netac", "OnlyDisk*",
"2000"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
- Add quirk for Sony DSC digital cameras. This umass devices fail to attach without these quirks applied. PR: usb/137035 URL: http://lists.freebsd.org/pipermail/freebsd-current/2009-August/010852.html Reported by: Henri Hennebert <hlh@restart.be>, Andrey V. Elsukov <bu7cher@yandex.ru> MFC after: 1 week
2009-08-26 21:14:28 +00:00
},
{
/*
* Sony Cyber-Shot DSC cameras
* PR: usb/137035
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "Sony", "Sony DSC", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE | DA_Q_NO_PREVENT
Merge to da driver quirks hinting 4K physical sector sizes for SATA disks connected via SAS or USB. Unluckily I've found that SAS (mps) and USB-SATA I have translate models in different ways, requiring twice more quirks. Unluckily for Hitachi, their model names are trimmed on SAS, making impossible to identify 4K sector drives that way.
2011-12-23 00:10:17 +00:00
},
Add DA_Q_NO_PREVENT quirk for Kingston DataTraveler G3 1.00 USB flash. PREVENT ALLOW MEDIUM REMOVAL commands return errors on these devices without returning sense data. In some cases unrelated following commands start to return errors too, that makes device to be dropped by CAM.
2013-02-22 17:45:32 +00:00
{
{T_DIRECT, SIP_MEDIA_REMOVABLE, "Kingston", "DataTraveler G3",
"1.00"}, /*quirks*/ DA_Q_NO_PREVENT
},
Merge to da driver quirks hinting 4K physical sector sizes for SATA disks connected via SAS or USB. Unluckily I've found that SAS (mps) and USB-SATA I have translate models in different ways, requiring twice more quirks. Unluckily for Hitachi, their model names are trimmed on SAS, making impossible to identify 4K sector drives that way.
2011-12-23 00:10:17 +00:00
/* ATA/SATA devices over SAS/USB/... */
{
/* Hitachi Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "Hitachi", "H??????????E3*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Samsung Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "SAMSUNG HD155UI*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Samsung Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "SAMSUNG", "HD155UI*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Samsung Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "SAMSUNG HD204UI*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Samsung Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "SAMSUNG", "HD204UI*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Barracuda Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "ST????DL*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Barracuda Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ST????DL", "*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Barracuda Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "ST???DM*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Barracuda Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ST???DM*", "*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Barracuda Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "ST????DM*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Barracuda Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ST????DM", "*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "ST9500423AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ST950042", "3AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "ST9500424AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ST950042", "4AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "ST9640423AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ST964042", "3AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "ST9640424AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ST964042", "4AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "ST9750420AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ST975042", "0AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "ST9750422AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ST975042", "2AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "ST9750423AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ST975042", "3AS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Thin Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "ST???LT*", "*" },
/*quirks*/DA_Q_4K
},
{
/* Seagate Momentus Thin Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ST???LT*", "*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Caviar Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "WDC WD????RS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Caviar Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "WDC WD??", "??RS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Caviar Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "WDC WD????RX*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Caviar Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "WDC WD??", "??RX*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Caviar Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "WDC WD??????RS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Caviar Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "WDC WD??", "????RS*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Caviar Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "WDC WD??????RX*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Caviar Green Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "WDC WD??", "????RX*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Scorpio Black Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "WDC WD???PKT*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Scorpio Black Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "WDC WD??", "?PKT*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Scorpio Black Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "WDC WD?????PKT*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Scorpio Black Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "WDC WD??", "???PKT*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Scorpio Blue Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "WDC WD???PVT*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Scorpio Blue Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "WDC WD??", "?PVT*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Scorpio Blue Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "ATA", "WDC WD?????PVT*", "*" },
/*quirks*/DA_Q_4K
},
{
/* WDC Scorpio Blue Advanced Format (4k) drives */
{ T_DIRECT, SIP_MEDIA_FIXED, "WDC WD??", "???PVT*", "*" },
/*quirks*/DA_Q_4K
},
Add support for: Olympus FE-210 camera LG UP3S MP3 player Laser MP3-2GA13 MP3 PR: usb/119201 Submitted by: Peter Jeremy <peterjeremy@optushome.com.au> Approved by: cperciva MFC after: 1 week
2012-04-29 22:28:15 +00:00
{
/*
* Olympus FE-210 camera
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "OLYMPUS", "FE210*",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
{
/*
* LG UP3S MP3 player
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "LG", "UP3S",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
{
/*
* Laser MP3-2GA13 MP3 player
*/
{T_DIRECT, SIP_MEDIA_REMOVABLE, "USB 2.0", "(HS) Flash Disk",
"*"}, /*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add support for samsung HM250JI PR: usb/121474 Submitted by: Ben Stuyts <ben@altesco.nl> Approved by: cperciva (implicit) MFC after: 3 days
2012-10-16 17:49:14 +00:00
{
/*
* LaCie external 250GB Hard drive des by Porsche
* Submitted by: Ben Stuyts <ben@altesco.nl>
* PR: 121474
*/
{T_DIRECT, SIP_MEDIA_FIXED, "SAMSUNG", "HM250JI", "*"},
/*quirks*/ DA_Q_NO_SYNC_CACHE
},
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
};
NO_GEOM cleanup: Move to "struct disk *" centric api.
2003-02-25 22:06:21 +00:00
static disk_strategy_t dastrategy;
NO_GEOM cleanup: Retire the "d_dump_t" and use the "dumper_t" type instead. Dumper_t takes a void * as first arg which is more general than the dev_t taken by d_dump_t. (Remember: we could have net-dumpers if somebody wrote us one!) Define the convention for GEOM controlled disk devices to be that the first argument to the dumper function is the struct disk pointer. Change device drivers accordingly.
2003-02-21 19:00:48 +00:00
static dumper_t dadump;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
static periph_init_t dainit;
static void daasync(void *callback_arg, u_int32_t code,
struct cam_path *path, void *arg);
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
static void dasysctlinit(void *context, int pending);
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
static int dacmdsizesysctl(SYSCTL_HANDLER_ARGS);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
static int dadeletemethodsysctl(SYSCTL_HANDLER_ARGS);
Updates delete_method sysctl changes to always maintain disk d_flags DISKFLAG_CANDELETE. While this change makes this layer consistent other layers such as UFS and ZFS BIO_DELETE support may not notice any change made manually via these device sysctls until the device is reopened via a mount. Also corrected var order in dadeletemethodsysctl PR: kern/169801 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:57:46 +00:00
static int dadeletemethodset(struct da_softc *softc,
da_delete_methods delete_method);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
static periph_ctor_t daregister;
static periph_dtor_t dacleanup;
static periph_start_t dastart;
Fix a problem with the way we handled device invalidation when attaching to a device failed. In theory, the same steps that happen when we get an AC_LOST_DEVICE async notification should have been taken when a driver fails to attach. In practice, that wasn't the case. This only affected the da, cd and ch drivers, but the fix affects all peripheral drivers. There were several possible problems: - In the da driver, we didn't remove the peripheral's softc from the da driver's linked list of softcs. Once the peripheral and softc got removed, we'd get a kernel panic the next time the timeout routine called dasendorderedtag(). - In the da, cd and possibly ch drivers, we didn't remove the peripheral's devstat structure from the devstat queue. Once the peripheral and softc were removed, this could cause a panic if anyone tried to access device statistics. (one component of the linked list wouldn't exist anymore) - In the cd driver, we didn't take the peripheral off the changer run queue if it was scheduled to run. In practice, it's highly unlikely, and maybe impossible that the peripheral would have been on the changer run queue at that stage of the probe process. The fix is: - Add a new peripheral callback function (the "oninvalidate" function) that is called the first time cam_periph_invalidate() is called for a peripheral. - Create new foooninvalidate() routines for each peripheral driver. This routine is always called at splsoftcam(), and contains all the stuff that used to be in the AC_LOST_DEVICE case of the async callback handler. - Move the devstat cleanup call to the destructor/cleanup routines, since some of the drivers do I/O in their close routines. - Make sure that when we're flushing the buffer queue, we traverse it at splbio(). - Add a check for the invalid flag in the pt driver's open routine. Reviewed by: gibbs
1998-10-22 22:16:56 +00:00
static periph_oninv_t daoninvalidate;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
static void dadone(struct cam_periph *periph,
union ccb *done_ccb);
static int daerror(union ccb *ccb, u_int32_t cam_flags,
u_int32_t sense_flags);
static void daprevent(struct cam_periph *periph, int action);
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
static void dareprobe(struct cam_periph *periph);
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
static void dasetgeom(struct cam_periph *periph, uint32_t block_len,
Add CAM infrastructure to allow reporting when a drive's long read capacity data changes. cam_ccb.h: Add a new advanced information type, CDAI_TYPE_RCAPLONG, for long read capacity data. cam_xpt_internal.h: Add a read capacity data pointer and length to struct cam_ed. cam_xpt.c: Free the read capacity buffer when a device goes away. While we're here, make sure we don't leak memory for other malloced fields in struct cam_ed. scsi_all.c: Update the scsi_read_capacity_16() to take a uint8_t * and a length instead of just a pointer to the parameter data structure. This will hopefully make this function somewhat immune to future changes in the parameter data. scsi_all.h: Add some extra bit definitions to struct scsi_read_capacity_data_long, and bump up the structure size to the full size specified by SBC-3. Change the prototype for scsi_read_capacity_16(). scsi_da.c: Register changes in read capacity data with the transport layer. This allows the transport layer to send out an async notification to interested parties. Update the dasetgeom() API. Use scsi_extract_sense_len() instead of scsi_extract_sense(). scsi_xpt.c: Add support for the new CDAI_TYPE_RCAPLONG advanced information type. Make sure we set the physpath pointer to NULL after freeing it. This allows blindly freeing it in the struct cam_ed destructor. sys/param.h: Bump __FreeBSD_version from 1000005 to 1000006 to make it easier for third party drivers to determine that the read capacity data async notification is available. camcontrol.c, mptutil/mpt_cam.c: Update these for the new scsi_read_capacity_16() argument structure. Sponsored by: Spectra Logic
2012-01-26 18:09:28 +00:00
uint64_t maxsector,
struct scsi_read_capacity_data_long *rcaplong,
size_t rcap_size);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
static timeout_t dasendorderedtag;
Implement a new generic mechanism for attaching handler functions to events, in order to pave the way for removing a number of the ad-hoc implementations currently in use. Retire the at_shutdown family of functions and replace them with new event handler lists. Rework kern_shutdown.c to take greater advantage of the use of event handlers. Reviewed by: green
1999-08-21 06:24:40 +00:00
static void dashutdown(void *arg, int howto);
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
static timeout_t damediapoll;
#ifndef DA_DEFAULT_POLL_PERIOD
#define DA_DEFAULT_POLL_PERIOD 3
#endif
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
#ifndef DA_DEFAULT_TIMEOUT
#define DA_DEFAULT_TIMEOUT 60 /* Timeout in seconds */
#endif
Add SYSCTL ints for default normal I/O timeout && retry counts. This is useful if you want to dynamically move into a Fibre Channel or Multi-initiator environment that happens to be particularly noisy and ugly that requires a lot of retries (with shorter I/O timeouts) for commands destried by LIPs or Bus Resets. Reviewed by: deafening silence on audit && scsi on the retry counts MFC after: 2 weeks
2001-07-09 19:18:00 +00:00
#ifndef DA_DEFAULT_RETRY
#define DA_DEFAULT_RETRY 4
#endif
Add a tunable that allows one to turn off the automatic sending of the ORDERED tag. This recoups significant performance gains for many arrays. The default is still to send out the ORDERED tag periodically. Reviewed by: scsi (justin+timeout)
2006-11-02 21:12:37 +00:00
#ifndef DA_DEFAULT_SEND_ORDERED
#define DA_DEFAULT_SEND_ORDERED 1
#endif
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
static int da_poll_period = DA_DEFAULT_POLL_PERIOD;
Add SYSCTL ints for default normal I/O timeout && retry counts. This is useful if you want to dynamically move into a Fibre Channel or Multi-initiator environment that happens to be particularly noisy and ugly that requires a lot of retries (with shorter I/O timeouts) for commands destried by LIPs or Bus Resets. Reviewed by: deafening silence on audit && scsi on the retry counts MFC after: 2 weeks
2001-07-09 19:18:00 +00:00
static int da_retry_count = DA_DEFAULT_RETRY;
static int da_default_timeout = DA_DEFAULT_TIMEOUT;
Add a tunable that allows one to turn off the automatic sending of the ORDERED tag. This recoups significant performance gains for many arrays. The default is still to send out the ORDERED tag periodically. Reviewed by: scsi (justin+timeout)
2006-11-02 21:12:37 +00:00
static int da_send_ordered = DA_DEFAULT_SEND_ORDERED;
Add SYSCTL ints for default normal I/O timeout && retry counts. This is useful if you want to dynamically move into a Fibre Channel or Multi-initiator environment that happens to be particularly noisy and ugly that requires a lot of retries (with shorter I/O timeouts) for commands destried by LIPs or Bus Resets. Reviewed by: deafening silence on audit && scsi on the retry counts MFC after: 2 weeks
2001-07-09 19:18:00 +00:00
Mark all SYSCTL_NODEs static that have no corresponding SYSCTL_DECLs. The SYSCTL_NODE macro defines a list that stores all child-elements of that node. If there's no SYSCTL_DECL macro anywhere else, there's no reason why it shouldn't be static.
2011-11-07 15:43:11 +00:00
static SYSCTL_NODE(_kern_cam, OID_AUTO, da, CTLFLAG_RD, 0,
Add SYSCTL ints for default normal I/O timeout && retry counts. This is useful if you want to dynamically move into a Fibre Channel or Multi-initiator environment that happens to be particularly noisy and ugly that requires a lot of retries (with shorter I/O timeouts) for commands destried by LIPs or Bus Resets. Reviewed by: deafening silence on audit && scsi on the retry counts MFC after: 2 weeks
2001-07-09 19:18:00 +00:00
"CAM Direct Access Disk driver");
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
SYSCTL_INT(_kern_cam_da, OID_AUTO, poll_period, CTLFLAG_RW,
&da_poll_period, 0, "Media polling period in seconds");
TUNABLE_INT("kern.cam.da.poll_period", &da_poll_period);
Add SYSCTL ints for default normal I/O timeout && retry counts. This is useful if you want to dynamically move into a Fibre Channel or Multi-initiator environment that happens to be particularly noisy and ugly that requires a lot of retries (with shorter I/O timeouts) for commands destried by LIPs or Bus Resets. Reviewed by: deafening silence on audit && scsi on the retry counts MFC after: 2 weeks
2001-07-09 19:18:00 +00:00
SYSCTL_INT(_kern_cam_da, OID_AUTO, retry_count, CTLFLAG_RW,
&da_retry_count, 0, "Normal I/O retry count");
Fix ATAPI/USB/Firewire CDROM drive handling in cd(4) and hopefully fix a number of related problems along the way. - Automatically detect CDROM drives that can't handle 6 byte mode sense and mode select, and adjust our command size accordingly. We have to handle this in the cd(4) driver (where the buffers are allocated), since the parameter list length is different for the 6 and 10 byte mode sense commands. - Remove MODE_SENSE and MODE_SELECT translation removed in ATAPICAM and in the umass(4) driver, since there's no way for that to work properly. - Add a quirk entry for CDROM drives that just hang when they get a 6 byte mode sense or mode select. The reason for the quirk must be documented in a PR, and all quirks must be approved by ken@FreeBSD.org. This is to make sure that we fully understand why each quirk is needed. Once the CAM_NEW_TRAN_CODE is finished, we should be able to remove any such quirks, since we'll know what protocol the drive speaks (SCSI, ATAPI, etc.) and therefore whether we should use 6 or 10 byte mode sense/select commands. - Change the way the da(4) handles the no_6_byte sysctl. There is now a per-drive sysctl to set the minimum command size for that particular disk. (Since you could have multiple disks with multiple requirements in one system.) - Loader tunable support for all the sysctls in the da(4) and cd(4) drivers. - Add a CDIOCCLOSE ioctl for cd(4) (bde pointed this out a long time ago). - Add a media validation routine (cdcheckmedia()) to the cd(4) driver, to fix some problems bde pointed out a long time ago. We now allow open() to succeed no matter what, but if we don't detect valid media, the user can only issue CDIOCCLOSE or CDIOCEJECT ioctls. - The media validation routine also reads the table of contents off the drive. We use the table of contents to implement the CDIOCPLAYTRACKS ioctl using the PLAY AUDIO MSF command. The PLAY AUDIO TRACK INDEX command that we previously used was deprecated after SCSI-2. It works in every SCSI CDROM I've tried, but doesn't seem to work on ATAPI CDROM drives. We still use the play audio track index command if we don't have a valid TOC, but I suppose it'll fail anyway in that case. - Add _len() versions of scsi_mode_sense() and scsi_mode_select() so that we can specify the minimum command length. - Fix a couple of formatting problems in the sense printing code. MFC after: 4 weeks
2003-02-21 06:19:38 +00:00
TUNABLE_INT("kern.cam.da.retry_count", &da_retry_count);
Add SYSCTL ints for default normal I/O timeout && retry counts. This is useful if you want to dynamically move into a Fibre Channel or Multi-initiator environment that happens to be particularly noisy and ugly that requires a lot of retries (with shorter I/O timeouts) for commands destried by LIPs or Bus Resets. Reviewed by: deafening silence on audit && scsi on the retry counts MFC after: 2 weeks
2001-07-09 19:18:00 +00:00
SYSCTL_INT(_kern_cam_da, OID_AUTO, default_timeout, CTLFLAG_RW,
&da_default_timeout, 0, "Normal I/O timeout (in seconds)");
Fix ATAPI/USB/Firewire CDROM drive handling in cd(4) and hopefully fix a number of related problems along the way. - Automatically detect CDROM drives that can't handle 6 byte mode sense and mode select, and adjust our command size accordingly. We have to handle this in the cd(4) driver (where the buffers are allocated), since the parameter list length is different for the 6 and 10 byte mode sense commands. - Remove MODE_SENSE and MODE_SELECT translation removed in ATAPICAM and in the umass(4) driver, since there's no way for that to work properly. - Add a quirk entry for CDROM drives that just hang when they get a 6 byte mode sense or mode select. The reason for the quirk must be documented in a PR, and all quirks must be approved by ken@FreeBSD.org. This is to make sure that we fully understand why each quirk is needed. Once the CAM_NEW_TRAN_CODE is finished, we should be able to remove any such quirks, since we'll know what protocol the drive speaks (SCSI, ATAPI, etc.) and therefore whether we should use 6 or 10 byte mode sense/select commands. - Change the way the da(4) handles the no_6_byte sysctl. There is now a per-drive sysctl to set the minimum command size for that particular disk. (Since you could have multiple disks with multiple requirements in one system.) - Loader tunable support for all the sysctls in the da(4) and cd(4) drivers. - Add a CDIOCCLOSE ioctl for cd(4) (bde pointed this out a long time ago). - Add a media validation routine (cdcheckmedia()) to the cd(4) driver, to fix some problems bde pointed out a long time ago. We now allow open() to succeed no matter what, but if we don't detect valid media, the user can only issue CDIOCCLOSE or CDIOCEJECT ioctls. - The media validation routine also reads the table of contents off the drive. We use the table of contents to implement the CDIOCPLAYTRACKS ioctl using the PLAY AUDIO MSF command. The PLAY AUDIO TRACK INDEX command that we previously used was deprecated after SCSI-2. It works in every SCSI CDROM I've tried, but doesn't seem to work on ATAPI CDROM drives. We still use the play audio track index command if we don't have a valid TOC, but I suppose it'll fail anyway in that case. - Add _len() versions of scsi_mode_sense() and scsi_mode_select() so that we can specify the minimum command length. - Fix a couple of formatting problems in the sense printing code. MFC after: 4 weeks
2003-02-21 06:19:38 +00:00
TUNABLE_INT("kern.cam.da.default_timeout", &da_default_timeout);
Renamed the kern.cam.da.da_send_ordered sysctl and tunable to kern.cam.da.send_ordered, more in line with the other da sysctls/tunables. PR: 169765 Submitted by: Steven Hartland <steven.hartland@multiplay.co.uk> Reviewed by: mav
2012-07-11 23:00:26 +00:00
SYSCTL_INT(_kern_cam_da, OID_AUTO, send_ordered, CTLFLAG_RW,
Add a tunable that allows one to turn off the automatic sending of the ORDERED tag. This recoups significant performance gains for many arrays. The default is still to send out the ORDERED tag periodically. Reviewed by: scsi (justin+timeout)
2006-11-02 21:12:37 +00:00
&da_send_ordered, 0, "Send Ordered Tags");
Renamed the kern.cam.da.da_send_ordered sysctl and tunable to kern.cam.da.send_ordered, more in line with the other da sysctls/tunables. PR: 169765 Submitted by: Steven Hartland <steven.hartland@multiplay.co.uk> Reviewed by: mav
2012-07-11 23:00:26 +00:00
TUNABLE_INT("kern.cam.da.send_ordered", &da_send_ordered);
Add SYSCTL ints for default normal I/O timeout && retry counts. This is useful if you want to dynamically move into a Fibre Channel or Multi-initiator environment that happens to be particularly noisy and ugly that requires a lot of retries (with shorter I/O timeouts) for commands destried by LIPs or Bus Resets. Reviewed by: deafening silence on audit && scsi on the retry counts MFC after: 2 weeks
2001-07-09 19:18:00 +00:00
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
* DA_ORDEREDTAG_INTERVAL determines how often, relative
* to the default timeout, we check to see whether an ordered
* tagged transaction is appropriate to prevent simple tag
* starvation. Since we'd like to ensure that there is at least
* 1/2 of the timeout length left for a starved transaction to
* complete after we've sent an ordered tag, we must poll at least
* four times in every timeout period. This takes care of the worst
* case where a starved transaction starts during an interval that
* meets the requirement "don't send an ordered tag" test so it takes
* us two intervals to determine that a tag must be sent.
*/
#ifndef DA_ORDEREDTAG_INTERVAL
#define DA_ORDEREDTAG_INTERVAL 4
#endif
static struct periph_driver dadriver =
{
dainit, "da",
TAILQ_HEAD_INITIALIZER(dadriver.units), /* generation */ 0
};
Change the peripheral driver list from a linker set to module driven driver registration. This should allow things like da, sa, cd etc to be in seperate KLD's to the cam core and make them preloadable.
2001-02-07 07:05:59 +00:00
PERIPHDRIVER_DECLARE(da, dadriver);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Mark MALLOC_DEFINEs static that have no corresponding MALLOC_DECLAREs. This means that their use is restricted to a single C file.
2011-11-07 06:44:47 +00:00
static MALLOC_DEFINE(M_SCSIDA, "scsi_da", "scsi_da buffers");
Eliminate the use of M_TEMP.
2007-05-14 21:48:53 +00:00
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
static int
NO_GEOM cleanup: Move to "struct disk *" centric api.
2003-02-25 22:06:21 +00:00
daopen(struct disk *dp)
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
{
struct cam_periph *periph;
struct da_softc *softc;
int unit;
int error;
NO_GEOM cleanup: Move to "struct disk *" centric api.
2003-02-25 22:06:21 +00:00
periph = (struct cam_periph *)dp->d_drv1;
Rewrite of the CAM error recovery code. Some of the major changes include: - The SCSI error handling portion of cam_periph_error() has been broken out into a number of subfunctions to better modularize the code that handles the hierarchy of SCSI errors. As a result, the code is now much easier to read. - String handling and error printing has been significantly revamped. We now use sbufs to do string formatting instead of using printfs (for the kernel) and snprintf/strncat (for userland) as before. There is a new catchall error printing routine, cam_error_print() and its string-based counterpart, cam_error_string() that allow the kernel and userland applications to pass in a CCB and have errors printed out properly, whether or not they're SCSI errors. Among other things, this helped eliminate a fair amount of duplicate code in camcontrol. We now print out more information than before, including the CAM status and SCSI status and the error recovery action taken to remedy the problem. - sbufs are now available in userland, via libsbuf. This change was necessary since most of the error printing code is shared between libcam and the kernel. - A new transfer settings interface is included in this checkin. This code is #ifdef'ed out, and is primarily intended to aid discussion with HBA driver authors on the final form the interface should take. There is example code in the ahc(4) driver that implements the HBA driver side of the new interface. The new transfer settings code won't be enabled until we're ready to switch all HBA drivers over to the new interface. src/Makefile.inc1, lib/Makefile: Add libsbuf. It must be built before libcam, since libcam uses sbuf routines. libcam/Makefile: libcam now depends on libsbuf. libsbuf/Makefile: Add a makefile for libsbuf. This pulls in the sbuf sources from sys/kern. bsd.libnames.mk: Add LIBSBUF. camcontrol/Makefile: Add -lsbuf. Since camcontrol is statically linked, we can't depend on the dynamic linker to pull in libsbuf. camcontrol.c: Use cam_error_print() instead of checking for CAM_SCSI_STATUS_ERROR on every failed CCB. sbuf.9: Change the prototypes for sbuf_cat() and sbuf_cpy() so that the source string is now a const char *. This is more in line wth the standard system string functions, and helps eliminate warnings when dealing with a const source buffer. Fix a typo. cam.c: Add description strings for the various CAM error status values, as well as routines to look up those strings. Add new cam_error_string() and cam_error_print() routines for userland and the kernel. cam.h: Add a new CAM flag, CAM_RETRY_SELTO. Add enumerated types for the various options available with cam_error_print() and cam_error_string(). cam_ccb.h: Add new transfer negotiation structures/types. Change inq_len in the ccb_getdev structure to be "reserved". This field has never been filled in, and will be removed when we next bump the CAM version. cam_debug.h: Fix typo. cam_periph.c: Modularize cam_periph_error(). The SCSI error handling part of cam_periph_error() is now in camperiphscsistatuserror() and camperiphscsisenseerror(). In cam_periph_lock(), increase the reference count on the periph while we wait for our lock attempt to succeed so that the periph won't go away while we're sleeping. cam_xpt.c: Add new transfer negotiation code. (ifdefed out) Add a new function, xpt_path_string(). This is a string/sbuf analog to xpt_print_path(). scsi_all.c: Revamp string handing and error printing code. We now use sbufs for much of the string formatting code. More of that code is shared between userland the kernel. scsi_all.h: Get rid of SS_TURSTART, it wasn't terribly useful in the first place. Add a new error action, SS_REQSENSE. (Send a request sense and then retry the command.) This is useful when the controller hasn't performed autosense for some reason. Change the default actions around a bit. scsi_cd.c, scsi_da.c, scsi_pt.c, scsi_ses.c: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Selection timeouts shouldn't be covered by a sense flag. scsi_pass.[ch]: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Get rid of the last vestiges of a read/write interface. libkern/bsearch.c, sys/libkern.h, conf/files: Add bsearch.c, which is needed for some of the new table lookup routines. aic7xxx_freebsd.c: Define AHC_NEW_TRAN_SETTINGS if CAM_NEW_TRAN_CODE is defined. sbuf.h, subr_sbuf.c: Add the appropriate #ifdefs so sbufs can compile and run in userland. Change sbuf_printf() to use vsnprintf() instead of kvprintf(), which is only available in the kernel. Change the source string for sbuf_cpy() and sbuf_cat() to be a const char *. Add __BEGIN_DECLS and __END_DECLS around function prototypes since they're now exported to userland. kdump/mkioctls: Include stdio.h before cam.h since cam.h now includes a function with a FILE * argument. Submitted by: gibbs (mostly) Reviewed by: jdp, marcel (libsbuf makefile changes) Reviewed by: des (sbuf changes) Reviewed by: ken
2001-03-27 05:45:52 +00:00
if (periph == NULL) {
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
return (ENXIO);
Rewrite of the CAM error recovery code. Some of the major changes include: - The SCSI error handling portion of cam_periph_error() has been broken out into a number of subfunctions to better modularize the code that handles the hierarchy of SCSI errors. As a result, the code is now much easier to read. - String handling and error printing has been significantly revamped. We now use sbufs to do string formatting instead of using printfs (for the kernel) and snprintf/strncat (for userland) as before. There is a new catchall error printing routine, cam_error_print() and its string-based counterpart, cam_error_string() that allow the kernel and userland applications to pass in a CCB and have errors printed out properly, whether or not they're SCSI errors. Among other things, this helped eliminate a fair amount of duplicate code in camcontrol. We now print out more information than before, including the CAM status and SCSI status and the error recovery action taken to remedy the problem. - sbufs are now available in userland, via libsbuf. This change was necessary since most of the error printing code is shared between libcam and the kernel. - A new transfer settings interface is included in this checkin. This code is #ifdef'ed out, and is primarily intended to aid discussion with HBA driver authors on the final form the interface should take. There is example code in the ahc(4) driver that implements the HBA driver side of the new interface. The new transfer settings code won't be enabled until we're ready to switch all HBA drivers over to the new interface. src/Makefile.inc1, lib/Makefile: Add libsbuf. It must be built before libcam, since libcam uses sbuf routines. libcam/Makefile: libcam now depends on libsbuf. libsbuf/Makefile: Add a makefile for libsbuf. This pulls in the sbuf sources from sys/kern. bsd.libnames.mk: Add LIBSBUF. camcontrol/Makefile: Add -lsbuf. Since camcontrol is statically linked, we can't depend on the dynamic linker to pull in libsbuf. camcontrol.c: Use cam_error_print() instead of checking for CAM_SCSI_STATUS_ERROR on every failed CCB. sbuf.9: Change the prototypes for sbuf_cat() and sbuf_cpy() so that the source string is now a const char *. This is more in line wth the standard system string functions, and helps eliminate warnings when dealing with a const source buffer. Fix a typo. cam.c: Add description strings for the various CAM error status values, as well as routines to look up those strings. Add new cam_error_string() and cam_error_print() routines for userland and the kernel. cam.h: Add a new CAM flag, CAM_RETRY_SELTO. Add enumerated types for the various options available with cam_error_print() and cam_error_string(). cam_ccb.h: Add new transfer negotiation structures/types. Change inq_len in the ccb_getdev structure to be "reserved". This field has never been filled in, and will be removed when we next bump the CAM version. cam_debug.h: Fix typo. cam_periph.c: Modularize cam_periph_error(). The SCSI error handling part of cam_periph_error() is now in camperiphscsistatuserror() and camperiphscsisenseerror(). In cam_periph_lock(), increase the reference count on the periph while we wait for our lock attempt to succeed so that the periph won't go away while we're sleeping. cam_xpt.c: Add new transfer negotiation code. (ifdefed out) Add a new function, xpt_path_string(). This is a string/sbuf analog to xpt_print_path(). scsi_all.c: Revamp string handing and error printing code. We now use sbufs for much of the string formatting code. More of that code is shared between userland the kernel. scsi_all.h: Get rid of SS_TURSTART, it wasn't terribly useful in the first place. Add a new error action, SS_REQSENSE. (Send a request sense and then retry the command.) This is useful when the controller hasn't performed autosense for some reason. Change the default actions around a bit. scsi_cd.c, scsi_da.c, scsi_pt.c, scsi_ses.c: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Selection timeouts shouldn't be covered by a sense flag. scsi_pass.[ch]: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Get rid of the last vestiges of a read/write interface. libkern/bsearch.c, sys/libkern.h, conf/files: Add bsearch.c, which is needed for some of the new table lookup routines. aic7xxx_freebsd.c: Define AHC_NEW_TRAN_SETTINGS if CAM_NEW_TRAN_CODE is defined. sbuf.h, subr_sbuf.c: Add the appropriate #ifdefs so sbufs can compile and run in userland. Change sbuf_printf() to use vsnprintf() instead of kvprintf(), which is only available in the kernel. Change the source string for sbuf_cpy() and sbuf_cat() to be a const char *. Add __BEGIN_DECLS and __END_DECLS around function prototypes since they're now exported to userland. kdump/mkioctls: Include stdio.h before cam.h since cam.h now includes a function with a FILE * argument. Submitted by: gibbs (mostly) Reviewed by: jdp, marcel (libsbuf makefile changes) Reviewed by: des (sbuf changes) Reviewed by: ken
2001-03-27 05:45:52 +00:00
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
if (cam_periph_acquire(periph) != CAM_REQ_CMP) {
Fix a race condition in CAM peripheral free handling, locking in the CAM XPT bus traversal code, and a number of other periph level issues. cam_periph.h, cam_periph.c: Modify cam_periph_acquire() to test the CAM_PERIPH_INVALID flag prior to allowing a reference count to be gained on a peripheral. Callers of this function will receive CAM_REQ_CMP_ERR status in the situation of attempting to reference an invalidated periph. This guarantees that a peripheral scheduled for a deferred free will not be accessed during its wait for destruction. Panic during attempts to drop a reference count on a peripheral that already has a zero reference count. In cam_periph_list(), use a local sbuf with SBUF_FIXEDLEN set so that mallocs do not occur while the xpt topology lock is held, regardless of the allocation policy of the passed in sbuf. Add a new routine, cam_periph_release_locked_buses(), that can be called when the caller already holds the CAM topology lock. Add some extra debugging for duplicate peripheral allocations in cam_periph_alloc(). Treat CAM_DEV_NOT_THERE much the same as a selection timeout (AC_LOST_DEVICE is emitted), but forgo retries. cam_xpt.c: Revamp the way the EDT traversal code does locking and reference counting. This was broken, since it assumed that the EDT would not change during traversal, but that assumption is no longer valid. So, to prevent devices from going away while we traverse the EDT, make sure we properly lock everything and hold references on devices that we are using. The two peripheral driver traversal routines should be examined. xptpdperiphtraverse() holds the topology lock for the entire time it runs. xptperiphtraverse() is now locked properly, but only holds the topology lock while it is traversing the list, and not while the traversal function is running. The bus locking code in xptbustraverse() should also be revisited at a later time, since it is complex and should probably be simplified. scsi_da.c: Pay attention to the return value from cam_periph_acquire(). Return 0 always from daclose() even if the disk is now gone. Add some rudimentary error injection support. scsi_sg.c: Fix reference counting in the sg(4) driver. The sg driver was calling cam_periph_release() on close, but never called cam_periph_acquire() (which increments the reference count) on open. The periph code correctly complained that the sg(4) driver was trying to decrement the refcount when it was already 0. Sponsored by: Spectra Logic MFC after: 2 weeks
2012-01-12 00:41:48 +00:00
return (ENXIO);
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
}
cam_periph_lock(periph);
if ((error = cam_periph_hold(periph, PRIBIO|PCATCH)) != 0) {
cam_periph_unlock(periph);
cam_periph_release(periph);
return (error);
}
unit = periph->unit_number;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
softc = (struct da_softc *)periph->softc;
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
softc->flags |= DA_FLAG_OPEN;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Tune and add some missing CAM_DEBUG() points for better consistency.
2012-06-05 09:45:42 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE | CAM_DEBUG_PERIPH,
("daopen\n"));
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if ((softc->flags & DA_FLAG_PACK_INVALID) != 0) {
Some fixes for the CD and DA drivers from bde. (and some of my own as well) Among them: [ cd driver ] 1. Old labeling code was still there. 2. Error handling for dsopen() was broken (no test for the `error' returned by dsopen(); bogus test of an `error' that is known to be 0). 3. cdopen() closed the physical device after certain errors although there may still be open partitions on it. 4. cdclose() closed the physical device although there may still be open partitions on it. 5. Some printf format fixes was incomplete or missing. 6. cdioctl() truncated unit numbers mod 256. 7. cdioctl() was missing locking. [ da driver ] 1. daclose() closed the physical device although there may still be open partitions on it. This was fixed many years ago in sd.c rev.1.57. 2. A minor optimization (the dk_slices != NULL test) in sdopen() became uglier in daopen(). It is not worth doing. da only regressed compared with od and my version of sd, since I never committed the change to sd. daopen() should probably do less if some partition is already open. This is not addressed by the diffs. [ ... ] 5. "opt_hw_wdog.h" was not included, so the HW_WDOG code was unreachable. - Added a getdev CCB call in the cdopen() and daopen() calls so that the vendor name and device name are available for the disklabel. (suggested by bde) - Removed vestigal devfs support in both drivers, since we can't properly work with devfs yet. (ask bde for details on this) - Cleaned up the probe code in both drivers in the failure cases. There were a number of things wrong here. The peripheral driver instances weren't getting properly cleaned up. Sometimes the wrong probe message would get printed out (with the failure message appended), so it wasn't very clear that we failed to attach. SCSI sense information was printed, even when the error in question wasn't a SCSI error. I put similar fixes into the changer driver in revision 1.2 of scsi_ch.c. Reviewed by: gibbs Submitted by: bde (partially)
1998-10-07 02:57:57 +00:00
/* Invalidate our pack information. */
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
softc->flags &= ~DA_FLAG_PACK_INVALID;
}
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
dareprobe(periph);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
/* Wait for the disk size update. */
error = msleep(&softc->disk->d_mediasize, periph->sim->mtx, PRIBIO,
"dareprobe", 0);
if (error != 0)
xpt_print(periph->path, "unable to retrieve capacity data");
Some fixes for the CD and DA drivers from bde. (and some of my own as well) Among them: [ cd driver ] 1. Old labeling code was still there. 2. Error handling for dsopen() was broken (no test for the `error' returned by dsopen(); bogus test of an `error' that is known to be 0). 3. cdopen() closed the physical device after certain errors although there may still be open partitions on it. 4. cdclose() closed the physical device although there may still be open partitions on it. 5. Some printf format fixes was incomplete or missing. 6. cdioctl() truncated unit numbers mod 256. 7. cdioctl() was missing locking. [ da driver ] 1. daclose() closed the physical device although there may still be open partitions on it. This was fixed many years ago in sd.c rev.1.57. 2. A minor optimization (the dk_slices != NULL test) in sdopen() became uglier in daopen(). It is not worth doing. da only regressed compared with od and my version of sd, since I never committed the change to sd. daopen() should probably do less if some partition is already open. This is not addressed by the diffs. [ ... ] 5. "opt_hw_wdog.h" was not included, so the HW_WDOG code was unreachable. - Added a getdev CCB call in the cdopen() and daopen() calls so that the vendor name and device name are available for the disklabel. (suggested by bde) - Removed vestigal devfs support in both drivers, since we can't properly work with devfs yet. (ask bde for details on this) - Cleaned up the probe code in both drivers in the failure cases. There were a number of things wrong here. The peripheral driver instances weren't getting properly cleaned up. Sometimes the wrong probe message would get printed out (with the failure message appended), so it wasn't very clear that we failed to attach. SCSI sense information was printed, even when the error in question wasn't a SCSI error. I put similar fixes into the changer driver in revision 1.2 of scsi_ch.c. Reviewed by: gibbs Submitted by: bde (partially)
1998-10-07 02:57:57 +00:00
Fix panic at boot with SD/MMC readers with no media present, introduced at r234177. Note that this is a temporary fix, until I come up with something prettier.
