d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
12d7eaa009
freebsd-dev/sys/cam/scsi/scsi_sa.c

3748 lines
100 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 Sequential Access Peripheral driver for CAM.
*
Per Justin's request- remove his name from Copyright.
2000-03-18 22:03:41 +00:00
* Copyright (c) 1999, 2000 Matthew Jacob
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
* 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>
#include <sys/queue.h>
Change #ifdef KERNEL to #ifdef _KERNEL in the public headers. "KERNEL" is an application space macro and the applications are supposed to be free to use it as they please (but cannot). This is consistant with the other BSD's who made this change quite some time ago. More commits to come.
1999-12-29 04:46:21 +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>
#endif
#include <sys/types.h>
Rename bio_linkage to the more obvious bio_parent. Add bio_t0 timestamp, and include <sys/time.h> where needed
2003-02-07 21:09:51 +00:00
#include <sys/time.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>
Deprecate machine/limits.h in favor of new sys/limits.h. Change all in-tree consumers to include <sys/limits.h> Discussed on: standards@ Partially submitted by: Craig Rodrigues <rodrigc@attbi.com>
2003-04-29 13:36:06 +00:00
#include <sys/limits.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>
#include <sys/mtio.h>
Quick fix. <sys/conf.h> now depends on <sys/time.h>, which is not present when libcam is building this in userland.
2000-08-24 19:23:09 +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/conf.h>
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
#include <sys/sysctl.h>
#include <sys/taskqueue.h>
Quick fix. <sys/conf.h> now depends on <sys/time.h>, which is not present when libcam is building this in userland.
2000-08-24 19:23:09 +00:00
#endif
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
#include <sys/fcntl.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/devicestat.h>
Change #ifdef KERNEL to #ifdef _KERNEL in the public headers. "KERNEL" is an application space macro and the applications are supposed to be free to use it as they please (but cannot). This is consistant with the other BSD's who made this change quite some time ago. More commits to come.
1999-12-29 04:46:21 +00:00
#ifndef _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 <stdio.h>
#include <string.h>
#endif
#include <cam/cam.h>
#include <cam/cam_ccb.h>
#include <cam/cam_periph.h>
#include <cam/cam_xpt_periph.h>
#include <cam/cam_debug.h>
#include <cam/scsi/scsi_all.h>
#include <cam/scsi/scsi_message.h>
#include <cam/scsi/scsi_sa.h>
Change #ifdef KERNEL to #ifdef _KERNEL in the public headers. "KERNEL" is an application space macro and the applications are supposed to be free to use it as they please (but cannot). This is consistant with the other BSD's who made this change quite some time ago. More commits to come.
1999-12-29 04:46:21 +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
Patches from DES to create three new kernel config options to control timeouts in the SA driver (timeouts for space, rewind and erase). Folks can lengthen the timeouts if their hardware is especially slow, or shorten them if they want to be notified of errors a little sooner. Also, get rid of two OD driver options. The od driver has been made obsolete by the da driver. Reviewed by: ken, gibbs Submitted by: Dag-Erling Coidan Smørgrav <des@FreeBSD.ORG>
1998-10-02 05:15:51 +00:00
#include <opt_sa.h>
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
#ifndef SA_IO_TIMEOUT
#define SA_IO_TIMEOUT 4
#endif
Patches from DES to create three new kernel config options to control timeouts in the SA driver (timeouts for space, rewind and erase). Folks can lengthen the timeouts if their hardware is especially slow, or shorten them if they want to be notified of errors a little sooner. Also, get rid of two OD driver options. The od driver has been made obsolete by the da driver. Reviewed by: ken, gibbs Submitted by: Dag-Erling Coidan Smørgrav <des@FreeBSD.ORG>
1998-10-02 05:15:51 +00:00
#ifndef SA_SPACE_TIMEOUT
#define SA_SPACE_TIMEOUT 1 * 60
#endif
#ifndef SA_REWIND_TIMEOUT
#define SA_REWIND_TIMEOUT 2 * 60
#endif
#ifndef SA_ERASE_TIMEOUT
#define SA_ERASE_TIMEOUT 4 * 60
#endif
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
#define SCSIOP_TIMEOUT (60 * 1000) /* not an option */
#define IO_TIMEOUT (SA_IO_TIMEOUT * 60 * 1000)
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
#define REWIND_TIMEOUT (SA_REWIND_TIMEOUT * 60 * 1000)
#define ERASE_TIMEOUT (SA_ERASE_TIMEOUT * 60 * 1000)
#define SPACE_TIMEOUT (SA_SPACE_TIMEOUT * 60 * 1000)
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
/*
Another 1FM@EOD quirk (from mike@sentex.net for a Seagate STT20000), and add a config option that allows one to default to 1FM@EOD for tapes otherwise unquirked or unknown as to which to prefer. Note that tcopy will be broken for these tapes until tcopy is fixed.
1999-10-02 20:17:16 +00:00
* Additional options that can be set for config: SA_1FM_AT_EOT
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
*/
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
#ifndef UNUSED_PARAMETER
#define UNUSED_PARAMETER(x) x = x
#endif
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
#define QFRLS(ccb) \
if (((ccb)->ccb_h.status & CAM_DEV_QFRZN) != 0) \
cam_release_devq((ccb)->ccb_h.path, 0, 0, 0, FALSE)
/*
* Driver states
*/
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_SCSISA, "SCSI sa", "SCSI sequential access buffers");
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +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
typedef enum {
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
SA_STATE_NORMAL, SA_STATE_ABNORMAL
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
} sa_state;
Fix PR 24220 by using the periph private field0 for a CCB for not only CCB type but also extra flags- one of which can be "position updated". In other changes: Add in a SA_QUIRK_NO_CPAGE quirk so that it's possible to avoid using a (broken) device's implementation of he DEVICE COMPRESSION page. Also do a couple of printout cleanups. As per some discussion on FreeBSD-scsi, skip doing tape flushing if we're reading tape logical block location (MTIOCRDSPOS).
2001-01-15 22:28:11 +00:00
#define ccb_pflags ppriv_field0
#define ccb_bp ppriv_ptr1
Fix breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +00:00
/* bits in ccb_pflags */
#define SA_POSITION_UPDATED 0x1
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 {
SA_FLAG_OPEN = 0x0001,
SA_FLAG_FIXED = 0x0002,
SA_FLAG_TAPE_LOCKED = 0x0004,
SA_FLAG_TAPE_MOUNTED = 0x0008,
SA_FLAG_TAPE_WP = 0x0010,
SA_FLAG_TAPE_WRITTEN = 0x0020,
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
SA_FLAG_EOM_PENDING = 0x0040,
SA_FLAG_EIO_PENDING = 0x0080,
SA_FLAG_EOF_PENDING = 0x0100,
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
SA_FLAG_ERR_PENDING = (SA_FLAG_EOM_PENDING|SA_FLAG_EIO_PENDING|
SA_FLAG_EOF_PENDING),
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
SA_FLAG_INVALID = 0x0200,
SA_FLAG_COMP_ENABLED = 0x0400,
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
SA_FLAG_COMP_SUPP = 0x0800,
SA_FLAG_COMP_UNSUPP = 0x1000,
SA_FLAG_TAPE_FROZEN = 0x2000
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
} sa_flags;
typedef enum {
SA_MODE_REWIND = 0x00,
SA_MODE_NOREWIND = 0x01,
SA_MODE_OFFLINE = 0x02
} sa_mode;
typedef enum {
SA_PARAM_NONE = 0x00,
SA_PARAM_BLOCKSIZE = 0x01,
SA_PARAM_DENSITY = 0x02,
SA_PARAM_COMPRESSION = 0x04,
SA_PARAM_BUFF_MODE = 0x08,
SA_PARAM_NUMBLOCKS = 0x10,
SA_PARAM_WP = 0x20,
SA_PARAM_SPEED = 0x40,
SA_PARAM_ALL = 0x7f
} sa_params;
typedef enum {
SA_QUIRK_NONE = 0x00,
Some prettifying for the quirk comments. Add a SA_QUIRK_NO_MODESEL type and use it for the OnStream real SCSI device (not the broken one). This one is still broken in that it can't be set to the same fixed block size it reports [ unflattering comments about this company elided ]. If we're unable to set buffered mode on, complain, but drive on. It's not a fatal error to not be in buffered mode.
2000-05-09 04:54:10 +00:00
SA_QUIRK_NOCOMP = 0x01, /* Can't deal with compression at all */
SA_QUIRK_FIXED = 0x02, /* Force fixed mode */
SA_QUIRK_VARIABLE = 0x04, /* Force variable mode */
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
SA_QUIRK_2FM = 0x08, /* Needs Two File Marks at EOD */
Add a quirk type (and one for ARCHIVE Python's) that disables the dummy read at sa_mount that attempts to latch up density. This breaks on at least one drive and sends it into the weeds. Approved: jkh
2000-02-03 18:29:25 +00:00
SA_QUIRK_1FM = 0x10, /* No more than 1 File Mark at EOD */
Some prettifying for the quirk comments. Add a SA_QUIRK_NO_MODESEL type and use it for the OnStream real SCSI device (not the broken one). This one is still broken in that it can't be set to the same fixed block size it reports [ unflattering comments about this company elided ]. If we're unable to set buffered mode on, complain, but drive on. It's not a fatal error to not be in buffered mode.
2000-05-09 04:54:10 +00:00
SA_QUIRK_NODREAD = 0x20, /* Don't try and dummy read density */
Fix PR 24220 by using the periph private field0 for a CCB for not only CCB type but also extra flags- one of which can be "position updated". In other changes: Add in a SA_QUIRK_NO_CPAGE quirk so that it's possible to avoid using a (broken) device's implementation of he DEVICE COMPRESSION page. Also do a couple of printout cleanups. As per some discussion on FreeBSD-scsi, skip doing tape flushing if we're reading tape logical block location (MTIOCRDSPOS).
2001-01-15 22:28:11 +00:00
SA_QUIRK_NO_MODESEL = 0x40, /* Don't do mode select at all */
SA_QUIRK_NO_CPAGE = 0x80 /* Don't use DEVICE COMPRESSION page */
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
} sa_quirks;
Added output of device QUIRKS for CAM and AHCI devices during boot. Reviewed by: mav Approved by: pjd (mentor) MFC after: 2 weeks
2013-05-18 23:36:21 +00:00
#define SA_QUIRK_BIT_STRING \
"\020" \
"\001NOCOMP" \
"\002FIXED" \
"\003VARIABLE" \
"\0042FM" \
"\0051FM" \
"\006NODREAD" \
"\007NO_MODESEL" \
"\010NO_CPAGE"
Get rid of the device index number stored in the sa(4) unit number. The device index number stored in the unit number of sa(4) devices is only used to print debug messages. Get rid of this index number and use devtoname() to just print the entire device name.
2009-04-20 10:40:42 +00:00
#define SAMODE(z) (dev2unit(z) & 0x3)
#define SADENSITY(z) ((dev2unit(z) >> 2) & 0x3)
#define SA_IS_CTRL(z) (dev2unit(z) & (1 << 4))
Fix a couple of boo-boos in the last delta.
1999-11-17 17:11:21 +00:00
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
#define SA_NOT_CTLDEV 0
#define SA_CTLDEV 1
#define SA_ATYPE_R 0
#define SA_ATYPE_NR 1
#define SA_ATYPE_ER 2
Get rid of the device index number stored in the sa(4) unit number. The device index number stored in the unit number of sa(4) devices is only used to print debug messages. Get rid of this index number and use devtoname() to just print the entire device name.
2009-04-20 10:40:42 +00:00
#define SAMINOR(ctl, mode, access) \
((ctl << 4) | (mode << 2) | (access & 0x3))
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
#define SA_NUM_MODES 4
struct sa_devs {
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
struct cdev *ctl_dev;
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
struct sa_mode_devs {
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
struct cdev *r_dev;
struct cdev *nr_dev;
struct cdev *er_dev;
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
} mode_devs[SA_NUM_MODES];
};
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 sa_softc {
sa_state state;
sa_flags flags;
sa_quirks quirks;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
u_int si_flags;
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 a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
int queue_count;
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
struct devstat *device_stats;
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
struct sa_devs devs;
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 blk_gran;
int blk_mask;
int blk_shift;
u_int32_t max_blk;
u_int32_t min_blk;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
u_int32_t maxio;
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
u_int32_t cpi_maxio;
int allow_io_split;
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_int32_t comp_algorithm;
u_int32_t saved_comp_algorithm;
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
u_int32_t media_blksize;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
u_int32_t last_media_blksize;
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
u_int32_t media_numblks;
u_int8_t media_density;
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 speed;
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
u_int8_t scsi_rev;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
u_int8_t dsreg; /* mtio mt_dsreg, redux */
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 buffer_mode;
int filemarks;
union ccb saved_ccb;
Make some attempt to accomodate kern/24221- propate residual to mt_resid even if it might overflow. If we do a read or set of hardware or logical block position, we also clear Tape Frozen status.
2001-01-19 21:08:15 +00:00
int last_resid_was_io;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
/*
* Relative to BOT Location.