2012-04-17 10:44:28 +00:00
if (periph->flags & CAM_PERIPH_INVALID ||
softc->disk->d_sectorsize == 0 ||
softc->disk->d_mediasize == 0)
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
error = ENXIO;
Fix a locking mistake in daopen(). If the open fails, which can happen because the media was removed, the periph would get its refcount dropped and ultimately freed before getting unlocked. This created a dangling pointer that was easy to trip over. This fixes a common source of crashes with removaable media, but problems remain and will get tracked down.
2008-08-29 04:39:46 +00:00
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
if (error == 0 && (softc->flags & DA_FLAG_PACK_REMOVABLE) != 0 &&
(softc->quirks & DA_Q_NO_PREVENT) == 0)
daprevent(periph, PR_PREVENT);
Fix a locking mistake in daopen(). If the open fails, which can happen because the media was removed, the periph would get its refcount dropped and ultimately freed before getting unlocked. This created a dangling pointer that was easy to trip over. This fixes a common source of crashes with removaable media, but problems remain and will get tracked down.
2008-08-29 04:39:46 +00:00
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
if (error == 0)
softc->flags |= DA_FLAG_SAW_MEDIA;
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
cam_periph_unhold(periph);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
cam_periph_unlock(periph);
Fix a locking mistake in daopen(). If the open fails, which can happen because the media was removed, the periph would get its refcount dropped and ultimately freed before getting unlocked. This created a dangling pointer that was easy to trip over. This fixes a common source of crashes with removaable media, but problems remain and will get tracked down.
2008-08-29 04:39:46 +00:00
if (error != 0) {
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
softc->flags &= ~DA_FLAG_OPEN;
Fix a locking mistake in daopen(). If the open fails, which can happen because the media was removed, the periph would get its refcount dropped and ultimately freed before getting unlocked. This created a dangling pointer that was easy to trip over. This fixes a common source of crashes with removaable media, but problems remain and will get tracked down.
2008-08-29 04:39:46 +00:00
cam_periph_release(periph);
}
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
return (error);
}
static int
NO_GEOM cleanup: Move to "struct disk *" centric api.
2003-02-25 22:06:21 +00:00
daclose(struct disk *dp)
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
{
struct cam_periph *periph;
struct da_softc *softc;
NO_GEOM cleanup: Move to "struct disk *" centric api.
2003-02-25 22:06:21 +00:00
periph = (struct cam_periph *)dp->d_drv1;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if (periph == NULL)
Fix a race condition in CAM peripheral free handling, locking in the CAM XPT bus traversal code, and a number of other periph level issues. cam_periph.h, cam_periph.c: Modify cam_periph_acquire() to test the CAM_PERIPH_INVALID flag prior to allowing a reference count to be gained on a peripheral. Callers of this function will receive CAM_REQ_CMP_ERR status in the situation of attempting to reference an invalidated periph. This guarantees that a peripheral scheduled for a deferred free will not be accessed during its wait for destruction. Panic during attempts to drop a reference count on a peripheral that already has a zero reference count. In cam_periph_list(), use a local sbuf with SBUF_FIXEDLEN set so that mallocs do not occur while the xpt topology lock is held, regardless of the allocation policy of the passed in sbuf. Add a new routine, cam_periph_release_locked_buses(), that can be called when the caller already holds the CAM topology lock. Add some extra debugging for duplicate peripheral allocations in cam_periph_alloc(). Treat CAM_DEV_NOT_THERE much the same as a selection timeout (AC_LOST_DEVICE is emitted), but forgo retries. cam_xpt.c: Revamp the way the EDT traversal code does locking and reference counting. This was broken, since it assumed that the EDT would not change during traversal, but that assumption is no longer valid. So, to prevent devices from going away while we traverse the EDT, make sure we properly lock everything and hold references on devices that we are using. The two peripheral driver traversal routines should be examined. xptpdperiphtraverse() holds the topology lock for the entire time it runs. xptperiphtraverse() is now locked properly, but only holds the topology lock while it is traversing the list, and not while the traversal function is running. The bus locking code in xptbustraverse() should also be revisited at a later time, since it is complex and should probably be simplified. scsi_da.c: Pay attention to the return value from cam_periph_acquire(). Return 0 always from daclose() even if the disk is now gone. Add some rudimentary error injection support. scsi_sg.c: Fix reference counting in the sg(4) driver. The sg driver was calling cam_periph_release() on close, but never called cam_periph_acquire() (which increments the reference count) on open. The periph code correctly complained that the sg(4) driver was trying to decrement the refcount when it was already 0. Sponsored by: Spectra Logic MFC after: 2 weeks
2012-01-12 00:41:48 +00:00
return (0);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
cam_periph_lock(periph);
Remove unused error variables in cdclose() and daclose().
2012-06-20 18:35:36 +00:00
if (cam_periph_hold(periph, PRIBIO) != 0) {
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
cam_periph_unlock(periph);
cam_periph_release(periph);
Fix a race condition in CAM peripheral free handling, locking in the CAM XPT bus traversal code, and a number of other periph level issues. cam_periph.h, cam_periph.c: Modify cam_periph_acquire() to test the CAM_PERIPH_INVALID flag prior to allowing a reference count to be gained on a peripheral. Callers of this function will receive CAM_REQ_CMP_ERR status in the situation of attempting to reference an invalidated periph. This guarantees that a peripheral scheduled for a deferred free will not be accessed during its wait for destruction. Panic during attempts to drop a reference count on a peripheral that already has a zero reference count. In cam_periph_list(), use a local sbuf with SBUF_FIXEDLEN set so that mallocs do not occur while the xpt topology lock is held, regardless of the allocation policy of the passed in sbuf. Add a new routine, cam_periph_release_locked_buses(), that can be called when the caller already holds the CAM topology lock. Add some extra debugging for duplicate peripheral allocations in cam_periph_alloc(). Treat CAM_DEV_NOT_THERE much the same as a selection timeout (AC_LOST_DEVICE is emitted), but forgo retries. cam_xpt.c: Revamp the way the EDT traversal code does locking and reference counting. This was broken, since it assumed that the EDT would not change during traversal, but that assumption is no longer valid. So, to prevent devices from going away while we traverse the EDT, make sure we properly lock everything and hold references on devices that we are using. The two peripheral driver traversal routines should be examined. xptpdperiphtraverse() holds the topology lock for the entire time it runs. xptperiphtraverse() is now locked properly, but only holds the topology lock while it is traversing the list, and not while the traversal function is running. The bus locking code in xptbustraverse() should also be revisited at a later time, since it is complex and should probably be simplified. scsi_da.c: Pay attention to the return value from cam_periph_acquire(). Return 0 always from daclose() even if the disk is now gone. Add some rudimentary error injection support. scsi_sg.c: Fix reference counting in the sg(4) driver. The sg driver was calling cam_periph_release() on close, but never called cam_periph_acquire() (which increments the reference count) on open. The periph code correctly complained that the sg(4) driver was trying to decrement the refcount when it was already 0. Sponsored by: Spectra Logic MFC after: 2 weeks
2012-01-12 00:41:48 +00:00
return (0);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
softc = (struct da_softc *)periph->softc;
Tune and add some missing CAM_DEBUG() points for better consistency.
2012-06-05 09:45:42 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE | CAM_DEBUG_PERIPH,
("daclose\n"));
sys/cam/scsi/scsi_da.c: - Only attempt the closing synchronize cache on a disk if it is still there. - When a device is lost, report the number of outstanding I/Os as they are drained. - When a device is lost, return any unprocessed bios with ENXIO instead of EIO. - Filter asynchronous events, but always allow cam_periph_async() to see them too. Sponsored by: Spectra Logic Corporation
2011-06-14 16:05:00 +00:00
if ((softc->quirks & DA_Q_NO_SYNC_CACHE) == 0
&& (softc->flags & DA_FLAG_PACK_INVALID) == 0) {
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
union ccb *ccb;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Replace most of priority numbers with defines. No logical changes.
2009-10-23 08:27:55 +00:00
ccb = cam_periph_getccb(periph, CAM_PRIORITY_NORMAL);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
scsi_synchronize_cache(&ccb->csio,
/*retries*/1,
/*cbfcnp*/dadone,
MSG_SIMPLE_Q_TAG,
/*begin_lba*/0,/* Cover the whole disk */
/*lb_count*/0,
SSD_FULL_SIZE,
5 * 60 * 1000);
Do not reinvent a wheel and let default error handler do its job.
2012-06-05 10:08:22 +00:00
cam_periph_runccb(ccb, daerror, /*cam_flags*/0,
/*sense_flags*/SF_RETRY_UA | SF_QUIET_IR,
Change the disk(9) API in order to make device removal more robust. Previously the "struct disk" were owned by the device driver and this gave us problems when the device disappared and the users of that device were not immediately disappearing. Now the struct disk is allocate with a new call, disk_alloc() and owned by geom_disk and just abandonned by the device driver when disk_create() is called. Unfortunately, this results in a ton of "s/\./->/" changes to device drivers. Since I'm doing the sweep anyway, a couple of other API improvements have been carried out at the same time: The Giant awareness flag has been flipped from DISKFLAG_NOGIANT to DISKFLAG_NEEDSGIANT A version number have been added to disk_create() so that we can detect, report and ignore binary drivers with old ABI in the future. Manual page update to follow shortly.
2004-02-18 21:36:53 +00:00
softc->disk->d_devstat);
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
xpt_release_ccb(ccb);
}
Don't leave the device queue in a frozen state if the Synchronize Cache command on close fails.
1998-09-19 04:59:35 +00:00
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if ((softc->flags & DA_FLAG_PACK_REMOVABLE) != 0) {
Add the DA_Q_NO_PREVENT quirk which keeps da(4) from sending PREVENT/ALLOW commands. Add a quirk for the Creative Nomad MuVo USB device that uses it as well as NO_SYNCHRONIZE_CACHE. PR: kern/53094 Submitted by: Richard Nyberg <rnyberg@it.su.se> MFC after: 3 days
2003-08-22 16:35:53 +00:00
if ((softc->quirks & DA_Q_NO_PREVENT) == 0)
daprevent(periph, PR_ALLOW);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
* If we've got removeable media, mark the blocksize as
* unavailable, since it could change when new media is
* inserted.
*/
Change the disk(9) API in order to make device removal more robust. Previously the "struct disk" were owned by the device driver and this gave us problems when the device disappared and the users of that device were not immediately disappearing. Now the struct disk is allocate with a new call, disk_alloc() and owned by geom_disk and just abandonned by the device driver when disk_create() is called. Unfortunately, this results in a ton of "s/\./->/" changes to device drivers. Since I'm doing the sweep anyway, a couple of other API improvements have been carried out at the same time: The Giant awareness flag has been flipped from DISKFLAG_NOGIANT to DISKFLAG_NEEDSGIANT A version number have been added to disk_create() so that we can detect, report and ignore binary drivers with old ABI in the future. Manual page update to follow shortly.
2004-02-18 21:36:53 +00:00
softc->disk->d_devstat->flags |= DEVSTAT_BS_UNAVAILABLE;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
softc->flags &= ~DA_FLAG_OPEN;
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
cam_periph_unhold(periph);
Revert r186186 for now; it breaks stuff. Approved by: rwatson (mentor)
2008-12-17 10:49:03 +00:00
cam_periph_unlock(periph);
Periph driver fixes, second try. Reviewed by: scottl Approved by: rwatson (mentor) Sponsored by: FreeBSD Foundation
2008-12-19 14:31:40 +00:00
cam_periph_release(periph);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
return (0);
}
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
static void
daschedule(struct cam_periph *periph)
{
struct da_softc *softc = (struct da_softc *)periph->softc;
uint32_t prio;
Do not schedule periph for payload/TUR requests if reprobe is in progress to avoid sending extra READ CAPACITY requests by dastart(). Schedule periph again on reprobe completion, or otherwise it may stuck indefinitely long. This should fix USB explore thread hanging on device unplug, waiting for periph destruction. Reported by: hselasky
2013-01-11 16:10:11 +00:00
if (softc->state != DA_STATE_NORMAL)
return;
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
/* Check if cam_periph_getccb() was called. */
prio = periph->immediate_priority;
/* Check if we have more work to do. */
if (bioq_first(&softc->bio_queue) ||
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
(!softc->delete_running && bioq_first(&softc->delete_queue)) ||
softc->tur) {
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
prio = CAM_PRIORITY_NORMAL;
}
/* Schedule CCB if any of above is true. */
if (prio != CAM_PRIORITY_NONE)
xpt_schedule(periph, prio);
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
* Actually translate the requested transfer into one the physical driver
* can understand. The transfer is described by a buf and will include
* only one physical transfer.
*/
static void
Complete the bio/buf divorce for all code below devfs::strategy Exceptions: Vinum untouched. This means that it cannot be compiled. Greg Lehey is on the case. CCD not converted yet, casts to struct buf (still safe) atapi-cd casts to struct buf to examine B_PHYS
2000-04-15 05:54:02 +00:00
dastrategy(struct bio *bp)
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
{
struct cam_periph *periph;
struct da_softc *softc;
NO_GEOM cleanup: Move to "struct disk *" centric api.
2003-02-25 22:06:21 +00:00
periph = (struct cam_periph *)bp->bio_disk->d_drv1;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if (periph == NULL) {
Exploit recent improvements in the disk minilayer to simplify error handling a bit. Dogmatic lingupurists can celebrate that a number of gotos got removed. Reviewed by: mjacob, ken
2001-05-08 08:30:48 +00:00
biofinish(bp, NULL, ENXIO);
return;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
softc = (struct da_softc *)periph->softc;
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
cam_periph_lock(periph);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
* If the device has been made invalid, error out
*/
if ((softc->flags & DA_FLAG_PACK_INVALID)) {
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
cam_periph_unlock(periph);
Exploit recent improvements in the disk minilayer to simplify error handling a bit. Dogmatic lingupurists can celebrate that a number of gotos got removed. Reviewed by: mjacob, ken
2001-05-08 08:30:48 +00:00
biofinish(bp, NULL, ENXIO);
return;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
Tune and add some missing CAM_DEBUG() points for better consistency.
2012-06-05 09:45:42 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE, ("dastrategy(%p)\n", bp));
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
* Place it in the queue of disk activities for this disk
Work around the other uses of M_WAITOK.
2007-06-16 18:20:29 +00:00
*/
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
if (bp->bio_cmd == BIO_DELETE) {
if (bp->bio_bcount == 0)
biodone(bp);
else
bioq_disksort(&softc->delete_queue, bp);
} else
bioq_disksort(&softc->bio_queue, bp);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
* Schedule ourselves for performing the work.
*/
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
daschedule(periph);
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
cam_periph_unlock(periph);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
return;
}
static int
NO_GEOM cleanup: Retire the "d_dump_t" and use the "dumper_t" type instead. Dumper_t takes a void * as first arg which is more general than the dev_t taken by d_dump_t. (Remember: we could have net-dumpers if somebody wrote us one!) Define the convention for GEOM controlled disk devices to be that the first argument to the dumper function is the struct disk pointer. Change device drivers accordingly.
2003-02-21 19:00:48 +00:00
dadump(void *arg, void *virtual, vm_offset_t physical, off_t offset, size_t length)
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
{
Convert dadump to use reasonable data types so that some casting is unecessary.
1998-12-11 03:54:43 +00:00
struct cam_periph *periph;
struct da_softc *softc;
Introduce the disk mini-layer and devstat_end_transaction_buf() in cam/scsi. Somewhat reviewed by: ken
1999-10-01 09:34:10 +00:00
u_int secsize;
Convert dadump to use reasonable data types so that some casting is unecessary.
1998-12-11 03:54:43 +00:00
struct ccb_scsiio csio;
NO_GEOM cleanup: Retire the "d_dump_t" and use the "dumper_t" type instead. Dumper_t takes a void * as first arg which is more general than the dev_t taken by d_dump_t. (Remember: we could have net-dumpers if somebody wrote us one!) Define the convention for GEOM controlled disk devices to be that the first argument to the dumper function is the struct disk pointer. Change device drivers accordingly.
2003-02-21 19:00:48 +00:00
struct disk *dp;
One more major cam_periph_error() rewrite to improve error handling and reporting. It includes: - removing of error messages controlled by bootverbose, replacing them with more universal and informative debugging on CAM_DEBUG_INFO level, that is now built into the kernel by default; - more close following to the arguments submitted by caller, such as SF_PRINT_ALWAYS, SF_QUIET_IR and SF_NO_PRINT; consumer knows better which errors are usual/expected at this point and which are really informative; - adding two new flags SF_NO_RECOVERY and SF_NO_RETRY to allow caller specify how much assistance it needs at this point; previously consumers controlled that by not calling cam_periph_error() at all, but that made behavior inconsistent and debugging complicated; - tuning debug messages and taken actions order to make debugging output more readable and cause-effect relationships visible; - making camperiphdone() (common device recovery completion handler) to also use cam_periph_error() in most cases, instead of own dumb code; - removing manual sense fetching code from cam_periph_error(); I was told by number of people that it is SIM obligation to fetch sense data, so this code is useless and only significantly complicates recovery logic; - making ada, da and pass driver to use cam_periph_error() with new limited recovery options to handle error recovery and debugging in common way; as one of results, CAM_REQUEUE_REQ and other retrying statuses are now working fine with pass driver, that caused many problems before. - reverting r186891 by raj@ to avoid burning few seconds in tight DELAY() loops on device probe, while device simply loads media; I think that problem may already be fixed in other way, and even if it is not, solution must be different. Sponsored by: iXsystems, Inc. MFC after: 2 weeks
2012-06-09 13:07:44 +00:00
int error = 0;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
NO_GEOM cleanup: Retire the "d_dump_t" and use the "dumper_t" type instead. Dumper_t takes a void * as first arg which is more general than the dev_t taken by d_dump_t. (Remember: we could have net-dumpers if somebody wrote us one!) Define the convention for GEOM controlled disk devices to be that the first argument to the dumper function is the struct disk pointer. Change device drivers accordingly.
2003-02-21 19:00:48 +00:00
dp = arg;
NO_GEOM cleanup: Move to "struct disk *" centric api.
2003-02-25 22:06:21 +00:00
periph = dp->d_drv1;
DA (scsi) and AD (ata) diskdrivers: Make the dump routine do just writing of data.
2002-03-31 22:28:03 +00:00
if (periph == NULL)
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
return (ENXIO);
softc = (struct da_softc *)periph->softc;
Missed locking the dump and shutdown entry points in the scsi_da driver.
2007-04-18 05:14:16 +00:00
cam_periph_lock(periph);
DA (scsi) and AD (ata) diskdrivers: Make the dump routine do just writing of data.
2002-03-31 22:28:03 +00:00
secsize = softc->params.secsize;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Missed locking the dump and shutdown entry points in the scsi_da driver.
2007-04-18 05:14:16 +00:00
if ((softc->flags & DA_FLAG_PACK_INVALID) != 0) {
cam_periph_unlock(periph);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
return (ENXIO);
Missed locking the dump and shutdown entry points in the scsi_da driver.
2007-04-18 05:14:16 +00:00
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
DA (scsi) and AD (ata) diskdrivers: Make the dump routine do just writing of data.
2002-03-31 22:28:03 +00:00
if (length > 0) {
Replace most of priority numbers with defines. No logical changes.
2009-10-23 08:27:55 +00:00
xpt_setup_ccb(&csio.ccb_h, periph->path, CAM_PRIORITY_NORMAL);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
csio.ccb_h.ccb_state = DA_CCB_DUMP;
scsi_read_write(&csio,
One more major cam_periph_error() rewrite to improve error handling and reporting. It includes: - removing of error messages controlled by bootverbose, replacing them with more universal and informative debugging on CAM_DEBUG_INFO level, that is now built into the kernel by default; - more close following to the arguments submitted by caller, such as SF_PRINT_ALWAYS, SF_QUIET_IR and SF_NO_PRINT; consumer knows better which errors are usual/expected at this point and which are really informative; - adding two new flags SF_NO_RECOVERY and SF_NO_RETRY to allow caller specify how much assistance it needs at this point; previously consumers controlled that by not calling cam_periph_error() at all, but that made behavior inconsistent and debugging complicated; - tuning debug messages and taken actions order to make debugging output more readable and cause-effect relationships visible; - making camperiphdone() (common device recovery completion handler) to also use cam_periph_error() in most cases, instead of own dumb code; - removing manual sense fetching code from cam_periph_error(); I was told by number of people that it is SIM obligation to fetch sense data, so this code is useless and only significantly complicates recovery logic; - making ada, da and pass driver to use cam_periph_error() with new limited recovery options to handle error recovery and debugging in common way; as one of results, CAM_REQUEUE_REQ and other retrying statuses are now working fine with pass driver, that caused many problems before. - reverting r186891 by raj@ to avoid burning few seconds in tight DELAY() loops on device probe, while device simply loads media; I think that problem may already be fixed in other way, and even if it is not, solution must be different. Sponsored by: iXsystems, Inc. MFC after: 2 weeks
2012-06-09 13:07:44 +00:00
/*retries*/0,
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
dadone,
MSG_ORDERED_Q_TAG,
/*read*/FALSE,
/*byte2*/0,
"Fix" a problem with the Quantum Viking. It appears that this drive does not like the 6-byte read and write commands! It returns illegal request, with the field pointer pointing to byte 9 of a 6 byte CDB. In any case, the work around is to put in a quirk mechanism that makes sure that we don't send 6-byte reads or writes to this device. It's rather sad that this is necessary. You'd think that they would be able to get something that basic to work right in their firmware... Reviewed by: gibbs Reported by: Adam McDougall <bsdx@spawnet.com>
1998-12-02 17:35:28 +00:00
/*minimum_cmd_size*/ softc->minimum_cmd_size,
DA (scsi) and AD (ata) diskdrivers: Make the dump routine do just writing of data.
2002-03-31 22:28:03 +00:00
offset / secsize,
length / secsize,
/*data_ptr*/(u_int8_t *) virtual,
/*dxfer_len*/length,
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*sense_len*/SSD_FULL_SIZE,
Insert ordered command every 1/4 of the current command timeout, not 1/4 of the default one. Without this change setting kern.cam.ada.default_timeout to 1 instead of 30 allowed me to trigger several false positive command timeouts under heavy ZFS load on a SiI3132 siis(4) controller with 5 HDDs on a port multiplier. MFC after: 1 week
2012-02-02 19:02:15 +00:00
da_default_timeout * 1000);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
xpt_polled_action((union ccb *)&csio);
One more major cam_periph_error() rewrite to improve error handling and reporting. It includes: - removing of error messages controlled by bootverbose, replacing them with more universal and informative debugging on CAM_DEBUG_INFO level, that is now built into the kernel by default; - more close following to the arguments submitted by caller, such as SF_PRINT_ALWAYS, SF_QUIET_IR and SF_NO_PRINT; consumer knows better which errors are usual/expected at this point and which are really informative; - adding two new flags SF_NO_RECOVERY and SF_NO_RETRY to allow caller specify how much assistance it needs at this point; previously consumers controlled that by not calling cam_periph_error() at all, but that made behavior inconsistent and debugging complicated; - tuning debug messages and taken actions order to make debugging output more readable and cause-effect relationships visible; - making camperiphdone() (common device recovery completion handler) to also use cam_periph_error() in most cases, instead of own dumb code; - removing manual sense fetching code from cam_periph_error(); I was told by number of people that it is SIM obligation to fetch sense data, so this code is useless and only significantly complicates recovery logic; - making ada, da and pass driver to use cam_periph_error() with new limited recovery options to handle error recovery and debugging in common way; as one of results, CAM_REQUEUE_REQ and other retrying statuses are now working fine with pass driver, that caused many problems before. - reverting r186891 by raj@ to avoid burning few seconds in tight DELAY() loops on device probe, while device simply loads media; I think that problem may already be fixed in other way, and even if it is not, solution must be different. Sponsored by: iXsystems, Inc. MFC after: 2 weeks
2012-06-09 13:07:44 +00:00
error = cam_periph_error((union ccb *)&csio,
0, SF_NO_RECOVERY | SF_NO_RETRY, NULL);
if ((csio.ccb_h.status & CAM_DEV_QFRZN) != 0)
cam_release_devq(csio.ccb_h.path, /*relsim_flags*/0,
/*reduction*/0, /*timeout*/0, /*getcount_only*/0);
if (error != 0)
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
printf("Aborting dump due to I/O error.\n");
One more major cam_periph_error() rewrite to improve error handling and reporting. It includes: - removing of error messages controlled by bootverbose, replacing them with more universal and informative debugging on CAM_DEBUG_INFO level, that is now built into the kernel by default; - more close following to the arguments submitted by caller, such as SF_PRINT_ALWAYS, SF_QUIET_IR and SF_NO_PRINT; consumer knows better which errors are usual/expected at this point and which are really informative; - adding two new flags SF_NO_RECOVERY and SF_NO_RETRY to allow caller specify how much assistance it needs at this point; previously consumers controlled that by not calling cam_periph_error() at all, but that made behavior inconsistent and debugging complicated; - tuning debug messages and taken actions order to make debugging output more readable and cause-effect relationships visible; - making camperiphdone() (common device recovery completion handler) to also use cam_periph_error() in most cases, instead of own dumb code; - removing manual sense fetching code from cam_periph_error(); I was told by number of people that it is SIM obligation to fetch sense data, so this code is useless and only significantly complicates recovery logic; - making ada, da and pass driver to use cam_periph_error() with new limited recovery options to handle error recovery and debugging in common way; as one of results, CAM_REQUEUE_REQ and other retrying statuses are now working fine with pass driver, that caused many problems before. - reverting r186891 by raj@ to avoid burning few seconds in tight DELAY() loops on device probe, while device simply loads media; I think that problem may already be fixed in other way, and even if it is not, solution must be different. Sponsored by: iXsystems, Inc. MFC after: 2 weeks
2012-06-09 13:07:44 +00:00
cam_periph_unlock(periph);
return (error);
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
/*
* Sync the disk cache contents to the physical media.
*/
if ((softc->quirks & DA_Q_NO_SYNC_CACHE) == 0) {
Replace most of priority numbers with defines. No logical changes.
2009-10-23 08:27:55 +00:00
xpt_setup_ccb(&csio.ccb_h, periph->path, CAM_PRIORITY_NORMAL);
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
csio.ccb_h.ccb_state = DA_CCB_DUMP;
scsi_synchronize_cache(&csio,
One more major cam_periph_error() rewrite to improve error handling and reporting. It includes: - removing of error messages controlled by bootverbose, replacing them with more universal and informative debugging on CAM_DEBUG_INFO level, that is now built into the kernel by default; - more close following to the arguments submitted by caller, such as SF_PRINT_ALWAYS, SF_QUIET_IR and SF_NO_PRINT; consumer knows better which errors are usual/expected at this point and which are really informative; - adding two new flags SF_NO_RECOVERY and SF_NO_RETRY to allow caller specify how much assistance it needs at this point; previously consumers controlled that by not calling cam_periph_error() at all, but that made behavior inconsistent and debugging complicated; - tuning debug messages and taken actions order to make debugging output more readable and cause-effect relationships visible; - making camperiphdone() (common device recovery completion handler) to also use cam_periph_error() in most cases, instead of own dumb code; - removing manual sense fetching code from cam_periph_error(); I was told by number of people that it is SIM obligation to fetch sense data, so this code is useless and only significantly complicates recovery logic; - making ada, da and pass driver to use cam_periph_error() with new limited recovery options to handle error recovery and debugging in common way; as one of results, CAM_REQUEUE_REQ and other retrying statuses are now working fine with pass driver, that caused many problems before. - reverting r186891 by raj@ to avoid burning few seconds in tight DELAY() loops on device probe, while device simply loads media; I think that problem may already be fixed in other way, and even if it is not, solution must be different. Sponsored by: iXsystems, Inc. MFC after: 2 weeks
2012-06-09 13:07:44 +00:00
/*retries*/0,
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
/*cbfcnp*/dadone,
MSG_SIMPLE_Q_TAG,
/*begin_lba*/0,/* Cover the whole disk */
/*lb_count*/0,
SSD_FULL_SIZE,
5 * 60 * 1000);
xpt_polled_action((union ccb *)&csio);
One more major cam_periph_error() rewrite to improve error handling and reporting. It includes: - removing of error messages controlled by bootverbose, replacing them with more universal and informative debugging on CAM_DEBUG_INFO level, that is now built into the kernel by default; - more close following to the arguments submitted by caller, such as SF_PRINT_ALWAYS, SF_QUIET_IR and SF_NO_PRINT; consumer knows better which errors are usual/expected at this point and which are really informative; - adding two new flags SF_NO_RECOVERY and SF_NO_RETRY to allow caller specify how much assistance it needs at this point; previously consumers controlled that by not calling cam_periph_error() at all, but that made behavior inconsistent and debugging complicated; - tuning debug messages and taken actions order to make debugging output more readable and cause-effect relationships visible; - making camperiphdone() (common device recovery completion handler) to also use cam_periph_error() in most cases, instead of own dumb code; - removing manual sense fetching code from cam_periph_error(); I was told by number of people that it is SIM obligation to fetch sense data, so this code is useless and only significantly complicates recovery logic; - making ada, da and pass driver to use cam_periph_error() with new limited recovery options to handle error recovery and debugging in common way; as one of results, CAM_REQUEUE_REQ and other retrying statuses are now working fine with pass driver, that caused many problems before. - reverting r186891 by raj@ to avoid burning few seconds in tight DELAY() loops on device probe, while device simply loads media; I think that problem may already be fixed in other way, and even if it is not, solution must be different. Sponsored by: iXsystems, Inc. MFC after: 2 weeks
2012-06-09 13:07:44 +00:00
error = cam_periph_error((union ccb *)&csio,
0, SF_NO_RECOVERY | SF_NO_RETRY | SF_QUIET_IR, NULL);
if ((csio.ccb_h.status & CAM_DEV_QFRZN) != 0)
cam_release_devq(csio.ccb_h.path, /*relsim_flags*/0,
/*reduction*/0, /*timeout*/0, /*getcount_only*/0);
if (error != 0)
xpt_print(periph->path, "Synchronize cache failed\n");
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
}
Missed locking the dump and shutdown entry points in the scsi_da driver.
2007-04-18 05:14:16 +00:00
cam_periph_unlock(periph);
One more major cam_periph_error() rewrite to improve error handling and reporting. It includes: - removing of error messages controlled by bootverbose, replacing them with more universal and informative debugging on CAM_DEBUG_INFO level, that is now built into the kernel by default; - more close following to the arguments submitted by caller, such as SF_PRINT_ALWAYS, SF_QUIET_IR and SF_NO_PRINT; consumer knows better which errors are usual/expected at this point and which are really informative; - adding two new flags SF_NO_RECOVERY and SF_NO_RETRY to allow caller specify how much assistance it needs at this point; previously consumers controlled that by not calling cam_periph_error() at all, but that made behavior inconsistent and debugging complicated; - tuning debug messages and taken actions order to make debugging output more readable and cause-effect relationships visible; - making camperiphdone() (common device recovery completion handler) to also use cam_periph_error() in most cases, instead of own dumb code; - removing manual sense fetching code from cam_periph_error(); I was told by number of people that it is SIM obligation to fetch sense data, so this code is useless and only significantly complicates recovery logic; - making ada, da and pass driver to use cam_periph_error() with new limited recovery options to handle error recovery and debugging in common way; as one of results, CAM_REQUEUE_REQ and other retrying statuses are now working fine with pass driver, that caused many problems before. - reverting r186891 by raj@ to avoid burning few seconds in tight DELAY() loops on device probe, while device simply loads media; I think that problem may already be fixed in other way, and even if it is not, solution must be different. Sponsored by: iXsystems, Inc. MFC after: 2 weeks
2012-06-09 13:07:44 +00:00
return (error);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
Plumb device physical path reporting from CAM devices, through GEOM and DEVFS, and make it accessible via the diskinfo utility. Extend GEOM's generic attribute query mechanism into generic disk consumers. sys/geom/geom_disk.c: sys/geom/geom_disk.h: sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Allow disk providers to implement a new method which can override the default BIO_GETATTR response, d_getattr(struct bio *). This function returns -1 if not handled, otherwise it returns 0 or an errno to be passed to g_io_deliver(). sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Don't copy the serial number to dp->d_ident anymore, as the CAM XPT is now responsible for returning this information via d_getattr()->(a)dagetattr()->xpt_getatr(). sys/geom/geom_dev.c: - Implement a new ioctl, DIOCGPHYSPATH, which returns the GEOM attribute "GEOM::physpath", if possible. If the attribute request returns a zero-length string, ENOENT is returned. usr.sbin/diskinfo/diskinfo.c: - If the DIOCGPHYSPATH ioctl is successful, report physical path data when diskinfo is executed with the '-v' option. Submitted by: will Reviewed by: gibbs Sponsored by: Spectra Logic Corporation Add generic attribute change notification support to GEOM. sys/sys/geom/geom.h: Add a new attrchanged method field to both g_class and g_geom. sys/sys/geom/geom.h: sys/geom/geom_event.c: - Provide the g_attr_changed() function that providers can use to advertise attribute changes. - Perform delivery of attribute change notifications from a thread context via the standard GEOM event mechanism. sys/geom/geom_subr.c: Inherit the attrchanged method from class to geom (class instance). sys/geom/geom_disk.c: Provide disk_attr_changed() to provide g_attr_changed() access to consumers of the disk API. sys/cam/scsi/scsi_pass.c: sys/cam/scsi/scsi_da.c: sys/geom/geom_dev.c: sys/geom/geom_disk.c: Use attribute changed events to track updates to physical path information. sys/cam/scsi/scsi_da.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, and the updated buffer type references our physical path attribute, emit a GEOM attribute changed event via the disk_attr_changed() API. sys/cam/scsi/scsi_pass.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, update the physical patch devfs alias for this pass instance. Submitted by: gibbs Sponsored by: Spectra Logic Corporation
2011-06-14 17:10:32 +00:00
static int
dagetattr(struct bio *bp)
{
Protect xpt_getattr() calls with the SIM lock and assert that. Submitted by: ken@ (earlier version)
2012-10-12 17:18:24 +00:00
int ret;
Plumb device physical path reporting from CAM devices, through GEOM and DEVFS, and make it accessible via the diskinfo utility. Extend GEOM's generic attribute query mechanism into generic disk consumers. sys/geom/geom_disk.c: sys/geom/geom_disk.h: sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Allow disk providers to implement a new method which can override the default BIO_GETATTR response, d_getattr(struct bio *). This function returns -1 if not handled, otherwise it returns 0 or an errno to be passed to g_io_deliver(). sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Don't copy the serial number to dp->d_ident anymore, as the CAM XPT is now responsible for returning this information via d_getattr()->(a)dagetattr()->xpt_getatr(). sys/geom/geom_dev.c: - Implement a new ioctl, DIOCGPHYSPATH, which returns the GEOM attribute "GEOM::physpath", if possible. If the attribute request returns a zero-length string, ENOENT is returned. usr.sbin/diskinfo/diskinfo.c: - If the DIOCGPHYSPATH ioctl is successful, report physical path data when diskinfo is executed with the '-v' option. Submitted by: will Reviewed by: gibbs Sponsored by: Spectra Logic Corporation Add generic attribute change notification support to GEOM. sys/sys/geom/geom.h: Add a new attrchanged method field to both g_class and g_geom. sys/sys/geom/geom.h: sys/geom/geom_event.c: - Provide the g_attr_changed() function that providers can use to advertise attribute changes. - Perform delivery of attribute change notifications from a thread context via the standard GEOM event mechanism. sys/geom/geom_subr.c: Inherit the attrchanged method from class to geom (class instance). sys/geom/geom_disk.c: Provide disk_attr_changed() to provide g_attr_changed() access to consumers of the disk API. sys/cam/scsi/scsi_pass.c: sys/cam/scsi/scsi_da.c: sys/geom/geom_dev.c: sys/geom/geom_disk.c: Use attribute changed events to track updates to physical path information. sys/cam/scsi/scsi_da.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, and the updated buffer type references our physical path attribute, emit a GEOM attribute changed event via the disk_attr_changed() API. sys/cam/scsi/scsi_pass.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, update the physical patch devfs alias for this pass instance. Submitted by: gibbs Sponsored by: Spectra Logic Corporation
2011-06-14 17:10:32 +00:00
struct cam_periph *periph;
periph = (struct cam_periph *)bp->bio_disk->d_drv1;
Protect xpt_getattr() calls with the SIM lock and assert that. Submitted by: ken@ (earlier version)
2012-10-12 17:18:24 +00:00
if (periph == NULL)
return (ENXIO);
Plumb device physical path reporting from CAM devices, through GEOM and DEVFS, and make it accessible via the diskinfo utility. Extend GEOM's generic attribute query mechanism into generic disk consumers. sys/geom/geom_disk.c: sys/geom/geom_disk.h: sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Allow disk providers to implement a new method which can override the default BIO_GETATTR response, d_getattr(struct bio *). This function returns -1 if not handled, otherwise it returns 0 or an errno to be passed to g_io_deliver(). sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Don't copy the serial number to dp->d_ident anymore, as the CAM XPT is now responsible for returning this information via d_getattr()->(a)dagetattr()->xpt_getatr(). sys/geom/geom_dev.c: - Implement a new ioctl, DIOCGPHYSPATH, which returns the GEOM attribute "GEOM::physpath", if possible. If the attribute request returns a zero-length string, ENOENT is returned. usr.sbin/diskinfo/diskinfo.c: - If the DIOCGPHYSPATH ioctl is successful, report physical path data when diskinfo is executed with the '-v' option. Submitted by: will Reviewed by: gibbs Sponsored by: Spectra Logic Corporation Add generic attribute change notification support to GEOM. sys/sys/geom/geom.h: Add a new attrchanged method field to both g_class and g_geom. sys/sys/geom/geom.h: sys/geom/geom_event.c: - Provide the g_attr_changed() function that providers can use to advertise attribute changes. - Perform delivery of attribute change notifications from a thread context via the standard GEOM event mechanism. sys/geom/geom_subr.c: Inherit the attrchanged method from class to geom (class instance). sys/geom/geom_disk.c: Provide disk_attr_changed() to provide g_attr_changed() access to consumers of the disk API. sys/cam/scsi/scsi_pass.c: sys/cam/scsi/scsi_da.c: sys/geom/geom_dev.c: sys/geom/geom_disk.c: Use attribute changed events to track updates to physical path information. sys/cam/scsi/scsi_da.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, and the updated buffer type references our physical path attribute, emit a GEOM attribute changed event via the disk_attr_changed() API. sys/cam/scsi/scsi_pass.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, update the physical patch devfs alias for this pass instance. Submitted by: gibbs Sponsored by: Spectra Logic Corporation
2011-06-14 17:10:32 +00:00
Protect xpt_getattr() calls with the SIM lock and assert that. Submitted by: ken@ (earlier version)
2012-10-12 17:18:24 +00:00
cam_periph_lock(periph);
Plumb device physical path reporting from CAM devices, through GEOM and DEVFS, and make it accessible via the diskinfo utility. Extend GEOM's generic attribute query mechanism into generic disk consumers. sys/geom/geom_disk.c: sys/geom/geom_disk.h: sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Allow disk providers to implement a new method which can override the default BIO_GETATTR response, d_getattr(struct bio *). This function returns -1 if not handled, otherwise it returns 0 or an errno to be passed to g_io_deliver(). sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Don't copy the serial number to dp->d_ident anymore, as the CAM XPT is now responsible for returning this information via d_getattr()->(a)dagetattr()->xpt_getatr(). sys/geom/geom_dev.c: - Implement a new ioctl, DIOCGPHYSPATH, which returns the GEOM attribute "GEOM::physpath", if possible. If the attribute request returns a zero-length string, ENOENT is returned. usr.sbin/diskinfo/diskinfo.c: - If the DIOCGPHYSPATH ioctl is successful, report physical path data when diskinfo is executed with the '-v' option. Submitted by: will Reviewed by: gibbs Sponsored by: Spectra Logic Corporation Add generic attribute change notification support to GEOM. sys/sys/geom/geom.h: Add a new attrchanged method field to both g_class and g_geom. sys/sys/geom/geom.h: sys/geom/geom_event.c: - Provide the g_attr_changed() function that providers can use to advertise attribute changes. - Perform delivery of attribute change notifications from a thread context via the standard GEOM event mechanism. sys/geom/geom_subr.c: Inherit the attrchanged method from class to geom (class instance). sys/geom/geom_disk.c: Provide disk_attr_changed() to provide g_attr_changed() access to consumers of the disk API. sys/cam/scsi/scsi_pass.c: sys/cam/scsi/scsi_da.c: sys/geom/geom_dev.c: sys/geom/geom_disk.c: Use attribute changed events to track updates to physical path information. sys/cam/scsi/scsi_da.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, and the updated buffer type references our physical path attribute, emit a GEOM attribute changed event via the disk_attr_changed() API. sys/cam/scsi/scsi_pass.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, update the physical patch devfs alias for this pass instance. Submitted by: gibbs Sponsored by: Spectra Logic Corporation
2011-06-14 17:10:32 +00:00
ret = xpt_getattr(bp->bio_data, bp->bio_length, bp->bio_attribute,
periph->path);
Protect xpt_getattr() calls with the SIM lock and assert that. Submitted by: ken@ (earlier version)
2012-10-12 17:18:24 +00:00
cam_periph_unlock(periph);
Plumb device physical path reporting from CAM devices, through GEOM and DEVFS, and make it accessible via the diskinfo utility. Extend GEOM's generic attribute query mechanism into generic disk consumers. sys/geom/geom_disk.c: sys/geom/geom_disk.h: sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Allow disk providers to implement a new method which can override the default BIO_GETATTR response, d_getattr(struct bio *). This function returns -1 if not handled, otherwise it returns 0 or an errno to be passed to g_io_deliver(). sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Don't copy the serial number to dp->d_ident anymore, as the CAM XPT is now responsible for returning this information via d_getattr()->(a)dagetattr()->xpt_getatr(). sys/geom/geom_dev.c: - Implement a new ioctl, DIOCGPHYSPATH, which returns the GEOM attribute "GEOM::physpath", if possible. If the attribute request returns a zero-length string, ENOENT is returned. usr.sbin/diskinfo/diskinfo.c: - If the DIOCGPHYSPATH ioctl is successful, report physical path data when diskinfo is executed with the '-v' option. Submitted by: will Reviewed by: gibbs Sponsored by: Spectra Logic Corporation Add generic attribute change notification support to GEOM. sys/sys/geom/geom.h: Add a new attrchanged method field to both g_class and g_geom. sys/sys/geom/geom.h: sys/geom/geom_event.c: - Provide the g_attr_changed() function that providers can use to advertise attribute changes. - Perform delivery of attribute change notifications from a thread context via the standard GEOM event mechanism. sys/geom/geom_subr.c: Inherit the attrchanged method from class to geom (class instance). sys/geom/geom_disk.c: Provide disk_attr_changed() to provide g_attr_changed() access to consumers of the disk API. sys/cam/scsi/scsi_pass.c: sys/cam/scsi/scsi_da.c: sys/geom/geom_dev.c: sys/geom/geom_disk.c: Use attribute changed events to track updates to physical path information. sys/cam/scsi/scsi_da.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, and the updated buffer type references our physical path attribute, emit a GEOM attribute changed event via the disk_attr_changed() API. sys/cam/scsi/scsi_pass.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, update the physical patch devfs alias for this pass instance. Submitted by: gibbs Sponsored by: Spectra Logic Corporation
2011-06-14 17:10:32 +00:00
if (ret == 0)
bp->bio_completed = bp->bio_length;
return ret;
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
static void
dainit(void)
{
cam_status status;
/*
* Install a global async callback. This callback will
* receive async callbacks like "new device found".
*/
Add a helper function for registering async callbacks. Besides eliminating a lot of duplicated code, this also fixes a locking edge case.
2007-05-16 16:54:23 +00:00
status = xpt_register_async(AC_FOUND_DEVICE, daasync, NULL, NULL);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if (status != CAM_REQ_CMP) {
printf("da: Failed to attach master async callback "
"due to status 0x%x!\n", status);
Add a tunable that allows one to turn off the automatic sending of the ORDERED tag. This recoups significant performance gains for many arrays. The default is still to send out the ORDERED tag periodically. Reviewed by: scsi (justin+timeout)
2006-11-02 21:12:37 +00:00
} else if (da_send_ordered) {
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
Implement a new generic mechanism for attaching handler functions to events, in order to pave the way for removing a number of the ad-hoc implementations currently in use. Retire the at_shutdown family of functions and replace them with new event handler lists. Rework kern_shutdown.c to take greater advantage of the use of event handlers. Reviewed by: green
1999-08-21 06:24:40 +00:00
/* Register our shutdown event handler */
if ((EVENTHANDLER_REGISTER(shutdown_post_sync, dashutdown,
Back out last commit.
2003-01-06 19:30:21 +00:00
NULL, SHUTDOWN_PRI_DEFAULT)) == NULL)
Implement a new generic mechanism for attaching handler functions to events, in order to pave the way for removing a number of the ad-hoc implementations currently in use. Retire the at_shutdown family of functions and replace them with new event handler lists. Rework kern_shutdown.c to take greater advantage of the use of event handlers. Reviewed by: green
1999-08-21 06:24:40 +00:00
printf("dainit: shutdown event registration failed!\n");
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
}
Fix a bug which causes a panic in daopen(). The panic is caused by a da(4) instance going away while GEOM is still probing it. In this case, the GEOM disk class instance has been created by disk_create(), and the taste of the disk is queued in the GEOM event queue. While that event is queued, the da(4) instance goes away. When the open call comes into the da(4) driver, it dereferences the freed (but non-NULL) peripheral pointer provided by GEOM, which results in a panic. The solution is to add a callback to the GEOM disk code that is called when all of its resources are cleaned up. This is implemented inside GEOM by adding an optional callback that is called when all consumers have detached from a provider, and the provider is about to be deleted. scsi_cd.c, scsi_da.c: In the register routine for the cd(4) and da(4) routines, acquire a reference to the CAM peripheral instance just before we call disk_create(). Use the new GEOM disk d_gone() callback to register a callback (dadiskgonecb()/cddiskgonecb()) that decrements the peripheral reference count once GEOM has finished cleaning up its resources. In the cd(4) driver, clean up open and close behavior slightly. GEOM makes sure we only get one open() and one close call, so there is no need to set an open flag and decrement the reference count if we are not the first open. In the cd(4) driver, use cam_periph_release_locked() in a couple of error scenarios to avoid extra mutex calls. geom.h: Add a new, optional, providergone callback that is called when a provider is about to be deleted. geom_disk.h: Add a new d_gone() callback to the GEOM disk interface. Bump the DISK_VERSION to version 2. This probably should have been done after a couple of previous changes, especially the addition of the d_getattr() callback. geom_disk.c: Add a providergone callback for the disk class, g_disk_providergone(), that calls the user's d_gone() callback if it exists. Bump the DISK_VERSION to 2. geom_subr.c: In g_destroy_provider(), call the providergone callback if it has been provided. In g_new_geomf(), propagate the class's providergone callback to the new geom instance. blkfront.c: Callers of disk_create() are supposed to pass in DISK_VERSION, not an explicit disk API version number. Update the blkfront driver to do that. disk.9: Update the disk(9) man page to include information on the new d_gone() callback, as well as the previously added d_getattr() callback, d_descr field, and HBA PCI ID fields. MFC after: 5 days
2012-06-24 04:29:03 +00:00
/*
* Callback from GEOM, called when it has finished cleaning up its
* resources.
*/
static void
dadiskgonecb(struct disk *dp)
{
struct cam_periph *periph;
periph = (struct cam_periph *)dp->d_drv1;
cam_periph_release(periph);
}
Fix a problem with the way we handled device invalidation when attaching to a device failed. In theory, the same steps that happen when we get an AC_LOST_DEVICE async notification should have been taken when a driver fails to attach. In practice, that wasn't the case. This only affected the da, cd and ch drivers, but the fix affects all peripheral drivers. There were several possible problems: - In the da driver, we didn't remove the peripheral's softc from the da driver's linked list of softcs. Once the peripheral and softc got removed, we'd get a kernel panic the next time the timeout routine called dasendorderedtag(). - In the da, cd and possibly ch drivers, we didn't remove the peripheral's devstat structure from the devstat queue. Once the peripheral and softc were removed, this could cause a panic if anyone tried to access device statistics. (one component of the linked list wouldn't exist anymore) - In the cd driver, we didn't take the peripheral off the changer run queue if it was scheduled to run. In practice, it's highly unlikely, and maybe impossible that the peripheral would have been on the changer run queue at that stage of the probe process. The fix is: - Add a new peripheral callback function (the "oninvalidate" function) that is called the first time cam_periph_invalidate() is called for a peripheral. - Create new foooninvalidate() routines for each peripheral driver. This routine is always called at splsoftcam(), and contains all the stuff that used to be in the AC_LOST_DEVICE case of the async callback handler. - Move the devstat cleanup call to the destructor/cleanup routines, since some of the drivers do I/O in their close routines. - Make sure that when we're flushing the buffer queue, we traverse it at splbio(). - Add a check for the invalid flag in the pt driver's open routine. Reviewed by: gibbs
1998-10-22 22:16:56 +00:00
static void
daoninvalidate(struct cam_periph *periph)
{
struct da_softc *softc;
softc = (struct da_softc *)periph->softc;
/*
* De-register any async callbacks.
*/
Add a helper function for registering async callbacks. Besides eliminating a lot of duplicated code, this also fixes a locking edge case.
2007-05-16 16:54:23 +00:00
xpt_register_async(0, daasync, periph, periph->path);
Fix a problem with the way we handled device invalidation when attaching to a device failed. In theory, the same steps that happen when we get an AC_LOST_DEVICE async notification should have been taken when a driver fails to attach. In practice, that wasn't the case. This only affected the da, cd and ch drivers, but the fix affects all peripheral drivers. There were several possible problems: - In the da driver, we didn't remove the peripheral's softc from the da driver's linked list of softcs. Once the peripheral and softc got removed, we'd get a kernel panic the next time the timeout routine called dasendorderedtag(). - In the da, cd and possibly ch drivers, we didn't remove the peripheral's devstat structure from the devstat queue. Once the peripheral and softc were removed, this could cause a panic if anyone tried to access device statistics. (one component of the linked list wouldn't exist anymore) - In the cd driver, we didn't take the peripheral off the changer run queue if it was scheduled to run. In practice, it's highly unlikely, and maybe impossible that the peripheral would have been on the changer run queue at that stage of the probe process. The fix is: - Add a new peripheral callback function (the "oninvalidate" function) that is called the first time cam_periph_invalidate() is called for a peripheral. - Create new foooninvalidate() routines for each peripheral driver. This routine is always called at splsoftcam(), and contains all the stuff that used to be in the AC_LOST_DEVICE case of the async callback handler. - Move the devstat cleanup call to the destructor/cleanup routines, since some of the drivers do I/O in their close routines. - Make sure that when we're flushing the buffer queue, we traverse it at splbio(). - Add a check for the invalid flag in the pt driver's open routine. Reviewed by: gibbs
1998-10-22 22:16:56 +00:00
softc->flags |= DA_FLAG_PACK_INVALID;
/*
* Return all queued I/O with ENXIO.
Revert r211434. Offline discussions have convinced me that this should be left alone for now.
2010-08-20 17:20:05 +00:00
* XXX Handle any transactions queued to the card
* with XPT_ABORT_CCB.
Fix a problem with the way we handled device invalidation when attaching to a device failed. In theory, the same steps that happen when we get an AC_LOST_DEVICE async notification should have been taken when a driver fails to attach. In practice, that wasn't the case. This only affected the da, cd and ch drivers, but the fix affects all peripheral drivers. There were several possible problems: - In the da driver, we didn't remove the peripheral's softc from the da driver's linked list of softcs. Once the peripheral and softc got removed, we'd get a kernel panic the next time the timeout routine called dasendorderedtag(). - In the da, cd and possibly ch drivers, we didn't remove the peripheral's devstat structure from the devstat queue. Once the peripheral and softc were removed, this could cause a panic if anyone tried to access device statistics. (one component of the linked list wouldn't exist anymore) - In the cd driver, we didn't take the peripheral off the changer run queue if it was scheduled to run. In practice, it's highly unlikely, and maybe impossible that the peripheral would have been on the changer run queue at that stage of the probe process. The fix is: - Add a new peripheral callback function (the "oninvalidate" function) that is called the first time cam_periph_invalidate() is called for a peripheral. - Create new foooninvalidate() routines for each peripheral driver. This routine is always called at splsoftcam(), and contains all the stuff that used to be in the AC_LOST_DEVICE case of the async callback handler. - Move the devstat cleanup call to the destructor/cleanup routines, since some of the drivers do I/O in their close routines. - Make sure that when we're flushing the buffer queue, we traverse it at splbio(). - Add a check for the invalid flag in the pt driver's open routine. Reviewed by: gibbs
1998-10-22 22:16:56 +00:00
*/
Use bioq_flush() to drain a bio queue with a specific error code. Retain the mistake of not updating the devstat API for now. Spell bioq_disksort() consistently with the remaining bioq_*(). #include <geom/geom_disk.h> where this is more appropriate.
2003-04-01 15:06:26 +00:00
bioq_flush(&softc->bio_queue, NULL, ENXIO);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
bioq_flush(&softc->delete_queue, NULL, ENXIO);
Fix a problem with the way we handled device invalidation when attaching to a device failed. In theory, the same steps that happen when we get an AC_LOST_DEVICE async notification should have been taken when a driver fails to attach. In practice, that wasn't the case. This only affected the da, cd and ch drivers, but the fix affects all peripheral drivers. There were several possible problems: - In the da driver, we didn't remove the peripheral's softc from the da driver's linked list of softcs. Once the peripheral and softc got removed, we'd get a kernel panic the next time the timeout routine called dasendorderedtag(). - In the da, cd and possibly ch drivers, we didn't remove the peripheral's devstat structure from the devstat queue. Once the peripheral and softc were removed, this could cause a panic if anyone tried to access device statistics. (one component of the linked list wouldn't exist anymore) - In the cd driver, we didn't take the peripheral off the changer run queue if it was scheduled to run. In practice, it's highly unlikely, and maybe impossible that the peripheral would have been on the changer run queue at that stage of the probe process. The fix is: - Add a new peripheral callback function (the "oninvalidate" function) that is called the first time cam_periph_invalidate() is called for a peripheral. - Create new foooninvalidate() routines for each peripheral driver. This routine is always called at splsoftcam(), and contains all the stuff that used to be in the AC_LOST_DEVICE case of the async callback handler. - Move the devstat cleanup call to the destructor/cleanup routines, since some of the drivers do I/O in their close routines. - Make sure that when we're flushing the buffer queue, we traverse it at splbio(). - Add a check for the invalid flag in the pt driver's open routine. Reviewed by: gibbs
1998-10-22 22:16:56 +00:00
Fix a bug which causes a panic in daopen(). The panic is caused by a da(4) instance going away while GEOM is still probing it. In this case, the GEOM disk class instance has been created by disk_create(), and the taste of the disk is queued in the GEOM event queue. While that event is queued, the da(4) instance goes away. When the open call comes into the da(4) driver, it dereferences the freed (but non-NULL) peripheral pointer provided by GEOM, which results in a panic. The solution is to add a callback to the GEOM disk code that is called when all of its resources are cleaned up. This is implemented inside GEOM by adding an optional callback that is called when all consumers have detached from a provider, and the provider is about to be deleted. scsi_cd.c, scsi_da.c: In the register routine for the cd(4) and da(4) routines, acquire a reference to the CAM peripheral instance just before we call disk_create(). Use the new GEOM disk d_gone() callback to register a callback (dadiskgonecb()/cddiskgonecb()) that decrements the peripheral reference count once GEOM has finished cleaning up its resources. In the cd(4) driver, clean up open and close behavior slightly. GEOM makes sure we only get one open() and one close call, so there is no need to set an open flag and decrement the reference count if we are not the first open. In the cd(4) driver, use cam_periph_release_locked() in a couple of error scenarios to avoid extra mutex calls. geom.h: Add a new, optional, providergone callback that is called when a provider is about to be deleted. geom_disk.h: Add a new d_gone() callback to the GEOM disk interface. Bump the DISK_VERSION to version 2. This probably should have been done after a couple of previous changes, especially the addition of the d_getattr() callback. geom_disk.c: Add a providergone callback for the disk class, g_disk_providergone(), that calls the user's d_gone() callback if it exists. Bump the DISK_VERSION to 2. geom_subr.c: In g_destroy_provider(), call the providergone callback if it has been provided. In g_new_geomf(), propagate the class's providergone callback to the new geom instance. blkfront.c: Callers of disk_create() are supposed to pass in DISK_VERSION, not an explicit disk API version number. Update the blkfront driver to do that. disk.9: Update the disk(9) man page to include information on the new d_gone() callback, as well as the previously added d_getattr() callback, d_descr field, and HBA PCI ID fields. MFC after: 5 days
2012-06-24 04:29:03 +00:00
/*
* Tell GEOM that we've gone away, we'll get a callback when it is
* done cleaning up its resources.
*/
Fix a bug that caused some /dev entries to continue to exist after the underlying drive had been hot-unplugged from the system. Here is a specific example. Filesystem code had opened /dev/da1s1e. Subsequently, the drive was hot-unplugged. This (correctly) caused all of the associated /dev/da1* entries to be deleted. When the filesystem later realized that the drive was gone it closed the device, reducing the write-access counts to 0 on the geom providers for da1s1e, da1s1, and da1. This caused geom to re-taste the providers, resulting in the devices being created again. When the drive was hot-plugged back in, it resulted in duplicate /dev entries for da1s1e, da1s1, and da1. This fix adds a new disk_gone() function which is called by CAM when a drive goes away. It orphans all of the providers associated with the drive, setting an error condition of ENXIO in each one. In addition, we prevent a re-taste on last close for writing if an error condition has been set in the provider. Sponsored by: Isilon Systems Reviewed by: phk MFC after: 1 week
2005-11-18 02:43:49 +00:00
disk_gone(softc->disk);
Fix a bug which causes a panic in daopen(). The panic is caused by a da(4) instance going away while GEOM is still probing it. In this case, the GEOM disk class instance has been created by disk_create(), and the taste of the disk is queued in the GEOM event queue. While that event is queued, the da(4) instance goes away. When the open call comes into the da(4) driver, it dereferences the freed (but non-NULL) peripheral pointer provided by GEOM, which results in a panic. The solution is to add a callback to the GEOM disk code that is called when all of its resources are cleaned up. This is implemented inside GEOM by adding an optional callback that is called when all consumers have detached from a provider, and the provider is about to be deleted. scsi_cd.c, scsi_da.c: In the register routine for the cd(4) and da(4) routines, acquire a reference to the CAM peripheral instance just before we call disk_create(). Use the new GEOM disk d_gone() callback to register a callback (dadiskgonecb()/cddiskgonecb()) that decrements the peripheral reference count once GEOM has finished cleaning up its resources. In the cd(4) driver, clean up open and close behavior slightly. GEOM makes sure we only get one open() and one close call, so there is no need to set an open flag and decrement the reference count if we are not the first open. In the cd(4) driver, use cam_periph_release_locked() in a couple of error scenarios to avoid extra mutex calls. geom.h: Add a new, optional, providergone callback that is called when a provider is about to be deleted. geom_disk.h: Add a new d_gone() callback to the GEOM disk interface. Bump the DISK_VERSION to version 2. This probably should have been done after a couple of previous changes, especially the addition of the d_getattr() callback. geom_disk.c: Add a providergone callback for the disk class, g_disk_providergone(), that calls the user's d_gone() callback if it exists. Bump the DISK_VERSION to 2. geom_subr.c: In g_destroy_provider(), call the providergone callback if it has been provided. In g_new_geomf(), propagate the class's providergone callback to the new geom instance. blkfront.c: Callers of disk_create() are supposed to pass in DISK_VERSION, not an explicit disk API version number. Update the blkfront driver to do that. disk.9: Update the disk(9) man page to include information on the new d_gone() callback, as well as the previously added d_getattr() callback, d_descr field, and HBA PCI ID fields. MFC after: 5 days
2012-06-24 04:29:03 +00:00
Fix a race condition in CAM peripheral free handling, locking in the CAM XPT bus traversal code, and a number of other periph level issues. cam_periph.h, cam_periph.c: Modify cam_periph_acquire() to test the CAM_PERIPH_INVALID flag prior to allowing a reference count to be gained on a peripheral. Callers of this function will receive CAM_REQ_CMP_ERR status in the situation of attempting to reference an invalidated periph. This guarantees that a peripheral scheduled for a deferred free will not be accessed during its wait for destruction. Panic during attempts to drop a reference count on a peripheral that already has a zero reference count. In cam_periph_list(), use a local sbuf with SBUF_FIXEDLEN set so that mallocs do not occur while the xpt topology lock is held, regardless of the allocation policy of the passed in sbuf. Add a new routine, cam_periph_release_locked_buses(), that can be called when the caller already holds the CAM topology lock. Add some extra debugging for duplicate peripheral allocations in cam_periph_alloc(). Treat CAM_DEV_NOT_THERE much the same as a selection timeout (AC_LOST_DEVICE is emitted), but forgo retries. cam_xpt.c: Revamp the way the EDT traversal code does locking and reference counting. This was broken, since it assumed that the EDT would not change during traversal, but that assumption is no longer valid. So, to prevent devices from going away while we traverse the EDT, make sure we properly lock everything and hold references on devices that we are using. The two peripheral driver traversal routines should be examined. xptpdperiphtraverse() holds the topology lock for the entire time it runs. xptperiphtraverse() is now locked properly, but only holds the topology lock while it is traversing the list, and not while the traversal function is running. The bus locking code in xptbustraverse() should also be revisited at a later time, since it is complex and should probably be simplified. scsi_da.c: Pay attention to the return value from cam_periph_acquire(). Return 0 always from daclose() even if the disk is now gone. Add some rudimentary error injection support. scsi_sg.c: Fix reference counting in the sg(4) driver. The sg driver was calling cam_periph_release() on close, but never called cam_periph_acquire() (which increments the reference count) on open. The periph code correctly complained that the sg(4) driver was trying to decrement the refcount when it was already 0. Sponsored by: Spectra Logic MFC after: 2 weeks
2012-01-12 00:41:48 +00:00
xpt_print(periph->path, "lost device - %d outstanding, %d refs\n",
softc->outstanding_cmds, periph->refcount);
Fix a problem with the way we handled device invalidation when attaching to a device failed. In theory, the same steps that happen when we get an AC_LOST_DEVICE async notification should have been taken when a driver fails to attach. In practice, that wasn't the case. This only affected the da, cd and ch drivers, but the fix affects all peripheral drivers. There were several possible problems: - In the da driver, we didn't remove the peripheral's softc from the da driver's linked list of softcs. Once the peripheral and softc got removed, we'd get a kernel panic the next time the timeout routine called dasendorderedtag(). - In the da, cd and possibly ch drivers, we didn't remove the peripheral's devstat structure from the devstat queue. Once the peripheral and softc were removed, this could cause a panic if anyone tried to access device statistics. (one component of the linked list wouldn't exist anymore) - In the cd driver, we didn't take the peripheral off the changer run queue if it was scheduled to run. In practice, it's highly unlikely, and maybe impossible that the peripheral would have been on the changer run queue at that stage of the probe process. The fix is: - Add a new peripheral callback function (the "oninvalidate" function) that is called the first time cam_periph_invalidate() is called for a peripheral. - Create new foooninvalidate() routines for each peripheral driver. This routine is always called at splsoftcam(), and contains all the stuff that used to be in the AC_LOST_DEVICE case of the async callback handler. - Move the devstat cleanup call to the destructor/cleanup routines, since some of the drivers do I/O in their close routines. - Make sure that when we're flushing the buffer queue, we traverse it at splbio(). - Add a check for the invalid flag in the pt driver's open routine. Reviewed by: gibbs
1998-10-22 22:16:56 +00:00
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
static void
dacleanup(struct cam_periph *periph)
{
Fix a problem with the way we handled device invalidation when attaching to a device failed. In theory, the same steps that happen when we get an AC_LOST_DEVICE async notification should have been taken when a driver fails to attach. In practice, that wasn't the case. This only affected the da, cd and ch drivers, but the fix affects all peripheral drivers. There were several possible problems: - In the da driver, we didn't remove the peripheral's softc from the da driver's linked list of softcs. Once the peripheral and softc got removed, we'd get a kernel panic the next time the timeout routine called dasendorderedtag(). - In the da, cd and possibly ch drivers, we didn't remove the peripheral's devstat structure from the devstat queue. Once the peripheral and softc were removed, this could cause a panic if anyone tried to access device statistics. (one component of the linked list wouldn't exist anymore) - In the cd driver, we didn't take the peripheral off the changer run queue if it was scheduled to run. In practice, it's highly unlikely, and maybe impossible that the peripheral would have been on the changer run queue at that stage of the probe process. The fix is: - Add a new peripheral callback function (the "oninvalidate" function) that is called the first time cam_periph_invalidate() is called for a peripheral. - Create new foooninvalidate() routines for each peripheral driver. This routine is always called at splsoftcam(), and contains all the stuff that used to be in the AC_LOST_DEVICE case of the async callback handler. - Move the devstat cleanup call to the destructor/cleanup routines, since some of the drivers do I/O in their close routines. - Make sure that when we're flushing the buffer queue, we traverse it at splbio(). - Add a check for the invalid flag in the pt driver's open routine. Reviewed by: gibbs
1998-10-22 22:16:56 +00:00
struct da_softc *softc;
softc = (struct da_softc *)periph->softc;
Add an xpt_print function to reduce most of the xpt_print_path/printf pairs. Convert the core code to use it. Reviewed by: scsi (various folks weighing in with good advice) MFC after: 2 weeks
2006-12-05 07:45:28 +00:00
xpt_print(periph->path, "removing device entry\n");
Reorder dacleanup() and cdcleanup() slightly so that the sysctl context is freed while the periph lock is not held. While here, wait until after freeing the softc before reacquiring the periph lock. Tested by: sbruno
2009-02-11 22:29:09 +00:00
cam_periph_unlock(periph);
Clean up dynamically allocated sysctl variables when we run dacleanup() and cdcleanup(). This fixes sysctl problems ("can't re-use a leaf") when someone adds another peripheral at the same unit number. (e.g. rescan da0, it goes away, then rescan again and da0 comes back, but since we haven't cleaned up the sysctl variables from the last da0 instance, we can't register the variables for the new instance under the same name.) Reported by: njl Tested by: njl
2003-03-26 04:38:39 +00:00
/*
* If we can't free the sysctl tree, oh well...
*/
(dacleanup, dasysctlinit): Defend against calling sysctl_ctx_free on an uninitialized sysctl_ctx, using flag DA_FLAG_SCTX_INIT. This prevents a panic encoutered with some umass units that probe correctly but fail to attach. Same problem, and same fix, as scsi_cd.c rev. 1.86. Reviewed by: njl, ken
2003-10-08 07:12:30 +00:00
if ((softc->flags & DA_FLAG_SCTX_INIT) != 0
&& sysctl_ctx_free(&softc->sysctl_ctx) != 0) {
Add an xpt_print function to reduce most of the xpt_print_path/printf pairs. Convert the core code to use it. Reviewed by: scsi (various folks weighing in with good advice) MFC after: 2 weeks
2006-12-05 07:45:28 +00:00
xpt_print(periph->path, "can't remove sysctl context\n");
Clean up dynamically allocated sysctl variables when we run dacleanup() and cdcleanup(). This fixes sysctl problems ("can't re-use a leaf") when someone adds another peripheral at the same unit number. (e.g. rescan da0, it goes away, then rescan again and da0 comes back, but since we haven't cleaned up the sysctl variables from the last da0 instance, we can't register the variables for the new instance under the same name.) Reported by: njl Tested by: njl
2003-03-26 04:38:39 +00:00
}
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
callout_drain(&softc->mediapoll_c);
Drop the periph/sim lock when calling disk_destroy().
2007-04-16 19:41:14 +00:00
disk_destroy(softc->disk);
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
callout_drain(&softc->sendordered_c);
Fix a problem with the way we handled device invalidation when attaching to a device failed. In theory, the same steps that happen when we get an AC_LOST_DEVICE async notification should have been taken when a driver fails to attach. In practice, that wasn't the case. This only affected the da, cd and ch drivers, but the fix affects all peripheral drivers. There were several possible problems: - In the da driver, we didn't remove the peripheral's softc from the da driver's linked list of softcs. Once the peripheral and softc got removed, we'd get a kernel panic the next time the timeout routine called dasendorderedtag(). - In the da, cd and possibly ch drivers, we didn't remove the peripheral's devstat structure from the devstat queue. Once the peripheral and softc were removed, this could cause a panic if anyone tried to access device statistics. (one component of the linked list wouldn't exist anymore) - In the cd driver, we didn't take the peripheral off the changer run queue if it was scheduled to run. In practice, it's highly unlikely, and maybe impossible that the peripheral would have been on the changer run queue at that stage of the probe process. The fix is: - Add a new peripheral callback function (the "oninvalidate" function) that is called the first time cam_periph_invalidate() is called for a peripheral. - Create new foooninvalidate() routines for each peripheral driver. This routine is always called at splsoftcam(), and contains all the stuff that used to be in the AC_LOST_DEVICE case of the async callback handler. - Move the devstat cleanup call to the destructor/cleanup routines, since some of the drivers do I/O in their close routines. - Make sure that when we're flushing the buffer queue, we traverse it at splbio(). - Add a check for the invalid flag in the pt driver's open routine. Reviewed by: gibbs
1998-10-22 22:16:56 +00:00
free(softc, M_DEVBUF);
Reorder dacleanup() and cdcleanup() slightly so that the sysctl context is freed while the periph lock is not held. While here, wait until after freeing the softc before reacquiring the periph lock. Tested by: sbruno
2009-02-11 22:29:09 +00:00
cam_periph_lock(periph);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
static void
daasync(void *callback_arg, u_int32_t code,
struct cam_path *path, void *arg)
{
struct cam_periph *periph;
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
struct da_softc *softc;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
periph = (struct cam_periph *)callback_arg;
switch (code) {
case AC_FOUND_DEVICE:
{
struct ccb_getdev *cgd;
cam_status status;
cgd = (struct ccb_getdev *)arg;
Check the void * argument in the AC_FOUND_DEV case against NULL. Whether correctly or not, this sometimes is propagated up via XPT.
2001-07-04 05:22:42 +00:00
if (cgd == NULL)
break;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Separate the parallel scsi knowledge out of the core of the XPT, and modularize it so that new transports can be created. Add a transport for SATA Add a periph+protocol layer for ATA Add a driver for AHCI-compliant hardware. Add a maxio field to CAM so that drivers can advertise their max I/O capability. Modify various drivers so that they are insulated from the value of MAXPHYS. The new ATA/SATA code supports AHCI-compliant hardware, and will override the classic ATA driver if it is loaded as a module at boot time or compiled into the kernel. The stack now support NCQ (tagged queueing) for increased performance on modern SATA drives. It also supports port multipliers. ATA drives are accessed via 'ada' device nodes. ATAPI drives are accessed via 'cd' device nodes. They can all be enumerated and manipulated via camcontrol, just like SCSI drives. SCSI commands are not translated to their ATA equivalents; ATA native commands are used throughout the entire stack, including camcontrol. See the camcontrol manpage for further details. Testing this code may require that you update your fstab, and possibly modify your BIOS to enable AHCI functionality, if available. This code is very experimental at the moment. The userland ABI/API has changed, so applications will need to be recompiled. It may change further in the near future. The 'ada' device name may also change as more infrastructure is completed in this project. The goal is to eventually put all CAM busses and devices until newbus, allowing for interesting topology and management options. Few functional changes will be seen with existing SCSI/SAS/FC drivers, though the userland ABI has still changed. In the future, transports specific modules for SAS and FC may appear in order to better support the topologies and capabilities of these technologies. The modularization of CAM and the addition of the ATA/SATA modules is meant to break CAM out of the mold of being specific to SCSI, letting it grow to be a framework for arbitrary transports and protocols. It also allows drivers to be written to support discrete hardware without jeopardizing the stability of non-related hardware. While only an AHCI driver is provided now, a Silicon Image driver is also in the works. Drivers for ICH1-4, ICH5-6, PIIX, classic IDE, and any other hardware is possible and encouraged. Help with new transports is also encouraged. Submitted by: scottl, mav Approved by: re
2009-07-10 08:18:08 +00:00
if (cgd->protocol != PROTO_SCSI)
break;
Do the minor changes needed because of change to ccb_getdev structure. JKH Trading Stamps applied. Reviewed by: gibbs@freebsd.org, ken@freebsd.org
2000-01-17 06:27:37 +00:00
if (SID_TYPE(&cgd->inq_data) != T_DIRECT
- Add support for Simplified Direct Access Device, mostly for Firewire/SBP-II devices. - Add quirk for Logitec USB/Firewire HDD. MFC after: 3 days.
2002-02-21 11:58:47 +00:00
&& SID_TYPE(&cgd->inq_data) != T_RBC
Do the minor changes needed because of change to ccb_getdev structure. JKH Trading Stamps applied. Reviewed by: gibbs@freebsd.org, ken@freebsd.org
2000-01-17 06:27:37 +00:00
&& SID_TYPE(&cgd->inq_data) != T_OPTICAL)
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
break;
/*
* Allocate a peripheral instance for
* this device and start the probe
* process.