*/
daddr_t fileno;
daddr_t blkno;
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
/*
* Latched Error Info
*/
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
struct {
struct scsi_sense_data _last_io_sense;
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
u_int64_t _last_io_resid;
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
u_int8_t _last_io_cdb[CAM_MAX_CDBLEN];
struct scsi_sense_data _last_ctl_sense;
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
u_int64_t _last_ctl_resid;
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
u_int8_t _last_ctl_cdb[CAM_MAX_CDBLEN];
#define last_io_sense errinfo._last_io_sense
#define last_io_resid errinfo._last_io_resid
#define last_io_cdb errinfo._last_io_cdb
#define last_ctl_sense errinfo._last_ctl_sense
#define last_ctl_resid errinfo._last_ctl_resid
#define last_ctl_cdb errinfo._last_ctl_cdb
} errinfo;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
/*
* Misc other flags/state
*/
u_int32_t
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
: 29,
open_rdonly : 1, /* open read-only */
open_pending_mount : 1, /* open pending mount */
ctrl_mode : 1; /* control device open */
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
struct task sysctl_task;
struct sysctl_ctx_list sysctl_ctx;
struct sysctl_oid *sysctl_tree;
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 sa_quirk_entry {
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
struct scsi_inquiry_pattern inq_pat; /* matching pattern */
sa_quirks quirks; /* specific quirk type */
u_int32_t prefblk; /* preferred blocksize when in fixed mode */
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 struct sa_quirk_entry sa_quirk_table[] =
{
Some prettifying for the quirk comments. Add a SA_QUIRK_NO_MODESEL type and use it for the OnStream real SCSI device (not the broken one). This one is still broken in that it can't be set to the same fixed block size it reports [ unflattering comments about this company elided ]. If we're unable to set buffered mode on, complain, but drive on. It's not a fatal error to not be in buffered mode.
2000-05-09 04:54:10 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "OnStream",
"ADR*", "*"}, SA_QUIRK_FIXED|SA_QUIRK_NODREAD |
SA_QUIRK_1FM|SA_QUIRK_NO_MODESEL, 32768
},
Add a quirk entry for ARCHIVE Python 06408. Approved by: mjacob
2001-06-01 12:36:24 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "ARCHIVE",
Sorry, an "ARCHIVE Python 06408" does not need SA_QUIRK_NOCOMP.
2001-06-06 13:01:44 +00:00
"Python 06408*", "*"}, SA_QUIRK_NODREAD, 0
Add a quirk entry for ARCHIVE Python 06408. Approved by: mjacob
2001-06-01 12:36:24 +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
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "ARCHIVE",
Add a quirk type (and one for ARCHIVE Python's) that disables the dummy read at sa_mount that attempts to latch up density. This breaks on at least one drive and sends it into the weeds. Approved: jkh
2000-02-03 18:29:25 +00:00
"Python 25601*", "*"}, SA_QUIRK_NOCOMP|SA_QUIRK_NODREAD, 0
},
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "ARCHIVE",
"Python*", "*"}, SA_QUIRK_NODREAD, 0
...nor does this old TDC3620 like to be asked for compression. But well, now it's running again!
1998-11-26 10:47:52 +00:00
},
Force ARCHIVE Vipers to be FIXED
1998-12-28 19:21:12 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "ARCHIVE",
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
"VIPER 150*", "*"}, SA_QUIRK_FIXED|SA_QUIRK_1FM, 512
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
},
Add special quirk for this ancient Viper drive. This is all creeping driver bloat -I really *have* to do a quirk table that can be loaded by the loader. Submitted by: Hellmuth Michaelis <hm@kts.org>
2000-11-08 18:37:12 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "ARCHIVE",
"VIPER 2525 25462", "-011"},
SA_QUIRK_NOCOMP|SA_QUIRK_1FM|SA_QUIRK_NODREAD, 0
},
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "ARCHIVE",
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
"VIPER 2525*", "*"}, SA_QUIRK_FIXED|SA_QUIRK_1FM, 1024
},
Fix PR 24220 by using the periph private field0 for a CCB for not only CCB type but also extra flags- one of which can be "position updated". In other changes: Add in a SA_QUIRK_NO_CPAGE quirk so that it's possible to avoid using a (broken) device's implementation of he DEVICE COMPRESSION page. Also do a couple of printout cleanups. As per some discussion on FreeBSD-scsi, skip doing tape flushing if we're reading tape logical block location (MTIOCRDSPOS).
2001-01-15 22:28:11 +00:00
#if 0
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "HP",
"C15*", "*"}, SA_QUIRK_VARIABLE|SA_QUIRK_NO_CPAGE, 0,
},
#endif
The HP DAT 40 tape drive should be able to handle variable block sizes. But for some reason the block size is different when a different type of tape is placed in the drive. This commit fixes that. PR: 46209 Submitted by: Alex Wang <alex@alexwang.com> Approved by: mjacob
2002-12-16 17:40:17 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "HP",
"C56*", "*"}, SA_QUIRK_VARIABLE|SA_QUIRK_2FM, 0
},
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "HP",
"T20*", "*"}, SA_QUIRK_FIXED|SA_QUIRK_1FM, 512
Force ARCHIVE Vipers to be FIXED
1998-12-28 19:21:12 +00:00
},
Make HP T4000S quirk to FIXED mode
1999-01-11 18:26:25 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "HP",
+Match against T4000* for HP QIC quirks (not T4000S* where it doesn't then catch a T4000s) + Set *some* kind of error at EOM if we're in fixed mode and have pending errs. Do not clear the ERR_PENDING bit if more buffers are queued. + Release the start_ccb in this case also, else we hang forever on rewinding. + Any kind of error for load to BOT in samount should then cause an attempt to use REWIND to come back to BOT. Do the initial load command quietly. + In samount, if we succeed, set the relative position markers.
1999-03-01 01:07:47 +00:00
"T4000*", "*"}, SA_QUIRK_FIXED|SA_QUIRK_1FM, 512
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
},
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "HP",
"HP-88780*", "*"}, SA_QUIRK_VARIABLE|SA_QUIRK_2FM, 0
},
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "KENNEDY",
"*", "*"}, SA_QUIRK_VARIABLE|SA_QUIRK_2FM, 0
},
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "M4 DATA",
"123107 SCSI*", "*"}, SA_QUIRK_VARIABLE|SA_QUIRK_2FM, 0
},
Ooops- forgot to commit this. PR: 14009 Submitted by: jreynold@primenet.com
1999-09-28 05:14:52 +00:00
{ /* jreynold@primenet.com */
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "Seagate",
"STT8000N*", "*"}, SA_QUIRK_1FM, 0
},
Another 1FM@EOD quirk (from mike@sentex.net for a Seagate STT20000), and add a config option that allows one to default to 1FM@EOD for tapes otherwise unquirked or unknown as to which to prefer. Note that tcopy will be broken for these tapes until tcopy is fixed.
1999-10-02 20:17:16 +00:00
{ /* mike@sentex.net */
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "Seagate",
"STT20000*", "*"}, SA_QUIRK_1FM, 0
},
- Add support for Support SEAGATE DAT Scopion 130 PR: kern/141934 Submitted by: HASHI Hiroaki <hashiz@stenmark.meridiani.jp> Approved by: sbruno@ MFC after: 1 week
2011-12-08 03:20:48 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "SEAGATE",
"DAT 06241-XXX", "*"}, SA_QUIRK_VARIABLE|SA_QUIRK_2FM, 0
},
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "TANDBERG",
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
" TDC 3600", "U07:"}, SA_QUIRK_NOCOMP|SA_QUIRK_1FM, 512
Make HP T4000S quirk to FIXED mode
1999-01-11 18:26:25 +00:00
},
Add TDC 3800 to quirk list for fixed mode device. PR:11866 Submitted by: Helge Oldach hmo@sep.hamburg.com
1999-05-25 23:10:54 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "TANDBERG",
" TDC 3800", "*"}, SA_QUIRK_NOCOMP|SA_QUIRK_1FM, 512
},
Quirk Tandberg 4100 like Tandberg 4200. PR: 11675 Reviewed by: msmith@FreeBSD.ORG Submitted by: blank@uni-trier.de
1999-06-24 15:21:10 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "TANDBERG",
" TDC 4100", "*"}, SA_QUIRK_NOCOMP|SA_QUIRK_1FM, 512
},
...nor does this old TDC3620 like to be asked for compression. But well, now it's running again!
1998-11-26 10:47:52 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "TANDBERG",
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
" TDC 4200", "*"}, SA_QUIRK_NOCOMP|SA_QUIRK_1FM, 512
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
},
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "TANDBERG",
" SLR*", "*"}, SA_QUIRK_1FM, 0
},
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "WANGTEK",
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
"5525ES*", "*"}, SA_QUIRK_FIXED|SA_QUIRK_1FM, 512
Make WangTek 51000 a fixed block device. PR: kern/11194 Obtained from:naddy@mips.rhein-neckar.de
1999-04-18 01:05:03 +00:00
},
{
{ T_SEQUENTIAL, SIP_MEDIA_REMOVABLE, "WANGTEK",
"51000*", "*"}, SA_QUIRK_FIXED|SA_QUIRK_1FM, 1024
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 d_open_t saopen;
static d_close_t saclose;
static d_strategy_t sastrategy;
static d_ioctl_t saioctl;
static periph_init_t sainit;
static periph_ctor_t saregister;
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 saoninvalidate;
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_dtor_t sacleanup;
static periph_start_t sastart;
static void saasync(void *callback_arg, u_int32_t code,
struct cam_path *path, void *arg);
static void sadone(struct cam_periph *periph,
union ccb *start_ccb);
static int saerror(union ccb *ccb, u_int32_t cam_flags,
u_int32_t sense_flags);
scsi_sa.h
2000-10-31 22:34:51 +00:00
static int samarkswanted(struct cam_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 int sacheckeod(struct cam_periph *periph);
static int sagetparams(struct cam_periph *periph,
sa_params params_to_get,
u_int32_t *blocksize, u_int8_t *density,
u_int32_t *numblocks, int *buff_mode,
u_int8_t *write_protect, u_int8_t *speed,
int *comp_supported, int *comp_enabled,
u_int32_t *comp_algorithm,
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
sa_comp_t *comp_page);
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 sasetparams(struct cam_periph *periph,
sa_params params_to_set,
u_int32_t blocksize, u_int8_t density,
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
u_int32_t comp_algorithm,
u_int32_t sense_flags);
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 saprevent(struct cam_periph *periph, int action);
static int sarewind(struct cam_periph *periph);
static int saspace(struct cam_periph *periph, int count,
scsi_space_code code);
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
static int samount(struct cam_periph *, int, struct cdev *);
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 saretension(struct cam_periph *periph);
static int sareservereleaseunit(struct cam_periph *periph,
int reserve);
static int saloadunload(struct cam_periph *periph, int load);
static int saerase(struct cam_periph *periph, int longerase);
static int sawritefilemarks(struct cam_periph *periph,
int nmarks, int setmarks);
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
static int sardpos(struct cam_periph *periph, int, u_int32_t *);
static int sasetpos(struct cam_periph *periph, int, u_int32_t *);
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 physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
#ifndef SA_DEFAULT_IO_SPLIT
#define SA_DEFAULT_IO_SPLIT 0
#endif
static int sa_allow_io_split = SA_DEFAULT_IO_SPLIT;
/*
* Tunable to allow the user to set a global allow_io_split value. Note
* that this WILL GO AWAY in FreeBSD 11.0. Silently splitting the I/O up
* is bad behavior, because it hides the true tape block size from the
* application.
*/
static SYSCTL_NODE(_kern_cam, OID_AUTO, sa, CTLFLAG_RD, 0,
"CAM Sequential Access Tape Driver");
Pull in r267961 and r267973 again. Fix for issues reported will follow.
2014-06-28 03:56:17 +00:00
SYSCTL_INT(_kern_cam_sa, OID_AUTO, allow_io_split, CTLFLAG_RDTUN,
&sa_allow_io_split, 0, "Default I/O split value");
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +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 struct periph_driver sadriver =
{
sainit, "sa",
TAILQ_HEAD_INITIALIZER(sadriver.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(sa, sadriver);
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
/* For 2.2-stable support */
#ifndef D_TAPE
#define D_TAPE 0
#endif
This commit should be a extensive NO-OP: Reformat and initialize correctly all "struct cdevsw". Initialize the d_maj and d_bmaj fields. The d_reset field was not removed, although it is never used. I used a program to do most of this, so all the files now use the same consistent format. Please keep it that way. Vinum and i4b not modified, patches emailed to respective authors.
1999-05-30 16:53:49 +00:00
static struct cdevsw sa_cdevsw = {
Device megapatch 4/6: Introduce d_version field in struct cdevsw, this must always be initialized to D_VERSION. Flip sense of D_NOGIANT flag to D_NEEDGIANT, this involves removing four D_NOGIANT flags and adding 145 D_NEEDGIANT flags.
2004-02-21 21:10:55 +00:00
.d_version = D_VERSION,
Gigacommit to improve device-driver source compatibility between branches: Initialize struct cdevsw using C99 sparse initializtion and remove all initializations to default values. This patch is automatically generated and has been tested by compiling LINT with all the fields in struct cdevsw in reverse order on alpha, sparc64 and i386. Approved by: re(scottl)
2003-03-03 12:15:54 +00:00
.d_open = saopen,
.d_close = saclose,
.d_read = physread,
.d_write = physwrite,
.d_ioctl = saioctl,
.d_strategy = sastrategy,
.d_name = "sa",
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
.d_flags = D_TAPE,
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
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
saopen(struct cdev *dev, int flags, int fmt, struct thread *td)
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 sa_softc *softc;
int error;
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
periph = (struct cam_periph *)dev->si_drv1;
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) {
return (ENXIO);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +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);
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 sa_softc *)periph->softc;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE|CAM_DEBUG_INFO,
Get rid of the device index number stored in the sa(4) unit number. The device index number stored in the unit number of sa(4) devices is only used to print debug messages. Get rid of this index number and use devtoname() to just print the entire device name.