*/
Fix a problem with the way we handled device invalidation when attaching to a device failed. In theory, the same steps that happen when we get an AC_LOST_DEVICE async notification should have been taken when a driver fails to attach. In practice, that wasn't the case. This only affected the da, cd and ch drivers, but the fix affects all peripheral drivers. There were several possible problems: - In the da driver, we didn't remove the peripheral's softc from the da driver's linked list of softcs. Once the peripheral and softc got removed, we'd get a kernel panic the next time the timeout routine called dasendorderedtag(). - In the da, cd and possibly ch drivers, we didn't remove the peripheral's devstat structure from the devstat queue. Once the peripheral and softc were removed, this could cause a panic if anyone tried to access device statistics. (one component of the linked list wouldn't exist anymore) - In the cd driver, we didn't take the peripheral off the changer run queue if it was scheduled to run. In practice, it's highly unlikely, and maybe impossible that the peripheral would have been on the changer run queue at that stage of the probe process. The fix is: - Add a new peripheral callback function (the "oninvalidate" function) that is called the first time cam_periph_invalidate() is called for a peripheral. - Create new foooninvalidate() routines for each peripheral driver. This routine is always called at splsoftcam(), and contains all the stuff that used to be in the AC_LOST_DEVICE case of the async callback handler. - Move the devstat cleanup call to the destructor/cleanup routines, since some of the drivers do I/O in their close routines. - Make sure that when we're flushing the buffer queue, we traverse it at splbio(). - Add a check for the invalid flag in the pt driver's open routine. Reviewed by: gibbs
1998-10-22 22:16:56 +00:00
status = cam_periph_alloc(daregister, daoninvalidate,
dacleanup, dastart,
"da", CAM_PERIPH_BIO,
cgd->ccb_h.path, daasync,
AC_FOUND_DEVICE, cgd);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if (status != CAM_REQ_CMP
&& status != CAM_REQ_INPROG)
printf("daasync: Unable to attach to new device "
"due to status 0x%x\n", status);
Plumb device physical path reporting from CAM devices, through GEOM and DEVFS, and make it accessible via the diskinfo utility. Extend GEOM's generic attribute query mechanism into generic disk consumers. sys/geom/geom_disk.c: sys/geom/geom_disk.h: sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Allow disk providers to implement a new method which can override the default BIO_GETATTR response, d_getattr(struct bio *). This function returns -1 if not handled, otherwise it returns 0 or an errno to be passed to g_io_deliver(). sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Don't copy the serial number to dp->d_ident anymore, as the CAM XPT is now responsible for returning this information via d_getattr()->(a)dagetattr()->xpt_getatr(). sys/geom/geom_dev.c: - Implement a new ioctl, DIOCGPHYSPATH, which returns the GEOM attribute "GEOM::physpath", if possible. If the attribute request returns a zero-length string, ENOENT is returned. usr.sbin/diskinfo/diskinfo.c: - If the DIOCGPHYSPATH ioctl is successful, report physical path data when diskinfo is executed with the '-v' option. Submitted by: will Reviewed by: gibbs Sponsored by: Spectra Logic Corporation Add generic attribute change notification support to GEOM. sys/sys/geom/geom.h: Add a new attrchanged method field to both g_class and g_geom. sys/sys/geom/geom.h: sys/geom/geom_event.c: - Provide the g_attr_changed() function that providers can use to advertise attribute changes. - Perform delivery of attribute change notifications from a thread context via the standard GEOM event mechanism. sys/geom/geom_subr.c: Inherit the attrchanged method from class to geom (class instance). sys/geom/geom_disk.c: Provide disk_attr_changed() to provide g_attr_changed() access to consumers of the disk API. sys/cam/scsi/scsi_pass.c: sys/cam/scsi/scsi_da.c: sys/geom/geom_dev.c: sys/geom/geom_disk.c: Use attribute changed events to track updates to physical path information. sys/cam/scsi/scsi_da.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, and the updated buffer type references our physical path attribute, emit a GEOM attribute changed event via the disk_attr_changed() API. sys/cam/scsi/scsi_pass.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, update the physical patch devfs alias for this pass instance. Submitted by: gibbs Sponsored by: Spectra Logic Corporation
2011-06-14 17:10:32 +00:00
return;
}
case AC_ADVINFO_CHANGED:
{
uintptr_t buftype;
buftype = (uintptr_t)arg;
if (buftype == CDAI_TYPE_PHYS_PATH) {
struct da_softc *softc;
softc = periph->softc;
disk_attr_changed(softc->disk, "GEOM::physpath",
M_NOWAIT);
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
break;
}
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
case AC_UNIT_ATTENTION:
{
union ccb *ccb;
int error_code, sense_key, asc, ascq;
softc = (struct da_softc *)periph->softc;
ccb = (union ccb *)arg;
/*
* Handle all UNIT ATTENTIONs except our own,
* as they will be handled by daerror().
*/
if (xpt_path_periph(ccb->ccb_h.path) != periph &&
scsi_extract_sense_ccb(ccb,
&error_code, &sense_key, &asc, &ascq)) {
if (asc == 0x2A && ascq == 0x09) {
xpt_print(ccb->ccb_h.path,
"capacity data has changed\n");
dareprobe(periph);
} else if (asc == 0x28 && ascq == 0x00)
disk_media_changed(softc->disk, M_NOWAIT);
}
cam_periph_async(periph, code, path, arg);
break;
}
case AC_SCSI_AEN:
softc = (struct da_softc *)periph->softc;
Do not schedule periph for payload/TUR requests if reprobe is in progress to avoid sending extra READ CAPACITY requests by dastart(). Schedule periph again on reprobe completion, or otherwise it may stuck indefinitely long. This should fix USB explore thread hanging on device unplug, waiting for periph destruction. Reported by: hselasky
2013-01-11 16:10:11 +00:00
if (!softc->tur) {
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
if (cam_periph_acquire(periph) == CAM_REQ_CMP) {
softc->tur = 1;
Do not schedule periph for payload/TUR requests if reprobe is in progress to avoid sending extra READ CAPACITY requests by dastart(). Schedule periph again on reprobe completion, or otherwise it may stuck indefinitely long. This should fix USB explore thread hanging on device unplug, waiting for periph destruction. Reported by: hselasky
2013-01-11 16:10:11 +00:00
daschedule(periph);
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
}
}
/* FALLTHROUGH */
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
case AC_SENT_BDR:
case AC_BUS_RESET:
{
struct ccb_hdr *ccbh;
softc = (struct da_softc *)periph->softc;
/*
* Don't fail on the expected unit attention
* that will occur.
*/
softc->flags |= DA_FLAG_RETRY_UA;
Mechanical change to use <sys/queue.h> macro API instead of fondling implementation details. Created with: sed(1) Reviewed by: md5(1)
2001-02-04 13:13:25 +00:00
LIST_FOREACH(ccbh, &softc->pending_ccbs, periph_links.le)
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
ccbh->ccb_state |= DA_CCB_RETRY_UA;
sys/cam/scsi/scsi_da.c: - Only attempt the closing synchronize cache on a disk if it is still there. - When a device is lost, report the number of outstanding I/Os as they are drained. - When a device is lost, return any unprocessed bios with ENXIO instead of EIO. - Filter asynchronous events, but always allow cam_periph_async() to see them too. Sponsored by: Spectra Logic Corporation
2011-06-14 16:05:00 +00:00
break;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
default:
break;
}
sys/cam/scsi/scsi_da.c: - Only attempt the closing synchronize cache on a disk if it is still there. - When a device is lost, report the number of outstanding I/Os as they are drained. - When a device is lost, return any unprocessed bios with ENXIO instead of EIO. - Filter asynchronous events, but always allow cam_periph_async() to see them too. Sponsored by: Spectra Logic Corporation
2011-06-14 16:05:00 +00:00
cam_periph_async(periph, code, path, arg);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
static void
dasysctlinit(void *context, int pending)
{
struct cam_periph *periph;
struct da_softc *softc;
char tmpstr[80], tmpstr2[80];
redfacedly noting a missing part of a patch. Pointy Hat To: Me MFC after: 2 Week X-MFC: 208896
2010-06-07 22:43:37 +00:00
struct ccb_trans_settings cts;
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
periph = (struct cam_periph *)context;
Rearrange how things are done to avoid dereferencing stale pointers in the case of immediate unconfigure after configure. Hold the periph an extra count while we have the task to create sysctl context outstanding so that the periph doesn't go away unexpectedly. Sponsored by: Panasas Reviewed by: scsi@ MFC after: 1 month
2010-06-08 22:46:44 +00:00
/*
* periph was held for us when this task was enqueued
*/
if (periph->flags & CAM_PERIPH_INVALID) {
cam_periph_release(periph);
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
return;
Rearrange how things are done to avoid dereferencing stale pointers in the case of immediate unconfigure after configure. Hold the periph an extra count while we have the task to create sysctl context outstanding so that the periph doesn't go away unexpectedly. Sponsored by: Panasas Reviewed by: scsi@ MFC after: 1 month
2010-06-08 22:46:44 +00:00
}
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
softc = (struct da_softc *)periph->softc;
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
snprintf(tmpstr, sizeof(tmpstr), "CAM DA unit %d", periph->unit_number);
snprintf(tmpstr2, sizeof(tmpstr2), "%d", periph->unit_number);
sysctl_ctx_init(&softc->sysctl_ctx);
(dacleanup, dasysctlinit): Defend against calling sysctl_ctx_free on an uninitialized sysctl_ctx, using flag DA_FLAG_SCTX_INIT. This prevents a panic encoutered with some umass units that probe correctly but fail to attach. Same problem, and same fix, as scsi_cd.c rev. 1.86. Reviewed by: njl, ken
2003-10-08 07:12:30 +00:00
softc->flags |= DA_FLAG_SCTX_INIT;
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
softc->sysctl_tree = SYSCTL_ADD_NODE(&softc->sysctl_ctx,
SYSCTL_STATIC_CHILDREN(_kern_cam_da), OID_AUTO, tmpstr2,
CTLFLAG_RD, 0, tmpstr);
if (softc->sysctl_tree == NULL) {
printf("dasysctlinit: unable to allocate sysctl tree\n");
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
cam_periph_release(periph);
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
return;
}
/*
Do a minor amount of stylifying. Also, get a Fibre Channel WWPN if one exists for a da unit and create a sysctl OID for it.
2010-06-07 17:41:34 +00:00
* Now register the sysctl handler, so the user can change the value on
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
* the fly.
*/
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
SYSCTL_ADD_PROC(&softc->sysctl_ctx, SYSCTL_CHILDREN(softc->sysctl_tree),
OID_AUTO, "delete_method", CTLTYPE_STRING | CTLFLAG_RW,
Updates delete_method sysctl changes to always maintain disk d_flags DISKFLAG_CANDELETE. While this change makes this layer consistent other layers such as UFS and ZFS BIO_DELETE support may not notice any change made manually via these device sysctls until the device is reopened via a mount. Also corrected var order in dadeletemethodsysctl PR: kern/169801 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:57:46 +00:00
softc, 0, dadeletemethodsysctl, "A",
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
"BIO_DELETE execution method");
Do a minor amount of stylifying. Also, get a Fibre Channel WWPN if one exists for a da unit and create a sysctl OID for it.
2010-06-07 17:41:34 +00:00
SYSCTL_ADD_PROC(&softc->sysctl_ctx, SYSCTL_CHILDREN(softc->sysctl_tree),
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
OID_AUTO, "minimum_cmd_size", CTLTYPE_INT | CTLFLAG_RW,
&softc->minimum_cmd_size, 0, dacmdsizesysctl, "I",
"Minimum CDB size");
Fix a race condition in CAM peripheral free handling, locking in the CAM XPT bus traversal code, and a number of other periph level issues. cam_periph.h, cam_periph.c: Modify cam_periph_acquire() to test the CAM_PERIPH_INVALID flag prior to allowing a reference count to be gained on a peripheral. Callers of this function will receive CAM_REQ_CMP_ERR status in the situation of attempting to reference an invalidated periph. This guarantees that a peripheral scheduled for a deferred free will not be accessed during its wait for destruction. Panic during attempts to drop a reference count on a peripheral that already has a zero reference count. In cam_periph_list(), use a local sbuf with SBUF_FIXEDLEN set so that mallocs do not occur while the xpt topology lock is held, regardless of the allocation policy of the passed in sbuf. Add a new routine, cam_periph_release_locked_buses(), that can be called when the caller already holds the CAM topology lock. Add some extra debugging for duplicate peripheral allocations in cam_periph_alloc(). Treat CAM_DEV_NOT_THERE much the same as a selection timeout (AC_LOST_DEVICE is emitted), but forgo retries. cam_xpt.c: Revamp the way the EDT traversal code does locking and reference counting. This was broken, since it assumed that the EDT would not change during traversal, but that assumption is no longer valid. So, to prevent devices from going away while we traverse the EDT, make sure we properly lock everything and hold references on devices that we are using. The two peripheral driver traversal routines should be examined. xptpdperiphtraverse() holds the topology lock for the entire time it runs. xptperiphtraverse() is now locked properly, but only holds the topology lock while it is traversing the list, and not while the traversal function is running. The bus locking code in xptbustraverse() should also be revisited at a later time, since it is complex and should probably be simplified. scsi_da.c: Pay attention to the return value from cam_periph_acquire(). Return 0 always from daclose() even if the disk is now gone. Add some rudimentary error injection support. scsi_sg.c: Fix reference counting in the sg(4) driver. The sg driver was calling cam_periph_release() on close, but never called cam_periph_acquire() (which increments the reference count) on open. The periph code correctly complained that the sg(4) driver was trying to decrement the refcount when it was already 0. Sponsored by: Spectra Logic MFC after: 2 weeks
2012-01-12 00:41:48 +00:00
SYSCTL_ADD_INT(&softc->sysctl_ctx,
SYSCTL_CHILDREN(softc->sysctl_tree),
OID_AUTO,
"error_inject",
CTLFLAG_RW,
&softc->error_inject,
0,
"error_inject leaf");
Do a minor amount of stylifying. Also, get a Fibre Channel WWPN if one exists for a da unit and create a sysctl OID for it.
2010-06-07 17:41:34 +00:00
/*
* Add some addressing info.
*/
memset(&cts, 0, sizeof (cts));
Remove several uses of numeric priorities from immediate CCB setups.
2012-10-27 09:40:29 +00:00
xpt_setup_ccb(&cts.ccb_h, periph->path, CAM_PRIORITY_NONE);
Do a minor amount of stylifying. Also, get a Fibre Channel WWPN if one exists for a da unit and create a sysctl OID for it.
2010-06-07 17:41:34 +00:00
cts.ccb_h.func_code = XPT_GET_TRAN_SETTINGS;
cts.type = CTS_TYPE_CURRENT_SETTINGS;
cam_periph_lock(periph);
xpt_action((union ccb *)&cts);
cam_periph_unlock(periph);
if (cts.ccb_h.status != CAM_REQ_CMP) {
cam_periph_release(periph);
return;
}
if (cts.protocol == PROTO_SCSI && cts.transport == XPORT_FC) {
struct ccb_trans_settings_fc *fc = &cts.xport_specific.fc;
if (fc->valid & CTS_FC_VALID_WWPN) {
softc->wwpn = fc->wwpn;
sysctl(8) should use the CTLTYPE to determine the type of data when reading. (This was already done for writing to a sysctl). This requires all SYSCTL setups to specify a type. Most of them are now checked at compile-time. Remove SYSCTL_*X* sysctl additions as the print being in hex should be controlled by the -x flag to sysctl(8). Succested by: bde
2011-01-19 17:04:07 +00:00
SYSCTL_ADD_UQUAD(&softc->sysctl_ctx,
Do a minor amount of stylifying. Also, get a Fibre Channel WWPN if one exists for a da unit and create a sysctl OID for it.
2010-06-07 17:41:34 +00:00
SYSCTL_CHILDREN(softc->sysctl_tree),
Add a 64-bit hex-printed sysctl(9) since there is at least one place in the code that wanted it. It is named X64 rather than XQUAD since the quad name is a historical abomination that should not be perpetuated.
2011-01-13 18:20:33 +00:00
OID_AUTO, "wwpn", CTLFLAG_RD,
Do a minor amount of stylifying. Also, get a Fibre Channel WWPN if one exists for a da unit and create a sysctl OID for it.
2010-06-07 17:41:34 +00:00
&softc->wwpn, "World Wide Port Name");
}
}
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
cam_periph_release(periph);
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
}
Fix ATAPI/USB/Firewire CDROM drive handling in cd(4) and hopefully fix a number of related problems along the way. - Automatically detect CDROM drives that can't handle 6 byte mode sense and mode select, and adjust our command size accordingly. We have to handle this in the cd(4) driver (where the buffers are allocated), since the parameter list length is different for the 6 and 10 byte mode sense commands. - Remove MODE_SENSE and MODE_SELECT translation removed in ATAPICAM and in the umass(4) driver, since there's no way for that to work properly. - Add a quirk entry for CDROM drives that just hang when they get a 6 byte mode sense or mode select. The reason for the quirk must be documented in a PR, and all quirks must be approved by ken@FreeBSD.org. This is to make sure that we fully understand why each quirk is needed. Once the CAM_NEW_TRAN_CODE is finished, we should be able to remove any such quirks, since we'll know what protocol the drive speaks (SCSI, ATAPI, etc.) and therefore whether we should use 6 or 10 byte mode sense/select commands. - Change the way the da(4) handles the no_6_byte sysctl. There is now a per-drive sysctl to set the minimum command size for that particular disk. (Since you could have multiple disks with multiple requirements in one system.) - Loader tunable support for all the sysctls in the da(4) and cd(4) drivers. - Add a CDIOCCLOSE ioctl for cd(4) (bde pointed this out a long time ago). - Add a media validation routine (cdcheckmedia()) to the cd(4) driver, to fix some problems bde pointed out a long time ago. We now allow open() to succeed no matter what, but if we don't detect valid media, the user can only issue CDIOCCLOSE or CDIOCEJECT ioctls. - The media validation routine also reads the table of contents off the drive. We use the table of contents to implement the CDIOCPLAYTRACKS ioctl using the PLAY AUDIO MSF command. The PLAY AUDIO TRACK INDEX command that we previously used was deprecated after SCSI-2. It works in every SCSI CDROM I've tried, but doesn't seem to work on ATAPI CDROM drives. We still use the play audio track index command if we don't have a valid TOC, but I suppose it'll fail anyway in that case. - Add _len() versions of scsi_mode_sense() and scsi_mode_select() so that we can specify the minimum command length. - Fix a couple of formatting problems in the sense printing code. MFC after: 4 weeks
2003-02-21 06:19:38 +00:00
static int
dacmdsizesysctl(SYSCTL_HANDLER_ARGS)
{
int error, value;
value = *(int *)arg1;
error = sysctl_handle_int(oidp, &value, 0, req);
if ((error != 0)
|| (req->newptr == NULL))
return (error);
/*
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
* Acceptable values here are 6, 10, 12 or 16.
Fix ATAPI/USB/Firewire CDROM drive handling in cd(4) and hopefully fix a number of related problems along the way. - Automatically detect CDROM drives that can't handle 6 byte mode sense and mode select, and adjust our command size accordingly. We have to handle this in the cd(4) driver (where the buffers are allocated), since the parameter list length is different for the 6 and 10 byte mode sense commands. - Remove MODE_SENSE and MODE_SELECT translation removed in ATAPICAM and in the umass(4) driver, since there's no way for that to work properly. - Add a quirk entry for CDROM drives that just hang when they get a 6 byte mode sense or mode select. The reason for the quirk must be documented in a PR, and all quirks must be approved by ken@FreeBSD.org. This is to make sure that we fully understand why each quirk is needed. Once the CAM_NEW_TRAN_CODE is finished, we should be able to remove any such quirks, since we'll know what protocol the drive speaks (SCSI, ATAPI, etc.) and therefore whether we should use 6 or 10 byte mode sense/select commands. - Change the way the da(4) handles the no_6_byte sysctl. There is now a per-drive sysctl to set the minimum command size for that particular disk. (Since you could have multiple disks with multiple requirements in one system.) - Loader tunable support for all the sysctls in the da(4) and cd(4) drivers. - Add a CDIOCCLOSE ioctl for cd(4) (bde pointed this out a long time ago). - Add a media validation routine (cdcheckmedia()) to the cd(4) driver, to fix some problems bde pointed out a long time ago. We now allow open() to succeed no matter what, but if we don't detect valid media, the user can only issue CDIOCCLOSE or CDIOCEJECT ioctls. - The media validation routine also reads the table of contents off the drive. We use the table of contents to implement the CDIOCPLAYTRACKS ioctl using the PLAY AUDIO MSF command. The PLAY AUDIO TRACK INDEX command that we previously used was deprecated after SCSI-2. It works in every SCSI CDROM I've tried, but doesn't seem to work on ATAPI CDROM drives. We still use the play audio track index command if we don't have a valid TOC, but I suppose it'll fail anyway in that case. - Add _len() versions of scsi_mode_sense() and scsi_mode_select() so that we can specify the minimum command length. - Fix a couple of formatting problems in the sense printing code. MFC after: 4 weeks
2003-02-21 06:19:38 +00:00
*/
if (value < 6)
value = 6;
else if ((value > 6)
&& (value <= 10))
value = 10;
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
else if ((value > 10)
&& (value <= 12))
Fix ATAPI/USB/Firewire CDROM drive handling in cd(4) and hopefully fix a number of related problems along the way. - Automatically detect CDROM drives that can't handle 6 byte mode sense and mode select, and adjust our command size accordingly. We have to handle this in the cd(4) driver (where the buffers are allocated), since the parameter list length is different for the 6 and 10 byte mode sense commands. - Remove MODE_SENSE and MODE_SELECT translation removed in ATAPICAM and in the umass(4) driver, since there's no way for that to work properly. - Add a quirk entry for CDROM drives that just hang when they get a 6 byte mode sense or mode select. The reason for the quirk must be documented in a PR, and all quirks must be approved by ken@FreeBSD.org. This is to make sure that we fully understand why each quirk is needed. Once the CAM_NEW_TRAN_CODE is finished, we should be able to remove any such quirks, since we'll know what protocol the drive speaks (SCSI, ATAPI, etc.) and therefore whether we should use 6 or 10 byte mode sense/select commands. - Change the way the da(4) handles the no_6_byte sysctl. There is now a per-drive sysctl to set the minimum command size for that particular disk. (Since you could have multiple disks with multiple requirements in one system.) - Loader tunable support for all the sysctls in the da(4) and cd(4) drivers. - Add a CDIOCCLOSE ioctl for cd(4) (bde pointed this out a long time ago). - Add a media validation routine (cdcheckmedia()) to the cd(4) driver, to fix some problems bde pointed out a long time ago. We now allow open() to succeed no matter what, but if we don't detect valid media, the user can only issue CDIOCCLOSE or CDIOCEJECT ioctls. - The media validation routine also reads the table of contents off the drive. We use the table of contents to implement the CDIOCPLAYTRACKS ioctl using the PLAY AUDIO MSF command. The PLAY AUDIO TRACK INDEX command that we previously used was deprecated after SCSI-2. It works in every SCSI CDROM I've tried, but doesn't seem to work on ATAPI CDROM drives. We still use the play audio track index command if we don't have a valid TOC, but I suppose it'll fail anyway in that case. - Add _len() versions of scsi_mode_sense() and scsi_mode_select() so that we can specify the minimum command length. - Fix a couple of formatting problems in the sense printing code. MFC after: 4 weeks
2003-02-21 06:19:38 +00:00
value = 12;
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
else if (value > 12)
value = 16;
Fix ATAPI/USB/Firewire CDROM drive handling in cd(4) and hopefully fix a number of related problems along the way. - Automatically detect CDROM drives that can't handle 6 byte mode sense and mode select, and adjust our command size accordingly. We have to handle this in the cd(4) driver (where the buffers are allocated), since the parameter list length is different for the 6 and 10 byte mode sense commands. - Remove MODE_SENSE and MODE_SELECT translation removed in ATAPICAM and in the umass(4) driver, since there's no way for that to work properly. - Add a quirk entry for CDROM drives that just hang when they get a 6 byte mode sense or mode select. The reason for the quirk must be documented in a PR, and all quirks must be approved by ken@FreeBSD.org. This is to make sure that we fully understand why each quirk is needed. Once the CAM_NEW_TRAN_CODE is finished, we should be able to remove any such quirks, since we'll know what protocol the drive speaks (SCSI, ATAPI, etc.) and therefore whether we should use 6 or 10 byte mode sense/select commands. - Change the way the da(4) handles the no_6_byte sysctl. There is now a per-drive sysctl to set the minimum command size for that particular disk. (Since you could have multiple disks with multiple requirements in one system.) - Loader tunable support for all the sysctls in the da(4) and cd(4) drivers. - Add a CDIOCCLOSE ioctl for cd(4) (bde pointed this out a long time ago). - Add a media validation routine (cdcheckmedia()) to the cd(4) driver, to fix some problems bde pointed out a long time ago. We now allow open() to succeed no matter what, but if we don't detect valid media, the user can only issue CDIOCCLOSE or CDIOCEJECT ioctls. - The media validation routine also reads the table of contents off the drive. We use the table of contents to implement the CDIOCPLAYTRACKS ioctl using the PLAY AUDIO MSF command. The PLAY AUDIO TRACK INDEX command that we previously used was deprecated after SCSI-2. It works in every SCSI CDROM I've tried, but doesn't seem to work on ATAPI CDROM drives. We still use the play audio track index command if we don't have a valid TOC, but I suppose it'll fail anyway in that case. - Add _len() versions of scsi_mode_sense() and scsi_mode_select() so that we can specify the minimum command length. - Fix a couple of formatting problems in the sense printing code. MFC after: 4 weeks
2003-02-21 06:19:38 +00:00
*(int *)arg1 = value;
return (0);
}
Updates delete_method sysctl changes to always maintain disk d_flags DISKFLAG_CANDELETE. While this change makes this layer consistent other layers such as UFS and ZFS BIO_DELETE support may not notice any change made manually via these device sysctls until the device is reopened via a mount. Also corrected var order in dadeletemethodsysctl PR: kern/169801 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:57:46 +00:00
static int
dadeletemethodset(struct da_softc *softc, da_delete_methods delete_method)
{
if (delete_method < 0 || delete_method > DA_DELETE_MAX)
return (EINVAL);
softc->delete_method = delete_method;
if (softc->delete_method > DA_DELETE_DISABLE)
softc->disk->d_flags |= DISKFLAG_CANDELETE;
else
softc->disk->d_flags &= ~DISKFLAG_CANDELETE;
return (0);
}
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
static int
dadeletemethodsysctl(SYSCTL_HANDLER_ARGS)
{
char buf[16];
const char *p;
Updates delete_method sysctl changes to always maintain disk d_flags DISKFLAG_CANDELETE. While this change makes this layer consistent other layers such as UFS and ZFS BIO_DELETE support may not notice any change made manually via these device sysctls until the device is reopened via a mount. Also corrected var order in dadeletemethodsysctl PR: kern/169801 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:57:46 +00:00
struct da_softc *softc;
int i, error, value;
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
Updates delete_method sysctl changes to always maintain disk d_flags DISKFLAG_CANDELETE. While this change makes this layer consistent other layers such as UFS and ZFS BIO_DELETE support may not notice any change made manually via these device sysctls until the device is reopened via a mount. Also corrected var order in dadeletemethodsysctl PR: kern/169801 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:57:46 +00:00
softc = (struct da_softc *)arg1;
value = softc->delete_method;
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
if (value < 0 || value > DA_DELETE_MAX)
p = "UNKNOWN";
else
p = da_delete_method_names[value];
strncpy(buf, p, sizeof(buf));
error = sysctl_handle_string(oidp, buf, sizeof(buf), req);
if (error != 0 || req->newptr == NULL)
return (error);
for (i = 0; i <= DA_DELETE_MAX; i++) {
if (strcmp(buf, da_delete_method_names[i]) != 0)
continue;
Updates delete_method sysctl changes to always maintain disk d_flags DISKFLAG_CANDELETE. While this change makes this layer consistent other layers such as UFS and ZFS BIO_DELETE support may not notice any change made manually via these device sysctls until the device is reopened via a mount. Also corrected var order in dadeletemethodsysctl PR: kern/169801 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:57:46 +00:00
return dadeletemethodset(softc, i);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
}
return (EINVAL);
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
static cam_status
daregister(struct cam_periph *periph, void *arg)
{
struct da_softc *softc;
Add a PATH_INQ flag, PIM_NO_6_BYTE, which indicates the SIM never wishes to receive 6 byte commands. Add a check for this flag to da(4) and cd(4) so that they honor it. This is a quick workaround for many devices (especially USB) that require da(4) quirks to operate. The more complete approach is to finish the new transport code which will be aware of the SCSI version a transport implements. MFC after: 1 day
2003-07-28 06:15:59 +00:00
struct ccb_pathinq cpi;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
struct ccb_getdev *cgd;
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
char tmpstr[80];
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
caddr_t match;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
cgd = (struct ccb_getdev *)arg;
if (cgd == NULL) {
printf("daregister: no getdev CCB, can't register device\n");
return(CAM_REQ_CMP_ERR);
}
Use malloc(9)'s M_ZERO rather than explicit bzero(9) call.
2002-10-11 09:59:22 +00:00
softc = (struct da_softc *)malloc(sizeof(*softc), M_DEVBUF,
M_NOWAIT|M_ZERO);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if (softc == NULL) {
printf("daregister: Unable to probe new device. "
"Unable to allocate softc\n");
return(CAM_REQ_CMP_ERR);
}
LIST_INIT(&softc->pending_ccbs);
softc->state = DA_STATE_PROBE;
Complete the bio/buf divorce for all code below devfs::strategy Exceptions: Vinum untouched. This means that it cannot be compiled. Greg Lehey is on the case. CCD not converted yet, casts to struct buf (still safe) atapi-cd casts to struct buf to examine B_PHYS
2000-04-15 05:54:02 +00:00
bioq_init(&softc->bio_queue);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
bioq_init(&softc->delete_queue);
bioq_init(&softc->delete_run_queue);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if (SID_IS_REMOVABLE(&cgd->inq_data))
softc->flags |= DA_FLAG_PACK_REMOVABLE;
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
softc->unmap_max_ranges = UNMAP_MAX_RANGES;
softc->unmap_max_lba = 1024*1024*2;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
periph->softc = softc;
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
/*
* See if this device has any quirks.
*/
match = cam_quirkmatch((caddr_t)&cgd->inq_data,
(caddr_t)da_quirk_table,
sizeof(da_quirk_table)/sizeof(*da_quirk_table),
sizeof(*da_quirk_table), scsi_inquiry_match);
if (match != NULL)
softc->quirks = ((struct da_quirk_entry *)match)->quirks;
else
softc->quirks = DA_Q_NONE;
Add a PATH_INQ flag, PIM_NO_6_BYTE, which indicates the SIM never wishes to receive 6 byte commands. Add a check for this flag to da(4) and cd(4) so that they honor it. This is a quick workaround for many devices (especially USB) that require da(4) quirks to operate. The more complete approach is to finish the new transport code which will be aware of the SCSI version a transport implements. MFC after: 1 day
2003-07-28 06:15:59 +00:00
/* Check if the SIM does not want 6 byte commands */
Separate the parallel scsi knowledge out of the core of the XPT, and modularize it so that new transports can be created. Add a transport for SATA Add a periph+protocol layer for ATA Add a driver for AHCI-compliant hardware. Add a maxio field to CAM so that drivers can advertise their max I/O capability. Modify various drivers so that they are insulated from the value of MAXPHYS. The new ATA/SATA code supports AHCI-compliant hardware, and will override the classic ATA driver if it is loaded as a module at boot time or compiled into the kernel. The stack now support NCQ (tagged queueing) for increased performance on modern SATA drives. It also supports port multipliers. ATA drives are accessed via 'ada' device nodes. ATAPI drives are accessed via 'cd' device nodes. They can all be enumerated and manipulated via camcontrol, just like SCSI drives. SCSI commands are not translated to their ATA equivalents; ATA native commands are used throughout the entire stack, including camcontrol. See the camcontrol manpage for further details. Testing this code may require that you update your fstab, and possibly modify your BIOS to enable AHCI functionality, if available. This code is very experimental at the moment. The userland ABI/API has changed, so applications will need to be recompiled. It may change further in the near future. The 'ada' device name may also change as more infrastructure is completed in this project. The goal is to eventually put all CAM busses and devices until newbus, allowing for interesting topology and management options. Few functional changes will be seen with existing SCSI/SAS/FC drivers, though the userland ABI has still changed. In the future, transports specific modules for SAS and FC may appear in order to better support the topologies and capabilities of these technologies. The modularization of CAM and the addition of the ATA/SATA modules is meant to break CAM out of the mold of being specific to SCSI, letting it grow to be a framework for arbitrary transports and protocols. It also allows drivers to be written to support discrete hardware without jeopardizing the stability of non-related hardware. While only an AHCI driver is provided now, a Silicon Image driver is also in the works. Drivers for ICH1-4, ICH5-6, PIIX, classic IDE, and any other hardware is possible and encouraged. Help with new transports is also encouraged. Submitted by: scottl, mav Approved by: re
2009-07-10 08:18:08 +00:00
bzero(&cpi, sizeof(cpi));
Replace most of priority numbers with defines. No logical changes.
2009-10-23 08:27:55 +00:00
xpt_setup_ccb(&cpi.ccb_h, periph->path, CAM_PRIORITY_NORMAL);
Add a PATH_INQ flag, PIM_NO_6_BYTE, which indicates the SIM never wishes to receive 6 byte commands. Add a check for this flag to da(4) and cd(4) so that they honor it. This is a quick workaround for many devices (especially USB) that require da(4) quirks to operate. The more complete approach is to finish the new transport code which will be aware of the SCSI version a transport implements. MFC after: 1 day
2003-07-28 06:15:59 +00:00
cpi.ccb_h.func_code = XPT_PATH_INQ;
xpt_action((union ccb *)&cpi);
if (cpi.ccb_h.status == CAM_REQ_CMP && (cpi.hba_misc & PIM_NO_6_BYTE))
softc->quirks |= DA_Q_NO_6_BYTE;
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
TASK_INIT(&softc->sysctl_task, 0, dasysctlinit, periph);
Fix ATAPI/USB/Firewire CDROM drive handling in cd(4) and hopefully fix a number of related problems along the way. - Automatically detect CDROM drives that can't handle 6 byte mode sense and mode select, and adjust our command size accordingly. We have to handle this in the cd(4) driver (where the buffers are allocated), since the parameter list length is different for the 6 and 10 byte mode sense commands. - Remove MODE_SENSE and MODE_SELECT translation removed in ATAPICAM and in the umass(4) driver, since there's no way for that to work properly. - Add a quirk entry for CDROM drives that just hang when they get a 6 byte mode sense or mode select. The reason for the quirk must be documented in a PR, and all quirks must be approved by ken@FreeBSD.org. This is to make sure that we fully understand why each quirk is needed. Once the CAM_NEW_TRAN_CODE is finished, we should be able to remove any such quirks, since we'll know what protocol the drive speaks (SCSI, ATAPI, etc.) and therefore whether we should use 6 or 10 byte mode sense/select commands. - Change the way the da(4) handles the no_6_byte sysctl. There is now a per-drive sysctl to set the minimum command size for that particular disk. (Since you could have multiple disks with multiple requirements in one system.) - Loader tunable support for all the sysctls in the da(4) and cd(4) drivers. - Add a CDIOCCLOSE ioctl for cd(4) (bde pointed this out a long time ago). - Add a media validation routine (cdcheckmedia()) to the cd(4) driver, to fix some problems bde pointed out a long time ago. We now allow open() to succeed no matter what, but if we don't detect valid media, the user can only issue CDIOCCLOSE or CDIOCEJECT ioctls. - The media validation routine also reads the table of contents off the drive. We use the table of contents to implement the CDIOCPLAYTRACKS ioctl using the PLAY AUDIO MSF command. The PLAY AUDIO TRACK INDEX command that we previously used was deprecated after SCSI-2. It works in every SCSI CDROM I've tried, but doesn't seem to work on ATAPI CDROM drives. We still use the play audio track index command if we don't have a valid TOC, but I suppose it'll fail anyway in that case. - Add _len() versions of scsi_mode_sense() and scsi_mode_select() so that we can specify the minimum command length. - Fix a couple of formatting problems in the sense printing code. MFC after: 4 weeks
2003-02-21 06:19:38 +00:00
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
* Take an exclusive refcount on the periph while dastart is called
* to finish the probe. The reference will be dropped in dadone at
* the end of probe.
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
*/
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
(void)cam_periph_hold(periph, PRIBIO);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
/*
* Schedule a periodic event to occasionally send an
* ordered tag to a device.
*/
callout_init_mtx(&softc->sendordered_c, periph->sim->mtx, 0);
callout_reset(&softc->sendordered_c,
Insert ordered command every 1/4 of the current command timeout, not 1/4 of the default one. Without this change setting kern.cam.ada.default_timeout to 1 instead of 30 allowed me to trigger several false positive command timeouts under heavy ZFS load on a SiI3132 siis(4) controller with 5 HDDs on a port multiplier. MFC after: 1 week
2012-02-02 19:02:15 +00:00
(da_default_timeout * hz) / DA_ORDEREDTAG_INTERVAL,
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
dasendorderedtag, softc);
Rearrange how things are done to avoid dereferencing stale pointers in the case of immediate unconfigure after configure. Hold the periph an extra count while we have the task to create sysctl context outstanding so that the periph doesn't go away unexpectedly. Sponsored by: Panasas Reviewed by: scsi@ MFC after: 1 month
2010-06-08 22:46:44 +00:00
mtx_unlock(periph->sim->mtx);
Move TUNABLE_INT_FETCH() out of the SIM lock to avoid sleeping while holding the lock. The fix is analogous to r220618 for ada(4). Reviewed by: mav
2011-04-16 06:54:41 +00:00
/*
* RBC devices don't have to support READ(6), only READ(10).
*/
if (softc->quirks & DA_Q_NO_6_BYTE || SID_TYPE(&cgd->inq_data) == T_RBC)
softc->minimum_cmd_size = 10;
else
softc->minimum_cmd_size = 6;
/*
* Load the user's default, if any.
*/
snprintf(tmpstr, sizeof(tmpstr), "kern.cam.da.%d.minimum_cmd_size",
periph->unit_number);
TUNABLE_INT_FETCH(tmpstr, &softc->minimum_cmd_size);
/*
* 6, 10, 12 and 16 are the currently permissible values.
*/
if (softc->minimum_cmd_size < 6)
softc->minimum_cmd_size = 6;
else if ((softc->minimum_cmd_size > 6)
&& (softc->minimum_cmd_size <= 10))
softc->minimum_cmd_size = 10;
else if ((softc->minimum_cmd_size > 10)
&& (softc->minimum_cmd_size <= 12))
softc->minimum_cmd_size = 12;
else if (softc->minimum_cmd_size > 12)
softc->minimum_cmd_size = 16;
Use READ CAPACITY(16) to get information about device physical sectors. As soon as not all devices support READ CAPACITY(16), automatically fall back to READ CAPACITY(10) if CAM_REQ_INVALID or SSD_KEY_ILLEGAL_REQUEST status returned. It also provides first bits of information about Logical Block Provisioning (aka UNMAP/TRIM) support by the device.
2011-12-23 19:12:02 +00:00
/* Predict whether device may support READ CAPACITY(16). */
Be more conservative in using READ CAPACITY(16) command. Previous code checked PROTECT bit in INQUIRY data for all SPC devices, while it is defined only since SPC-3. But there are some SPC-2 USB devices were reported, that have PROTECT bit set, return no error for READ CAPACITY(16) command, but return wrong sector count value in response. MFC after: 3 days
2012-03-31 11:23:09 +00:00
if (SID_ANSI_REV(&cgd->inq_data) >= SCSI_REV_SPC3) {
Use READ CAPACITY(16) to get information about device physical sectors. As soon as not all devices support READ CAPACITY(16), automatically fall back to READ CAPACITY(10) if CAM_REQ_INVALID or SSD_KEY_ILLEGAL_REQUEST status returned. It also provides first bits of information about Logical Block Provisioning (aka UNMAP/TRIM) support by the device.
2011-12-23 19:12:02 +00:00
softc->flags |= DA_FLAG_CAN_RC16;
softc->state = DA_STATE_PROBE2;
}
Move a comment to the right place. The rearrange done in r208928 left the comment to a wrong place.
2011-04-16 08:38:11 +00:00
/*
* Register this media as a disk.
*/
Rearrange how things are done to avoid dereferencing stale pointers in the case of immediate unconfigure after configure. Hold the periph an extra count while we have the task to create sysctl context outstanding so that the periph doesn't go away unexpectedly. Sponsored by: Panasas Reviewed by: scsi@ MFC after: 1 month
2010-06-08 22:46:44 +00:00
softc->disk = disk_alloc();
Make CAM report devices with ATA/SATA transport to devstat(9) as IDE.
2011-04-14 21:25:32 +00:00
softc->disk->d_devstat = devstat_new_entry(periph->periph_name,
periph->unit_number, 0,
DEVSTAT_BS_UNAVAILABLE,
SID_TYPE(&cgd->inq_data) |
XPORT_DEVSTAT_TYPE(cpi.transport),
DEVSTAT_PRIORITY_DISK);
Rearrange how things are done to avoid dereferencing stale pointers in the case of immediate unconfigure after configure. Hold the periph an extra count while we have the task to create sysctl context outstanding so that the periph doesn't go away unexpectedly. Sponsored by: Panasas Reviewed by: scsi@ MFC after: 1 month
2010-06-08 22:46:44 +00:00
softc->disk->d_open = daopen;
softc->disk->d_close = daclose;
softc->disk->d_strategy = dastrategy;
softc->disk->d_dump = dadump;
Plumb device physical path reporting from CAM devices, through GEOM and DEVFS, and make it accessible via the diskinfo utility. Extend GEOM's generic attribute query mechanism into generic disk consumers. sys/geom/geom_disk.c: sys/geom/geom_disk.h: sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Allow disk providers to implement a new method which can override the default BIO_GETATTR response, d_getattr(struct bio *). This function returns -1 if not handled, otherwise it returns 0 or an errno to be passed to g_io_deliver(). sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Don't copy the serial number to dp->d_ident anymore, as the CAM XPT is now responsible for returning this information via d_getattr()->(a)dagetattr()->xpt_getatr(). sys/geom/geom_dev.c: - Implement a new ioctl, DIOCGPHYSPATH, which returns the GEOM attribute "GEOM::physpath", if possible. If the attribute request returns a zero-length string, ENOENT is returned. usr.sbin/diskinfo/diskinfo.c: - If the DIOCGPHYSPATH ioctl is successful, report physical path data when diskinfo is executed with the '-v' option. Submitted by: will Reviewed by: gibbs Sponsored by: Spectra Logic Corporation Add generic attribute change notification support to GEOM. sys/sys/geom/geom.h: Add a new attrchanged method field to both g_class and g_geom. sys/sys/geom/geom.h: sys/geom/geom_event.c: - Provide the g_attr_changed() function that providers can use to advertise attribute changes. - Perform delivery of attribute change notifications from a thread context via the standard GEOM event mechanism. sys/geom/geom_subr.c: Inherit the attrchanged method from class to geom (class instance). sys/geom/geom_disk.c: Provide disk_attr_changed() to provide g_attr_changed() access to consumers of the disk API. sys/cam/scsi/scsi_pass.c: sys/cam/scsi/scsi_da.c: sys/geom/geom_dev.c: sys/geom/geom_disk.c: Use attribute changed events to track updates to physical path information. sys/cam/scsi/scsi_da.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, and the updated buffer type references our physical path attribute, emit a GEOM attribute changed event via the disk_attr_changed() API. sys/cam/scsi/scsi_pass.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, update the physical patch devfs alias for this pass instance. Submitted by: gibbs Sponsored by: Spectra Logic Corporation
2011-06-14 17:10:32 +00:00
softc->disk->d_getattr = dagetattr;
Fix a bug which causes a panic in daopen(). The panic is caused by a da(4) instance going away while GEOM is still probing it. In this case, the GEOM disk class instance has been created by disk_create(), and the taste of the disk is queued in the GEOM event queue. While that event is queued, the da(4) instance goes away. When the open call comes into the da(4) driver, it dereferences the freed (but non-NULL) peripheral pointer provided by GEOM, which results in a panic. The solution is to add a callback to the GEOM disk code that is called when all of its resources are cleaned up. This is implemented inside GEOM by adding an optional callback that is called when all consumers have detached from a provider, and the provider is about to be deleted. scsi_cd.c, scsi_da.c: In the register routine for the cd(4) and da(4) routines, acquire a reference to the CAM peripheral instance just before we call disk_create(). Use the new GEOM disk d_gone() callback to register a callback (dadiskgonecb()/cddiskgonecb()) that decrements the peripheral reference count once GEOM has finished cleaning up its resources. In the cd(4) driver, clean up open and close behavior slightly. GEOM makes sure we only get one open() and one close call, so there is no need to set an open flag and decrement the reference count if we are not the first open. In the cd(4) driver, use cam_periph_release_locked() in a couple of error scenarios to avoid extra mutex calls. geom.h: Add a new, optional, providergone callback that is called when a provider is about to be deleted. geom_disk.h: Add a new d_gone() callback to the GEOM disk interface. Bump the DISK_VERSION to version 2. This probably should have been done after a couple of previous changes, especially the addition of the d_getattr() callback. geom_disk.c: Add a providergone callback for the disk class, g_disk_providergone(), that calls the user's d_gone() callback if it exists. Bump the DISK_VERSION to 2. geom_subr.c: In g_destroy_provider(), call the providergone callback if it has been provided. In g_new_geomf(), propagate the class's providergone callback to the new geom instance. blkfront.c: Callers of disk_create() are supposed to pass in DISK_VERSION, not an explicit disk API version number. Update the blkfront driver to do that. disk.9: Update the disk(9) man page to include information on the new d_gone() callback, as well as the previously added d_getattr() callback, d_descr field, and HBA PCI ID fields. MFC after: 5 days
2012-06-24 04:29:03 +00:00
softc->disk->d_gone = dadiskgonecb;
Rearrange how things are done to avoid dereferencing stale pointers in the case of immediate unconfigure after configure. Hold the periph an extra count while we have the task to create sysctl context outstanding so that the periph doesn't go away unexpectedly. Sponsored by: Panasas Reviewed by: scsi@ MFC after: 1 month
2010-06-08 22:46:44 +00:00
softc->disk->d_name = "da";
softc->disk->d_drv1 = periph;
if (cpi.maxio == 0)
softc->disk->d_maxsize = DFLTPHYS; /* traditional default */
else if (cpi.maxio > MAXPHYS)
softc->disk->d_maxsize = MAXPHYS; /* for safety */
else
softc->disk->d_maxsize = cpi.maxio;
softc->disk->d_unit = periph->unit_number;
softc->disk->d_flags = 0;
if ((softc->quirks & DA_Q_NO_SYNC_CACHE) == 0)
softc->disk->d_flags |= DISKFLAG_CANFLUSHCACHE;
Missed a file in r219056: add disk description for da(4).
2011-02-26 23:30:32 +00:00
cam_strvis(softc->disk->d_descr, cgd->inq_data.vendor,
sizeof(cgd->inq_data.vendor), sizeof(softc->disk->d_descr));
strlcat(softc->disk->d_descr, " ", sizeof(softc->disk->d_descr));
cam_strvis(&softc->disk->d_descr[strlen(softc->disk->d_descr)],
cgd->inq_data.product, sizeof(cgd->inq_data.product),
sizeof(softc->disk->d_descr) - strlen(softc->disk->d_descr));
Export PCI IDs of ATA/SATA controllers through CAM and ata(4) layers to GEOM. This information needed for proper soft-RAID's on-disk metadata reading and writing.
2010-07-25 15:43:52 +00:00
softc->disk->d_hba_vendor = cpi.hba_vendor;
softc->disk->d_hba_device = cpi.hba_device;
softc->disk->d_hba_subvendor = cpi.hba_subvendor;
softc->disk->d_hba_subdevice = cpi.hba_subdevice;
Fix a bug which causes a panic in daopen(). The panic is caused by a da(4) instance going away while GEOM is still probing it. In this case, the GEOM disk class instance has been created by disk_create(), and the taste of the disk is queued in the GEOM event queue. While that event is queued, the da(4) instance goes away. When the open call comes into the da(4) driver, it dereferences the freed (but non-NULL) peripheral pointer provided by GEOM, which results in a panic. The solution is to add a callback to the GEOM disk code that is called when all of its resources are cleaned up. This is implemented inside GEOM by adding an optional callback that is called when all consumers have detached from a provider, and the provider is about to be deleted. scsi_cd.c, scsi_da.c: In the register routine for the cd(4) and da(4) routines, acquire a reference to the CAM peripheral instance just before we call disk_create(). Use the new GEOM disk d_gone() callback to register a callback (dadiskgonecb()/cddiskgonecb()) that decrements the peripheral reference count once GEOM has finished cleaning up its resources. In the cd(4) driver, clean up open and close behavior slightly. GEOM makes sure we only get one open() and one close call, so there is no need to set an open flag and decrement the reference count if we are not the first open. In the cd(4) driver, use cam_periph_release_locked() in a couple of error scenarios to avoid extra mutex calls. geom.h: Add a new, optional, providergone callback that is called when a provider is about to be deleted. geom_disk.h: Add a new d_gone() callback to the GEOM disk interface. Bump the DISK_VERSION to version 2. This probably should have been done after a couple of previous changes, especially the addition of the d_getattr() callback. geom_disk.c: Add a providergone callback for the disk class, g_disk_providergone(), that calls the user's d_gone() callback if it exists. Bump the DISK_VERSION to 2. geom_subr.c: In g_destroy_provider(), call the providergone callback if it has been provided. In g_new_geomf(), propagate the class's providergone callback to the new geom instance. blkfront.c: Callers of disk_create() are supposed to pass in DISK_VERSION, not an explicit disk API version number. Update the blkfront driver to do that. disk.9: Update the disk(9) man page to include information on the new d_gone() callback, as well as the previously added d_getattr() callback, d_descr field, and HBA PCI ID fields. MFC after: 5 days
2012-06-24 04:29:03 +00:00
/*
* Acquire a reference to the periph before we register with GEOM.
* We'll release this reference once GEOM calls us back (via
* dadiskgonecb()) telling us that our provider has been freed.
*/
if (cam_periph_acquire(periph) != CAM_REQ_CMP) {
xpt_print(periph->path, "%s: lost periph during "
"registration!\n", __func__);
mtx_lock(periph->sim->mtx);
return (CAM_REQ_CMP_ERR);
}
Rearrange how things are done to avoid dereferencing stale pointers in the case of immediate unconfigure after configure. Hold the periph an extra count while we have the task to create sysctl context outstanding so that the periph doesn't go away unexpectedly. Sponsored by: Panasas Reviewed by: scsi@ MFC after: 1 month
2010-06-08 22:46:44 +00:00
disk_create(softc->disk, DISK_VERSION);
mtx_lock(periph->sim->mtx);
Plumb device physical path reporting from CAM devices, through GEOM and DEVFS, and make it accessible via the diskinfo utility. Extend GEOM's generic attribute query mechanism into generic disk consumers. sys/geom/geom_disk.c: sys/geom/geom_disk.h: sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Allow disk providers to implement a new method which can override the default BIO_GETATTR response, d_getattr(struct bio *). This function returns -1 if not handled, otherwise it returns 0 or an errno to be passed to g_io_deliver(). sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Don't copy the serial number to dp->d_ident anymore, as the CAM XPT is now responsible for returning this information via d_getattr()->(a)dagetattr()->xpt_getatr(). sys/geom/geom_dev.c: - Implement a new ioctl, DIOCGPHYSPATH, which returns the GEOM attribute "GEOM::physpath", if possible. If the attribute request returns a zero-length string, ENOENT is returned. usr.sbin/diskinfo/diskinfo.c: - If the DIOCGPHYSPATH ioctl is successful, report physical path data when diskinfo is executed with the '-v' option. Submitted by: will Reviewed by: gibbs Sponsored by: Spectra Logic Corporation Add generic attribute change notification support to GEOM. sys/sys/geom/geom.h: Add a new attrchanged method field to both g_class and g_geom. sys/sys/geom/geom.h: sys/geom/geom_event.c: - Provide the g_attr_changed() function that providers can use to advertise attribute changes. - Perform delivery of attribute change notifications from a thread context via the standard GEOM event mechanism. sys/geom/geom_subr.c: Inherit the attrchanged method from class to geom (class instance). sys/geom/geom_disk.c: Provide disk_attr_changed() to provide g_attr_changed() access to consumers of the disk API. sys/cam/scsi/scsi_pass.c: sys/cam/scsi/scsi_da.c: sys/geom/geom_dev.c: sys/geom/geom_disk.c: Use attribute changed events to track updates to physical path information. sys/cam/scsi/scsi_da.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, and the updated buffer type references our physical path attribute, emit a GEOM attribute changed event via the disk_attr_changed() API. sys/cam/scsi/scsi_pass.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, update the physical patch devfs alias for this pass instance. Submitted by: gibbs Sponsored by: Spectra Logic Corporation
2011-06-14 17:10:32 +00:00
/*
* Add async callbacks for events of interest.
* I don't bother checking if this fails as,
* in most cases, the system will function just
* fine without them and the only alternative
* would be to not attach the device on failure.
*/
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
xpt_register_async(AC_SENT_BDR | AC_BUS_RESET | AC_LOST_DEVICE |
AC_ADVINFO_CHANGED | AC_SCSI_AEN | AC_UNIT_ATTENTION,
daasync, periph, periph->path);
Plumb device physical path reporting from CAM devices, through GEOM and DEVFS, and make it accessible via the diskinfo utility. Extend GEOM's generic attribute query mechanism into generic disk consumers. sys/geom/geom_disk.c: sys/geom/geom_disk.h: sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Allow disk providers to implement a new method which can override the default BIO_GETATTR response, d_getattr(struct bio *). This function returns -1 if not handled, otherwise it returns 0 or an errno to be passed to g_io_deliver(). sys/cam/scsi/scsi_da.c: sys/cam/ata/ata_da.c: - Don't copy the serial number to dp->d_ident anymore, as the CAM XPT is now responsible for returning this information via d_getattr()->(a)dagetattr()->xpt_getatr(). sys/geom/geom_dev.c: - Implement a new ioctl, DIOCGPHYSPATH, which returns the GEOM attribute "GEOM::physpath", if possible. If the attribute request returns a zero-length string, ENOENT is returned. usr.sbin/diskinfo/diskinfo.c: - If the DIOCGPHYSPATH ioctl is successful, report physical path data when diskinfo is executed with the '-v' option. Submitted by: will Reviewed by: gibbs Sponsored by: Spectra Logic Corporation Add generic attribute change notification support to GEOM. sys/sys/geom/geom.h: Add a new attrchanged method field to both g_class and g_geom. sys/sys/geom/geom.h: sys/geom/geom_event.c: - Provide the g_attr_changed() function that providers can use to advertise attribute changes. - Perform delivery of attribute change notifications from a thread context via the standard GEOM event mechanism. sys/geom/geom_subr.c: Inherit the attrchanged method from class to geom (class instance). sys/geom/geom_disk.c: Provide disk_attr_changed() to provide g_attr_changed() access to consumers of the disk API. sys/cam/scsi/scsi_pass.c: sys/cam/scsi/scsi_da.c: sys/geom/geom_dev.c: sys/geom/geom_disk.c: Use attribute changed events to track updates to physical path information. sys/cam/scsi/scsi_da.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, and the updated buffer type references our physical path attribute, emit a GEOM attribute changed event via the disk_attr_changed() API. sys/cam/scsi/scsi_pass.c: Add AC_ADVINFO_CHANGED to the registered asynchronous CAM events for this driver. When this event occurs, update the physical patch devfs alias for this pass instance. Submitted by: gibbs Sponsored by: Spectra Logic Corporation
2011-06-14 17:10:32 +00:00
/*
* Emit an attribute changed notification just in case
* physical path information arrived before our async
* event handler was registered, but after anyone attaching
* to our disk device polled it.
*/
disk_attr_changed(softc->disk, "GEOM::physpath", M_NOWAIT);
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
/*
* Schedule a periodic media polling events.
*/
callout_init_mtx(&softc->mediapoll_c, periph->sim->mtx, 0);
if ((softc->flags & DA_FLAG_PACK_REMOVABLE) &&
(cgd->inq_flags & SID_AEN) == 0 &&
da_poll_period != 0)
callout_reset(&softc->mediapoll_c, da_poll_period * hz,
damediapoll, periph);
Rearrange how things are done to avoid dereferencing stale pointers in the case of immediate unconfigure after configure. Hold the periph an extra count while we have the task to create sysctl context outstanding so that the periph doesn't go away unexpectedly. Sponsored by: Panasas Reviewed by: scsi@ MFC after: 1 month
2010-06-08 22:46:44 +00:00
xpt_schedule(periph, CAM_PRIORITY_DEV);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
return(CAM_REQ_CMP);
}
static void
dastart(struct cam_periph *periph, union ccb *start_ccb)
{
struct da_softc *softc;
softc = (struct da_softc *)periph->softc;
Tune and add some missing CAM_DEBUG() points for better consistency.
2012-06-05 09:45:42 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE, ("dastart\n"));
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
switch (softc->state) {
case DA_STATE_NORMAL:
{
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
struct bio *bp, *bp1;
uint8_t tag_code;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
/* Execute immediate CCB if waiting. */
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if (periph->immediate_priority <= periph->pinfo.priority) {
Tune and add some missing CAM_DEBUG() points for better consistency.
2012-06-05 09:45:42 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_SUBTRACE,
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
("queuing for immediate ccb\n"));
start_ccb->ccb_h.ccb_state = DA_CCB_WAITING;
SLIST_INSERT_HEAD(&periph->ccb_list, &start_ccb->ccb_h,
periph_links.sle);
periph->immediate_priority = CAM_PRIORITY_NONE;
wakeup(&periph->ccb_list);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
/* May have more work to do, so ensure we stay scheduled */
daschedule(periph);
break;
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
/* Run BIO_DELETE if not running yet. */
if (!softc->delete_running &&
(bp = bioq_first(&softc->delete_queue)) != NULL) {
uint64_t lba;
u_int count;
if (softc->delete_method == DA_DELETE_UNMAP) {
uint8_t *buf = softc->unmap_buf;
uint64_t lastlba = (uint64_t)-1;
uint32_t lastcount = 0;
int blocks = 0, off, ranges = 0;
softc->delete_running = 1;
bzero(softc->unmap_buf, sizeof(softc->unmap_buf));
bp1 = bp;
do {
bioq_remove(&softc->delete_queue, bp1);
if (bp1 != bp)
bioq_insert_tail(&softc->delete_run_queue, bp1);
lba = bp1->bio_pblkno;
count = bp1->bio_bcount / softc->params.secsize;
/* Try to extend the previous range. */
if (lba == lastlba) {
lastcount += count;
off = (ranges - 1) * 16 + 8;
scsi_ulto4b(lastcount, &buf[off + 8]);
} else if (count > 0) {
off = ranges * 16 + 8;
scsi_u64to8b(lba, &buf[off + 0]);
scsi_ulto4b(count, &buf[off + 8]);
lastcount = count;
ranges++;
}
blocks += count;
lastlba = lba + count;
bp1 = bioq_first(&softc->delete_queue);
if (bp1 == NULL ||
ranges >= softc->unmap_max_ranges ||
blocks + bp1->bio_bcount /
softc->params.secsize > softc->unmap_max_lba)
break;
} while (1);
Fix scsi_da's BIO_DELETE->SCSI_UNMAP translation to use correct local variable when determining various sizes related to SCSI UNMAP block descriptor lists. Sponsored by: Intel Reviewed by: mav MFC after: 3 days
2012-08-24 17:08:02 +00:00
scsi_ulto2b(ranges * 16 + 6, &buf[0]);
scsi_ulto2b(ranges * 16, &buf[2]);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
scsi_unmap(&start_ccb->csio,
/*retries*/da_retry_count,
/*cbfcnp*/dadone,
/*tag_action*/MSG_SIMPLE_Q_TAG,
/*byte2*/0,
/*data_ptr*/ buf,
Fix scsi_da's BIO_DELETE->SCSI_UNMAP translation to use correct local variable when determining various sizes related to SCSI UNMAP block descriptor lists. Sponsored by: Intel Reviewed by: mav MFC after: 3 days
2012-08-24 17:08:02 +00:00
/*dxfer_len*/ ranges * 16 + 8,
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
/*sense_len*/SSD_FULL_SIZE,
da_default_timeout * 1000);
start_ccb->ccb_h.ccb_state = DA_CCB_DELETE;
goto out;
} else if (softc->delete_method == DA_DELETE_ZERO ||
softc->delete_method == DA_DELETE_WS10 ||
softc->delete_method == DA_DELETE_WS16) {
softc->delete_running = 1;
lba = bp->bio_pblkno;
count = 0;
bp1 = bp;
do {
bioq_remove(&softc->delete_queue, bp1);
if (bp1 != bp)
bioq_insert_tail(&softc->delete_run_queue, bp1);
count += bp1->bio_bcount / softc->params.secsize;
bp1 = bioq_first(&softc->delete_queue);
if (bp1 == NULL ||
lba + count != bp1->bio_pblkno ||
count + bp1->bio_bcount /
softc->params.secsize > 0xffff)
break;
} while (1);
scsi_write_same(&start_ccb->csio,
/*retries*/da_retry_count,
/*cbfcnp*/dadone,
/*tag_action*/MSG_SIMPLE_Q_TAG,
/*byte2*/softc->delete_method ==
DA_DELETE_ZERO ? 0 : SWS_UNMAP,
softc->delete_method ==
DA_DELETE_WS16 ? 16 : 10,
/*lba*/lba,
/*block_count*/count,
/*data_ptr*/ __DECONST(void *,
zero_region),
/*dxfer_len*/ softc->params.secsize,
/*sense_len*/SSD_FULL_SIZE,
da_default_timeout * 1000);
start_ccb->ccb_h.ccb_state = DA_CCB_DELETE;
goto out;
} else {
bioq_flush(&softc->delete_queue, NULL, 0);
/* FALLTHROUGH */
}
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
/* Run regular command. */
bp = bioq_takefirst(&softc->bio_queue);
if (bp == NULL) {
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
if (softc->tur) {
softc->tur = 0;
scsi_test_unit_ready(&start_ccb->csio,
/*retries*/ da_retry_count,
dadone,
MSG_SIMPLE_Q_TAG,
SSD_FULL_SIZE,
da_default_timeout * 1000);
start_ccb->ccb_h.ccb_bp = NULL;
start_ccb->ccb_h.ccb_state = DA_CCB_TUR;
xpt_action(start_ccb);
} else
xpt_release_ccb(start_ccb);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
break;
}
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
if (softc->tur) {
softc->tur = 0;
cam_periph_release_locked(periph);
}
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
if ((bp->bio_flags & BIO_ORDERED) != 0 ||
(softc->flags & DA_FLAG_NEED_OTAG) != 0) {
softc->flags &= ~DA_FLAG_NEED_OTAG;
softc->ordered_tag_count++;
tag_code = MSG_ORDERED_Q_TAG;
} else {
tag_code = MSG_SIMPLE_Q_TAG;
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
switch (bp->bio_cmd) {
case BIO_READ:
case BIO_WRITE:
scsi_read_write(&start_ccb->csio,
/*retries*/da_retry_count,
/*cbfcnp*/dadone,
/*tag_action*/tag_code,
/*read_op*/bp->bio_cmd
== BIO_READ,
/*byte2*/0,
softc->minimum_cmd_size,
/*lba*/bp->bio_pblkno,
/*block_count*/bp->bio_bcount /
softc->params.secsize,
/*data_ptr*/ bp->bio_data,
/*dxfer_len*/ bp->bio_bcount,
/*sense_len*/SSD_FULL_SIZE,
da_default_timeout * 1000);
break;
case BIO_FLUSH:
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
* BIO_FLUSH doesn't currently communicate
* range data, so we synchronize the cache
* over the whole disk. We also force
* ordered tag semantics the flush applies
* to all previously queued I/O.
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
*/
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
scsi_synchronize_cache(&start_ccb->csio,
/*retries*/1,
/*cbfcnp*/dadone,
MSG_ORDERED_Q_TAG,
/*begin_lba*/0,
/*lb_count*/0,
SSD_FULL_SIZE,
da_default_timeout*1000);
break;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
start_ccb->ccb_h.ccb_state = DA_CCB_BUFFER_IO;
out:
/*
* Block out any asyncronous callbacks
* while we touch the pending ccb list.
*/
LIST_INSERT_HEAD(&softc->pending_ccbs,
&start_ccb->ccb_h, periph_links.le);
softc->outstanding_cmds++;
/* We expect a unit attention from this device */
if ((softc->flags & DA_FLAG_RETRY_UA) != 0) {
start_ccb->ccb_h.ccb_state |= DA_CCB_RETRY_UA;
softc->flags &= ~DA_FLAG_RETRY_UA;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
start_ccb->ccb_h.ccb_bp = bp;
xpt_action(start_ccb);
/* May have more work to do, so ensure we stay scheduled */
daschedule(periph);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
break;
}
case DA_STATE_PROBE:
{
struct scsi_read_capacity_data *rcap;
Zero scsi_readcapacity allocations so we can really tell if there has been data returned (for some broken residual checking points). MFC after: 3 days
2007-05-18 21:58:07 +00:00
rcap = (struct scsi_read_capacity_data *)
malloc(sizeof(*rcap), M_SCSIDA, M_NOWAIT|M_ZERO);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if (rcap == NULL) {
printf("dastart: Couldn't malloc read_capacity data\n");
/* da_free_periph??? */
break;
}
Removes essentially unused variables from scsi_da probe setups PR: kern/169835 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:28:12 +00:00
scsi_read_capacity(&start_ccb->csio,
Changed scsi_da device requests to use the sysctl tunable value for retry_count and da_default_timeout where their current hardcoded values matched the current default value for said tunables. PR: kern/169976 Reviewed by: pjd (mentor) Approved by: mav
2013-01-10 12:25:00 +00:00
/*retries*/da_retry_count,
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
dadone,
MSG_SIMPLE_Q_TAG,
rcap,
SSD_FULL_SIZE,
/*timeout*/5000);
start_ccb->ccb_h.ccb_bp = NULL;
start_ccb->ccb_h.ccb_state = DA_CCB_PROBE;
xpt_action(start_ccb);
break;
}
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
case DA_STATE_PROBE2:
{
struct scsi_read_capacity_data_long *rcaplong;
rcaplong = (struct scsi_read_capacity_data_long *)
Zero scsi_readcapacity allocations so we can really tell if there has been data returned (for some broken residual checking points). MFC after: 3 days
2007-05-18 21:58:07 +00:00
malloc(sizeof(*rcaplong), M_SCSIDA, M_NOWAIT|M_ZERO);
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
if (rcaplong == NULL) {
printf("dastart: Couldn't malloc read_capacity data\n");
/* da_free_periph??? */
break;
}
Removes essentially unused variables from scsi_da probe setups PR: kern/169835 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:28:12 +00:00
scsi_read_capacity_16(&start_ccb->csio,
Changed scsi_da device requests to use the sysctl tunable value for retry_count and da_default_timeout where their current hardcoded values matched the current default value for said tunables. PR: kern/169976 Reviewed by: pjd (mentor) Approved by: mav
2013-01-10 12:25:00 +00:00
/*retries*/ da_retry_count,
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
/*cbfcnp*/ dadone,
/*tag_action*/ MSG_SIMPLE_Q_TAG,
/*lba*/ 0,
/*reladr*/ 0,
/*pmi*/ 0,
Add CAM infrastructure to allow reporting when a drive's long read capacity data changes. cam_ccb.h: Add a new advanced information type, CDAI_TYPE_RCAPLONG, for long read capacity data. cam_xpt_internal.h: Add a read capacity data pointer and length to struct cam_ed. cam_xpt.c: Free the read capacity buffer when a device goes away. While we're here, make sure we don't leak memory for other malloced fields in struct cam_ed. scsi_all.c: Update the scsi_read_capacity_16() to take a uint8_t * and a length instead of just a pointer to the parameter data structure. This will hopefully make this function somewhat immune to future changes in the parameter data. scsi_all.h: Add some extra bit definitions to struct scsi_read_capacity_data_long, and bump up the structure size to the full size specified by SBC-3. Change the prototype for scsi_read_capacity_16(). scsi_da.c: Register changes in read capacity data with the transport layer. This allows the transport layer to send out an async notification to interested parties. Update the dasetgeom() API. Use scsi_extract_sense_len() instead of scsi_extract_sense(). scsi_xpt.c: Add support for the new CDAI_TYPE_RCAPLONG advanced information type. Make sure we set the physpath pointer to NULL after freeing it. This allows blindly freeing it in the struct cam_ed destructor. sys/param.h: Bump __FreeBSD_version from 1000005 to 1000006 to make it easier for third party drivers to determine that the read capacity data async notification is available. camcontrol.c, mptutil/mpt_cam.c: Update these for the new scsi_read_capacity_16() argument structure. Sponsored by: Spectra Logic
2012-01-26 18:09:28 +00:00
/*rcap_buf*/ (uint8_t *)rcaplong,
/*rcap_buf_len*/ sizeof(*rcaplong),
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
/*sense_len*/ SSD_FULL_SIZE,
Changed scsi_da device requests to use the sysctl tunable value for retry_count and da_default_timeout where their current hardcoded values matched the current default value for said tunables. PR: kern/169976 Reviewed by: pjd (mentor) Approved by: mav
2013-01-10 12:25:00 +00:00
/*timeout*/ da_default_timeout * 1000);
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
start_ccb->ccb_h.ccb_bp = NULL;
start_ccb->ccb_h.ccb_state = DA_CCB_PROBE2;
xpt_action(start_ccb);
break;
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
}
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
static int
cmd6workaround(union ccb *ccb)
{
struct scsi_rw_6 cmd6;
struct scsi_rw_10 *cmd10;
struct da_softc *softc;
Updates to cmd6workaround: * Only update cdb in place if not CDB_POINTER * Correctly check for QFRZ before restarting CCB * More accurate printf message * style(9) changes at end Patch tested successfully on Maxtor 4 G120J6 GAK8.
2002-08-16 22:05:19 +00:00
u_int8_t *cdb;
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
struct bio *bp;
Updates to cmd6workaround: * Only update cdb in place if not CDB_POINTER * Correctly check for QFRZ before restarting CCB * More accurate printf message * style(9) changes at end Patch tested successfully on Maxtor 4 G120J6 GAK8.
2002-08-16 22:05:19 +00:00
int frozen;
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
Updates to cmd6workaround: * Only update cdb in place if not CDB_POINTER * Correctly check for QFRZ before restarting CCB * More accurate printf message * style(9) changes at end Patch tested successfully on Maxtor 4 G120J6 GAK8.
2002-08-16 22:05:19 +00:00
cdb = ccb->csio.cdb_io.cdb_bytes;
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
softc = (struct da_softc *)xpt_path_periph(ccb->ccb_h.path)->softc;
if (ccb->ccb_h.ccb_state == DA_CCB_DELETE) {
if (softc->delete_method == DA_DELETE_UNMAP) {
xpt_print(ccb->ccb_h.path, "UNMAP is not supported, "
"switching to WRITE SAME(16) with UNMAP.\n");
Updates delete_method sysctl changes to always maintain disk d_flags DISKFLAG_CANDELETE. While this change makes this layer consistent other layers such as UFS and ZFS BIO_DELETE support may not notice any change made manually via these device sysctls until the device is reopened via a mount. Also corrected var order in dadeletemethodsysctl PR: kern/169801 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:57:46 +00:00
dadeletemethodset(softc, DA_DELETE_WS16);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
} else if (softc->delete_method == DA_DELETE_WS16) {
xpt_print(ccb->ccb_h.path,
"WRITE SAME(16) with UNMAP is not supported, "
"disabling BIO_DELETE.\n");
Updates delete_method sysctl changes to always maintain disk d_flags DISKFLAG_CANDELETE. While this change makes this layer consistent other layers such as UFS and ZFS BIO_DELETE support may not notice any change made manually via these device sysctls until the device is reopened via a mount. Also corrected var order in dadeletemethodsysctl PR: kern/169801 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:57:46 +00:00
dadeletemethodset(softc, DA_DELETE_DISABLE);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
} else if (softc->delete_method == DA_DELETE_WS10) {
xpt_print(ccb->ccb_h.path,
"WRITE SAME(10) with UNMAP is not supported, "
"disabling BIO_DELETE.\n");
Updates delete_method sysctl changes to always maintain disk d_flags DISKFLAG_CANDELETE. While this change makes this layer consistent other layers such as UFS and ZFS BIO_DELETE support may not notice any change made manually via these device sysctls until the device is reopened via a mount. Also corrected var order in dadeletemethodsysctl PR: kern/169801 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:57:46 +00:00
dadeletemethodset(softc, DA_DELETE_DISABLE);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
} else if (softc->delete_method == DA_DELETE_ZERO) {
xpt_print(ccb->ccb_h.path,
"WRITE SAME(10) is not supported, "
"disabling BIO_DELETE.\n");
Updates delete_method sysctl changes to always maintain disk d_flags DISKFLAG_CANDELETE. While this change makes this layer consistent other layers such as UFS and ZFS BIO_DELETE support may not notice any change made manually via these device sysctls until the device is reopened via a mount. Also corrected var order in dadeletemethodsysctl PR: kern/169801 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:57:46 +00:00
dadeletemethodset(softc, DA_DELETE_DISABLE);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
} else
Updates delete_method sysctl changes to always maintain disk d_flags DISKFLAG_CANDELETE. While this change makes this layer consistent other layers such as UFS and ZFS BIO_DELETE support may not notice any change made manually via these device sysctls until the device is reopened via a mount. Also corrected var order in dadeletemethodsysctl PR: kern/169801 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:57:46 +00:00
dadeletemethodset(softc, DA_DELETE_DISABLE);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
while ((bp = bioq_takefirst(&softc->delete_run_queue))
!= NULL)
bioq_disksort(&softc->delete_queue, bp);
bioq_insert_tail(&softc->delete_queue,
(struct bio *)ccb->ccb_h.ccb_bp);
ccb->ccb_h.ccb_bp = NULL;
return (0);
}
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
Updates to cmd6workaround: * Only update cdb in place if not CDB_POINTER * Correctly check for QFRZ before restarting CCB * More accurate printf message * style(9) changes at end Patch tested successfully on Maxtor 4 G120J6 GAK8.
2002-08-16 22:05:19 +00:00
/* Translation only possible if CDB is an array and cmd is R/W6 */
if ((ccb->ccb_h.flags & CAM_CDB_POINTER) != 0 ||
(*cdb != READ_6 && *cdb != WRITE_6))
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
return 0;
Add an xpt_print function to reduce most of the xpt_print_path/printf pairs. Convert the core code to use it. Reviewed by: scsi (various folks weighing in with good advice) MFC after: 2 weeks
2006-12-05 07:45:28 +00:00
xpt_print(ccb->ccb_h.path, "READ(6)/WRITE(6) not supported, "
"increasing minimum_cmd_size to 10.\n");
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
softc->minimum_cmd_size = 10;
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
Updates to cmd6workaround: * Only update cdb in place if not CDB_POINTER * Correctly check for QFRZ before restarting CCB * More accurate printf message * style(9) changes at end Patch tested successfully on Maxtor 4 G120J6 GAK8.