2009-04-20 10:40:42 +00:00
("saopen(%s): softc=0x%x\n", devtoname(dev), softc->flags));
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
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
if (SA_IS_CTRL(dev)) {
softc->ctrl_mode = 1;
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
cam_periph_unlock(periph);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
return (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
if ((error = cam_periph_hold(periph, PRIBIO|PCATCH)) != 0) {
cam_periph_unlock(periph);
cam_periph_release(periph);
return (error);
}
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
if (softc->flags & SA_FLAG_OPEN) {
error = EBUSY;
} else if (softc->flags & SA_FLAG_INVALID) {
error = ENXIO;
} else {
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
/*
* Preserve whether this is a read_only open.
*/
softc->open_rdonly = (flags & O_RDWR) == O_RDONLY;
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
/*
* The function samount ensures media is loaded and ready.
* It also does a device RESERVE if the tape isn't yet mounted.
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
*
* If the mount fails and this was a non-blocking open,
* make this a 'open_pending_mount' action.
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
*/
error = samount(periph, flags, dev);
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
if (error && (flags & O_NONBLOCK)) {
softc->flags |= SA_FLAG_OPEN;
softc->open_pending_mount = 1;
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);
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
cam_periph_unlock(periph);
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
}
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
if (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
cam_periph_unhold(periph);
cam_periph_unlock(periph);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
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 (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
}
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
saprevent(periph, PR_PREVENT);
softc->flags |= SA_FLAG_OPEN;
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);
return (error);
}
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
saclose(struct cdev *dev, int flag, int fmt, struct thread *td)
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 sa_softc *softc;
Get rid of the device index number stored in the sa(4) unit number. The device index number stored in the unit number of sa(4) devices is only used to print debug messages. Get rid of this index number and use devtoname() to just print the entire device name.
2009-04-20 10:40:42 +00:00
int mode, error, writing, tmp;
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
int closedbits = SA_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
mode = SAMODE(dev);
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
periph = (struct cam_periph *)dev->si_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)
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);
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 sa_softc *)periph->softc;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE|CAM_DEBUG_INFO,
Get rid of the device index number stored in the sa(4) unit number. The device index number stored in the unit number of sa(4) devices is only used to print debug messages. Get rid of this index number and use devtoname() to just print the entire device name.
2009-04-20 10:40:42 +00:00
("saclose(%s): softc=0x%x\n", devtoname(dev), softc->flags));
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
softc->open_rdonly = 0;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
if (SA_IS_CTRL(dev)) {
softc->ctrl_mode = 0;
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
cam_periph_unlock(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_release(periph);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
return (0);
}
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
if (softc->open_pending_mount) {
softc->flags &= ~SA_FLAG_OPEN;
softc->open_pending_mount = 0;
cam_periph_unlock(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_release(periph);
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
return (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
if ((error = cam_periph_hold(periph, PRIBIO)) != 0) {
cam_periph_unlock(periph);
return (error);
}
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
* Were we writing the tape?
*/
writing = (softc->flags & SA_FLAG_TAPE_WRITTEN) != 0;
/*
* See whether or not we need to write filemarks. If this
* fails, we probably have to assume we've lost tape
* position.
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
*/
error = sacheckeod(periph);
if (error) {
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,
"failed to write terminating filemark(s)\n");
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->flags |= SA_FLAG_TAPE_FROZEN;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +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
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
* Whatever we end up doing, allow users to eject tapes from here on.
*/
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
saprevent(periph, PR_ALLOW);
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
* Decide how to end...
*/
Don't clear the SA_FLAG_TAPE_LOCKED prior to calling saprevent when we're doing an OFFLINE ioctl- saprevent won't unlock the door, which then causes the unload to fail on some units. If we've already unloaded the tape, don't try and rewind and unload in saclose. This is a slightly riskier change because we're now going to depend on SA_FLAG_TAPE_MOUNTED to say whether we've really unloaded the tape. This involved changing the setting in sadone for tape errors to SA_FLAG_TAPE_FROZEN (which is more accurate anyway-if you get an EIO you've probably lost tape position anyway) where it used to just clear the mounted flag.
1999-11-21 20:23:58 +00:00
if ((softc->flags & SA_FLAG_TAPE_MOUNTED) == 0) {
closedbits |= SA_FLAG_TAPE_FROZEN;
} else switch (mode) {
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 SA_MODE_OFFLINE:
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/*
* An 'offline' close is an unconditional release of
* frozen && mount conditions, irrespective of whether
* these operations succeeded. The reason for this is
* to allow at least some kind of programmatic way
* around our state getting all fouled up. If somebody
* issues an 'offline' command, that will be allowed
* to clear state.
*/
(void) sarewind(periph);
(void) saloadunload(periph, FALSE);
closedbits |= SA_FLAG_TAPE_MOUNTED|SA_FLAG_TAPE_FROZEN;
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;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
case SA_MODE_REWIND:
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/*
* If the rewind fails, return an error- if anyone cares,
* but not overwriting any previous error.
*
* We don't clear the notion of mounted here, but we do
* clear the notion of frozen if we successfully rewound.
*/
tmp = sarewind(periph);
if (tmp) {
if (error != 0)
error = tmp;
} else {
closedbits |= SA_FLAG_TAPE_FROZEN;
}
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +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
case SA_MODE_NOREWIND:
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
* If we're not rewinding/unloading the tape, find out
* whether we need to back up over one of two filemarks
* we wrote (if we wrote two filemarks) so that appends
* from this point on will be sane.
*/
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (error == 0 && writing && (softc->quirks & SA_QUIRK_2FM)) {
tmp = saspace(periph, -1, SS_FILEMARKS);
if (tmp) {
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, "unable to backspace "
"over one of double filemarks at end of "
"tape\n");
xpt_print(periph->path, "it is possible that "
"this device needs a SA_QUIRK_1FM quirk set"
"for it\n");
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->flags |= SA_FLAG_TAPE_FROZEN;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +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
break;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
default:
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, "unknown mode 0x%x in saclose\n", mode);
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/* NOTREACHED */
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
}
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
* We wish to note here that there are no more filemarks to be written.
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
*/
softc->filemarks = 0;
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
softc->flags &= ~SA_FLAG_TAPE_WRITTEN;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
* And we are no longer open for business.
*/
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
softc->flags &= ~closedbits;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/*
* Inform users if tape state if frozen....
*/
if (softc->flags & SA_FLAG_TAPE_FROZEN) {
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, "tape is now frozen- use an OFFLINE, "
"REWIND or MTEOM command to clear this state.\n");
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +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
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
/* release the device if it is no longer mounted */
if ((softc->flags & SA_FLAG_TAPE_MOUNTED) == 0)
sareservereleaseunit(periph, 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
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);
cam_periph_release(periph);
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +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
}
/*
* 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
sastrategy(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 sa_softc *softc;
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
bp->bio_resid = bp->bio_bcount;
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
if (SA_IS_CTRL(bp->bio_dev)) {
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, EINVAL);
return;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
}
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
periph = (struct cam_periph *)bp->bio_dev->si_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
}
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
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 sa_softc *)periph->softc;
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
if (softc->flags & SA_FLAG_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;
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
}
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (softc->flags & SA_FLAG_TAPE_FROZEN) {
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, EPERM);
return;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
}
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
/*
* This should actually never occur as the write(2)
* system call traps attempts to write to a read-only
* file descriptor.
*/
if (bp->bio_cmd == BIO_WRITE && softc->open_rdonly) {
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);
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
biofinish(bp, NULL, EBADF);
return;
}
if (softc->open_pending_mount) {
int error = samount(periph, 0, bp->bio_dev);
if (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
cam_periph_unlock(periph);
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
biofinish(bp, NULL, ENXIO);
return;
}
saprevent(periph, PR_PREVENT);
softc->open_pending_mount = 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
/*
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
* If it's a null transfer, return immediately
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
*/
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
if (bp->bio_bcount == 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);
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
biodone(bp);
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
/* valid request? */
if (softc->flags & SA_FLAG_FIXED) {
/*
* Fixed block device. The byte count must
* be a multiple of our block size.
*/
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
if (((softc->blk_mask != ~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_bcount & softc->blk_mask) != 0)) ||
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
((softc->blk_mask == ~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_bcount % softc->min_blk) != 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, "Invalid request. Fixed block "
"device requests must be a multiple of %d bytes\n",
softc->min_blk);
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, EINVAL);
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
}
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
} else if ((bp->bio_bcount > softc->max_blk) ||
(bp->bio_bcount < softc->min_blk) ||
(bp->bio_bcount & softc->blk_mask) != 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_print_path(periph->path);
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
printf("Invalid request. Variable block "
"device requests must be ");
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->blk_mask != 0) {
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
printf("a multiple of %d ", (0x1 << softc->blk_gran));
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
}
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
printf("between %d and %d bytes\n", softc->min_blk,
softc->max_blk);
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, EINVAL);
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
}
/*
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
* Place it at the end of the 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
*/
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_insert_tail(&softc->bio_queue, bp);
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->queue_count++;
Check in some months long pending minor debug output changes. Clarify that the implicit fallthrough was *not* intentional (thanks, Poul!) and reorganize the code so a correct fallthrough (with /* FALLTHROUGH */) occurs.
2003-06-02 00:37:42 +00:00
#if 0
CAM_DEBUG(periph->path, CAM_DEBUG_INFO,
("sastrategy: queuing a %ld %s byte %s\n", bp->bio_bcount,
(softc->flags & SA_FLAG_FIXED)? "fixed" : "variable",
(bp->bio_cmd == BIO_READ)? "read" : "write"));
#endif
if (softc->queue_count > 1) {
CAM_DEBUG(periph->path, CAM_DEBUG_INFO,
("sastrategy: queue count now %d\n", softc->queue_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
/*
* Schedule ourselves for performing the work.
*/
Replace most of priority numbers with defines. No logical changes.
2009-10-23 08:27:55 +00:00
xpt_schedule(periph, CAM_PRIORITY_NORMAL);
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;
}
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
#define PENDING_MOUNT_CHECK(softc, periph, dev) \
if (softc->open_pending_mount) { \
error = samount(periph, 0, dev); \
if (error) { \
break; \
} \
saprevent(periph, PR_PREVENT); \
softc->open_pending_mount = 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
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
saioctl(struct cdev *dev, u_long cmd, caddr_t arg, int flag, struct thread *td)
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 sa_softc *softc;
A minor tweak to shut up gcc (on the Alpha) for two (false) warnings. How it can think that something that is initialized at declaration can be used while uninitialized is beyond me.
1999-07-03 16:29:32 +00:00
scsi_space_code spaceop;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
int didlockperiph = 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
int mode;
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +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
mode = SAMODE(dev);
A minor tweak to shut up gcc (on the Alpha) for two (false) warnings. How it can think that something that is initialized at declaration can be used while uninitialized is beyond me.
1999-07-03 16:29:32 +00:00
error = 0; /* shut up gcc */
spaceop = 0; /* shut up gcc */
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 usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
periph = (struct cam_periph *)dev->si_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)
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);
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 sa_softc *)periph->softc;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
/*
* Check for control mode accesses. We allow MTIOCGET and
* MTIOCERRSTAT (but need to be the only one open in order
* to clear latched status), and MTSETBSIZE, MTSETDNSTY
* and MTCOMP (but need to be the only one accessing this
* device to run those).
*/
if (SA_IS_CTRL(dev)) {
switch (cmd) {
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
case MTIOCGETEOTMODEL:
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
case MTIOCGET:
break;
case MTIOCERRSTAT:
/*
* If the periph isn't already locked, lock it
* so our MTIOCERRSTAT can reset latched error stats.
*
* If the periph is already locked, skip it because
* we're just getting status and it'll be up to the
* other thread that has this device open to do
* an MTIOCERRSTAT that would clear latched status.
*/
if ((periph->flags & CAM_PERIPH_LOCKED) == 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
error = cam_periph_hold(periph, PRIBIO|PCATCH);
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
if (error != 0) {
cam_periph_unlock(periph);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
return (error);
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
}
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
didlockperiph = 1;
}
break;
Make sure we look at the correct sub op codes when deciding whether it's an operation we can perform via the control device. PR: kern/72010 MFC after: 1 week
2005-04-14 04:51:18 +00:00
case MTIOCTOP:
{
struct mtop *mt = (struct mtop *) arg;
/*
* Check to make sure it's an OP we can perform
* with no media inserted.
*/
switch (mt->mt_op) {
case MTSETBSIZ:
case MTSETDNSTY:
case MTCOMP:
mt = NULL;
/* FALLTHROUGH */
default:
break;
}
if (mt != NULL) {
break;
}
/* FALLTHROUGH */
}
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
case MTIOCSETEOTMODEL:
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
/*
* We need to acquire the peripheral here rather
* than at open time because we are sharing writable
* access to data structures.
*/
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
error = cam_periph_hold(periph, PRIBIO|PCATCH);
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
if (error != 0) {
cam_periph_unlock(periph);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
return (error);
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
}
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
didlockperiph = 1;
break;
default:
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
cam_periph_unlock(periph);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +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
/*
* Find the device that the user is talking about
*/
switch (cmd) {
case MTIOCGET:
{
struct mtget *g = (struct mtget *)arg;
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
/*
* If this isn't the control mode device, actually go out
* and ask the drive again what it's set to.