2002-08-16 22:05:19 +00:00
bcopy(cdb, &cmd6, sizeof(struct scsi_rw_6));
cmd10 = (struct scsi_rw_10 *)cdb;
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
cmd10->opcode = (cmd6.opcode == READ_6) ? READ_10 : WRITE_10;
cmd10->byte2 = 0;
scsi_ulto4b(scsi_3btoul(cmd6.addr), cmd10->addr);
cmd10->reserved = 0;
scsi_ulto2b(cmd6.length, cmd10->length);
cmd10->control = cmd6.control;
Updates to cmd6workaround: * Only update cdb in place if not CDB_POINTER * Correctly check for QFRZ before restarting CCB * More accurate printf message * style(9) changes at end Patch tested successfully on Maxtor 4 G120J6 GAK8.
2002-08-16 22:05:19 +00:00
ccb->csio.cdb_len = sizeof(*cmd10);
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
Updates to cmd6workaround: * Only update cdb in place if not CDB_POINTER * Correctly check for QFRZ before restarting CCB * More accurate printf message * style(9) changes at end Patch tested successfully on Maxtor 4 G120J6 GAK8.
2002-08-16 22:05:19 +00:00
/* Requeue request, unfreezing queue if necessary */
frozen = (ccb->ccb_h.status & CAM_DEV_QFRZN) != 0;
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
ccb->ccb_h.status = CAM_REQUEUE_REQ;
xpt_action(ccb);
Updates to cmd6workaround: * Only update cdb in place if not CDB_POINTER * Correctly check for QFRZ before restarting CCB * More accurate printf message * style(9) changes at end Patch tested successfully on Maxtor 4 G120J6 GAK8.
2002-08-16 22:05:19 +00:00
if (frozen) {
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
cam_release_devq(ccb->ccb_h.path,
/*relsim_flags*/0,
/*reduction*/0,
/*timeout*/0,
/*getcount_only*/0);
Updates to cmd6workaround: * Only update cdb in place if not CDB_POINTER * Correctly check for QFRZ before restarting CCB * More accurate printf message * style(9) changes at end Patch tested successfully on Maxtor 4 G120J6 GAK8.
2002-08-16 22:05:19 +00:00
}
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
return (ERESTART);
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
static void
dadone(struct cam_periph *periph, union ccb *done_ccb)
{
struct da_softc *softc;
struct ccb_scsiio *csio;
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
u_int32_t priority;
Do not schedule periph for payload/TUR requests if reprobe is in progress to avoid sending extra READ CAPACITY requests by dastart(). Schedule periph again on reprobe completion, or otherwise it may stuck indefinitely long. This should fix USB explore thread hanging on device unplug, waiting for periph destruction. Reported by: hselasky
2013-01-11 16:10:11 +00:00
da_ccb_state state;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
softc = (struct da_softc *)periph->softc;
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
priority = done_ccb->ccb_h.pinfo.priority;
Tune and add some missing CAM_DEBUG() points for better consistency.
2012-06-05 09:45:42 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE, ("dadone\n"));
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
csio = &done_ccb->csio;
Do not schedule periph for payload/TUR requests if reprobe is in progress to avoid sending extra READ CAPACITY requests by dastart(). Schedule periph again on reprobe completion, or otherwise it may stuck indefinitely long. This should fix USB explore thread hanging on device unplug, waiting for periph destruction. Reported by: hselasky
2013-01-11 16:10:11 +00:00
state = csio->ccb_h.ccb_state & DA_CCB_TYPE_MASK;
switch (state) {
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
case DA_CCB_BUFFER_IO:
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
case DA_CCB_DELETE:
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
{
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
struct bio *bp, *bp1;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Complete the bio/buf divorce for all code below devfs::strategy Exceptions: Vinum untouched. This means that it cannot be compiled. Greg Lehey is on the case. CCD not converted yet, casts to struct buf (still safe) atapi-cd casts to struct buf to examine B_PHYS
2000-04-15 05:54:02 +00:00
bp = (struct bio *)done_ccb->ccb_h.ccb_bp;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if ((done_ccb->ccb_h.status & CAM_STATUS_MASK) != CAM_REQ_CMP) {
int error;
int sf;
sys/cam/scsi/scsi_da.c: - Only attempt the closing synchronize cache on a disk if it is still there. - When a device is lost, report the number of outstanding I/Os as they are drained. - When a device is lost, return any unprocessed bios with ENXIO instead of EIO. - Filter asynchronous events, but always allow cam_periph_async() to see them too. Sponsored by: Spectra Logic Corporation
2011-06-14 16:05:00 +00:00
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if ((csio->ccb_h.ccb_state & DA_CCB_RETRY_UA) != 0)
sf = SF_RETRY_UA;
else
sf = 0;
Rewrite of the CAM error recovery code. Some of the major changes include: - The SCSI error handling portion of cam_periph_error() has been broken out into a number of subfunctions to better modularize the code that handles the hierarchy of SCSI errors. As a result, the code is now much easier to read. - String handling and error printing has been significantly revamped. We now use sbufs to do string formatting instead of using printfs (for the kernel) and snprintf/strncat (for userland) as before. There is a new catchall error printing routine, cam_error_print() and its string-based counterpart, cam_error_string() that allow the kernel and userland applications to pass in a CCB and have errors printed out properly, whether or not they're SCSI errors. Among other things, this helped eliminate a fair amount of duplicate code in camcontrol. We now print out more information than before, including the CAM status and SCSI status and the error recovery action taken to remedy the problem. - sbufs are now available in userland, via libsbuf. This change was necessary since most of the error printing code is shared between libcam and the kernel. - A new transfer settings interface is included in this checkin. This code is #ifdef'ed out, and is primarily intended to aid discussion with HBA driver authors on the final form the interface should take. There is example code in the ahc(4) driver that implements the HBA driver side of the new interface. The new transfer settings code won't be enabled until we're ready to switch all HBA drivers over to the new interface. src/Makefile.inc1, lib/Makefile: Add libsbuf. It must be built before libcam, since libcam uses sbuf routines. libcam/Makefile: libcam now depends on libsbuf. libsbuf/Makefile: Add a makefile for libsbuf. This pulls in the sbuf sources from sys/kern. bsd.libnames.mk: Add LIBSBUF. camcontrol/Makefile: Add -lsbuf. Since camcontrol is statically linked, we can't depend on the dynamic linker to pull in libsbuf. camcontrol.c: Use cam_error_print() instead of checking for CAM_SCSI_STATUS_ERROR on every failed CCB. sbuf.9: Change the prototypes for sbuf_cat() and sbuf_cpy() so that the source string is now a const char *. This is more in line wth the standard system string functions, and helps eliminate warnings when dealing with a const source buffer. Fix a typo. cam.c: Add description strings for the various CAM error status values, as well as routines to look up those strings. Add new cam_error_string() and cam_error_print() routines for userland and the kernel. cam.h: Add a new CAM flag, CAM_RETRY_SELTO. Add enumerated types for the various options available with cam_error_print() and cam_error_string(). cam_ccb.h: Add new transfer negotiation structures/types. Change inq_len in the ccb_getdev structure to be "reserved". This field has never been filled in, and will be removed when we next bump the CAM version. cam_debug.h: Fix typo. cam_periph.c: Modularize cam_periph_error(). The SCSI error handling part of cam_periph_error() is now in camperiphscsistatuserror() and camperiphscsisenseerror(). In cam_periph_lock(), increase the reference count on the periph while we wait for our lock attempt to succeed so that the periph won't go away while we're sleeping. cam_xpt.c: Add new transfer negotiation code. (ifdefed out) Add a new function, xpt_path_string(). This is a string/sbuf analog to xpt_print_path(). scsi_all.c: Revamp string handing and error printing code. We now use sbufs for much of the string formatting code. More of that code is shared between userland the kernel. scsi_all.h: Get rid of SS_TURSTART, it wasn't terribly useful in the first place. Add a new error action, SS_REQSENSE. (Send a request sense and then retry the command.) This is useful when the controller hasn't performed autosense for some reason. Change the default actions around a bit. scsi_cd.c, scsi_da.c, scsi_pt.c, scsi_ses.c: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Selection timeouts shouldn't be covered by a sense flag. scsi_pass.[ch]: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Get rid of the last vestiges of a read/write interface. libkern/bsearch.c, sys/libkern.h, conf/files: Add bsearch.c, which is needed for some of the new table lookup routines. aic7xxx_freebsd.c: Define AHC_NEW_TRAN_SETTINGS if CAM_NEW_TRAN_CODE is defined. sbuf.h, subr_sbuf.c: Add the appropriate #ifdefs so sbufs can compile and run in userland. Change sbuf_printf() to use vsnprintf() instead of kvprintf(), which is only available in the kernel. Change the source string for sbuf_cpy() and sbuf_cat() to be a const char *. Add __BEGIN_DECLS and __END_DECLS around function prototypes since they're now exported to userland. kdump/mkioctls: Include stdio.h before cam.h since cam.h now includes a function with a FILE * argument. Submitted by: gibbs (mostly) Reviewed by: jdp, marcel (libsbuf makefile changes) Reviewed by: des (sbuf changes) Reviewed by: ken
2001-03-27 05:45:52 +00:00
error = daerror(done_ccb, CAM_RETRY_SELTO, sf);
if (error == ERESTART) {
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
* A retry was scheuled, so
* just return.
*/
return;
}
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
bp = (struct bio *)done_ccb->ccb_h.ccb_bp;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if (error != 0) {
sys/cam/scsi/scsi_da.c: - Only attempt the closing synchronize cache on a disk if it is still there. - When a device is lost, report the number of outstanding I/Os as they are drained. - When a device is lost, return any unprocessed bios with ENXIO instead of EIO. - Filter asynchronous events, but always allow cam_periph_async() to see them too. Sponsored by: Spectra Logic Corporation
2011-06-14 16:05:00 +00:00
int queued_error;
/*
* return all queued I/O with EIO, so that
* the client can retry these I/Os in the
* proper order should it attempt to recover.
*/
queued_error = EIO;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
sys/cam/scsi/scsi_da.c: - Only attempt the closing synchronize cache on a disk if it is still there. - When a device is lost, report the number of outstanding I/Os as they are drained. - When a device is lost, return any unprocessed bios with ENXIO instead of EIO. - Filter asynchronous events, but always allow cam_periph_async() to see them too. Sponsored by: Spectra Logic Corporation
2011-06-14 16:05:00 +00:00
if (error == ENXIO
&& (softc->flags & DA_FLAG_PACK_INVALID)== 0) {
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
* Catastrophic error. Mark our pack as
* invalid.
*/
Add an xpt_print function to reduce most of the xpt_print_path/printf pairs. Convert the core code to use it. Reviewed by: scsi (various folks weighing in with good advice) MFC after: 2 weeks
2006-12-05 07:45:28 +00:00
/*
* XXX See if this is really a media
* XXX change first?
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
*/
Add an xpt_print function to reduce most of the xpt_print_path/printf pairs. Convert the core code to use it. Reviewed by: scsi (various folks weighing in with good advice) MFC after: 2 weeks
2006-12-05 07:45:28 +00:00
xpt_print(periph->path,
"Invalidating pack\n");
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
softc->flags |= DA_FLAG_PACK_INVALID;
sys/cam/scsi/scsi_da.c: - Only attempt the closing synchronize cache on a disk if it is still there. - When a device is lost, report the number of outstanding I/Os as they are drained. - When a device is lost, return any unprocessed bios with ENXIO instead of EIO. - Filter asynchronous events, but always allow cam_periph_async() to see them too. Sponsored by: Spectra Logic Corporation
2011-06-14 16:05:00 +00:00
queued_error = ENXIO;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
sys/cam/scsi/scsi_da.c: - Only attempt the closing synchronize cache on a disk if it is still there. - When a device is lost, report the number of outstanding I/Os as they are drained. - When a device is lost, return any unprocessed bios with ENXIO instead of EIO. - Filter asynchronous events, but always allow cam_periph_async() to see them too. Sponsored by: Spectra Logic Corporation
2011-06-14 16:05:00 +00:00
bioq_flush(&softc->bio_queue, NULL,
queued_error);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
if (bp != NULL) {
bp->bio_error = error;
bp->bio_resid = bp->bio_bcount;
bp->bio_flags |= BIO_ERROR;
}
} else if (bp != NULL) {
Complete the bio/buf divorce for all code below devfs::strategy Exceptions: Vinum untouched. This means that it cannot be compiled. Greg Lehey is on the case. CCD not converted yet, casts to struct buf (still safe) atapi-cd casts to struct buf to examine B_PHYS
2000-04-15 05:54:02 +00:00
bp->bio_resid = csio->resid;
bp->bio_error = 0;
Remove unused cmd6workaround() calls. MFC after: 1 day
2003-08-06 14:24:16 +00:00
if (bp->bio_resid != 0)
Complete the bio/buf divorce for all code below devfs::strategy Exceptions: Vinum untouched. This means that it cannot be compiled. Greg Lehey is on the case. CCD not converted yet, casts to struct buf (still safe) atapi-cd casts to struct buf to examine B_PHYS
2000-04-15 05:54:02 +00:00
bp->bio_flags |= BIO_ERROR;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
if ((done_ccb->ccb_h.status & CAM_DEV_QFRZN) != 0)
cam_release_devq(done_ccb->ccb_h.path,
/*relsim_flags*/0,
/*reduction*/0,
/*timeout*/0,
/*getcount_only*/0);
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
} else if (bp != NULL) {
Rewrite of the CAM error recovery code. Some of the major changes include: - The SCSI error handling portion of cam_periph_error() has been broken out into a number of subfunctions to better modularize the code that handles the hierarchy of SCSI errors. As a result, the code is now much easier to read. - String handling and error printing has been significantly revamped. We now use sbufs to do string formatting instead of using printfs (for the kernel) and snprintf/strncat (for userland) as before. There is a new catchall error printing routine, cam_error_print() and its string-based counterpart, cam_error_string() that allow the kernel and userland applications to pass in a CCB and have errors printed out properly, whether or not they're SCSI errors. Among other things, this helped eliminate a fair amount of duplicate code in camcontrol. We now print out more information than before, including the CAM status and SCSI status and the error recovery action taken to remedy the problem. - sbufs are now available in userland, via libsbuf. This change was necessary since most of the error printing code is shared between libcam and the kernel. - A new transfer settings interface is included in this checkin. This code is #ifdef'ed out, and is primarily intended to aid discussion with HBA driver authors on the final form the interface should take. There is example code in the ahc(4) driver that implements the HBA driver side of the new interface. The new transfer settings code won't be enabled until we're ready to switch all HBA drivers over to the new interface. src/Makefile.inc1, lib/Makefile: Add libsbuf. It must be built before libcam, since libcam uses sbuf routines. libcam/Makefile: libcam now depends on libsbuf. libsbuf/Makefile: Add a makefile for libsbuf. This pulls in the sbuf sources from sys/kern. bsd.libnames.mk: Add LIBSBUF. camcontrol/Makefile: Add -lsbuf. Since camcontrol is statically linked, we can't depend on the dynamic linker to pull in libsbuf. camcontrol.c: Use cam_error_print() instead of checking for CAM_SCSI_STATUS_ERROR on every failed CCB. sbuf.9: Change the prototypes for sbuf_cat() and sbuf_cpy() so that the source string is now a const char *. This is more in line wth the standard system string functions, and helps eliminate warnings when dealing with a const source buffer. Fix a typo. cam.c: Add description strings for the various CAM error status values, as well as routines to look up those strings. Add new cam_error_string() and cam_error_print() routines for userland and the kernel. cam.h: Add a new CAM flag, CAM_RETRY_SELTO. Add enumerated types for the various options available with cam_error_print() and cam_error_string(). cam_ccb.h: Add new transfer negotiation structures/types. Change inq_len in the ccb_getdev structure to be "reserved". This field has never been filled in, and will be removed when we next bump the CAM version. cam_debug.h: Fix typo. cam_periph.c: Modularize cam_periph_error(). The SCSI error handling part of cam_periph_error() is now in camperiphscsistatuserror() and camperiphscsisenseerror(). In cam_periph_lock(), increase the reference count on the periph while we wait for our lock attempt to succeed so that the periph won't go away while we're sleeping. cam_xpt.c: Add new transfer negotiation code. (ifdefed out) Add a new function, xpt_path_string(). This is a string/sbuf analog to xpt_print_path(). scsi_all.c: Revamp string handing and error printing code. We now use sbufs for much of the string formatting code. More of that code is shared between userland the kernel. scsi_all.h: Get rid of SS_TURSTART, it wasn't terribly useful in the first place. Add a new error action, SS_REQSENSE. (Send a request sense and then retry the command.) This is useful when the controller hasn't performed autosense for some reason. Change the default actions around a bit. scsi_cd.c, scsi_da.c, scsi_pt.c, scsi_ses.c: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Selection timeouts shouldn't be covered by a sense flag. scsi_pass.[ch]: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Get rid of the last vestiges of a read/write interface. libkern/bsearch.c, sys/libkern.h, conf/files: Add bsearch.c, which is needed for some of the new table lookup routines. aic7xxx_freebsd.c: Define AHC_NEW_TRAN_SETTINGS if CAM_NEW_TRAN_CODE is defined. sbuf.h, subr_sbuf.c: Add the appropriate #ifdefs so sbufs can compile and run in userland. Change sbuf_printf() to use vsnprintf() instead of kvprintf(), which is only available in the kernel. Change the source string for sbuf_cpy() and sbuf_cat() to be a const char *. Add __BEGIN_DECLS and __END_DECLS around function prototypes since they're now exported to userland. kdump/mkioctls: Include stdio.h before cam.h since cam.h now includes a function with a FILE * argument. Submitted by: gibbs (mostly) Reviewed by: jdp, marcel (libsbuf makefile changes) Reviewed by: des (sbuf changes) Reviewed by: ken
2001-03-27 05:45:52 +00:00
if ((done_ccb->ccb_h.status & CAM_DEV_QFRZN) != 0)
panic("REQ_CMP with QFRZN");
Complete the bio/buf divorce for all code below devfs::strategy Exceptions: Vinum untouched. This means that it cannot be compiled. Greg Lehey is on the case. CCD not converted yet, casts to struct buf (still safe) atapi-cd casts to struct buf to examine B_PHYS
2000-04-15 05:54:02 +00:00
bp->bio_resid = csio->resid;
Remove unused cmd6workaround() calls. MFC after: 1 day
2003-08-06 14:24:16 +00:00
if (csio->resid > 0)
Complete the bio/buf divorce for all code below devfs::strategy Exceptions: Vinum untouched. This means that it cannot be compiled. Greg Lehey is on the case. CCD not converted yet, casts to struct buf (still safe) atapi-cd casts to struct buf to examine B_PHYS
2000-04-15 05:54:02 +00:00
bp->bio_flags |= BIO_ERROR;
Fix a race condition in CAM peripheral free handling, locking in the CAM XPT bus traversal code, and a number of other periph level issues. cam_periph.h, cam_periph.c: Modify cam_periph_acquire() to test the CAM_PERIPH_INVALID flag prior to allowing a reference count to be gained on a peripheral. Callers of this function will receive CAM_REQ_CMP_ERR status in the situation of attempting to reference an invalidated periph. This guarantees that a peripheral scheduled for a deferred free will not be accessed during its wait for destruction. Panic during attempts to drop a reference count on a peripheral that already has a zero reference count. In cam_periph_list(), use a local sbuf with SBUF_FIXEDLEN set so that mallocs do not occur while the xpt topology lock is held, regardless of the allocation policy of the passed in sbuf. Add a new routine, cam_periph_release_locked_buses(), that can be called when the caller already holds the CAM topology lock. Add some extra debugging for duplicate peripheral allocations in cam_periph_alloc(). Treat CAM_DEV_NOT_THERE much the same as a selection timeout (AC_LOST_DEVICE is emitted), but forgo retries. cam_xpt.c: Revamp the way the EDT traversal code does locking and reference counting. This was broken, since it assumed that the EDT would not change during traversal, but that assumption is no longer valid. So, to prevent devices from going away while we traverse the EDT, make sure we properly lock everything and hold references on devices that we are using. The two peripheral driver traversal routines should be examined. xptpdperiphtraverse() holds the topology lock for the entire time it runs. xptperiphtraverse() is now locked properly, but only holds the topology lock while it is traversing the list, and not while the traversal function is running. The bus locking code in xptbustraverse() should also be revisited at a later time, since it is complex and should probably be simplified. scsi_da.c: Pay attention to the return value from cam_periph_acquire(). Return 0 always from daclose() even if the disk is now gone. Add some rudimentary error injection support. scsi_sg.c: Fix reference counting in the sg(4) driver. The sg driver was calling cam_periph_release() on close, but never called cam_periph_acquire() (which increments the reference count) on open. The periph code correctly complained that the sg(4) driver was trying to decrement the refcount when it was already 0. Sponsored by: Spectra Logic MFC after: 2 weeks
2012-01-12 00:41:48 +00:00
if (softc->error_inject != 0) {
bp->bio_error = softc->error_inject;
bp->bio_resid = bp->bio_bcount;
bp->bio_flags |= BIO_ERROR;
softc->error_inject = 0;
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
/*
* Block out any asyncronous callbacks
* while we touch the pending ccb list.
*/
LIST_REMOVE(&done_ccb->ccb_h, periph_links.le);
Don't use the devstat->busy_count for state decisions in the device drivers. Doing so imposes atomicity and locking constraints on the devstat API. By: ken
2003-03-15 11:00:56 +00:00
softc->outstanding_cmds--;
if (softc->outstanding_cmds == 0)
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
softc->flags |= DA_FLAG_WENT_IDLE;
sys/cam/scsi/scsi_da.c: - Only attempt the closing synchronize cache on a disk if it is still there. - When a device is lost, report the number of outstanding I/Os as they are drained. - When a device is lost, return any unprocessed bios with ENXIO instead of EIO. - Filter asynchronous events, but always allow cam_periph_async() to see them too. Sponsored by: Spectra Logic Corporation
2011-06-14 16:05:00 +00:00
if ((softc->flags & DA_FLAG_PACK_INVALID) != 0) {
xpt_print(periph->path, "oustanding %d\n",
softc->outstanding_cmds);
}
Do not schedule periph for payload/TUR requests if reprobe is in progress to avoid sending extra READ CAPACITY requests by dastart(). Schedule periph again on reprobe completion, or otherwise it may stuck indefinitely long. This should fix USB explore thread hanging on device unplug, waiting for periph destruction. Reported by: hselasky
2013-01-11 16:10:11 +00:00
if (state == DA_CCB_DELETE) {
Add BIO_DELETE support for SCSI Direct Access devices (da). Depending on device capabilities use different methods to implement it. Currently used method can be read/set via kern.cam.da.X.delete_method sysctls. Possible values are: NONE - no provisioning support reported by the device; DISABLE - provisioning support was disabled because of errors; ZERO - use WRITE SAME (10) command to write zeroes; WS10 - use WRITE SAME (10) command with UNMAP bit set; WS16 - use WRITE SAME (16) command with UNMAP bit set; UNMAP - use UNMAP command (equivalent of the ATA DSM TRIM command). The last two methods (UNMAP and WS16) are defined by SBC specification and the UNMAP method is the most advanced one. The rest of methods I've found supported in Linux, and as soon as they were trivial to implement, then why not? Hope they will be useful in some cases. Unluckily I have no devices properly reporting parameters of the logical block provisioning support via respective VPD pages (0xB0 and 0xB2). So all info I have/use now is the flag telling whether logical block provisioning is supported or not. As result, specific methods chosen now by trying different ones in order (UNMAP, WS16, DISABLE) and checking completion status to fallback if needed. I don't expect problems from this, as if something go wrong, it should just disable itself. It may disable even too aggressively if only some command parameter misfit. Unlike Linux, which executes each delete with separate request, I've implemented here the same request aggregation as implemented in ada driver. Tests on SSDs I have show much better results doing it this way: above 8GB/s of the linear delete on Intel SATA SSD on LSI SAS HBA (mps). Reviewed by: silence on scsi@ MFC after: 2 month Sponsored by: iXsystems, Inc.
2012-01-13 10:21:17 +00:00
while ((bp1 = bioq_takefirst(&softc->delete_run_queue))
!= NULL) {
bp1->bio_resid = bp->bio_resid;
bp1->bio_error = bp->bio_error;
if (bp->bio_flags & BIO_ERROR)
bp1->bio_flags |= BIO_ERROR;
biodone(bp1);
}
softc->delete_running = 0;
if (bp != NULL)
biodone(bp);
daschedule(periph);
} else if (bp != NULL)
biodone(bp);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
break;
}
case DA_CCB_PROBE:
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
case DA_CCB_PROBE2:
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
{
struct scsi_read_capacity_data *rdcap;
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
struct scsi_read_capacity_data_long *rcaplong;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
char announce_buf[80];
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
rdcap = NULL;
rcaplong = NULL;
Do not schedule periph for payload/TUR requests if reprobe is in progress to avoid sending extra READ CAPACITY requests by dastart(). Schedule periph again on reprobe completion, or otherwise it may stuck indefinitely long. This should fix USB explore thread hanging on device unplug, waiting for periph destruction. Reported by: hselasky
2013-01-11 16:10:11 +00:00
if (state == DA_CCB_PROBE)
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
rdcap =(struct scsi_read_capacity_data *)csio->data_ptr;
else
rcaplong = (struct scsi_read_capacity_data_long *)
csio->data_ptr;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if ((csio->ccb_h.status & CAM_STATUS_MASK) == CAM_REQ_CMP) {
struct disk_params *dp;
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
uint32_t block_size;
uint64_t maxsector;
Use READ CAPACITY(16) to get information about device physical sectors. As soon as not all devices support READ CAPACITY(16), automatically fall back to READ CAPACITY(10) if CAM_REQ_INVALID or SSD_KEY_ILLEGAL_REQUEST status returned. It also provides first bits of information about Logical Block Provisioning (aka UNMAP/TRIM) support by the device.
2011-12-23 19:12:02 +00:00
u_int lbppbe; /* LB per physical block exponent. */
u_int lalba; /* Lowest aligned LBA. */
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
Do not schedule periph for payload/TUR requests if reprobe is in progress to avoid sending extra READ CAPACITY requests by dastart(). Schedule periph again on reprobe completion, or otherwise it may stuck indefinitely long. This should fix USB explore thread hanging on device unplug, waiting for periph destruction. Reported by: hselasky
2013-01-11 16:10:11 +00:00
if (state == DA_CCB_PROBE) {
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
block_size = scsi_4btoul(rdcap->length);
maxsector = scsi_4btoul(rdcap->addr);
Use READ CAPACITY(16) to get information about device physical sectors. As soon as not all devices support READ CAPACITY(16), automatically fall back to READ CAPACITY(10) if CAM_REQ_INVALID or SSD_KEY_ILLEGAL_REQUEST status returned. It also provides first bits of information about Logical Block Provisioning (aka UNMAP/TRIM) support by the device.
2011-12-23 19:12:02 +00:00
lbppbe = 0;
lalba = 0;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
/*
* According to SBC-2, if the standard 10
* byte READ CAPACITY command returns 2^32,
* we should issue the 16 byte version of
* the command, since the device in question
* has more sectors than can be represented
* with the short version of the command.
*/
if (maxsector == 0xffffffff) {
softc->state = DA_STATE_PROBE2;
Eliminate the use of M_TEMP.
2007-05-14 21:48:53 +00:00
free(rdcap, M_SCSIDA);
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
xpt_release_ccb(done_ccb);
MFp4: - Reduce code duplication in ATA XPT and PMP driver. - Move PIO size setting from ada driver to ATA XPT. It is XPT business to negotiate transfer details. ada driver is now stateless. - Report PIO size to SIM. It is required for correct PATA SIM operation. - Tune PMP scan timings. It workarounds some problems with SiI. - If reset hapens during PMP initialization - restart it. - Introduce early-initialized periph drivers, which are used during initial scan process. Use it for xpt, probe, aprobe and pmp. It gives pmp chance to finish scan before mountroot and numerate devices in right order.
2009-10-31 10:43:38 +00:00
xpt_schedule(periph, priority);
Fix three problems in large (>2TB) device handling: - Make sure we don't release the READ CAPACITY CCB twice - If we have a device that needs a 16 byte READ CAPACITY command, make sure we call xpt_schedule() so we can get a CCB. - Don't unlock the peripheral until we're fully probed. Many thanks to Julian Elischer for providing hardware and testing this. Tested by: julian
2003-05-03 00:21:40 +00:00
return;
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
}
} else {
block_size = scsi_4btoul(rcaplong->length);
maxsector = scsi_8btou64(rcaplong->addr);
Use READ CAPACITY(16) to get information about device physical sectors. As soon as not all devices support READ CAPACITY(16), automatically fall back to READ CAPACITY(10) if CAM_REQ_INVALID or SSD_KEY_ILLEGAL_REQUEST status returned. It also provides first bits of information about Logical Block Provisioning (aka UNMAP/TRIM) support by the device.
2011-12-23 19:12:02 +00:00
lbppbe = rcaplong->prot_lbppbe & SRC16_LBPPBE;
lalba = scsi_2btoul(rcaplong->lalba_lbp);
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
}
Don't allow attachment of disks that could cause GEOM to panic.
2006-09-16 21:21:07 +00:00
/*
* Because GEOM code just will panic us if we
* give them an 'illegal' value we'll avoid that
* here.
*/
Make da driver to handle some probably broken Android devices, returning zero media and sector size instead of "Medium not present" error, until some confirmation button is tapped on device.
2010-10-24 18:53:16 +00:00
if (block_size == 0 && maxsector == 0) {
snprintf(announce_buf, sizeof(announce_buf),
"0MB (no media?)");
} else if (block_size >= MAXPHYS || block_size == 0) {
Add an xpt_print function to reduce most of the xpt_print_path/printf pairs. Convert the core code to use it. Reviewed by: scsi (various folks weighing in with good advice) MFC after: 2 weeks
2006-12-05 07:45:28 +00:00
xpt_print(periph->path,
"unsupportable block size %ju\n",
(uintmax_t) block_size);
Don't allow attachment of disks that could cause GEOM to panic.
2006-09-16 21:21:07 +00:00
announce_buf[0] = '\0';
cam_periph_invalidate(periph);
} else {
Add CAM infrastructure to allow reporting when a drive's long read capacity data changes. cam_ccb.h: Add a new advanced information type, CDAI_TYPE_RCAPLONG, for long read capacity data. cam_xpt_internal.h: Add a read capacity data pointer and length to struct cam_ed. cam_xpt.c: Free the read capacity buffer when a device goes away. While we're here, make sure we don't leak memory for other malloced fields in struct cam_ed. scsi_all.c: Update the scsi_read_capacity_16() to take a uint8_t * and a length instead of just a pointer to the parameter data structure. This will hopefully make this function somewhat immune to future changes in the parameter data. scsi_all.h: Add some extra bit definitions to struct scsi_read_capacity_data_long, and bump up the structure size to the full size specified by SBC-3. Change the prototype for scsi_read_capacity_16(). scsi_da.c: Register changes in read capacity data with the transport layer. This allows the transport layer to send out an async notification to interested parties. Update the dasetgeom() API. Use scsi_extract_sense_len() instead of scsi_extract_sense(). scsi_xpt.c: Add support for the new CDAI_TYPE_RCAPLONG advanced information type. Make sure we set the physpath pointer to NULL after freeing it. This allows blindly freeing it in the struct cam_ed destructor. sys/param.h: Bump __FreeBSD_version from 1000005 to 1000006 to make it easier for third party drivers to determine that the read capacity data async notification is available. camcontrol.c, mptutil/mpt_cam.c: Update these for the new scsi_read_capacity_16() argument structure. Sponsored by: Spectra Logic
2012-01-26 18:09:28 +00:00
/*
* We pass rcaplong into dasetgeom(),
* because it will only use it if it is
* non-NULL.
*/
Use READ CAPACITY(16) to get information about device physical sectors. As soon as not all devices support READ CAPACITY(16), automatically fall back to READ CAPACITY(10) if CAM_REQ_INVALID or SSD_KEY_ILLEGAL_REQUEST status returned. It also provides first bits of information about Logical Block Provisioning (aka UNMAP/TRIM) support by the device.
2011-12-23 19:12:02 +00:00
dasetgeom(periph, block_size, maxsector,
Add CAM infrastructure to allow reporting when a drive's long read capacity data changes. cam_ccb.h: Add a new advanced information type, CDAI_TYPE_RCAPLONG, for long read capacity data. cam_xpt_internal.h: Add a read capacity data pointer and length to struct cam_ed. cam_xpt.c: Free the read capacity buffer when a device goes away. While we're here, make sure we don't leak memory for other malloced fields in struct cam_ed. scsi_all.c: Update the scsi_read_capacity_16() to take a uint8_t * and a length instead of just a pointer to the parameter data structure. This will hopefully make this function somewhat immune to future changes in the parameter data. scsi_all.h: Add some extra bit definitions to struct scsi_read_capacity_data_long, and bump up the structure size to the full size specified by SBC-3. Change the prototype for scsi_read_capacity_16(). scsi_da.c: Register changes in read capacity data with the transport layer. This allows the transport layer to send out an async notification to interested parties. Update the dasetgeom() API. Use scsi_extract_sense_len() instead of scsi_extract_sense(). scsi_xpt.c: Add support for the new CDAI_TYPE_RCAPLONG advanced information type. Make sure we set the physpath pointer to NULL after freeing it. This allows blindly freeing it in the struct cam_ed destructor. sys/param.h: Bump __FreeBSD_version from 1000005 to 1000006 to make it easier for third party drivers to determine that the read capacity data async notification is available. camcontrol.c, mptutil/mpt_cam.c: Update these for the new scsi_read_capacity_16() argument structure. Sponsored by: Spectra Logic
2012-01-26 18:09:28 +00:00
rcaplong, sizeof(*rcaplong));
if ((lalba & SRC16_LBPME_A)
&& softc->delete_method == DA_DELETE_NONE)
Updates delete_method sysctl changes to always maintain disk d_flags DISKFLAG_CANDELETE. While this change makes this layer consistent other layers such as UFS and ZFS BIO_DELETE support may not notice any change made manually via these device sysctls until the device is reopened via a mount. Also corrected var order in dadeletemethodsysctl PR: kern/169801 Reviewed by: pjd (mentor) Approved by: mav MFC after: 2 weeks
2013-01-10 11:57:46 +00:00
dadeletemethodset(softc, DA_DELETE_UNMAP);
Don't allow attachment of disks that could cause GEOM to panic.
2006-09-16 21:21:07 +00:00
dp = &softc->params;
snprintf(announce_buf, sizeof(announce_buf),
"%juMB (%ju %u byte sectors: %dH %dS/T "
"%dC)", (uintmax_t)
(((uintmax_t)dp->secsize *
dp->sectors) / (1024*1024)),
(uintmax_t)dp->sectors,
dp->secsize, dp->heads,
dp->secs_per_track, dp->cylinders);
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
} else {
int error;
Move initialization of announce_buf up to make sure it is initialized in all cases.
1999-08-15 23:34:40 +00:00
announce_buf[0] = '\0';
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
* Retry any UNIT ATTENTION type errors. They
* are expected at boot.
*/
Rewrite of the CAM error recovery code. Some of the major changes include: - The SCSI error handling portion of cam_periph_error() has been broken out into a number of subfunctions to better modularize the code that handles the hierarchy of SCSI errors. As a result, the code is now much easier to read. - String handling and error printing has been significantly revamped. We now use sbufs to do string formatting instead of using printfs (for the kernel) and snprintf/strncat (for userland) as before. There is a new catchall error printing routine, cam_error_print() and its string-based counterpart, cam_error_string() that allow the kernel and userland applications to pass in a CCB and have errors printed out properly, whether or not they're SCSI errors. Among other things, this helped eliminate a fair amount of duplicate code in camcontrol. We now print out more information than before, including the CAM status and SCSI status and the error recovery action taken to remedy the problem. - sbufs are now available in userland, via libsbuf. This change was necessary since most of the error printing code is shared between libcam and the kernel. - A new transfer settings interface is included in this checkin. This code is #ifdef'ed out, and is primarily intended to aid discussion with HBA driver authors on the final form the interface should take. There is example code in the ahc(4) driver that implements the HBA driver side of the new interface. The new transfer settings code won't be enabled until we're ready to switch all HBA drivers over to the new interface. src/Makefile.inc1, lib/Makefile: Add libsbuf. It must be built before libcam, since libcam uses sbuf routines. libcam/Makefile: libcam now depends on libsbuf. libsbuf/Makefile: Add a makefile for libsbuf. This pulls in the sbuf sources from sys/kern. bsd.libnames.mk: Add LIBSBUF. camcontrol/Makefile: Add -lsbuf. Since camcontrol is statically linked, we can't depend on the dynamic linker to pull in libsbuf. camcontrol.c: Use cam_error_print() instead of checking for CAM_SCSI_STATUS_ERROR on every failed CCB. sbuf.9: Change the prototypes for sbuf_cat() and sbuf_cpy() so that the source string is now a const char *. This is more in line wth the standard system string functions, and helps eliminate warnings when dealing with a const source buffer. Fix a typo. cam.c: Add description strings for the various CAM error status values, as well as routines to look up those strings. Add new cam_error_string() and cam_error_print() routines for userland and the kernel. cam.h: Add a new CAM flag, CAM_RETRY_SELTO. Add enumerated types for the various options available with cam_error_print() and cam_error_string(). cam_ccb.h: Add new transfer negotiation structures/types. Change inq_len in the ccb_getdev structure to be "reserved". This field has never been filled in, and will be removed when we next bump the CAM version. cam_debug.h: Fix typo. cam_periph.c: Modularize cam_periph_error(). The SCSI error handling part of cam_periph_error() is now in camperiphscsistatuserror() and camperiphscsisenseerror(). In cam_periph_lock(), increase the reference count on the periph while we wait for our lock attempt to succeed so that the periph won't go away while we're sleeping. cam_xpt.c: Add new transfer negotiation code. (ifdefed out) Add a new function, xpt_path_string(). This is a string/sbuf analog to xpt_print_path(). scsi_all.c: Revamp string handing and error printing code. We now use sbufs for much of the string formatting code. More of that code is shared between userland the kernel. scsi_all.h: Get rid of SS_TURSTART, it wasn't terribly useful in the first place. Add a new error action, SS_REQSENSE. (Send a request sense and then retry the command.) This is useful when the controller hasn't performed autosense for some reason. Change the default actions around a bit. scsi_cd.c, scsi_da.c, scsi_pt.c, scsi_ses.c: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Selection timeouts shouldn't be covered by a sense flag. scsi_pass.[ch]: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Get rid of the last vestiges of a read/write interface. libkern/bsearch.c, sys/libkern.h, conf/files: Add bsearch.c, which is needed for some of the new table lookup routines. aic7xxx_freebsd.c: Define AHC_NEW_TRAN_SETTINGS if CAM_NEW_TRAN_CODE is defined. sbuf.h, subr_sbuf.c: Add the appropriate #ifdefs so sbufs can compile and run in userland. Change sbuf_printf() to use vsnprintf() instead of kvprintf(), which is only available in the kernel. Change the source string for sbuf_cpy() and sbuf_cat() to be a const char *. Add __BEGIN_DECLS and __END_DECLS around function prototypes since they're now exported to userland. kdump/mkioctls: Include stdio.h before cam.h since cam.h now includes a function with a FILE * argument. Submitted by: gibbs (mostly) Reviewed by: jdp, marcel (libsbuf makefile changes) Reviewed by: des (sbuf changes) Reviewed by: ken
2001-03-27 05:45:52 +00:00
error = daerror(done_ccb, CAM_RETRY_SELTO,
SF_RETRY_UA|SF_NO_PRINT);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if (error == ERESTART) {
/*
* A retry was scheuled, so
* just return.
*/
return;
} else if (error != 0) {
int asc, ascq;
int sense_key, error_code;
int have_sense;
cam_status status;
struct ccb_getdev cgd;
/* Don't wedge this device's queue */
status = done_ccb->ccb_h.status;
Rewrite of the CAM error recovery code. Some of the major changes include: - The SCSI error handling portion of cam_periph_error() has been broken out into a number of subfunctions to better modularize the code that handles the hierarchy of SCSI errors. As a result, the code is now much easier to read. - String handling and error printing has been significantly revamped. We now use sbufs to do string formatting instead of using printfs (for the kernel) and snprintf/strncat (for userland) as before. There is a new catchall error printing routine, cam_error_print() and its string-based counterpart, cam_error_string() that allow the kernel and userland applications to pass in a CCB and have errors printed out properly, whether or not they're SCSI errors. Among other things, this helped eliminate a fair amount of duplicate code in camcontrol. We now print out more information than before, including the CAM status and SCSI status and the error recovery action taken to remedy the problem. - sbufs are now available in userland, via libsbuf. This change was necessary since most of the error printing code is shared between libcam and the kernel. - A new transfer settings interface is included in this checkin. This code is #ifdef'ed out, and is primarily intended to aid discussion with HBA driver authors on the final form the interface should take. There is example code in the ahc(4) driver that implements the HBA driver side of the new interface. The new transfer settings code won't be enabled until we're ready to switch all HBA drivers over to the new interface. src/Makefile.inc1, lib/Makefile: Add libsbuf. It must be built before libcam, since libcam uses sbuf routines. libcam/Makefile: libcam now depends on libsbuf. libsbuf/Makefile: Add a makefile for libsbuf. This pulls in the sbuf sources from sys/kern. bsd.libnames.mk: Add LIBSBUF. camcontrol/Makefile: Add -lsbuf. Since camcontrol is statically linked, we can't depend on the dynamic linker to pull in libsbuf. camcontrol.c: Use cam_error_print() instead of checking for CAM_SCSI_STATUS_ERROR on every failed CCB. sbuf.9: Change the prototypes for sbuf_cat() and sbuf_cpy() so that the source string is now a const char *. This is more in line wth the standard system string functions, and helps eliminate warnings when dealing with a const source buffer. Fix a typo. cam.c: Add description strings for the various CAM error status values, as well as routines to look up those strings. Add new cam_error_string() and cam_error_print() routines for userland and the kernel. cam.h: Add a new CAM flag, CAM_RETRY_SELTO. Add enumerated types for the various options available with cam_error_print() and cam_error_string(). cam_ccb.h: Add new transfer negotiation structures/types. Change inq_len in the ccb_getdev structure to be "reserved". This field has never been filled in, and will be removed when we next bump the CAM version. cam_debug.h: Fix typo. cam_periph.c: Modularize cam_periph_error(). The SCSI error handling part of cam_periph_error() is now in camperiphscsistatuserror() and camperiphscsisenseerror(). In cam_periph_lock(), increase the reference count on the periph while we wait for our lock attempt to succeed so that the periph won't go away while we're sleeping. cam_xpt.c: Add new transfer negotiation code. (ifdefed out) Add a new function, xpt_path_string(). This is a string/sbuf analog to xpt_print_path(). scsi_all.c: Revamp string handing and error printing code. We now use sbufs for much of the string formatting code. More of that code is shared between userland the kernel. scsi_all.h: Get rid of SS_TURSTART, it wasn't terribly useful in the first place. Add a new error action, SS_REQSENSE. (Send a request sense and then retry the command.) This is useful when the controller hasn't performed autosense for some reason. Change the default actions around a bit. scsi_cd.c, scsi_da.c, scsi_pt.c, scsi_ses.c: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Selection timeouts shouldn't be covered by a sense flag. scsi_pass.[ch]: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Get rid of the last vestiges of a read/write interface. libkern/bsearch.c, sys/libkern.h, conf/files: Add bsearch.c, which is needed for some of the new table lookup routines. aic7xxx_freebsd.c: Define AHC_NEW_TRAN_SETTINGS if CAM_NEW_TRAN_CODE is defined. sbuf.h, subr_sbuf.c: Add the appropriate #ifdefs so sbufs can compile and run in userland. Change sbuf_printf() to use vsnprintf() instead of kvprintf(), which is only available in the kernel. Change the source string for sbuf_cpy() and sbuf_cat() to be a const char *. Add __BEGIN_DECLS and __END_DECLS around function prototypes since they're now exported to userland. kdump/mkioctls: Include stdio.h before cam.h since cam.h now includes a function with a FILE * argument. Submitted by: gibbs (mostly) Reviewed by: jdp, marcel (libsbuf makefile changes) Reviewed by: des (sbuf changes) Reviewed by: ken
2001-03-27 05:45:52 +00:00
if ((status & CAM_DEV_QFRZN) != 0)
cam_release_devq(done_ccb->ccb_h.path,
/*relsim_flags*/0,
/*reduction*/0,
/*timeout*/0,
/*getcount_only*/0);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
xpt_setup_ccb(&cgd.ccb_h,
done_ccb->ccb_h.path,
Replace most of priority numbers with defines. No logical changes.
2009-10-23 08:27:55 +00:00
CAM_PRIORITY_NORMAL);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
cgd.ccb_h.func_code = XPT_GDEV_TYPE;
xpt_action((union ccb *)&cgd);
Add scsi_extract_sense_ccb() -- wrapper around scsi_extract_sense_len(). It allows to remove number of duplicate checks from several places.
2012-06-23 12:32:53 +00:00
if (scsi_extract_sense_ccb(done_ccb,
&error_code, &sense_key, &asc, &ascq))
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
have_sense = TRUE;
Add scsi_extract_sense_ccb() -- wrapper around scsi_extract_sense_len(). It allows to remove number of duplicate checks from several places.
2012-06-23 12:32:53 +00:00
else
have_sense = FALSE;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Use READ CAPACITY(16) to get information about device physical sectors. As soon as not all devices support READ CAPACITY(16), automatically fall back to READ CAPACITY(10) if CAM_REQ_INVALID or SSD_KEY_ILLEGAL_REQUEST status returned. It also provides first bits of information about Logical Block Provisioning (aka UNMAP/TRIM) support by the device.
2011-12-23 19:12:02 +00:00
/*
* If we tried READ CAPACITY(16) and failed,
* fallback to READ CAPACITY(10).
*/
Do not schedule periph for payload/TUR requests if reprobe is in progress to avoid sending extra READ CAPACITY requests by dastart(). Schedule periph again on reprobe completion, or otherwise it may stuck indefinitely long. This should fix USB explore thread hanging on device unplug, waiting for periph destruction. Reported by: hselasky
2013-01-11 16:10:11 +00:00
if ((state == DA_CCB_PROBE2) &&
Use READ CAPACITY(16) to get information about device physical sectors. As soon as not all devices support READ CAPACITY(16), automatically fall back to READ CAPACITY(10) if CAM_REQ_INVALID or SSD_KEY_ILLEGAL_REQUEST status returned. It also provides first bits of information about Logical Block Provisioning (aka UNMAP/TRIM) support by the device.
2011-12-23 19:12:02 +00:00
(softc->flags & DA_FLAG_CAN_RC16) &&
(((csio->ccb_h.status & CAM_STATUS_MASK) ==
CAM_REQ_INVALID) ||
((have_sense) &&
(error_code == SSD_CURRENT_ERROR) &&
(sense_key == SSD_KEY_ILLEGAL_REQUEST)))) {
softc->flags &= ~DA_FLAG_CAN_RC16;
softc->state = DA_STATE_PROBE;
free(rdcap, M_SCSIDA);
xpt_release_ccb(done_ccb);
xpt_schedule(periph, priority);
return;
} else
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
Add a number of interrelated CAM feature enhancements and bug fixes. NOTE: These changes will require recompilation of any userland applications, like cdrecord, xmcd, etc., that use the CAM passthrough interface. A make world is recommended. camcontrol.[c8]: - We now support two new commands, "tags" and "negotiate". - The tags commands allows users to view the number of tagged openings for a device as well as a number of other related parameters, and it allows users to set tagged openings for a device. - The negotiate command allows users to enable and disable disconnection and tagged queueing, set sync rates, offsets and bus width. Note that not all of those features are available for all controllers. Only the adv, ahc, and ncr drivers fully support all of the features at this point. Some cards do not allow the setting of sync rates, offsets and the like, and some of the drivers don't have any facilities to do so. Some drivers, like the adw driver, only support enabling or disabling sync negotiation, but do not support setting sync rates. - new description in the camcontrol man page of how to format a disk - cleanup of the camcontrol inquiry command - add support in the 'devlist' command for skipping unconfigured devices if -v was not specified on the command line. - make use of the new base_transfer_speed in the path inquiry CCB. - fix CCB bzero cases cam_xpt.c, cam_sim.[ch], cam_ccb.h: - new flags on many CCB function codes to designate whether they're non-immediate, use a user-supplied CCB, and can only be passed from userland programs via the xpt device. Use these flags in the transport layer and pass driver to categorize CCBs. - new flag in the transport layer device matching code for device nodes that indicates whether a device is unconfigured - bump the CAM version from 0x10 to 0x11 - Change the CAM ioctls to use the version as their group code, so we can force users to recompile code even when the CCB size doesn't change. - add + fill in a new value in the path inquiry CCB, base_transfer_speed. Remove a corresponding field from the cam_sim structure, and add code to every SIM to set this field to the proper value. - Fix the set transfer settings code in the transport layer. scsi_cd.c: - make some variables volatile instead of just casting them in various places - fix a race condition in the changer code - attach unless we get a "logical unit not supported" error. This should fix all of the cases where people have devices that return weird errors when they don't have media in the drive. scsi_da.c: - attach unless we get a "logical unit not supported" error scsi_pass.c: - for immediate CCBs, just malloc a CCB to send the user request in. This gets rid of the 'held' count problem in camcontrol tags. scsi_pass.h: - change the CAM ioctls to use the CAM version as their group code. adv driver: - Allow changing the sync rate and offset separately. adw driver - Allow changing the sync rate and offset separately. aha driver: - Don't return CAM_REQ_CMP for SET_TRAN_SETTINGS CCBs. ahc driver: - Allow setting offset and sync rate separately bt driver: - Don't return CAM_REQ_CMP for SET_TRAN_SETTINGS CCBs. NCR driver: - Fix the ultra/ultra 2 negotiation bug - allow setting both the sync rate and offset separately Other HBA drivers: - Put code in to set the base_transfer_speed field for XPT_GET_TRAN_SETTINGS CCBs. Reviewed by: gibbs, mjacob (isp), imp (aha)
1999-05-06 20:16:39 +00:00
* Attach to anything that claims to be a
* direct access or optical disk device,
* as long as it doesn't return a "Logical
* unit not supported" (0x25) error.
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
*/
Add a number of interrelated CAM feature enhancements and bug fixes. NOTE: These changes will require recompilation of any userland applications, like cdrecord, xmcd, etc., that use the CAM passthrough interface. A make world is recommended. camcontrol.[c8]: - We now support two new commands, "tags" and "negotiate". - The tags commands allows users to view the number of tagged openings for a device as well as a number of other related parameters, and it allows users to set tagged openings for a device. - The negotiate command allows users to enable and disable disconnection and tagged queueing, set sync rates, offsets and bus width. Note that not all of those features are available for all controllers. Only the adv, ahc, and ncr drivers fully support all of the features at this point. Some cards do not allow the setting of sync rates, offsets and the like, and some of the drivers don't have any facilities to do so. Some drivers, like the adw driver, only support enabling or disabling sync negotiation, but do not support setting sync rates. - new description in the camcontrol man page of how to format a disk - cleanup of the camcontrol inquiry command - add support in the 'devlist' command for skipping unconfigured devices if -v was not specified on the command line. - make use of the new base_transfer_speed in the path inquiry CCB. - fix CCB bzero cases cam_xpt.c, cam_sim.[ch], cam_ccb.h: - new flags on many CCB function codes to designate whether they're non-immediate, use a user-supplied CCB, and can only be passed from userland programs via the xpt device. Use these flags in the transport layer and pass driver to categorize CCBs. - new flag in the transport layer device matching code for device nodes that indicates whether a device is unconfigured - bump the CAM version from 0x10 to 0x11 - Change the CAM ioctls to use the version as their group code, so we can force users to recompile code even when the CCB size doesn't change. - add + fill in a new value in the path inquiry CCB, base_transfer_speed. Remove a corresponding field from the cam_sim structure, and add code to every SIM to set this field to the proper value. - Fix the set transfer settings code in the transport layer. scsi_cd.c: - make some variables volatile instead of just casting them in various places - fix a race condition in the changer code - attach unless we get a "logical unit not supported" error. This should fix all of the cases where people have devices that return weird errors when they don't have media in the drive. scsi_da.c: - attach unless we get a "logical unit not supported" error scsi_pass.c: - for immediate CCBs, just malloc a CCB to send the user request in. This gets rid of the 'held' count problem in camcontrol tags. scsi_pass.h: - change the CAM ioctls to use the CAM version as their group code. adv driver: - Allow changing the sync rate and offset separately. adw driver - Allow changing the sync rate and offset separately. aha driver: - Don't return CAM_REQ_CMP for SET_TRAN_SETTINGS CCBs. ahc driver: - Allow setting offset and sync rate separately bt driver: - Don't return CAM_REQ_CMP for SET_TRAN_SETTINGS CCBs. NCR driver: - Fix the ultra/ultra 2 negotiation bug - allow setting both the sync rate and offset separately Other HBA drivers: - Put code in to set the base_transfer_speed field for XPT_GET_TRAN_SETTINGS CCBs. Reviewed by: gibbs, mjacob (isp), imp (aha)
1999-05-06 20:16:39 +00:00
if ((have_sense) && (asc != 0x25)
Rewrite of the CAM error recovery code. Some of the major changes include: - The SCSI error handling portion of cam_periph_error() has been broken out into a number of subfunctions to better modularize the code that handles the hierarchy of SCSI errors. As a result, the code is now much easier to read. - String handling and error printing has been significantly revamped. We now use sbufs to do string formatting instead of using printfs (for the kernel) and snprintf/strncat (for userland) as before. There is a new catchall error printing routine, cam_error_print() and its string-based counterpart, cam_error_string() that allow the kernel and userland applications to pass in a CCB and have errors printed out properly, whether or not they're SCSI errors. Among other things, this helped eliminate a fair amount of duplicate code in camcontrol. We now print out more information than before, including the CAM status and SCSI status and the error recovery action taken to remedy the problem. - sbufs are now available in userland, via libsbuf. This change was necessary since most of the error printing code is shared between libcam and the kernel. - A new transfer settings interface is included in this checkin. This code is #ifdef'ed out, and is primarily intended to aid discussion with HBA driver authors on the final form the interface should take. There is example code in the ahc(4) driver that implements the HBA driver side of the new interface. The new transfer settings code won't be enabled until we're ready to switch all HBA drivers over to the new interface. src/Makefile.inc1, lib/Makefile: Add libsbuf. It must be built before libcam, since libcam uses sbuf routines. libcam/Makefile: libcam now depends on libsbuf. libsbuf/Makefile: Add a makefile for libsbuf. This pulls in the sbuf sources from sys/kern. bsd.libnames.mk: Add LIBSBUF. camcontrol/Makefile: Add -lsbuf. Since camcontrol is statically linked, we can't depend on the dynamic linker to pull in libsbuf. camcontrol.c: Use cam_error_print() instead of checking for CAM_SCSI_STATUS_ERROR on every failed CCB. sbuf.9: Change the prototypes for sbuf_cat() and sbuf_cpy() so that the source string is now a const char *. This is more in line wth the standard system string functions, and helps eliminate warnings when dealing with a const source buffer. Fix a typo. cam.c: Add description strings for the various CAM error status values, as well as routines to look up those strings. Add new cam_error_string() and cam_error_print() routines for userland and the kernel. cam.h: Add a new CAM flag, CAM_RETRY_SELTO. Add enumerated types for the various options available with cam_error_print() and cam_error_string(). cam_ccb.h: Add new transfer negotiation structures/types. Change inq_len in the ccb_getdev structure to be "reserved". This field has never been filled in, and will be removed when we next bump the CAM version. cam_debug.h: Fix typo. cam_periph.c: Modularize cam_periph_error(). The SCSI error handling part of cam_periph_error() is now in camperiphscsistatuserror() and camperiphscsisenseerror(). In cam_periph_lock(), increase the reference count on the periph while we wait for our lock attempt to succeed so that the periph won't go away while we're sleeping. cam_xpt.c: Add new transfer negotiation code. (ifdefed out) Add a new function, xpt_path_string(). This is a string/sbuf analog to xpt_print_path(). scsi_all.c: Revamp string handing and error printing code. We now use sbufs for much of the string formatting code. More of that code is shared between userland the kernel. scsi_all.h: Get rid of SS_TURSTART, it wasn't terribly useful in the first place. Add a new error action, SS_REQSENSE. (Send a request sense and then retry the command.) This is useful when the controller hasn't performed autosense for some reason. Change the default actions around a bit. scsi_cd.c, scsi_da.c, scsi_pt.c, scsi_ses.c: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Selection timeouts shouldn't be covered by a sense flag. scsi_pass.[ch]: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Get rid of the last vestiges of a read/write interface. libkern/bsearch.c, sys/libkern.h, conf/files: Add bsearch.c, which is needed for some of the new table lookup routines. aic7xxx_freebsd.c: Define AHC_NEW_TRAN_SETTINGS if CAM_NEW_TRAN_CODE is defined. sbuf.h, subr_sbuf.c: Add the appropriate #ifdefs so sbufs can compile and run in userland. Change sbuf_printf() to use vsnprintf() instead of kvprintf(), which is only available in the kernel. Change the source string for sbuf_cpy() and sbuf_cat() to be a const char *. Add __BEGIN_DECLS and __END_DECLS around function prototypes since they're now exported to userland. kdump/mkioctls: Include stdio.h before cam.h since cam.h now includes a function with a FILE * argument. Submitted by: gibbs (mostly) Reviewed by: jdp, marcel (libsbuf makefile changes) Reviewed by: des (sbuf changes) Reviewed by: ken
2001-03-27 05:45:52 +00:00
&& (error_code == SSD_CURRENT_ERROR)) {
const char *sense_key_desc;
const char *asc_desc;
scsi_sense_desc(sense_key, asc, ascq,
&cgd.inq_data,
&sense_key_desc,
&asc_desc);
Examine all occurrences of sprintf(), strcat(), and str[n]cpy() for possible buffer overflow problems. Replaced most sprintf()'s with snprintf(); for others cases, added terminating NUL bytes where appropriate, replaced constants like "16" with sizeof(), etc. These changes include several bug fixes, but most changes are for maintainability's sake. Any instance where it wasn't "immediately obvious" that a buffer overflow could not occur was made safer. Reviewed by: Bruce Evans <bde@zeta.org.au> Reviewed by: Matthew Dillon <dillon@apollo.backplane.com> Reviewed by: Mike Spengler <mks@networkcs.com>
1998-12-04 22:54:57 +00:00
snprintf(announce_buf,
sizeof(announce_buf),
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
"Attempt to query device "
"size failed: %s, %s",
Rewrite of the CAM error recovery code. Some of the major changes include: - The SCSI error handling portion of cam_periph_error() has been broken out into a number of subfunctions to better modularize the code that handles the hierarchy of SCSI errors. As a result, the code is now much easier to read. - String handling and error printing has been significantly revamped. We now use sbufs to do string formatting instead of using printfs (for the kernel) and snprintf/strncat (for userland) as before. There is a new catchall error printing routine, cam_error_print() and its string-based counterpart, cam_error_string() that allow the kernel and userland applications to pass in a CCB and have errors printed out properly, whether or not they're SCSI errors. Among other things, this helped eliminate a fair amount of duplicate code in camcontrol. We now print out more information than before, including the CAM status and SCSI status and the error recovery action taken to remedy the problem. - sbufs are now available in userland, via libsbuf. This change was necessary since most of the error printing code is shared between libcam and the kernel. - A new transfer settings interface is included in this checkin. This code is #ifdef'ed out, and is primarily intended to aid discussion with HBA driver authors on the final form the interface should take. There is example code in the ahc(4) driver that implements the HBA driver side of the new interface. The new transfer settings code won't be enabled until we're ready to switch all HBA drivers over to the new interface. src/Makefile.inc1, lib/Makefile: Add libsbuf. It must be built before libcam, since libcam uses sbuf routines. libcam/Makefile: libcam now depends on libsbuf. libsbuf/Makefile: Add a makefile for libsbuf. This pulls in the sbuf sources from sys/kern. bsd.libnames.mk: Add LIBSBUF. camcontrol/Makefile: Add -lsbuf. Since camcontrol is statically linked, we can't depend on the dynamic linker to pull in libsbuf. camcontrol.c: Use cam_error_print() instead of checking for CAM_SCSI_STATUS_ERROR on every failed CCB. sbuf.9: Change the prototypes for sbuf_cat() and sbuf_cpy() so that the source string is now a const char *. This is more in line wth the standard system string functions, and helps eliminate warnings when dealing with a const source buffer. Fix a typo. cam.c: Add description strings for the various CAM error status values, as well as routines to look up those strings. Add new cam_error_string() and cam_error_print() routines for userland and the kernel. cam.h: Add a new CAM flag, CAM_RETRY_SELTO. Add enumerated types for the various options available with cam_error_print() and cam_error_string(). cam_ccb.h: Add new transfer negotiation structures/types. Change inq_len in the ccb_getdev structure to be "reserved". This field has never been filled in, and will be removed when we next bump the CAM version. cam_debug.h: Fix typo. cam_periph.c: Modularize cam_periph_error(). The SCSI error handling part of cam_periph_error() is now in camperiphscsistatuserror() and camperiphscsisenseerror(). In cam_periph_lock(), increase the reference count on the periph while we wait for our lock attempt to succeed so that the periph won't go away while we're sleeping. cam_xpt.c: Add new transfer negotiation code. (ifdefed out) Add a new function, xpt_path_string(). This is a string/sbuf analog to xpt_print_path(). scsi_all.c: Revamp string handing and error printing code. We now use sbufs for much of the string formatting code. More of that code is shared between userland the kernel. scsi_all.h: Get rid of SS_TURSTART, it wasn't terribly useful in the first place. Add a new error action, SS_REQSENSE. (Send a request sense and then retry the command.) This is useful when the controller hasn't performed autosense for some reason. Change the default actions around a bit. scsi_cd.c, scsi_da.c, scsi_pt.c, scsi_ses.c: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Selection timeouts shouldn't be covered by a sense flag. scsi_pass.[ch]: SF_RETRY_SELTO -> CAM_RETRY_SELTO. Get rid of the last vestiges of a read/write interface. libkern/bsearch.c, sys/libkern.h, conf/files: Add bsearch.c, which is needed for some of the new table lookup routines. aic7xxx_freebsd.c: Define AHC_NEW_TRAN_SETTINGS if CAM_NEW_TRAN_CODE is defined. sbuf.h, subr_sbuf.c: Add the appropriate #ifdefs so sbufs can compile and run in userland. Change sbuf_printf() to use vsnprintf() instead of kvprintf(), which is only available in the kernel. Change the source string for sbuf_cpy() and sbuf_cat() to be a const char *. Add __BEGIN_DECLS and __END_DECLS around function prototypes since they're now exported to userland. kdump/mkioctls: Include stdio.h before cam.h since cam.h now includes a function with a FILE * argument. Submitted by: gibbs (mostly) Reviewed by: jdp, marcel (libsbuf makefile changes) Reviewed by: des (sbuf changes) Reviewed by: ken
2001-03-27 05:45:52 +00:00
sense_key_desc,
asc_desc);
} else {
Some fixes for the CD and DA drivers from bde. (and some of my own as well) Among them: [ cd driver ] 1. Old labeling code was still there. 2. Error handling for dsopen() was broken (no test for the `error' returned by dsopen(); bogus test of an `error' that is known to be 0). 3. cdopen() closed the physical device after certain errors although there may still be open partitions on it. 4. cdclose() closed the physical device although there may still be open partitions on it. 5. Some printf format fixes was incomplete or missing. 6. cdioctl() truncated unit numbers mod 256. 7. cdioctl() was missing locking. [ da driver ] 1. daclose() closed the physical device although there may still be open partitions on it. This was fixed many years ago in sd.c rev.1.57. 2. A minor optimization (the dk_slices != NULL test) in sdopen() became uglier in daopen(). It is not worth doing. da only regressed compared with od and my version of sd, since I never committed the change to sd. daopen() should probably do less if some partition is already open. This is not addressed by the diffs. [ ... ] 5. "opt_hw_wdog.h" was not included, so the HW_WDOG code was unreachable. - Added a getdev CCB call in the cdopen() and daopen() calls so that the vendor name and device name are available for the disklabel. (suggested by bde) - Removed vestigal devfs support in both drivers, since we can't properly work with devfs yet. (ask bde for details on this) - Cleaned up the probe code in both drivers in the failure cases. There were a number of things wrong here. The peripheral driver instances weren't getting properly cleaned up. Sometimes the wrong probe message would get printed out (with the failure message appended), so it wasn't very clear that we failed to attach. SCSI sense information was printed, even when the error in question wasn't a SCSI error. I put similar fixes into the changer driver in revision 1.2 of scsi_ch.c. Reviewed by: gibbs Submitted by: bde (partially)
1998-10-07 02:57:57 +00:00
if (have_sense)
scsi_sense_print(
&done_ccb->csio);
else {
Add an xpt_print function to reduce most of the xpt_print_path/printf pairs. Convert the core code to use it. Reviewed by: scsi (various folks weighing in with good advice) MFC after: 2 weeks
2006-12-05 07:45:28 +00:00
xpt_print(periph->path,
"got CAM status %#x\n",
done_ccb->ccb_h.status);
Some fixes for the CD and DA drivers from bde. (and some of my own as well) Among them: [ cd driver ] 1. Old labeling code was still there. 2. Error handling for dsopen() was broken (no test for the `error' returned by dsopen(); bogus test of an `error' that is known to be 0). 3. cdopen() closed the physical device after certain errors although there may still be open partitions on it. 4. cdclose() closed the physical device although there may still be open partitions on it. 5. Some printf format fixes was incomplete or missing. 6. cdioctl() truncated unit numbers mod 256. 7. cdioctl() was missing locking. [ da driver ] 1. daclose() closed the physical device although there may still be open partitions on it. This was fixed many years ago in sd.c rev.1.57. 2. A minor optimization (the dk_slices != NULL test) in sdopen() became uglier in daopen(). It is not worth doing. da only regressed compared with od and my version of sd, since I never committed the change to sd. daopen() should probably do less if some partition is already open. This is not addressed by the diffs. [ ... ] 5. "opt_hw_wdog.h" was not included, so the HW_WDOG code was unreachable. - Added a getdev CCB call in the cdopen() and daopen() calls so that the vendor name and device name are available for the disklabel. (suggested by bde) - Removed vestigal devfs support in both drivers, since we can't properly work with devfs yet. (ask bde for details on this) - Cleaned up the probe code in both drivers in the failure cases. There were a number of things wrong here. The peripheral driver instances weren't getting properly cleaned up. Sometimes the wrong probe message would get printed out (with the failure message appended), so it wasn't very clear that we failed to attach. SCSI sense information was printed, even when the error in question wasn't a SCSI error. I put similar fixes into the changer driver in revision 1.2 of scsi_ch.c. Reviewed by: gibbs Submitted by: bde (partially)
1998-10-07 02:57:57 +00:00
}
A better fix to avoid race conditions between failed probes and peripheral removal. Obtained from:gibbs@freebsd.org
1999-01-07 20:19:09 +00:00
Add an xpt_print function to reduce most of the xpt_print_path/printf pairs. Convert the core code to use it. Reviewed by: scsi (various folks weighing in with good advice) MFC after: 2 weeks
2006-12-05 07:45:28 +00:00
xpt_print(periph->path, "fatal error, "
"failed to attach to device\n");
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
* Free up resources.
*/
cam_periph_invalidate(periph);
}
}
}
Eliminate the use of M_TEMP.
2007-05-14 21:48:53 +00:00
free(csio->data_ptr, M_SCSIDA);
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
if (announce_buf[0] != '\0' && ((softc->flags & DA_FLAG_PROBED) == 0)) {
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
/*
* Create our sysctl variables, now that we know
* we have successfully attached.
*/
Fix a race condition in CAM peripheral free handling, locking in the CAM XPT bus traversal code, and a number of other periph level issues. cam_periph.h, cam_periph.c: Modify cam_periph_acquire() to test the CAM_PERIPH_INVALID flag prior to allowing a reference count to be gained on a peripheral. Callers of this function will receive CAM_REQ_CMP_ERR status in the situation of attempting to reference an invalidated periph. This guarantees that a peripheral scheduled for a deferred free will not be accessed during its wait for destruction. Panic during attempts to drop a reference count on a peripheral that already has a zero reference count. In cam_periph_list(), use a local sbuf with SBUF_FIXEDLEN set so that mallocs do not occur while the xpt topology lock is held, regardless of the allocation policy of the passed in sbuf. Add a new routine, cam_periph_release_locked_buses(), that can be called when the caller already holds the CAM topology lock. Add some extra debugging for duplicate peripheral allocations in cam_periph_alloc(). Treat CAM_DEV_NOT_THERE much the same as a selection timeout (AC_LOST_DEVICE is emitted), but forgo retries. cam_xpt.c: Revamp the way the EDT traversal code does locking and reference counting. This was broken, since it assumed that the EDT would not change during traversal, but that assumption is no longer valid. So, to prevent devices from going away while we traverse the EDT, make sure we properly lock everything and hold references on devices that we are using. The two peripheral driver traversal routines should be examined. xptpdperiphtraverse() holds the topology lock for the entire time it runs. xptperiphtraverse() is now locked properly, but only holds the topology lock while it is traversing the list, and not while the traversal function is running. The bus locking code in xptbustraverse() should also be revisited at a later time, since it is complex and should probably be simplified. scsi_da.c: Pay attention to the return value from cam_periph_acquire(). Return 0 always from daclose() even if the disk is now gone. Add some rudimentary error injection support. scsi_sg.c: Fix reference counting in the sg(4) driver. The sg driver was calling cam_periph_release() on close, but never called cam_periph_acquire() (which increments the reference count) on open. The periph code correctly complained that the sg(4) driver was trying to decrement the refcount when it was already 0. Sponsored by: Spectra Logic MFC after: 2 weeks
2012-01-12 00:41:48 +00:00
/* increase the refcount */
if (cam_periph_acquire(periph) == CAM_REQ_CMP) {
taskqueue_enqueue(taskqueue_thread,
&softc->sysctl_task);
xpt_announce_periph(periph, announce_buf);
} else {
xpt_print(periph->path, "fatal error, "
"could not acquire reference count\n");
}
Move dynamic sysctl(8) variable creation for the cd(4) and da(4) drivers out of cdregister() and daregister(), which are run from interrupt context. The sysctl code does blocking mallocs (M_WAITOK), which causes problems if malloc(9) actually needs to sleep. The eventual fix for this issue will involve moving the CAM probe process inside a kernel thread. For now, though, I have fixed the issue by moving dynamic sysctl variable creation for these two drivers to a task queue running in a kernel thread. The existing task queues (taskqueue_swi and taskqueue_swi_giant) run in software interrupt handlers, which wouldn't fix the problem at hand. So I have created a new task queue, taskqueue_thread, that runs inside a kernel thread. (It also runs outside of Giant -- clients must explicitly acquire and release Giant in their taskqueue functions.) scsi_cd.c: Remove sysctl variable creation code from cdregister(), and move it to a new function, cdsysctlinit(). Queue cdsysctlinit() to the taskqueue_thread taskqueue once we have fully registered the cd(4) driver instance. scsi_da.c: Remove sysctl variable creation code from daregister(), and move it to move it to a new function, dasysctlinit(). Queue dasysctlinit() to the taskqueue_thread taskqueue once we have fully registered the da(4) instance. taskqueue.h: Declare the new taskqueue_thread taskqueue, update some comments. subr_taskqueue.c: Create the new kernel thread taskqueue. This taskqueue runs outside of Giant, so any functions queued to it would need to explicitly acquire/release Giant if they need it. cd.4: Update the cd(4) man page to talk about the minimum command size sysctl/loader tunable. Also note that the changer variables are available as loader tunables as well. da.4: Update the da(4) man page to cover the retry_count, default_timeout and minimum_cmd_size sysctl variables/loader tunables. Remove references to /dev/r???, they aren't used any longer. cd.9: Update the cd(9) man page to describe the CD_Q_10_BYTE_ONLY quirk. taskqueue.9: Update the taskqueue(9) man page to describe the new thread task queue, and the taskqueue_swi_giant queue. MFC after: 3 days
2003-09-03 04:46:28 +00:00
}
A better fix to avoid race conditions between failed probes and peripheral removal. Obtained from:gibbs@freebsd.org
1999-01-07 20:19:09 +00:00
/*
* Since our peripheral may be invalidated by an error
* above or an external event, we must release our CCB
* before releasing the probe lock on the peripheral.
* The peripheral will only go away once the last lock
* is removed, and we need it around for the CCB release
* operation.
*/
xpt_release_ccb(done_ccb);
Do not schedule periph for payload/TUR requests if reprobe is in progress to avoid sending extra READ CAPACITY requests by dastart(). Schedule periph again on reprobe completion, or otherwise it may stuck indefinitely long. This should fix USB explore thread hanging on device unplug, waiting for periph destruction. Reported by: hselasky
2013-01-11 16:10:11 +00:00
softc->state = DA_STATE_NORMAL;
daschedule(periph);
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
wakeup(&softc->disk->d_mediasize);
if ((softc->flags & DA_FLAG_PROBED) == 0) {
softc->flags |= DA_FLAG_PROBED;
cam_periph_unhold(periph);
} else
cam_periph_release_locked(periph);
A better fix to avoid race conditions between failed probes and peripheral removal. Obtained from:gibbs@freebsd.org
1999-01-07 20:19:09 +00:00
return;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
case DA_CCB_WAITING:
{
/* Caller will release the CCB */
wakeup(&done_ccb->ccb_h.cbfcnp);
return;
}
case DA_CCB_DUMP:
/* No-op. We're polling */
return;
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
case DA_CCB_TUR:
{
if ((done_ccb->ccb_h.status & CAM_STATUS_MASK) != CAM_REQ_CMP) {
if (daerror(done_ccb, CAM_RETRY_SELTO,
SF_RETRY_UA | SF_NO_RECOVERY | SF_NO_PRINT) ==
ERESTART)
return;
if ((done_ccb->ccb_h.status & CAM_DEV_QFRZN) != 0)
cam_release_devq(done_ccb->ccb_h.path,
/*relsim_flags*/0,
/*reduction*/0,
/*timeout*/0,
/*getcount_only*/0);
}
xpt_release_ccb(done_ccb);
cam_periph_release_locked(periph);
return;
}
A better fix to avoid race conditions between failed probes and peripheral removal. Obtained from:gibbs@freebsd.org
1999-01-07 20:19:09 +00:00
default:
break;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
A better fix to avoid race conditions between failed probes and peripheral removal. Obtained from:gibbs@freebsd.org
1999-01-07 20:19:09 +00:00
xpt_release_ccb(done_ccb);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
static void
dareprobe(struct cam_periph *periph)
{
struct da_softc *softc;
cam_status status;
softc = (struct da_softc *)periph->softc;
/* Probe in progress; don't interfere. */
if ((softc->flags & DA_FLAG_PROBED) == 0)
return;
status = cam_periph_acquire(periph);
KASSERT(status == CAM_REQ_CMP,
("dareprobe: cam_periph_acquire failed"));
if (softc->flags & DA_FLAG_CAN_RC16)
softc->state = DA_STATE_PROBE2;
else
softc->state = DA_STATE_PROBE;
xpt_schedule(periph, CAM_PRIORITY_DEV);
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
static int
daerror(union ccb *ccb, u_int32_t cam_flags, u_int32_t sense_flags)
{
struct da_softc *softc;
struct cam_periph *periph;
Add scsi_extract_sense_ccb() -- wrapper around scsi_extract_sense_len(). It allows to remove number of duplicate checks from several places.