*/
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
if (!SA_IS_CTRL(dev) && !softc->open_pending_mount) {
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
u_int8_t write_protect;
int comp_enabled, comp_supported;
error = sagetparams(periph, SA_PARAM_ALL,
&softc->media_blksize, &softc->media_density,
&softc->media_numblks, &softc->buffer_mode,
&write_protect, &softc->speed, &comp_supported,
&comp_enabled, &softc->comp_algorithm, NULL);
if (error)
break;
if (write_protect)
softc->flags |= SA_FLAG_TAPE_WP;
else
softc->flags &= ~SA_FLAG_TAPE_WP;
softc->flags &= ~(SA_FLAG_COMP_SUPP|
SA_FLAG_COMP_ENABLED|SA_FLAG_COMP_UNSUPP);
if (comp_supported) {
if (softc->saved_comp_algorithm == 0)
softc->saved_comp_algorithm =
softc->comp_algorithm;
softc->flags |= SA_FLAG_COMP_SUPP;
if (comp_enabled)
softc->flags |= SA_FLAG_COMP_ENABLED;
} else
softc->flags |= SA_FLAG_COMP_UNSUPP;
}
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
bzero(g, sizeof(struct mtget));
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
g->mt_type = MT_ISAR;
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 & SA_FLAG_COMP_UNSUPP) {
g->mt_comp = MT_COMP_UNSUPP;
g->mt_comp0 = MT_COMP_UNSUPP;
g->mt_comp1 = MT_COMP_UNSUPP;
g->mt_comp2 = MT_COMP_UNSUPP;
g->mt_comp3 = MT_COMP_UNSUPP;
} else {
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if ((softc->flags & SA_FLAG_COMP_ENABLED) == 0) {
g->mt_comp = MT_COMP_DISABLED;
} else {
g->mt_comp = softc->comp_algorithm;
}
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
g->mt_comp0 = softc->comp_algorithm;
g->mt_comp1 = softc->comp_algorithm;
g->mt_comp2 = softc->comp_algorithm;
g->mt_comp3 = softc->comp_algorithm;
}
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
g->mt_density = softc->media_density;
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
g->mt_density0 = softc->media_density;
g->mt_density1 = softc->media_density;
g->mt_density2 = softc->media_density;
g->mt_density3 = softc->media_density;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
g->mt_blksiz = softc->media_blksize;
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
g->mt_blksiz0 = softc->media_blksize;
g->mt_blksiz1 = softc->media_blksize;
g->mt_blksiz2 = softc->media_blksize;
g->mt_blksiz3 = softc->media_blksize;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
g->mt_fileno = softc->fileno;
g->mt_blkno = softc->blkno;
g->mt_dsreg = (short) softc->dsreg;
Make some attempt to accomodate kern/24221- propate residual to mt_resid even if it might overflow. If we do a read or set of hardware or logical block position, we also clear Tape Frozen status.
2001-01-19 21:08:15 +00:00
/*
* Yes, we know that this is likely to overflow
*/
if (softc->last_resid_was_io) {
if ((g->mt_resid = (short) softc->last_io_resid) != 0) {
if (SA_IS_CTRL(dev) == 0 || didlockperiph) {
softc->last_io_resid = 0;
}
}
} else {
if ((g->mt_resid = (short)softc->last_ctl_resid) != 0) {
if (SA_IS_CTRL(dev) == 0 || didlockperiph) {
softc->last_ctl_resid = 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
error = 0;
break;
}
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
case MTIOCERRSTAT:
{
struct scsi_tape_errors *sep =
&((union mterrstat *)arg)->scsi_errstat;
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE,
("saioctl: MTIOCERRSTAT\n"));
bzero(sep, sizeof(*sep));
sep->io_resid = softc->last_io_resid;
bcopy((caddr_t) &softc->last_io_sense, sep->io_sense,
sizeof (sep->io_sense));
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
bcopy((caddr_t) &softc->last_io_cdb, sep->io_cdb,
sizeof (sep->io_cdb));
sep->ctl_resid = softc->last_ctl_resid;
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
bcopy((caddr_t) &softc->last_ctl_sense, sep->ctl_sense,
sizeof (sep->ctl_sense));
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
bcopy((caddr_t) &softc->last_ctl_cdb, sep->ctl_cdb,
sizeof (sep->ctl_cdb));
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
if ((SA_IS_CTRL(dev) == 0 && softc->open_pending_mount) ||
didlockperiph)
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
bzero((caddr_t) &softc->errinfo,
sizeof (softc->errinfo));
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
error = 0;
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
case MTIOCTOP:
{
struct mtop *mt;
int count;
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
PENDING_MOUNT_CHECK(softc, periph, 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
mt = (struct mtop *)arg;
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +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
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE,
("saioctl: op=0x%x count=0x%x\n",
mt->mt_op, mt->mt_count));
count = mt->mt_count;
switch (mt->mt_op) {
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
case MTWEOF: /* write an end-of-file marker */
scsi_sa.h
2000-10-31 22:34:51 +00:00
/*
* We don't need to clear the SA_FLAG_TAPE_WRITTEN
* flag because by keeping track of filemarks
* we have last written we know ehether or not
* we need to write more when we close the device.
*/
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
error = sawritefilemarks(periph, count, 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
break;
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
case MTWSS: /* write a setmark */
error = sawritefilemarks(periph, count, TRUE);
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
case MTBSR: /* backward space record */
case MTFSR: /* forward space record */
case MTBSF: /* backward space file */
case MTFSF: /* forward space file */
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
case MTBSS: /* backward space setmark */
case MTFSS: /* forward space setmark */
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 MTEOD: /* space to end of recorded medium */
{
int nmarks;
A minor tweak to shut up gcc (on the Alpha) for two (false) warnings. How it can think that something that is initialized at declaration can be used while uninitialized is beyond me.
1999-07-03 16:29:32 +00:00
spaceop = SS_FILEMARKS;
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
nmarks = softc->filemarks;
error = sacheckeod(periph);
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
if (error) {
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,
"EOD check prior to spacing failed\n");
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
softc->flags |= SA_FLAG_EIO_PENDING;
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
nmarks -= softc->filemarks;
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
switch(mt->mt_op) {
case MTBSR:
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
count = -count;
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
/* FALLTHROUGH */
case MTFSR:
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
spaceop = SS_BLOCKS;
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
break;
case MTBSF:
count = -count;
/* FALLTHROUGH */
case MTFSF:
break;
case MTBSS:
count = -count;
/* FALLTHROUGH */
case MTFSS:
spaceop = SS_SETMARKS;
break;
case MTEOD:
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
spaceop = SS_EOD;
count = 0;
nmarks = 0;
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
break;
default:
error = EINVAL;
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 a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
if (error)
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
nmarks = softc->filemarks;
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
/*
* XXX: Why are we checking again?
*/
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
error = sacheckeod(periph);
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
if (error)
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
nmarks -= softc->filemarks;
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
error = saspace(periph, count - nmarks, spaceop);
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
* At this point, clear that we've written the tape
* and that we've written any filemarks. We really
* don't know what the applications wishes to do next-
* the sacheckeod's will make sure we terminated the
* tape correctly if we'd been writing, but the next
* action the user application takes will set again
* whether we need to write filemarks.
*/
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->flags &=
~(SA_FLAG_TAPE_WRITTEN|SA_FLAG_TAPE_FROZEN);
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
softc->filemarks = 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
break;
}
case MTREW: /* rewind */
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
PENDING_MOUNT_CHECK(softc, periph, dev);
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
(void) sacheckeod(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
error = sarewind(periph);
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/* see above */
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
softc->flags &=
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
~(SA_FLAG_TAPE_WRITTEN|SA_FLAG_TAPE_FROZEN);
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
softc->flags &= ~SA_FLAG_ERR_PENDING;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
softc->filemarks = 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
break;
case MTERASE: /* erase */
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
PENDING_MOUNT_CHECK(softc, periph, 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
error = saerase(periph, count);
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->flags &=
~(SA_FLAG_TAPE_WRITTEN|SA_FLAG_TAPE_FROZEN);
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
softc->flags &= ~SA_FLAG_ERR_PENDING;
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 MTRETENS: /* re-tension tape */
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
PENDING_MOUNT_CHECK(softc, periph, 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
error = saretension(periph);
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->flags &=
~(SA_FLAG_TAPE_WRITTEN|SA_FLAG_TAPE_FROZEN);
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
softc->flags &= ~SA_FLAG_ERR_PENDING;
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 MTOFFL: /* rewind and put the drive offline */
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
PENDING_MOUNT_CHECK(softc, periph, dev);
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
(void) sacheckeod(periph);
/* see above */
softc->flags &= ~SA_FLAG_TAPE_WRITTEN;
softc->filemarks = 0;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
error = sarewind(periph);
Don't clear the SA_FLAG_TAPE_LOCKED prior to calling saprevent when we're doing an OFFLINE ioctl- saprevent won't unlock the door, which then causes the unload to fail on some units. If we've already unloaded the tape, don't try and rewind and unload in saclose. This is a slightly riskier change because we're now going to depend on SA_FLAG_TAPE_MOUNTED to say whether we've really unloaded the tape. This involved changing the setting in sadone for tape errors to SA_FLAG_TAPE_FROZEN (which is more accurate anyway-if you get an EIO you've probably lost tape position anyway) where it used to just clear the mounted flag.
1999-11-21 20:23:58 +00:00
/* clear the frozen flag anyway */
softc->flags &= ~SA_FLAG_TAPE_FROZEN;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +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
/*
Don't clear the SA_FLAG_TAPE_LOCKED prior to calling saprevent when we're doing an OFFLINE ioctl- saprevent won't unlock the door, which then causes the unload to fail on some units. If we've already unloaded the tape, don't try and rewind and unload in saclose. This is a slightly riskier change because we're now going to depend on SA_FLAG_TAPE_MOUNTED to say whether we've really unloaded the tape. This involved changing the setting in sadone for tape errors to SA_FLAG_TAPE_FROZEN (which is more accurate anyway-if you get an EIO you've probably lost tape position anyway) where it used to just clear the mounted flag.
1999-11-21 20:23:58 +00:00
* Be sure to allow media removal before ejecting.
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
saprevent(periph, PR_ALLOW);
Don't clear the SA_FLAG_TAPE_LOCKED prior to calling saprevent when we're doing an OFFLINE ioctl- saprevent won't unlock the door, which then causes the unload to fail on some units. If we've already unloaded the tape, don't try and rewind and unload in saclose. This is a slightly riskier change because we're now going to depend on SA_FLAG_TAPE_MOUNTED to say whether we've really unloaded the tape. This involved changing the setting in sadone for tape errors to SA_FLAG_TAPE_FROZEN (which is more accurate anyway-if you get an EIO you've probably lost tape position anyway) where it used to just clear the mounted flag.
1999-11-21 20:23:58 +00:00
if (error == 0) {
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
error = saloadunload(periph, FALSE);
Don't clear the SA_FLAG_TAPE_LOCKED prior to calling saprevent when we're doing an OFFLINE ioctl- saprevent won't unlock the door, which then causes the unload to fail on some units. If we've already unloaded the tape, don't try and rewind and unload in saclose. This is a slightly riskier change because we're now going to depend on SA_FLAG_TAPE_MOUNTED to say whether we've really unloaded the tape. This involved changing the setting in sadone for tape errors to SA_FLAG_TAPE_FROZEN (which is more accurate anyway-if you get an EIO you've probably lost tape position anyway) where it used to just clear the mounted flag.
1999-11-21 20:23:58 +00:00
if (error == 0) {
softc->flags &= ~SA_FLAG_TAPE_MOUNTED;
}
}
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;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +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
case MTNOP: /* no operation, sets status only */
case MTCACHE: /* enable controller cache */
case MTNOCACHE: /* disable controller cache */
error = 0;
break;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +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
case MTSETBSIZ: /* Set block size for device */
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
PENDING_MOUNT_CHECK(softc, periph, 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
error = sasetparams(periph, SA_PARAM_BLOCKSIZE, count,
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
0, 0, 0);
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
if (error == 0) {
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
softc->last_media_blksize =
softc->media_blksize;
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
softc->media_blksize = count;
if (count) {
softc->flags |= SA_FLAG_FIXED;
if (powerof2(count)) {
softc->blk_shift =
ffs(count) - 1;
softc->blk_mask = count - 1;
} else {
softc->blk_mask = ~0;
softc->blk_shift = 0;
}
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
* Make the user's desire 'persistent'.
*/
softc->quirks &= ~SA_QUIRK_VARIABLE;
softc->quirks |= SA_QUIRK_FIXED;
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
} else {
softc->flags &= ~SA_FLAG_FIXED;
if (softc->max_blk == 0) {
softc->max_blk = ~0;
}
softc->blk_shift = 0;
if (softc->blk_gran != 0) {
softc->blk_mask =
softc->blk_gran - 1;
} else {
softc->blk_mask = 0;
}
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
* Make the user's desire 'persistent'.
*/
softc->quirks |= SA_QUIRK_VARIABLE;
softc->quirks &= ~SA_QUIRK_FIXED;
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +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
break;
case MTSETDNSTY: /* Set density for device and mode */
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
PENDING_MOUNT_CHECK(softc, periph, 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
if (count > UCHAR_MAX) {
error = EINVAL;
break;
} else {
error = sasetparams(periph, SA_PARAM_DENSITY,
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
0, count, 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
}
break;
case MTCOMP: /* enable compression */
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
PENDING_MOUNT_CHECK(softc, periph, 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
/*
* Some devices don't support compression, and
* don't like it if you ask them for the
* compression page.
*/
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
if ((softc->quirks & SA_QUIRK_NOCOMP) ||
(softc->flags & SA_FLAG_COMP_UNSUPP)) {
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
error = ENODEV;
break;
}
error = sasetparams(periph, SA_PARAM_COMPRESSION,
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
0, 0, count, 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
break;
default:
error = EINVAL;
}
break;
}
case MTIOCIEOT:
case MTIOCEEOT:
error = 0;
break;
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
case MTIOCRDSPOS:
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
PENDING_MOUNT_CHECK(softc, periph, dev);
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
error = sardpos(periph, 0, (u_int32_t *) arg);
break;
case MTIOCRDHPOS:
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
PENDING_MOUNT_CHECK(softc, periph, dev);
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
error = sardpos(periph, 1, (u_int32_t *) arg);
break;
case MTIOCSLOCATE:
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
PENDING_MOUNT_CHECK(softc, periph, dev);
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
error = sasetpos(periph, 0, (u_int32_t *) arg);
break;
case MTIOCHLOCATE:
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
PENDING_MOUNT_CHECK(softc, periph, dev);
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
error = sasetpos(periph, 1, (u_int32_t *) arg);
break;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
case MTIOCGETEOTMODEL:
error = 0;
if (softc->quirks & SA_QUIRK_1FM)
mode = 1;
else
mode = 2;
*((u_int32_t *) arg) = mode;
break;
case MTIOCSETEOTMODEL:
error = 0;
switch (*((u_int32_t *) arg)) {
case 1:
softc->quirks &= ~SA_QUIRK_2FM;
softc->quirks |= SA_QUIRK_1FM;
break;
case 2:
softc->quirks &= ~SA_QUIRK_1FM;
softc->quirks |= SA_QUIRK_2FM;
break;
default:
error = EINVAL;
break;
}
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:
error = cam_periph_ioctl(periph, cmd, arg, saerror);
break;
}
Make some attempt to accomodate kern/24221- propate residual to mt_resid even if it might overflow. If we do a read or set of hardware or logical block position, we also clear Tape Frozen status.
2001-01-19 21:08:15 +00:00
/*
* Check to see if we cleared a frozen state
*/
if (error == 0 && (softc->flags & SA_FLAG_TAPE_FROZEN)) {
switch(cmd) {
case MTIOCRDSPOS:
case MTIOCRDHPOS:
case MTIOCSLOCATE:
case MTIOCHLOCATE:
softc->fileno = (daddr_t) -1;
softc->blkno = (daddr_t) -1;
softc->flags &= ~SA_FLAG_TAPE_FROZEN;
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,
"tape state now unfrozen.\n");
Make some attempt to accomodate kern/24221- propate residual to mt_resid even if it might overflow. If we do a read or set of hardware or logical block position, we also clear Tape Frozen status.
2001-01-19 21:08:15 +00:00
break;
default:
break;
}
}
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
if (didlockperiph) {
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);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +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_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 (error);
}
static void
sainit(void)
{
cam_status status;
/*
* Install a global async callback.
*/
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, saasync, 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("sa: Failed to attach master async callback "
"due to status 0x%x!\n", status);
}
}
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
saoninvalidate(struct cam_periph *periph)
{
struct sa_softc *softc;
softc = (struct sa_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, saasync, 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 |= SA_FLAG_INVALID;
/*
* Return all queued I/O with ENXIO.
* XXX Handle any transactions queued to the card
* with XPT_ABORT_CCB.
*/
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 a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->queue_count = 0;
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
sacleanup(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 sa_softc *softc;
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
int i;
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 = (struct sa_softc *)periph->softc;
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
devstat_remove_entry(softc->device_stats);
Don't call destroy_dev(9) with a mutex held. While here, shuffle things around so the periph destructors look alike. Based on a patch by Jaakko Heinonen. Submitted by: Jaakko Heinonen Reviewed by: scottl Approved by: rwatson (mentor) Sponsored by: FreeBSD Foundation
2009-01-10 17:22:49 +00:00
cam_periph_unlock(periph);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
destroy_dev(softc->devs.ctl_dev);
for (i = 0; i < SA_NUM_MODES; i++) {
destroy_dev(softc->devs.mode_devs[i].r_dev);
destroy_dev(softc->devs.mode_devs[i].nr_dev);
destroy_dev(softc->devs.mode_devs[i].er_dev);
}
Don't call destroy_dev(9) with a mutex held. While here, shuffle things around so the periph destructors look alike. Based on a patch by Jaakko Heinonen. Submitted by: Jaakko Heinonen Reviewed by: scottl Approved by: rwatson (mentor) Sponsored by: FreeBSD Foundation
2009-01-10 17:22:49 +00:00
cam_periph_lock(periph);
- Providing fine-grained malloc statistic by replacing M_DEVBUF with module-specific malloc types. These should help us to pinpoint the possible memory leakage in the future. - Implementing xpt_alloc_ccb_nowait() and replacing all malloc/free based CCB memory management with xpt_alloc_ccb[_nowait]/xpt_free_ccb. Hopefully this would be helpful if someday we move the CCB allocator to use UMA instead of malloc(). Encouraged by: jeffr, rwatson Reviewed by: gibbs, scottl Approved by: re (scottl)
2005-07-01 15:21:30 +00:00
free(softc, M_SCSISA);
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
saasync(void *callback_arg, u_int32_t code,
struct cam_path *path, void *arg)
{
struct cam_periph *periph;
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_SEQUENTIAL)
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(saregister, saoninvalidate,
sacleanup, sastart,
Merge CAM locking changes from the projects/camlock branch to radically reduce lock congestion and improve SMP scalability of the SCSI/ATA stack, preparing the ground for the coming next GEOM direct dispatch support. Replace big per-SIM locks with bunch of smaller ones: - per-LUN locks to protect device and peripheral drivers state; - per-target locks to protect list of LUNs on target; - per-bus locks to protect reference counting; - per-send queue locks to protect queue of CCBs to be sent; - per-done queue locks to protect queue of completed CCBs; - remaining per-SIM locks now protect only HBA driver internals. While holding LUN lock it is allowed (while not recommended for performance reasons) to take SIM lock. The opposite acquisition order is forbidden. All the other locks are leaf locks, that can be taken anywhere, but should not be cascaded. Many functions, such as: xpt_action(), xpt_done(), xpt_async(), xpt_create_path(), etc. are no longer require (but allow) SIM lock to be held. To keep compatibility and solve cases where SIM lock can't be dropped, all xpt_async() calls in addition to xpt_done() calls are queued to completion threads for async processing in clean environment without SIM lock held. Instead of single CAM SWI thread, used for commands completion processing before, use multiple (depending on number of CPUs) threads. Load balanced between them using "hash" of the device B:T:L address. HBA drivers that can drop SIM lock during completion processing and have sufficient number of completion threads to efficiently scale to multiple CPUs can use new function xpt_done_direct() to avoid extra context switch. Make ahci(4) driver to use this mechanism depending on hardware setup. Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-21 12:00:26 +00:00
"sa", CAM_PERIPH_BIO, path,
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
saasync, AC_FOUND_DEVICE, cgd);
if (status != CAM_REQ_CMP
&& status != CAM_REQ_INPROG)
printf("saasync: Unable to probe new device "
"due to status 0x%x\n", status);
break;
}
default:
Add a default async handler funstion to cam_periph.c to remove duplicated code in all initiator type peripheral drivers. scsi_target.c: Release ATIO structures that wind up in the 'unkown command queue' for consumption by our userland counterpart, back to the controller when the exception for that command is cleared.
1999-05-22 22:00:24 +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
break;
}
}
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
static void
sasysctlinit(void *context, int pending)
{
struct cam_periph *periph;
struct sa_softc *softc;
char tmpstr[80], tmpstr2[80];
periph = (struct cam_periph *)context;
/*
* If the periph is invalid, no need to setup the sysctls.
*/
if (periph->flags & CAM_PERIPH_INVALID)
goto bailout;
softc = (struct sa_softc *)periph->softc;
snprintf(tmpstr, sizeof(tmpstr), "CAM SA unit %d", periph->unit_number);
snprintf(tmpstr2, sizeof(tmpstr2), "%u", periph->unit_number);
sysctl_ctx_init(&softc->sysctl_ctx);
softc->sysctl_tree = SYSCTL_ADD_NODE(&softc->sysctl_ctx,
SYSCTL_STATIC_CHILDREN(_kern_cam_sa), OID_AUTO, tmpstr2,
CTLFLAG_RD, 0, tmpstr);
if (softc->sysctl_tree == NULL)
goto bailout;
SYSCTL_ADD_INT(&softc->sysctl_ctx, SYSCTL_CHILDREN(softc->sysctl_tree),
Fix multiple incorrect SYSCTL arguments in the kernel: - Wrong integer type was specified. - Wrong or missing "access" specifier. The "access" specifier sometimes included the SYSCTL type, which it should not, except for procedural SYSCTL nodes. - Logical OR where binary OR was expected. - Properly assert the "access" argument passed to all SYSCTL macros, using the CTASSERT macro. This applies to both static- and dynamically created SYSCTLs. - Properly assert the the data type for both static and dynamic SYSCTLs. In the case of static SYSCTLs we only assert that the data pointed to by the SYSCTL data pointer has the correct size, hence there is no easy way to assert types in the C language outside a C-function. - Rewrote some code which doesn't pass a constant "access" specifier when creating dynamic SYSCTL nodes, which is now a requirement. - Updated "EXAMPLES" section in SYSCTL manual page. MFC after: 3 days Sponsored by: Mellanox Technologies
2014-10-21 07:31:21 +00:00
OID_AUTO, "allow_io_split", CTLFLAG_RDTUN | CTLFLAG_NOFETCH,
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
&softc->allow_io_split, 0, "Allow Splitting I/O");
SYSCTL_ADD_INT(&softc->sysctl_ctx, SYSCTL_CHILDREN(softc->sysctl_tree),
Fix multiple incorrect SYSCTL arguments in the kernel: - Wrong integer type was specified. - Wrong or missing "access" specifier. The "access" specifier sometimes included the SYSCTL type, which it should not, except for procedural SYSCTL nodes. - Logical OR where binary OR was expected. - Properly assert the "access" argument passed to all SYSCTL macros, using the CTASSERT macro. This applies to both static- and dynamically created SYSCTLs. - Properly assert the the data type for both static and dynamic SYSCTLs. In the case of static SYSCTLs we only assert that the data pointed to by the SYSCTL data pointer has the correct size, hence there is no easy way to assert types in the C language outside a C-function. - Rewrote some code which doesn't pass a constant "access" specifier when creating dynamic SYSCTL nodes, which is now a requirement. - Updated "EXAMPLES" section in SYSCTL manual page. MFC after: 3 days Sponsored by: Mellanox Technologies
2014-10-21 07:31:21 +00:00
OID_AUTO, "maxio", CTLFLAG_RD,
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
&softc->maxio, 0, "Maximum I/O size");
SYSCTL_ADD_INT(&softc->sysctl_ctx, SYSCTL_CHILDREN(softc->sysctl_tree),
Fix multiple incorrect SYSCTL arguments in the kernel: - Wrong integer type was specified. - Wrong or missing "access" specifier. The "access" specifier sometimes included the SYSCTL type, which it should not, except for procedural SYSCTL nodes. - Logical OR where binary OR was expected. - Properly assert the "access" argument passed to all SYSCTL macros, using the CTASSERT macro. This applies to both static- and dynamically created SYSCTLs. - Properly assert the the data type for both static and dynamic SYSCTLs. In the case of static SYSCTLs we only assert that the data pointed to by the SYSCTL data pointer has the correct size, hence there is no easy way to assert types in the C language outside a C-function. - Rewrote some code which doesn't pass a constant "access" specifier when creating dynamic SYSCTL nodes, which is now a requirement. - Updated "EXAMPLES" section in SYSCTL manual page. MFC after: 3 days Sponsored by: Mellanox Technologies
2014-10-21 07:31:21 +00:00
OID_AUTO, "cpi_maxio", CTLFLAG_RD,
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
&softc->cpi_maxio, 0, "Maximum Controller I/O size");
bailout:
/*
* Release the reference that was held when this task was enqueued.
*/
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
static cam_status
saregister(struct cam_periph *periph, void *arg)
{
struct sa_softc *softc;
struct ccb_getdev *cgd;
Make CAM report devices with ATA/SATA transport to devstat(9) as IDE.
2011-04-14 21:25:32 +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
caddr_t match;
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
char tmpstr[80];
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
int i;
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("saregister: no getdev CCB, can't register device\n");
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
return (CAM_REQ_CMP_ERR);
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 usage of M_ZERO to malloc calls where the result was just bzeroed. PR: 22186 Submitted (partially) by: josh@zipperup.org
2000-10-27 16:40:57 +00:00
softc = (struct sa_softc *)
- Providing fine-grained malloc statistic by replacing M_DEVBUF with module-specific malloc types. These should help us to pinpoint the possible memory leakage in the future. - Implementing xpt_alloc_ccb_nowait() and replacing all malloc/free based CCB memory management with xpt_alloc_ccb[_nowait]/xpt_free_ccb. Hopefully this would be helpful if someday we move the CCB allocator to use UMA instead of malloc(). Encouraged by: jeffr, rwatson Reviewed by: gibbs, scottl Approved by: re (scottl)
2005-07-01 15:21:30 +00:00
malloc(sizeof (*softc), M_SCSISA, 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("saregister: Unable to probe new device. "
"Unable to allocate softc\n");
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
return (CAM_REQ_CMP_ERR);
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
}
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
softc->scsi_rev = SID_ANSI_REV(&cgd->inq_data);
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->state = SA_STATE_NORMAL;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->fileno = (daddr_t) -1;
softc->blkno = (daddr_t) -1;
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);
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;
/*
* See if this device has any quirks.