2012-06-23 12:32:53 +00:00
int error, error_code, sense_key, asc, ascq;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
periph = xpt_path_periph(ccb->ccb_h.path);
softc = (struct da_softc *)periph->softc;
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
/*
* Automatically detect devices that do not support
* READ(6)/WRITE(6) and upgrade to using 10 byte cdbs.
*/
error = 0;
Fix ATAPI/USB/Firewire CDROM drive handling in cd(4) and hopefully fix a number of related problems along the way. - Automatically detect CDROM drives that can't handle 6 byte mode sense and mode select, and adjust our command size accordingly. We have to handle this in the cd(4) driver (where the buffers are allocated), since the parameter list length is different for the 6 and 10 byte mode sense commands. - Remove MODE_SENSE and MODE_SELECT translation removed in ATAPICAM and in the umass(4) driver, since there's no way for that to work properly. - Add a quirk entry for CDROM drives that just hang when they get a 6 byte mode sense or mode select. The reason for the quirk must be documented in a PR, and all quirks must be approved by ken@FreeBSD.org. This is to make sure that we fully understand why each quirk is needed. Once the CAM_NEW_TRAN_CODE is finished, we should be able to remove any such quirks, since we'll know what protocol the drive speaks (SCSI, ATAPI, etc.) and therefore whether we should use 6 or 10 byte mode sense/select commands. - Change the way the da(4) handles the no_6_byte sysctl. There is now a per-drive sysctl to set the minimum command size for that particular disk. (Since you could have multiple disks with multiple requirements in one system.) - Loader tunable support for all the sysctls in the da(4) and cd(4) drivers. - Add a CDIOCCLOSE ioctl for cd(4) (bde pointed this out a long time ago). - Add a media validation routine (cdcheckmedia()) to the cd(4) driver, to fix some problems bde pointed out a long time ago. We now allow open() to succeed no matter what, but if we don't detect valid media, the user can only issue CDIOCCLOSE or CDIOCEJECT ioctls. - The media validation routine also reads the table of contents off the drive. We use the table of contents to implement the CDIOCPLAYTRACKS ioctl using the PLAY AUDIO MSF command. The PLAY AUDIO TRACK INDEX command that we previously used was deprecated after SCSI-2. It works in every SCSI CDROM I've tried, but doesn't seem to work on ATAPI CDROM drives. We still use the play audio track index command if we don't have a valid TOC, but I suppose it'll fail anyway in that case. - Add _len() versions of scsi_mode_sense() and scsi_mode_select() so that we can specify the minimum command length. - Fix a couple of formatting problems in the sense printing code. MFC after: 4 weeks
2003-02-21 06:19:38 +00:00
if ((ccb->ccb_h.status & CAM_STATUS_MASK) == CAM_REQ_INVALID) {
error = cmd6workaround(ccb);
Add scsi_extract_sense_ccb() -- wrapper around scsi_extract_sense_len(). It allows to remove number of duplicate checks from several places.
2012-06-23 12:32:53 +00:00
} else if (scsi_extract_sense_ccb(ccb,
&error_code, &sense_key, &asc, &ascq)) {
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
if (sense_key == SSD_KEY_ILLEGAL_REQUEST)
error = cmd6workaround(ccb);
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
/*
* If the target replied with CAPACITY DATA HAS CHANGED UA,
* query the capacity and notify upper layers.
*/
else if (sense_key == SSD_KEY_UNIT_ATTENTION &&
asc == 0x2A && ascq == 0x09) {
xpt_print(periph->path, "capacity data has changed\n");
dareprobe(periph);
sense_flags |= SF_NO_PRINT;
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
} else if (sense_key == SSD_KEY_UNIT_ATTENTION &&
asc == 0x28 && ascq == 0x00)
disk_media_changed(softc->disk, M_NOWAIT);
else if (sense_key == SSD_KEY_NOT_READY &&
asc == 0x3a && (softc->flags & DA_FLAG_SAW_MEDIA)) {
softc->flags &= ~DA_FLAG_SAW_MEDIA;
disk_media_gone(softc->disk, M_NOWAIT);
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
}
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
}
if (error == ERESTART)
Updates to cmd6workaround: * Only update cdb in place if not CDB_POINTER * Correctly check for QFRZ before restarting CCB * More accurate printf message * style(9) changes at end Patch tested successfully on Maxtor 4 G120J6 GAK8.
2002-08-16 22:05:19 +00:00
return (ERESTART);
Automatically detect devices that do not support READ(6)/WRITE(6) and upgrade to using 10 byte cdbs. As far as I tested, this works efficiently for most of the SBP-II/Firewire devices but most of the umass devices still need ad-hoc work around because umass-sim doesn't return any SCSI errors. A sysctl nob is also added for the last resort. I hope we don't need DA_Q_NO_6_BYTE quirks anymore. Reviewed by: gibbs MFC after: 1 week
2002-03-23 18:18:02 +00:00
Don't invalidate devices due to unexpected unit attention errors. In a perfect world, we'd notice the UA and do some device validation to ensure that the device hasn't changed. We may get this before the year ends, but not before 3.0R. This change gives the adminstrator ample ammunition to take off a foot or two, but hey this *is* UN*X.
1998-09-20 07:17:11 +00:00
/*
* XXX
* Until we have a better way of doing pack validation,
* don't treat UAs as errors.
*/
sense_flags |= SF_RETRY_UA;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
return(cam_periph_error(ccb, cam_flags, sense_flags,
&softc->saved_ccb));
}
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
static void
damediapoll(void *arg)
{
struct cam_periph *periph = arg;
struct da_softc *softc = periph->softc;
Do not schedule periph for payload/TUR requests if reprobe is in progress to avoid sending extra READ CAPACITY requests by dastart(). Schedule periph again on reprobe completion, or otherwise it may stuck indefinitely long. This should fix USB explore thread hanging on device unplug, waiting for periph destruction. Reported by: hselasky
2013-01-11 16:10:11 +00:00
if (!softc->tur && softc->outstanding_cmds == 0) {
Implement media change notification for DA and CD removable media devices. It includes three parts: 1) Modifications to CAM to detect media media changes and report them to disk(9) layer. For modern SATA (and potentially UAS) devices it utilizes Asynchronous Notification mechanism to receive events from hardware. Active polling with TEST UNIT READY commands with 3 seconds period is used for incapable hardware. After that both CD and DA drivers work the same way, detecting two conditions: "NOT READY: Medium not present" after medium was detected previously, and "UNIT ATTENTION: Not ready to ready change, medium may have changed". First one reported to disk(9) as media removal, second as media insert/change. To reliably receive second event new AC_UNIT_ATTENTION async added to make UAs broadcasted to all periphs by generic error handling code in cam_periph_error(). 2) Modifications to GEOM core to handle media remove and change events. Media removal handled by spoiling all consumers attached to the provider. Media change event also schedules provider retaste after spoiling to probe new media. New flag G_CF_ORPHAN was added to consumers to reflect that consumer is in process of destruction. It allows retaste to create new geom instance of the same class, while previous one is still dying. 3) Modifications to some GEOM classes: DEV -- to report media change events to devd; VFS -- to handle spoiling same as orphan to prevent accessing replaced media. PART class already handles spoiling alike to orphan. Reviewed by: silence on geom@ and scsi@ Tested by: avg Sponsored by: iXsystems, Inc. / PC-BSD MFC after: 2 months
2012-07-29 11:51:48 +00:00
if (cam_periph_acquire(periph) == CAM_REQ_CMP) {
softc->tur = 1;
daschedule(periph);
}
}
/* Queue us up again */
if (da_poll_period != 0)
callout_schedule(&softc->mediapoll_c, da_poll_period * hz);
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
static void
daprevent(struct cam_periph *periph, int action)
{
struct da_softc *softc;
union ccb *ccb;
int error;
softc = (struct da_softc *)periph->softc;
if (((action == PR_ALLOW)
&& (softc->flags & DA_FLAG_PACK_LOCKED) == 0)
|| ((action == PR_PREVENT)
&& (softc->flags & DA_FLAG_PACK_LOCKED) != 0)) {
return;
}
Replace most of priority numbers with defines. No logical changes.
2009-10-23 08:27:55 +00:00
ccb = cam_periph_getccb(periph, CAM_PRIORITY_NORMAL);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
scsi_prevent(&ccb->csio,
/*retries*/1,
/*cbcfp*/dadone,
MSG_SIMPLE_Q_TAG,
action,
SSD_FULL_SIZE,
5000);
Do not reinvent a wheel and let default error handler do its job.
2012-06-05 10:08:22 +00:00
error = cam_periph_runccb(ccb, daerror, CAM_RETRY_SELTO,
SF_RETRY_UA | SF_QUIET_IR, softc->disk->d_devstat);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
if (error == 0) {
if (action == PR_ALLOW)
softc->flags &= ~DA_FLAG_PACK_LOCKED;
else
softc->flags |= DA_FLAG_PACK_LOCKED;
}
xpt_release_ccb(ccb);
}
static void
Merge to da driver quirks hinting 4K physical sector sizes for SATA disks connected via SAS or USB. Unluckily I've found that SAS (mps) and USB-SATA I have translate models in different ways, requiring twice more quirks. Unluckily for Hitachi, their model names are trimmed on SAS, making impossible to identify 4K sector drives that way.
2011-12-23 00:10:17 +00:00
dasetgeom(struct cam_periph *periph, uint32_t block_len, uint64_t maxsector,
Add CAM infrastructure to allow reporting when a drive's long read capacity data changes. cam_ccb.h: Add a new advanced information type, CDAI_TYPE_RCAPLONG, for long read capacity data. cam_xpt_internal.h: Add a read capacity data pointer and length to struct cam_ed. cam_xpt.c: Free the read capacity buffer when a device goes away. While we're here, make sure we don't leak memory for other malloced fields in struct cam_ed. scsi_all.c: Update the scsi_read_capacity_16() to take a uint8_t * and a length instead of just a pointer to the parameter data structure. This will hopefully make this function somewhat immune to future changes in the parameter data. scsi_all.h: Add some extra bit definitions to struct scsi_read_capacity_data_long, and bump up the structure size to the full size specified by SBC-3. Change the prototype for scsi_read_capacity_16(). scsi_da.c: Register changes in read capacity data with the transport layer. This allows the transport layer to send out an async notification to interested parties. Update the dasetgeom() API. Use scsi_extract_sense_len() instead of scsi_extract_sense(). scsi_xpt.c: Add support for the new CDAI_TYPE_RCAPLONG advanced information type. Make sure we set the physpath pointer to NULL after freeing it. This allows blindly freeing it in the struct cam_ed destructor. sys/param.h: Bump __FreeBSD_version from 1000005 to 1000006 to make it easier for third party drivers to determine that the read capacity data async notification is available. camcontrol.c, mptutil/mpt_cam.c: Update these for the new scsi_read_capacity_16() argument structure. Sponsored by: Spectra Logic
2012-01-26 18:09:28 +00:00
struct scsi_read_capacity_data_long *rcaplong, size_t rcap_len)
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
{
struct ccb_calc_geometry ccg;
struct da_softc *softc;
struct disk_params *dp;
Add CAM infrastructure to allow reporting when a drive's long read capacity data changes. cam_ccb.h: Add a new advanced information type, CDAI_TYPE_RCAPLONG, for long read capacity data. cam_xpt_internal.h: Add a read capacity data pointer and length to struct cam_ed. cam_xpt.c: Free the read capacity buffer when a device goes away. While we're here, make sure we don't leak memory for other malloced fields in struct cam_ed. scsi_all.c: Update the scsi_read_capacity_16() to take a uint8_t * and a length instead of just a pointer to the parameter data structure. This will hopefully make this function somewhat immune to future changes in the parameter data. scsi_all.h: Add some extra bit definitions to struct scsi_read_capacity_data_long, and bump up the structure size to the full size specified by SBC-3. Change the prototype for scsi_read_capacity_16(). scsi_da.c: Register changes in read capacity data with the transport layer. This allows the transport layer to send out an async notification to interested parties. Update the dasetgeom() API. Use scsi_extract_sense_len() instead of scsi_extract_sense(). scsi_xpt.c: Add support for the new CDAI_TYPE_RCAPLONG advanced information type. Make sure we set the physpath pointer to NULL after freeing it. This allows blindly freeing it in the struct cam_ed destructor. sys/param.h: Bump __FreeBSD_version from 1000005 to 1000006 to make it easier for third party drivers to determine that the read capacity data async notification is available. camcontrol.c, mptutil/mpt_cam.c: Update these for the new scsi_read_capacity_16() argument structure. Sponsored by: Spectra Logic
2012-01-26 18:09:28 +00:00
u_int lbppbe, lalba;
Fix locking problem in disk_resize(); previously it would run without topology lock, resulting in assertion when running with DIAGNOSTIC. Reviewed by: mav (earlier version)
2012-10-29 17:52:43 +00:00
int error;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
softc = (struct da_softc *)periph->softc;
dp = &softc->params;
Add support to CAM for devices with more than 2^32 blocks. (2TB if you're using 512 byte blocks). cam_ccb.h: Bump up volume_size and cylinders in ccb_calc_geometry to 64 bits and 32 bits respectively, so we can hold larger device sizes. cylinders would overflow at about 500GB. Bump CAM_VERSION for this change. Note that this will require a recompile of all applications that talk to the pass(4) driver. scsi_all.c: Add descriptions for READ/WRITE(16), update READ/WRITE(12) descriptions, add descriptions for SERVICE ACTION IN/OUT. Add a new function, scsi_read_capacity_16(), that issues the read capacity service action. (Necessary for arrays larger than 2^32 sectors.) Update scsi_read_write() to use a 64 bit LBA and issue READ(16) or WRITE(16) if necessary. NOTE the API change. This should be largely transparnet to most userland applications at compile time, but will break binary compatibility. The CAM_VERSION bump, above, also serves the purpose of forcing a recompile for any applications that talk to CAM. scsi_all.h: Add 16 byte READ/WRITE structures, structures for 16 byte READ CAPACITY/SERVICE ACTION IN. Add scsi_u64to8b() and scsi_8btou64. scsi_da.c: The da(4) driver probe now has two stages for devices larger than 2TB. If a standard READ CAPACITY(10) returns 0xffffffff, we issue the 16 byte version of read capacity to determine the true array capacity. We also do the same thing in daopen() -- use the 16 byte read capacity if the device is large enough. The sysctl/loader code has also been updated to accept 16 bytes as a minimum command size.
2003-04-30 00:35:22 +00:00
dp->secsize = block_len;
dp->sectors = maxsector + 1;
Add CAM infrastructure to allow reporting when a drive's long read capacity data changes. cam_ccb.h: Add a new advanced information type, CDAI_TYPE_RCAPLONG, for long read capacity data. cam_xpt_internal.h: Add a read capacity data pointer and length to struct cam_ed. cam_xpt.c: Free the read capacity buffer when a device goes away. While we're here, make sure we don't leak memory for other malloced fields in struct cam_ed. scsi_all.c: Update the scsi_read_capacity_16() to take a uint8_t * and a length instead of just a pointer to the parameter data structure. This will hopefully make this function somewhat immune to future changes in the parameter data. scsi_all.h: Add some extra bit definitions to struct scsi_read_capacity_data_long, and bump up the structure size to the full size specified by SBC-3. Change the prototype for scsi_read_capacity_16(). scsi_da.c: Register changes in read capacity data with the transport layer. This allows the transport layer to send out an async notification to interested parties. Update the dasetgeom() API. Use scsi_extract_sense_len() instead of scsi_extract_sense(). scsi_xpt.c: Add support for the new CDAI_TYPE_RCAPLONG advanced information type. Make sure we set the physpath pointer to NULL after freeing it. This allows blindly freeing it in the struct cam_ed destructor. sys/param.h: Bump __FreeBSD_version from 1000005 to 1000006 to make it easier for third party drivers to determine that the read capacity data async notification is available. camcontrol.c, mptutil/mpt_cam.c: Update these for the new scsi_read_capacity_16() argument structure. Sponsored by: Spectra Logic
2012-01-26 18:09:28 +00:00
if (rcaplong != NULL) {
lbppbe = rcaplong->prot_lbppbe & SRC16_LBPPBE;
lalba = scsi_2btoul(rcaplong->lalba_lbp);
lalba &= SRC16_LALBA_A;
} else {
lbppbe = 0;
lalba = 0;
}
Merge to da driver quirks hinting 4K physical sector sizes for SATA disks connected via SAS or USB. Unluckily I've found that SAS (mps) and USB-SATA I have translate models in different ways, requiring twice more quirks. Unluckily for Hitachi, their model names are trimmed on SAS, making impossible to identify 4K sector drives that way.
2011-12-23 00:10:17 +00:00
if (lbppbe > 0) {
dp->stripesize = block_len << lbppbe;
Report stripeoffset as zero not stripesize if physical block is zero aligned, same as it is done for ATA.
2011-12-23 20:59:13 +00:00
dp->stripeoffset = (dp->stripesize - block_len * lalba) %
dp->stripesize;
Merge to da driver quirks hinting 4K physical sector sizes for SATA disks connected via SAS or USB. Unluckily I've found that SAS (mps) and USB-SATA I have translate models in different ways, requiring twice more quirks. Unluckily for Hitachi, their model names are trimmed on SAS, making impossible to identify 4K sector drives that way.
2011-12-23 00:10:17 +00:00
} else if (softc->quirks & DA_Q_4K) {
dp->stripesize = 4096;
dp->stripeoffset = 0;
} else {
dp->stripesize = 0;
dp->stripeoffset = 0;
}
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
/*
* Have the controller provide us with a geometry
* for this disk. The only time the geometry
* matters is when we boot and the controller
* is the only one knowledgeable enough to come
* up with something that will make this a bootable
* device.
*/
Replace most of priority numbers with defines. No logical changes.
2009-10-23 08:27:55 +00:00
xpt_setup_ccb(&ccg.ccb_h, periph->path, CAM_PRIORITY_NORMAL);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
ccg.ccb_h.func_code = XPT_CALC_GEOMETRY;
ccg.block_size = dp->secsize;
ccg.volume_size = dp->sectors;
ccg.heads = 0;
ccg.secs_per_track = 0;
ccg.cylinders = 0;
xpt_action((union ccb*)&ccg);
Actually check to see if XPT_CALC_GEOMETRY had an error and pick a bogus geometry if it had so we don't later get a divide by zero trap.
2006-04-18 22:01:59 +00:00
if ((ccg.ccb_h.status & CAM_STATUS_MASK) != CAM_REQ_CMP) {
/*
* We don't know what went wrong here- but just pick
* a geometry so we don't have nasty things like divide
* by zero.
*/
dp->heads = 255;
dp->secs_per_track = 255;
dp->cylinders = dp->sectors / (255 * 255);
if (dp->cylinders == 0) {
dp->cylinders = 1;
}
} else {
dp->heads = ccg.heads;
dp->secs_per_track = ccg.secs_per_track;
dp->cylinders = ccg.cylinders;
}
Add CAM infrastructure to allow reporting when a drive's long read capacity data changes. cam_ccb.h: Add a new advanced information type, CDAI_TYPE_RCAPLONG, for long read capacity data. cam_xpt_internal.h: Add a read capacity data pointer and length to struct cam_ed. cam_xpt.c: Free the read capacity buffer when a device goes away. While we're here, make sure we don't leak memory for other malloced fields in struct cam_ed. scsi_all.c: Update the scsi_read_capacity_16() to take a uint8_t * and a length instead of just a pointer to the parameter data structure. This will hopefully make this function somewhat immune to future changes in the parameter data. scsi_all.h: Add some extra bit definitions to struct scsi_read_capacity_data_long, and bump up the structure size to the full size specified by SBC-3. Change the prototype for scsi_read_capacity_16(). scsi_da.c: Register changes in read capacity data with the transport layer. This allows the transport layer to send out an async notification to interested parties. Update the dasetgeom() API. Use scsi_extract_sense_len() instead of scsi_extract_sense(). scsi_xpt.c: Add support for the new CDAI_TYPE_RCAPLONG advanced information type. Make sure we set the physpath pointer to NULL after freeing it. This allows blindly freeing it in the struct cam_ed destructor. sys/param.h: Bump __FreeBSD_version from 1000005 to 1000006 to make it easier for third party drivers to determine that the read capacity data async notification is available. camcontrol.c, mptutil/mpt_cam.c: Update these for the new scsi_read_capacity_16() argument structure. Sponsored by: Spectra Logic
2012-01-26 18:09:28 +00:00
/*
* If the user supplied a read capacity buffer, and if it is
* different than the previous buffer, update the data in the EDT.
* If it's the same, we don't bother. This avoids sending an
* update every time someone opens this device.
*/
if ((rcaplong != NULL)
&& (bcmp(rcaplong, &softc->rcaplong,
min(sizeof(softc->rcaplong), rcap_len)) != 0)) {
struct ccb_dev_advinfo cdai;
xpt_setup_ccb(&cdai.ccb_h, periph->path, CAM_PRIORITY_NORMAL);
cdai.ccb_h.func_code = XPT_DEV_ADVINFO;
cdai.buftype = CDAI_TYPE_RCAPLONG;
cdai.flags |= CDAI_FLAG_STORE;
cdai.bufsiz = rcap_len;
cdai.buf = (uint8_t *)rcaplong;
xpt_action((union ccb *)&cdai);
if ((cdai.ccb_h.status & CAM_DEV_QFRZN) != 0)
cam_release_devq(cdai.ccb_h.path, 0, 0, 0, FALSE);
if (cdai.ccb_h.status != CAM_REQ_CMP) {
xpt_print(periph->path, "%s: failed to set read "
"capacity advinfo\n", __func__);
/* Use cam_error_print() to decode the status */
cam_error_print((union ccb *)&cdai, CAM_ESF_CAM_STATUS,
CAM_EPF_ALL);
} else {
bcopy(rcaplong, &softc->rcaplong,
min(sizeof(softc->rcaplong), rcap_len));
}
}
Refactor da(4) to remove one of two code paths used to query capacity data. Reviewed by: ken, mav (earlier version) Sponsored by: The FreeBSD Foundation
2012-04-12 12:58:14 +00:00
softc->disk->d_sectorsize = softc->params.secsize;
softc->disk->d_mediasize = softc->params.secsize * (off_t)softc->params.sectors;
softc->disk->d_stripesize = softc->params.stripesize;
softc->disk->d_stripeoffset = softc->params.stripeoffset;
/* XXX: these are not actually "firmware" values, so they may be wrong */
softc->disk->d_fwsectors = softc->params.secs_per_track;
softc->disk->d_fwheads = softc->params.heads;
softc->disk->d_devstat->block_size = softc->params.secsize;
softc->disk->d_devstat->flags &= ~DEVSTAT_BS_UNAVAILABLE;
if (softc->delete_method > DA_DELETE_DISABLE)
softc->disk->d_flags |= DISKFLAG_CANDELETE;
else
softc->disk->d_flags &= ~DISKFLAG_CANDELETE;
Make the da(4) driver notify GEOM about LUN size change. Reviewed by: mav Sponsored by: FreeBSD Foundation
2012-07-07 22:19:51 +00:00
Fix locking problem in disk_resize(); previously it would run without topology lock, resulting in assertion when running with DIAGNOSTIC. Reviewed by: mav (earlier version)
2012-10-29 17:52:43 +00:00
error = disk_resize(softc->disk, M_NOWAIT);
if (error != 0)
xpt_print(periph->path, "disk_resize(9) failed, error = %d\n", error);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
static void
dasendorderedtag(void *arg)
{
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
struct da_softc *softc = arg;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
if (da_send_ordered) {
if ((softc->ordered_tag_count == 0)
&& ((softc->flags & DA_FLAG_WENT_IDLE) == 0)) {
softc->flags |= DA_FLAG_NEED_OTAG;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
if (softc->outstanding_cmds > 0)
softc->flags &= ~DA_FLAG_WENT_IDLE;
softc->ordered_tag_count = 0;
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
/* Queue us up again */
callout_reset(&softc->sendordered_c,
Insert ordered command every 1/4 of the current command timeout, not 1/4 of the default one. Without this change setting kern.cam.ada.default_timeout to 1 instead of 30 allowed me to trigger several false positive command timeouts under heavy ZFS load on a SiI3132 siis(4) controller with 5 HDDs on a port multiplier. MFC after: 1 week
2012-02-02 19:02:15 +00:00
(da_default_timeout * hz) / DA_ORDEREDTAG_INTERVAL,
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
dasendorderedtag, softc);
SCSI Peripheral drivers for CAM: da - Direct Access Devices (disks, optical devices, SS disks) cd - CDROM (or devices that can act like them, WORM, CD-RW, etc) ch - Medium Changer devices. sa - Sequential Access Devices (tape drives) pass - Application pass-thru driver targ - Target Mode "Processor Target" Emulator pt - Processor Target Devices (scanners, cpus, etc.) Submitted by: The CAM Team
1998-09-15 06:36:34 +00:00
}
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
/*
* Step through all DA peripheral drivers, and if the device is still open,
* sync the disk cache to physical media.
*/
static void
Implement a new generic mechanism for attaching handler functions to events, in order to pave the way for removing a number of the ad-hoc implementations currently in use. Retire the at_shutdown family of functions and replace them with new event handler lists. Rework kern_shutdown.c to take greater advantage of the use of event handlers. Reviewed by: green
1999-08-21 06:24:40 +00:00
dashutdown(void * arg, int howto)
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
{
struct cam_periph *periph;
struct da_softc *softc;
One more major cam_periph_error() rewrite to improve error handling and reporting. It includes: - removing of error messages controlled by bootverbose, replacing them with more universal and informative debugging on CAM_DEBUG_INFO level, that is now built into the kernel by default; - more close following to the arguments submitted by caller, such as SF_PRINT_ALWAYS, SF_QUIET_IR and SF_NO_PRINT; consumer knows better which errors are usual/expected at this point and which are really informative; - adding two new flags SF_NO_RECOVERY and SF_NO_RETRY to allow caller specify how much assistance it needs at this point; previously consumers controlled that by not calling cam_periph_error() at all, but that made behavior inconsistent and debugging complicated; - tuning debug messages and taken actions order to make debugging output more readable and cause-effect relationships visible; - making camperiphdone() (common device recovery completion handler) to also use cam_periph_error() in most cases, instead of own dumb code; - removing manual sense fetching code from cam_periph_error(); I was told by number of people that it is SIM obligation to fetch sense data, so this code is useless and only significantly complicates recovery logic; - making ada, da and pass driver to use cam_periph_error() with new limited recovery options to handle error recovery and debugging in common way; as one of results, CAM_REQUEUE_REQ and other retrying statuses are now working fine with pass driver, that caused many problems before. - reverting r186891 by raj@ to avoid burning few seconds in tight DELAY() loops on device probe, while device simply loads media; I think that problem may already be fixed in other way, and even if it is not, solution must be different. Sponsored by: iXsystems, Inc. MFC after: 2 weeks
2012-06-09 13:07:44 +00:00
int error;
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
Another round of the <sys/queue.h> FOREACH transmogriffer. Created with: sed(1) Reviewed by: md5(1)
2001-02-04 16:08:18 +00:00
TAILQ_FOREACH(periph, &dadriver.units, unit_links) {
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
union ccb ccb;
Missed locking the dump and shutdown entry points in the scsi_da driver.
2007-04-18 05:14:16 +00:00
cam_periph_lock(periph);
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
softc = (struct da_softc *)periph->softc;
/*
* We only sync the cache if the drive is still open, and
* if the drive is capable of it..
*/
if (((softc->flags & DA_FLAG_OPEN) == 0)
Fix a leaked lock in dashutdown.
2007-04-19 22:18:15 +00:00
|| (softc->quirks & DA_Q_NO_SYNC_CACHE)) {
cam_periph_unlock(periph);
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
continue;
Fix a leaked lock in dashutdown.
2007-04-19 22:18:15 +00:00
}
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
Replace most of priority numbers with defines. No logical changes.
2009-10-23 08:27:55 +00:00
xpt_setup_ccb(&ccb.ccb_h, periph->path, CAM_PRIORITY_NORMAL);
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
ccb.ccb_h.ccb_state = DA_CCB_DUMP;
scsi_synchronize_cache(&ccb.csio,
One more major cam_periph_error() rewrite to improve error handling and reporting. It includes: - removing of error messages controlled by bootverbose, replacing them with more universal and informative debugging on CAM_DEBUG_INFO level, that is now built into the kernel by default; - more close following to the arguments submitted by caller, such as SF_PRINT_ALWAYS, SF_QUIET_IR and SF_NO_PRINT; consumer knows better which errors are usual/expected at this point and which are really informative; - adding two new flags SF_NO_RECOVERY and SF_NO_RETRY to allow caller specify how much assistance it needs at this point; previously consumers controlled that by not calling cam_periph_error() at all, but that made behavior inconsistent and debugging complicated; - tuning debug messages and taken actions order to make debugging output more readable and cause-effect relationships visible; - making camperiphdone() (common device recovery completion handler) to also use cam_periph_error() in most cases, instead of own dumb code; - removing manual sense fetching code from cam_periph_error(); I was told by number of people that it is SIM obligation to fetch sense data, so this code is useless and only significantly complicates recovery logic; - making ada, da and pass driver to use cam_periph_error() with new limited recovery options to handle error recovery and debugging in common way; as one of results, CAM_REQUEUE_REQ and other retrying statuses are now working fine with pass driver, that caused many problems before. - reverting r186891 by raj@ to avoid burning few seconds in tight DELAY() loops on device probe, while device simply loads media; I think that problem may already be fixed in other way, and even if it is not, solution must be different. Sponsored by: iXsystems, Inc. MFC after: 2 weeks
2012-06-09 13:07:44 +00:00
/*retries*/0,
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
/*cbfcnp*/dadone,
MSG_SIMPLE_Q_TAG,
/*begin_lba*/0, /* whole disk */
/*lb_count*/0,
SSD_FULL_SIZE,
Be very generous with timeouts for synchronize cache. We may wait a very long time in the cases where it really sends the drive out to lunch, but it also allows us to catch very wierd edge cases of strange drives that might take a very long time (emulated disk drives over a network, e.g.).
2001-11-17 18:26:00 +00:00
60 * 60 * 1000);
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
xpt_polled_action(&ccb);
One more major cam_periph_error() rewrite to improve error handling and reporting. It includes: - removing of error messages controlled by bootverbose, replacing them with more universal and informative debugging on CAM_DEBUG_INFO level, that is now built into the kernel by default; - more close following to the arguments submitted by caller, such as SF_PRINT_ALWAYS, SF_QUIET_IR and SF_NO_PRINT; consumer knows better which errors are usual/expected at this point and which are really informative; - adding two new flags SF_NO_RECOVERY and SF_NO_RETRY to allow caller specify how much assistance it needs at this point; previously consumers controlled that by not calling cam_periph_error() at all, but that made behavior inconsistent and debugging complicated; - tuning debug messages and taken actions order to make debugging output more readable and cause-effect relationships visible; - making camperiphdone() (common device recovery completion handler) to also use cam_periph_error() in most cases, instead of own dumb code; - removing manual sense fetching code from cam_periph_error(); I was told by number of people that it is SIM obligation to fetch sense data, so this code is useless and only significantly complicates recovery logic; - making ada, da and pass driver to use cam_periph_error() with new limited recovery options to handle error recovery and debugging in common way; as one of results, CAM_REQUEUE_REQ and other retrying statuses are now working fine with pass driver, that caused many problems before. - reverting r186891 by raj@ to avoid burning few seconds in tight DELAY() loops on device probe, while device simply loads media; I think that problem may already be fixed in other way, and even if it is not, solution must be different. Sponsored by: iXsystems, Inc. MFC after: 2 weeks
2012-06-09 13:07:44 +00:00
error = cam_periph_error(&ccb,
0, SF_NO_RECOVERY | SF_NO_RETRY | SF_QUIET_IR, NULL);
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
if ((ccb.ccb_h.status & CAM_DEV_QFRZN) != 0)
One more major cam_periph_error() rewrite to improve error handling and reporting. It includes: - removing of error messages controlled by bootverbose, replacing them with more universal and informative debugging on CAM_DEBUG_INFO level, that is now built into the kernel by default; - more close following to the arguments submitted by caller, such as SF_PRINT_ALWAYS, SF_QUIET_IR and SF_NO_PRINT; consumer knows better which errors are usual/expected at this point and which are really informative; - adding two new flags SF_NO_RECOVERY and SF_NO_RETRY to allow caller specify how much assistance it needs at this point; previously consumers controlled that by not calling cam_periph_error() at all, but that made behavior inconsistent and debugging complicated; - tuning debug messages and taken actions order to make debugging output more readable and cause-effect relationships visible; - making camperiphdone() (common device recovery completion handler) to also use cam_periph_error() in most cases, instead of own dumb code; - removing manual sense fetching code from cam_periph_error(); I was told by number of people that it is SIM obligation to fetch sense data, so this code is useless and only significantly complicates recovery logic; - making ada, da and pass driver to use cam_periph_error() with new limited recovery options to handle error recovery and debugging in common way; as one of results, CAM_REQUEUE_REQ and other retrying statuses are now working fine with pass driver, that caused many problems before. - reverting r186891 by raj@ to avoid burning few seconds in tight DELAY() loops on device probe, while device simply loads media; I think that problem may already be fixed in other way, and even if it is not, solution must be different. Sponsored by: iXsystems, Inc. MFC after: 2 weeks
2012-06-09 13:07:44 +00:00
cam_release_devq(ccb.ccb_h.path, /*relsim_flags*/0,
/*reduction*/0, /*timeout*/0, /*getcount_only*/0);
if (error != 0)
xpt_print(periph->path, "Synchronize cache failed\n");
Missed locking the dump and shutdown entry points in the scsi_da driver.
2007-04-18 05:14:16 +00:00
cam_periph_unlock(periph);
Add the quirk entry framework to handle disabling the synchronize cache command on drives that don't like it. Right now, there's just a bogus quirk entry in the table that doesn't do anything, but that should be changed once we get actual inquiry data for drives that don't like the synchronize cache command. Also, add a shutdown hook that runs through all direct access peripherals and runs a synchronize cache on them if they're still open, and if synchronize cache isn't disabled via a quirk entry. Add a synchronize cache call at the end of dadump() (again, conditionalized on the quirk entry), so we can insure that the disk cache contents get flushed to physical media after a dump. Check the new quirk entry in daclose() to decide whether or not to synchronize the cache for a disk at final close. Reviewed by: gibbs
1998-10-08 05:46:38 +00:00
}
}
Implement a new camcontrol function, 'camcontrol format'. libcam/Makefile: Add scsi_da.c to libcam for the new scsi_format_unit() function. camcontrol.8: Update the man page for the new format functionality, and take out the examples section describing how to do it with 'camcontrol cmd'. camcontrol.c: New format functionality. Note that unlike the rest of the camcontrol subcommands, this one is interactive by default. Because of the potential destructiveness of the format command, I thought it necessary to get confirmation from the user before spamming a disk. You can disable the interactive behavior, and the status meter with command line arguments. scsi_da.c: Add the new scsi_format_unit() cdb building function and use #ifdef _KERNEL to make this file compile in both the kernel and userland. The format unit function is currently only defined in the non-kernel case, because nothing in the kernel is using it. If that changes, it should be un-ifdefed and compiled in both cases. scsi_da.h: New function declaration, CDB structure and format data structures. Thanks to Nick Hibma for providing some valuable input on these changes.
2000-05-21 23:57:52 +00:00
#else /* !_KERNEL */
/*
* XXX This is only left out of the kernel build to silence warnings. If,
* for some reason this function is used in the kernel, the ifdefs should
* be moved so it is included both in the kernel and userland.
*/
void
scsi_format_unit(struct ccb_scsiio *csio, u_int32_t retries,
void (*cbfcnp)(struct cam_periph *, union ccb *),
u_int8_t tag_action, u_int8_t byte2, u_int16_t ileave,
u_int8_t *data_ptr, u_int32_t dxfer_len, u_int8_t sense_len,
u_int32_t timeout)
{
struct scsi_format_unit *scsi_cmd;
scsi_cmd = (struct scsi_format_unit *)&csio->cdb_io.cdb_bytes;
scsi_cmd->opcode = FORMAT_UNIT;
scsi_cmd->byte2 = byte2;
scsi_ulto2b(ileave, scsi_cmd->interleave);
cam_fill_csio(csio,
retries,
cbfcnp,
/*flags*/ (dxfer_len > 0) ? CAM_DIR_OUT : CAM_DIR_NONE,
tag_action,
data_ptr,
dxfer_len,
sense_len,
sizeof(*scsi_cmd),
timeout);
}
#endif /* _KERNEL */
Reference in New Issue Copy Permalink