*/
match = cam_quirkmatch((caddr_t)&cgd->inq_data,
(caddr_t)sa_quirk_table,
sizeof(sa_quirk_table)/sizeof(*sa_quirk_table),
sizeof(*sa_quirk_table), scsi_inquiry_match);
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
if (match != 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
softc->quirks = ((struct sa_quirk_entry *)match)->quirks;
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
softc->last_media_blksize =
((struct sa_quirk_entry *)match)->prefblk;
} 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
softc->quirks = SA_QUIRK_NONE;
Make CAM report devices with ATA/SATA transport to devstat(9) as IDE.
2011-04-14 21:25:32 +00:00
bzero(&cpi, sizeof(cpi));
xpt_setup_ccb(&cpi.ccb_h, periph->path, CAM_PRIORITY_NORMAL);
cpi.ccb_h.func_code = XPT_PATH_INQ;
xpt_action((union ccb *)&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
/*
Make CAM report devices with ATA/SATA transport to devstat(9) as IDE.
2011-04-14 21:25:32 +00:00
* The SA driver supports a blocksize, but we don't know the
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
* blocksize until we media is inserted. So, set a flag to
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
* indicate that the blocksize is unavailable right now.
*/
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
cam_periph_unlock(periph);
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
softc->device_stats = devstat_new_entry("sa", periph->unit_number, 0,
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
DEVSTAT_BS_UNAVAILABLE, SID_TYPE(&cgd->inq_data) |
Make CAM report devices with ATA/SATA transport to devstat(9) as IDE.
2011-04-14 21:25:32 +00:00
XPORT_DEVSTAT_TYPE(cpi.transport), DEVSTAT_PRIORITY_TAPE);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
/*
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
* Load the default value that is either compiled in, or loaded
* in the global kern.cam.sa.allow_io_split tunable.
*/
softc->allow_io_split = sa_allow_io_split;
/*
* Load a per-instance tunable, if it exists. NOTE that this
* tunable WILL GO AWAY in FreeBSD 11.0.
*/
snprintf(tmpstr, sizeof(tmpstr), "kern.cam.sa.%u.allow_io_split",
periph->unit_number);
TUNABLE_INT_FETCH(tmpstr, &softc->allow_io_split);
/*
* If maxio isn't set, we fall back to DFLTPHYS. Otherwise we take
* the smaller of cpi.maxio or MAXPHYS.
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
*/
if (cpi.maxio == 0)
softc->maxio = DFLTPHYS;
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
else if (cpi.maxio > MAXPHYS)
softc->maxio = MAXPHYS;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
else
softc->maxio = cpi.maxio;
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
/*
* Record the controller's maximum I/O size so we can report it to
* the user later.
*/
softc->cpi_maxio = cpi.maxio;
/*
* By default we tell physio that we do not want our I/O split.
* The user needs to have a 1:1 mapping between the size of his
* write to a tape character device and the size of the write
* that actually goes down to the drive.
*/
if (softc->allow_io_split == 0)
softc->si_flags = SI_NOSPLIT;
else
softc->si_flags = 0;
TASK_INIT(&softc->sysctl_task, 0, sasysctlinit, periph);
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
/*
* If the SIM supports unmapped I/O, let physio know that we can
* handle unmapped buffers.
*/
if (cpi.hba_misc & PIM_UNMAPPED)
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
softc->si_flags |= SI_UNMAPPED;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
softc->devs.ctl_dev = make_dev(&sa_cdevsw, SAMINOR(SA_CTLDEV,
Get rid of the device index number stored in the sa(4) unit number. The device index number stored in the unit number of sa(4) devices is only used to print debug messages. Get rid of this index number and use devtoname() to just print the entire device name.
2009-04-20 10:40:42 +00:00
0, SA_ATYPE_R), UID_ROOT, GID_OPERATOR,
Use the correct device names (now that we have devfs to embarrass us). PR: 25254
2001-02-21 17:29:01 +00:00
0660, "%s%d.ctl", periph->periph_name, periph->unit_number);
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
softc->devs.ctl_dev->si_drv1 = periph;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
softc->devs.ctl_dev->si_iosize_max = softc->maxio;
softc->devs.ctl_dev->si_flags |= softc->si_flags;
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
for (i = 0; i < SA_NUM_MODES; i++) {
Fix a couple of boo-boos in the last delta.
1999-11-17 17:11:21 +00:00
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
softc->devs.mode_devs[i].r_dev = make_dev(&sa_cdevsw,
Get rid of the device index number stored in the sa(4) unit number. The device index number stored in the unit number of sa(4) devices is only used to print debug messages. Get rid of this index number and use devtoname() to just print the entire device name.
2009-04-20 10:40:42 +00:00
SAMINOR(SA_NOT_CTLDEV, i, SA_ATYPE_R),
Use the correct device names (now that we have devfs to embarrass us). PR: 25254
2001-02-21 17:29:01 +00:00
UID_ROOT, GID_OPERATOR, 0660, "%s%d.%d",
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
periph->periph_name, periph->unit_number, i);
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
softc->devs.mode_devs[i].r_dev->si_drv1 = periph;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
softc->devs.mode_devs[i].r_dev->si_iosize_max = softc->maxio;
softc->devs.mode_devs[i].r_dev->si_flags |= softc->si_flags;
Fix a couple of boo-boos in the last delta.
1999-11-17 17:11:21 +00:00
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
softc->devs.mode_devs[i].nr_dev = make_dev(&sa_cdevsw,
Get rid of the device index number stored in the sa(4) unit number. The device index number stored in the unit number of sa(4) devices is only used to print debug messages. Get rid of this index number and use devtoname() to just print the entire device name.
2009-04-20 10:40:42 +00:00
SAMINOR(SA_NOT_CTLDEV, i, SA_ATYPE_NR),
Use the correct device names (now that we have devfs to embarrass us). PR: 25254
2001-02-21 17:29:01 +00:00
UID_ROOT, GID_OPERATOR, 0660, "n%s%d.%d",
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
periph->periph_name, periph->unit_number, i);
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
softc->devs.mode_devs[i].nr_dev->si_drv1 = periph;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
softc->devs.mode_devs[i].nr_dev->si_iosize_max = softc->maxio;
softc->devs.mode_devs[i].nr_dev->si_flags |= softc->si_flags;
Fix a couple of boo-boos in the last delta.
1999-11-17 17:11:21 +00:00
softc->devs.mode_devs[i].er_dev = make_dev(&sa_cdevsw,
Get rid of the device index number stored in the sa(4) unit number. The device index number stored in the unit number of sa(4) devices is only used to print debug messages. Get rid of this index number and use devtoname() to just print the entire device name.
2009-04-20 10:40:42 +00:00
SAMINOR(SA_NOT_CTLDEV, i, SA_ATYPE_ER),
Use the correct device names (now that we have devfs to embarrass us). PR: 25254
2001-02-21 17:29:01 +00:00
UID_ROOT, GID_OPERATOR, 0660, "e%s%d.%d",
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
periph->periph_name, periph->unit_number, i);
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
softc->devs.mode_devs[i].er_dev->si_drv1 = periph;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
softc->devs.mode_devs[i].er_dev->si_iosize_max = softc->maxio;
softc->devs.mode_devs[i].er_dev->si_flags |= softc->si_flags;
Follow the hints from PHK's new messages- only make_dev for a device once. Alias names use the make_dev_alias function.
2000-09-14 14:53:57 +00:00
/*
* Make the (well known) aliases for the first mode.
*/
if (i == 0) {
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
struct cdev *alias;
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
alias = make_dev_alias(softc->devs.mode_devs[i].r_dev,
Use the correct device names (now that we have devfs to embarrass us). PR: 25254
2001-02-21 17:29:01 +00:00
"%s%d", periph->periph_name, periph->unit_number);
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
alias->si_drv1 = periph;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
alias->si_iosize_max = softc->maxio;
alias->si_flags |= softc->si_flags;
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
alias = make_dev_alias(softc->devs.mode_devs[i].nr_dev,
Use the correct device names (now that we have devfs to embarrass us). PR: 25254
2001-02-21 17:29:01 +00:00
"n%s%d", periph->periph_name, periph->unit_number);
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
alias->si_drv1 = periph;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
alias->si_iosize_max = softc->maxio;
alias->si_flags |= softc->si_flags;
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
alias = make_dev_alias(softc->devs.mode_devs[i].er_dev,
Use the correct device names (now that we have devfs to embarrass us). PR: 25254
2001-02-21 17:29:01 +00:00
"e%s%d", periph->periph_name, periph->unit_number);
Remove usage of cam_extend.c, replace with dev->si_drv1 PR: kern/39809 Approved by: gibbs
2002-08-15 20:54:03 +00:00
alias->si_drv1 = periph;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
alias->si_iosize_max = softc->maxio;
alias->si_flags |= softc->si_flags;
Follow the hints from PHK's new messages- only make_dev for a device once. Alias names use the make_dev_alias function.
2000-09-14 14:53:57 +00:00
}
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
}
Avoid problems with make_dev.
2007-04-19 18:14:33 +00:00
cam_periph_lock(periph);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
Add support to physio(9) for devices that don't want I/O split and configure sa(4) to request no I/O splitting by default. For tape devices, the user needs to be able to clearly understand what blocksize is actually being used when writing to a tape device. The previous behavior of physio(9) was that it would split up any I/O that was too large for the device, or too large to fit into MAXPHYS. This means that if, for instance, the user wrote a 1MB block to a tape device, and MAXPHYS was 128KB, the 1MB write would be split into 8 128K chunks. This would be done without informing the user. This has suboptimal effects, especially when trying to communicate status to the user. In the event of an error writing to a tape (e.g. physical end of tape) in the middle of a 1MB block that has been split into 8 pieces, the user could have the first two 128K pieces written successfully, the third returned with an error, and the last 5 returned with 0 bytes written. If the user is using a standard write(2) system call, all he will see is the ENOSPC error. He won't have a clue how much actually got written. (With a writev(2) system call, he should be able to determine how much got written in addition to the error.) The solution is to prevent physio(9) from splitting the I/O. The new cdev flag, SI_NOSPLIT, tells physio that the driver does not want I/O to be split beforehand. Although the sa(4) driver now enables SI_NOSPLIT by default, that can be disabled by two loader tunables for now. It will not be configurable starting in FreeBSD 11.0. kern.cam.sa.allow_io_split allows the user to configure I/O splitting for all sa(4) driver instances. kern.cam.sa.%d.allow_io_split allows the user to configure I/O splitting for a specific sa(4) instance. There are also now three sa(4) driver sysctl variables that let the users see some sa(4) driver values. kern.cam.sa.%d.allow_io_split shows whether I/O splitting is turned on. kern.cam.sa.%d.maxio shows the maximum I/O size allowed by kernel configuration parameters (e.g. MAXPHYS, DFLTPHYS) and the capabilities of the controller. kern.cam.sa.%d.cpi_maxio shows the maximum I/O size supported by the controller. Note that a better long term solution would be to implement support for chaining buffers, so that that MAXPHYS is no longer a limiting factor for I/O size to tape and disk devices. At that point, the controller and the tape drive would become the limiting factors. sys/conf.h: Add a new cdev flag, SI_NOSPLIT, that allows a driver to tell physio not to split up I/O. sys/param.h: Bump __FreeBSD_version to 1000049 for the addition of the SI_NOSPLIT cdev flag. kern_physio.c: If the SI_NOSPLIT flag is set on the cdev, return any I/O that is larger than si_iosize_max or MAXPHYS, has more than one segment, or would have to be split because of misalignment with EFBIG. (File too large). In the event of an error, print a console message to give the user a clue about what happened. scsi_sa.c: Set the SI_NOSPLIT cdev flag on the devices created for the sa(4) driver by default. Add tunables to control whether we allow I/O splitting in physio(9). Explain in the comments that allowing I/O splitting will be deprecated for the sa(4) driver in FreeBSD 11.0. Add sysctl variables to display the maximum I/O size we can do (which could be further limited by read block limits) and the maximum I/O size that the controller can do. Limit our maximum I/O size (recorded in the cdev's si_iosize_max) by MAXPHYS. This isn't strictly necessary, because physio(9) will limit it to MAXPHYS, but it will provide some clarity for the application. Record the controller's maximum I/O size reported in the Path Inquiry CCB. sa.4: Document the block size behavior, and explain that the option of allowing physio(9) to split the I/O will disappear in FreeBSD 11.0. Sponsored by: Spectra Logic
2013-08-24 04:52:22 +00:00
/*
* Bump the peripheral refcount for the sysctl thread, in case we
* get invalidated before the thread has a chance to run.
*/
cam_periph_acquire(periph);
taskqueue_enqueue(taskqueue_thread, &softc->sysctl_task);
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 async callback so that we get
* notified if this device goes away.
*/
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(AC_LOST_DEVICE, saasync, periph, periph->path);
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_announce_periph(periph, NULL);
Added output of device QUIRKS for CAM and AHCI devices during boot. Reviewed by: mav Approved by: pjd (mentor) MFC after: 2 weeks
2013-05-18 23:36:21 +00:00
xpt_announce_quirks(periph, softc->quirks, SA_QUIRK_BIT_STRING);
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
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
return (CAM_REQ_CMP);
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
sastart(struct cam_periph *periph, union ccb *start_ccb)
{
struct sa_softc *softc;
softc = (struct sa_softc *)periph->softc;
Check in some months long pending minor debug output changes. Clarify that the implicit fallthrough was *not* intentional (thanks, Poul!) and reorganize the code so a correct fallthrough (with /* FALLTHROUGH */) occurs.
2003-06-02 00:37:42 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_TRACE, ("sastart\n"));
Fix PR 24220 by using the periph private field0 for a CCB for not only CCB type but also extra flags- one of which can be "position updated". In other changes: Add in a SA_QUIRK_NO_CPAGE quirk so that it's possible to avoid using a (broken) device's implementation of he DEVICE COMPRESSION page. Also do a couple of printout cleanups. As per some discussion on FreeBSD-scsi, skip doing tape flushing if we're reading tape logical block location (MTIOCRDSPOS).
2001-01-15 22:28:11 +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
switch (softc->state) {
case SA_STATE_NORMAL:
{
/* Pull a buffer from the queue and get going on it */
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 *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
/*
* See if there is a buf with work for us to do..
*/
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 = bioq_first(&softc->bio_queue);
Merge CAM locking changes from the projects/camlock branch to radically reduce lock congestion and improve SMP scalability of the SCSI/ATA stack, preparing the ground for the coming next GEOM direct dispatch support. Replace big per-SIM locks with bunch of smaller ones: - per-LUN locks to protect device and peripheral drivers state; - per-target locks to protect list of LUNs on target; - per-bus locks to protect reference counting; - per-send queue locks to protect queue of CCBs to be sent; - per-done queue locks to protect queue of completed CCBs; - remaining per-SIM locks now protect only HBA driver internals. While holding LUN lock it is allowed (while not recommended for performance reasons) to take SIM lock. The opposite acquisition order is forbidden. All the other locks are leaf locks, that can be taken anywhere, but should not be cascaded. Many functions, such as: xpt_action(), xpt_done(), xpt_async(), xpt_create_path(), etc. are no longer require (but allow) SIM lock to be held. To keep compatibility and solve cases where SIM lock can't be dropped, all xpt_async() calls in addition to xpt_done() calls are queued to completion threads for async processing in clean environment without SIM lock held. Instead of single CAM SWI thread, used for commands completion processing before, use multiple (depending on number of CPUs) threads. Load balanced between them using "hash" of the device B:T:L address. HBA drivers that can drop SIM lock during completion processing and have sufficient number of completion threads to efficiently scale to multiple CPUs can use new function xpt_done_direct() to avoid extra context switch. Make ahci(4) driver to use this mechanism depending on hardware setup. Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-21 12:00:26 +00:00
if (bp == 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
xpt_release_ccb(start_ccb);
} else if ((softc->flags & SA_FLAG_ERR_PENDING) != 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
struct bio *done_bp;
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
again:
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->queue_count--;
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_remove(&softc->bio_queue, bp);
bp->bio_resid = bp->bio_bcount;
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
done_bp = 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 ((softc->flags & SA_FLAG_EOM_PENDING) != 0) {
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
/*
* We now just clear errors in this case
* and let the residual be the notifier.
*/
bp->bio_error = 0;
} else if ((softc->flags & SA_FLAG_EOF_PENDING) != 0) {
/*
* This can only happen if we're reading
* in fixed length mode. In this case,
* we dump the rest of the list the
* same way.
*/
bp->bio_error = 0;
if (bioq_first(&softc->bio_queue) != NULL) {
biodone(done_bp);
goto again;
}
} else if ((softc->flags & SA_FLAG_EIO_PENDING) != 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_error = EIO;
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +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
}
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 = bioq_first(&softc->bio_queue);
+Match against T4000* for HP QIC quirks (not T4000S* where it doesn't then catch a T4000s) + Set *some* kind of error at EOM if we're in fixed mode and have pending errs. Do not clear the ERR_PENDING bit if more buffers are queued. + Release the start_ccb in this case also, else we hang forever on rewinding. + Any kind of error for load to BOT in samount should then cause an attempt to use REWIND to come back to BOT. Do the initial load command quietly. + In samount, if we succeed, set the relative position markers.
1999-03-01 01:07:47 +00:00
/*
* Only if we have no other buffers queued up
* do we clear the pending error flag.
*/
if (bp == NULL)
softc->flags &= ~SA_FLAG_ERR_PENDING;
CAM_DEBUG(periph->path, CAM_DEBUG_INFO,
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
("sastart- ERR_PENDING now 0x%x, bp is %sNULL, "
"%d more buffers queued up\n",
+Match against T4000* for HP QIC quirks (not T4000S* where it doesn't then catch a T4000s) + Set *some* kind of error at EOM if we're in fixed mode and have pending errs. Do not clear the ERR_PENDING bit if more buffers are queued. + Release the start_ccb in this case also, else we hang forever on rewinding. + Any kind of error for load to BOT in samount should then cause an attempt to use REWIND to come back to BOT. Do the initial load command quietly. + In samount, if we succeed, set the relative position markers.
1999-03-01 01:07:47 +00:00
(softc->flags & SA_FLAG_ERR_PENDING),
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
(bp != NULL)? "not " : " ", softc->queue_count));
+Match against T4000* for HP QIC quirks (not T4000S* where it doesn't then catch a T4000s) + Set *some* kind of error at EOM if we're in fixed mode and have pending errs. Do not clear the ERR_PENDING bit if more buffers are queued. + Release the start_ccb in this case also, else we hang forever on rewinding. + Any kind of error for load to BOT in samount should then cause an attempt to use REWIND to come back to BOT. Do the initial load command quietly. + In samount, if we succeed, set the relative position markers.
1999-03-01 01:07:47 +00:00
xpt_release_ccb(start_ccb);
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
biodone(done_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
} else {
u_int32_t length;
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_remove(&softc->bio_queue, bp);
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->queue_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
if ((softc->flags & SA_FLAG_FIXED) != 0) {
if (softc->blk_shift != 0) {
length =
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_bcount >> softc->blk_shift;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
} else if (softc->media_blksize != 0) {
Fix PR 24220 by using the periph private field0 for a CCB for not only CCB type but also extra flags- one of which can be "position updated". In other changes: Add in a SA_QUIRK_NO_CPAGE quirk so that it's possible to avoid using a (broken) device's implementation of he DEVICE COMPRESSION page. Also do a couple of printout cleanups. As per some discussion on FreeBSD-scsi, skip doing tape flushing if we're reading tape logical block location (MTIOCRDSPOS).
2001-01-15 22:28:11 +00:00
length = bp->bio_bcount /
softc->media_blksize;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
} else {
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_error = EIO;
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, "zero blocksize"
" for FIXED length writes?\n");
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
biodone(bp);
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
}
Check in some months long pending minor debug output changes. Clarify that the implicit fallthrough was *not* intentional (thanks, Poul!) and reorganize the code so a correct fallthrough (with /* FALLTHROUGH */) occurs.
2003-06-02 00:37:42 +00:00
#if 0
CAM_DEBUG(start_ccb->ccb_h.path, CAM_DEBUG_INFO,
("issuing a %d fixed record %s\n",
length, (bp->bio_cmd == BIO_READ)? "read" :
"write"));
#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
} else {
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
length = bp->bio_bcount;
Check in some months long pending minor debug output changes. Clarify that the implicit fallthrough was *not* intentional (thanks, Poul!) and reorganize the code so a correct fallthrough (with /* FALLTHROUGH */) occurs.
2003-06-02 00:37:42 +00:00
#if 0
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
CAM_DEBUG(start_ccb->ccb_h.path, CAM_DEBUG_INFO,
Check in some months long pending minor debug output changes. Clarify that the implicit fallthrough was *not* intentional (thanks, Poul!) and reorganize the code so a correct fallthrough (with /* FALLTHROUGH */) occurs.
2003-06-02 00:37:42 +00:00
("issuing a %d variable byte %s\n",
length, (bp->bio_cmd == BIO_READ)? "read" :
"write"));
#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
}
Call devstat_start_transaction_bio() instead of devstat_start_transaction()
2003-03-15 10:50:44 +00:00
devstat_start_transaction_bio(softc->device_stats, 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
/*
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
* Some people have theorized that we should
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
* suppress illegal length indication if we are
* running in variable block mode so that we don't
* have to request sense every time our requested
* block size is larger than the written block.
* The residual information from the ccb allows
* us to identify this situation anyway. The only
* problem with this is that we will not get
* information about blocks that are larger than
* our read buffer unless we set the block size
* in the mode page to something other than 0.
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
*
* I believe that this is a non-issue. If user apps
* don't adjust their read size to match our record
* size, that's just life. Anyway, the typical usage
* would be to issue, e.g., 64KB reads and occasionally
* have to do deal with 512 byte or 1KB intermediate
* records.
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
softc->dsreg = (bp->bio_cmd == BIO_READ)?
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
MTIO_DSREG_RD : MTIO_DSREG_WR;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
scsi_sa_read_write(&start_ccb->csio, 0, sadone,
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
MSG_SIMPLE_Q_TAG, (bp->bio_cmd == BIO_READ ?
SCSI_RW_READ : SCSI_RW_WRITE) |
((bp->bio_flags & BIO_UNMAPPED) != 0 ?
SCSI_RW_BIO : 0), FALSE,
(softc->flags & SA_FLAG_FIXED) != 0, length,
(bp->bio_flags & BIO_UNMAPPED) != 0 ? (void *)bp :
bp->bio_data, bp->bio_bcount, SSD_FULL_SIZE,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
IO_TIMEOUT);
Fix PR 24220 by using the periph private field0 for a CCB for not only CCB type but also extra flags- one of which can be "position updated". In other changes: Add in a SA_QUIRK_NO_CPAGE quirk so that it's possible to avoid using a (broken) device's implementation of he DEVICE COMPRESSION page. Also do a couple of printout cleanups. As per some discussion on FreeBSD-scsi, skip doing tape flushing if we're reading tape logical block location (MTIOCRDSPOS).
2001-01-15 22:28:11 +00:00
start_ccb->ccb_h.ccb_pflags &= ~SA_POSITION_UPDATED;
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
start_ccb->ccb_h.ccb_bp = bp;
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 = bioq_first(&softc->bio_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
xpt_action(start_ccb);
}
if (bp != NULL) {
/* Have more work to do, so ensure we stay scheduled */
Replace most of priority numbers with defines. No logical changes.
2009-10-23 08:27:55 +00:00
xpt_schedule(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
}
break;
}
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
case SA_STATE_ABNORMAL:
default:
panic("state 0x%x in sastart", softc->state);
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
}
}
static void
sadone(struct cam_periph *periph, union ccb *done_ccb)
{
struct sa_softc *softc;
struct ccb_scsiio *csio;
Fix breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +00:00
struct bio *bp;
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 sa_softc *)periph->softc;
csio = &done_ccb->csio;
Fix breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +00:00
softc->dsreg = MTIO_DSREG_REST;
bp = (struct bio *)done_ccb->ccb_h.ccb_bp;
error = 0;
if ((done_ccb->ccb_h.status & CAM_STATUS_MASK) != CAM_REQ_CMP) {
if ((error = saerror(done_ccb, 0, 0)) == ERESTART) {
/*
* A retry was scheduled, so just return.
*/
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
}
Fix breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +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
Fix breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +00:00
if (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
Fix breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +00:00
/*
* Catastrophic error. Mark the tape as frozen
* (we no longer know tape position).
*
* Return all queued I/O with EIO, and unfreeze
* our queue so that future transactions that
* attempt to fix this problem can get to the
* device.
*
*/
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 breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +00:00
softc->flags |= SA_FLAG_TAPE_FROZEN;
bioq_flush(&softc->bio_queue, NULL, EIO);
}
if (error != 0) {
bp->bio_resid = bp->bio_bcount;
bp->bio_error = error;
bp->bio_flags |= BIO_ERROR;
/*
* In the error case, position is updated in saerror.
*/
} else {
bp->bio_resid = csio->resid;
bp->bio_error = 0;
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 breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +00:00
}
if (bp->bio_cmd == BIO_WRITE) {
softc->flags |= SA_FLAG_TAPE_WRITTEN;
softc->filemarks = 0;
}
if (!(csio->ccb_h.ccb_pflags & SA_POSITION_UPDATED) &&
(softc->blkno != (daddr_t) -1)) {
if ((softc->flags & SA_FLAG_FIXED) != 0) {
u_int32_t l;
if (softc->blk_shift != 0) {
l = bp->bio_bcount >>
softc->blk_shift;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
} else {
Fix breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +00:00
l = bp->bio_bcount /
softc->media_blksize;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
}
Fix breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +00:00
softc->blkno += (daddr_t) l;
} else {
softc->blkno++;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +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
}
}
Fix breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +00:00
/*
* If we had an error (immediate or pending),
* release the device queue now.
*/
if (error || (softc->flags & SA_FLAG_ERR_PENDING))
cam_release_devq(done_ccb->ccb_h.path, 0, 0, 0, 0);
if (error || bp->bio_resid) {
CAM_DEBUG(periph->path, CAM_DEBUG_INFO,
("error %d resid %ld count %ld\n", error,
bp->bio_resid, bp->bio_bcount));
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 breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +00:00
biofinish(bp, softc->device_stats, 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_release_ccb(done_ccb);
}
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
* Mount the tape (make sure it's ready for 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
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
samount(struct cam_periph *periph, int oflags, struct cdev *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
{
struct sa_softc *softc;
union ccb *ccb;
int error;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
* oflags can be checked for 'kind' of open (read-only check) - later
* dev can be checked for a control-mode or compression open - later
*/
UNUSED_PARAMETER(oflags);
UNUSED_PARAMETER(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
softc = (struct sa_softc *)periph->softc;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/*
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
* This should determine if something has happend since the last
* open/mount that would invalidate the mount. We do *not* want
* to retry this command- we just want the status. But we only
* do this if we're mounted already- if we're not mounted,
* we don't care about the unit read state and can instead use
* this opportunity to attempt to reserve the tape unit.
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
*/
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
if (softc->flags & SA_FLAG_TAPE_MOUNTED) {
ccb = cam_periph_getccb(periph, 1);
scsi_test_unit_ready(&ccb->csio, 0, sadone,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
MSG_SIMPLE_Q_TAG, SSD_FULL_SIZE, IO_TIMEOUT);
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
error = cam_periph_runccb(ccb, saerror, 0, SF_NO_PRINT,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
softc->device_stats);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
if (error == ENXIO) {
softc->flags &= ~SA_FLAG_TAPE_MOUNTED;
scsi_test_unit_ready(&ccb->csio, 0, sadone,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
MSG_SIMPLE_Q_TAG, SSD_FULL_SIZE, IO_TIMEOUT);
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
error = cam_periph_runccb(ccb, saerror, 0, SF_NO_PRINT,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
softc->device_stats);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
} else if (error) {
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
/*
* We don't need to freeze the tape because we
* will now attempt to rewind/load it.
*/
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
softc->flags &= ~SA_FLAG_TAPE_MOUNTED;
avoid use after free Submitted by: gibbs Reviewed by: mdodd Approved by: Coverity Prevent analysis tool
2005-03-31 21:43:19 +00:00
if (CAM_DEBUGGED(periph->path, CAM_DEBUG_INFO)) {
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,
"error %d on TUR in samount\n", error);
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
}
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
}
} else {
error = sareservereleaseunit(periph, TRUE);
if (error) {
return (error);
}
ccb = cam_periph_getccb(periph, 1);
Add back in a prefatory TUR when the tape is *not* mounted. This seems to help with some older tape drives.
1999-12-04 01:13:59 +00:00
scsi_test_unit_ready(&ccb->csio, 0, sadone,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
MSG_SIMPLE_Q_TAG, SSD_FULL_SIZE, IO_TIMEOUT);
Add back in a prefatory TUR when the tape is *not* mounted. This seems to help with some older tape drives.
1999-12-04 01:13:59 +00:00
error = cam_periph_runccb(ccb, saerror, 0, SF_NO_PRINT,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
softc->device_stats);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +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 & SA_FLAG_TAPE_MOUNTED) == 0) {
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
struct scsi_read_block_limits_data *rblim = NULL;
int comp_enabled, comp_supported;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
u_int8_t write_protect, guessing = 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
/*
* Clear out old state.
*/
softc->flags &= ~(SA_FLAG_TAPE_WP|SA_FLAG_TAPE_WRITTEN|
SA_FLAG_ERR_PENDING|SA_FLAG_COMP_ENABLED|
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
SA_FLAG_COMP_SUPP|SA_FLAG_COMP_UNSUPP);
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->filemarks = 0;
/*
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
* *Very* first off, make sure we're loaded to BOT.
*/
scsi_load_unload(&ccb->csio, 2, sadone, MSG_SIMPLE_Q_TAG, FALSE,
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
FALSE, FALSE, 1, SSD_FULL_SIZE, REWIND_TIMEOUT);
error = cam_periph_runccb(ccb, saerror, 0, SF_NO_PRINT,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
softc->device_stats);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
/*
+Match against T4000* for HP QIC quirks (not T4000S* where it doesn't then catch a T4000s) + Set *some* kind of error at EOM if we're in fixed mode and have pending errs. Do not clear the ERR_PENDING bit if more buffers are queued. + Release the start_ccb in this case also, else we hang forever on rewinding. + Any kind of error for load to BOT in samount should then cause an attempt to use REWIND to come back to BOT. Do the initial load command quietly. + In samount, if we succeed, set the relative position markers.
1999-03-01 01:07:47 +00:00
* In case this doesn't work, do a REWIND instead
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
*/
+Match against T4000* for HP QIC quirks (not T4000S* where it doesn't then catch a T4000s) + Set *some* kind of error at EOM if we're in fixed mode and have pending errs. Do not clear the ERR_PENDING bit if more buffers are queued. + Release the start_ccb in this case also, else we hang forever on rewinding. + Any kind of error for load to BOT in samount should then cause an attempt to use REWIND to come back to BOT. Do the initial load command quietly. + In samount, if we succeed, set the relative position markers.
1999-03-01 01:07:47 +00:00
if (error) {
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
scsi_rewind(&ccb->csio, 2, sadone, MSG_SIMPLE_Q_TAG,
FALSE, SSD_FULL_SIZE, REWIND_TIMEOUT);
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
error = cam_periph_runccb(ccb, saerror, 0, SF_NO_PRINT,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
softc->device_stats);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
}
if (error) {
xpt_release_ccb(ccb);
goto exit;
}
/*
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
* Do a dummy test read to force access to the
* media so that the drive will really know what's
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
* there. We actually don't really care what the
* blocksize on tape is and don't expect to really
* read a full record.
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
*/
rblim = (struct scsi_read_block_limits_data *)
Work around the other uses of M_WAITOK.
2007-06-16 18:20:29 +00:00
malloc(8192, M_SCSISA, M_NOWAIT);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
if (rblim == NULL) {
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, "no memory for test read\n");
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
xpt_release_ccb(ccb);
error = ENOMEM;
goto exit;
}
Add a quirk type (and one for ARCHIVE Python's) that disables the dummy read at sa_mount that attempts to latch up density. This breaks on at least one drive and sends it into the weeds. Approved: jkh
2000-02-03 18:29:25 +00:00
if ((softc->quirks & SA_QUIRK_NODREAD) == 0) {
scsi_sa_read_write(&ccb->csio, 0, sadone,
MSG_SIMPLE_Q_TAG, 1, FALSE, 0, 8192,
(void *) rblim, 8192, SSD_FULL_SIZE,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
IO_TIMEOUT);
Add a quirk type (and one for ARCHIVE Python's) that disables the dummy read at sa_mount that attempts to latch up density. This breaks on at least one drive and sends it into the weeds. Approved: jkh
2000-02-03 18:29:25 +00:00
(void) cam_periph_runccb(ccb, saerror, 0, SF_NO_PRINT,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
softc->device_stats);
Add a quirk type (and one for ARCHIVE Python's) that disables the dummy read at sa_mount that attempts to latch up density. This breaks on at least one drive and sends it into the weeds. Approved: jkh
2000-02-03 18:29:25 +00:00
scsi_rewind(&ccb->csio, 1, sadone, MSG_SIMPLE_Q_TAG,
FALSE, SSD_FULL_SIZE, REWIND_TIMEOUT);
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 = cam_periph_runccb(ccb, saerror, CAM_RETRY_SELTO,
SF_NO_PRINT | SF_RETRY_UA,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
softc->device_stats);
Add a quirk type (and one for ARCHIVE Python's) that disables the dummy read at sa_mount that attempts to latch up density. This breaks on at least one drive and sends it into the weeds. Approved: jkh
2000-02-03 18:29:25 +00:00
if (error) {
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,
"unable to rewind after test read\n");
Add a quirk type (and one for ARCHIVE Python's) that disables the dummy read at sa_mount that attempts to latch up density. This breaks on at least one drive and sends it into the weeds. Approved: jkh
2000-02-03 18:29:25 +00:00
xpt_release_ccb(ccb);
goto exit;
}
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +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
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
/*
* Next off, determine block limits.
*/
scsi_read_block_limits(&ccb->csio, 5, sadone, MSG_SIMPLE_Q_TAG,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
rblim, SSD_FULL_SIZE, SCSIOP_TIMEOUT);
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
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 = cam_periph_runccb(ccb, saerror, CAM_RETRY_SELTO,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
SF_NO_PRINT | SF_RETRY_UA, softc->device_stats);
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
xpt_release_ccb(ccb);
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
if (error != 0) {
/*
* If it's less than SCSI-2, READ BLOCK LIMITS is not
* a MANDATORY command. Anyway- it doesn't matter-
* we can proceed anyway.
*/
softc->blk_gran = 0;
softc->max_blk = ~0;
softc->min_blk = 0;
} else {
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 (softc->scsi_rev >= SCSI_REV_SPC) {
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
softc->blk_gran = RBL_GRAN(rblim);
} else {
softc->blk_gran = 0;
}
/*
* We take max_blk == min_blk to mean a default to
* fixed mode- but note that whatever we get out of
* sagetparams below will actually determine whether
* we are actually *in* fixed mode.
*/
softc->max_blk = scsi_3btoul(rblim->maximum);
softc->min_blk = scsi_2btoul(rblim->minimum);
}
/*
* Next, perform a mode sense to determine
* current density, blocksize, compression etc.
*/
error = sagetparams(periph, SA_PARAM_ALL,
&softc->media_blksize,
&softc->media_density,
&softc->media_numblks,
&softc->buffer_mode, &write_protect,
&softc->speed, &comp_supported,
&comp_enabled, &softc->comp_algorithm,
NULL);
if (error != 0) {
/*
* We could work a little harder here. We could
* adjust our attempts to get information. It
* might be an ancient tape drive. If someone
* nudges us, we'll do that.
*/
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
goto exit;
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +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
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
* If no quirk has determined that this is a device that is
* preferred to be in fixed or variable mode, now is the time
* to find out.
*/
if ((softc->quirks & (SA_QUIRK_FIXED|SA_QUIRK_VARIABLE)) == 0) {
guessing = 1;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
/*
* This could be expensive to find out. Luckily we
* only need to do this once. If we start out in
* 'default' mode, try and set ourselves to one
* of the densities that would determine a wad
* of other stuff. Go from highest to lowest.
*/
if (softc->media_density == SCSI_DEFAULT_DENSITY) {
int i;
static u_int8_t ctry[] = {
SCSI_DENSITY_HALFINCH_PE,
SCSI_DENSITY_HALFINCH_6250C,
SCSI_DENSITY_HALFINCH_6250,
SCSI_DENSITY_HALFINCH_1600,
SCSI_DENSITY_HALFINCH_800,
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
SCSI_DENSITY_QIC_4GB,
SCSI_DENSITY_QIC_2GB,
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
SCSI_DENSITY_QIC_525_320,
SCSI_DENSITY_QIC_150,
SCSI_DENSITY_QIC_120,
SCSI_DENSITY_QIC_24,
SCSI_DENSITY_QIC_11_9TRK,
SCSI_DENSITY_QIC_11_4TRK,
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
SCSI_DENSITY_QIC_1320,
SCSI_DENSITY_QIC_3080,
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
0
};
for (i = 0; ctry[i]; i++) {
error = sasetparams(periph,
SA_PARAM_DENSITY, 0, ctry[i],
0, SF_NO_PRINT);
if (error == 0) {
softc->media_density = ctry[i];
break;
}
}
}
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
switch (softc->media_density) {
case SCSI_DENSITY_QIC_11_4TRK:
case SCSI_DENSITY_QIC_11_9TRK:
case SCSI_DENSITY_QIC_24:
case SCSI_DENSITY_QIC_120:
case SCSI_DENSITY_QIC_150:
With the help of 'Eric Christeson <echriste@ssesco.com>', determined that QIC 525 really should be 512 byte fixed blocksize.
2000-09-14 21:38:44 +00:00
case SCSI_DENSITY_QIC_525_320:
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
case SCSI_DENSITY_QIC_1320:
case SCSI_DENSITY_QIC_3080:
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->quirks &= ~SA_QUIRK_2FM;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->quirks |= SA_QUIRK_FIXED|SA_QUIRK_1FM;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
softc->last_media_blksize = 512;
break;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
case SCSI_DENSITY_QIC_4GB:
case SCSI_DENSITY_QIC_2GB:
softc->quirks &= ~SA_QUIRK_2FM;
softc->quirks |= SA_QUIRK_FIXED|SA_QUIRK_1FM;
softc->last_media_blksize = 1024;
break;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
default:
softc->last_media_blksize =
softc->media_blksize;
softc->quirks |= SA_QUIRK_VARIABLE;
break;
}
}
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
* If no quirk has determined that this is a device that needs
* to have 2 Filemarks at EOD, now is the time to find out.
*/
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
Sascha Blank <blank@uni-trier.de> convinced me I was an idiot about testing SA_QUIRK_2FM in samount. Fixed. Removed the NORRLS quirk (to save quirk space) and left the behaviour of being quiet about failed reserve/release (failed due Illegal Request) the same. Added a SF_QUIET_IR for prevent/allow for the same purposes.
1999-01-16 19:20:30 +00:00
if ((softc->quirks & SA_QUIRK_2FM) == 0) {
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
switch (softc->media_density) {
case SCSI_DENSITY_HALFINCH_800:
case SCSI_DENSITY_HALFINCH_1600:
case SCSI_DENSITY_HALFINCH_6250:
case SCSI_DENSITY_HALFINCH_6250C:
case SCSI_DENSITY_HALFINCH_PE:
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->quirks &= ~SA_QUIRK_1FM;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
softc->quirks |= SA_QUIRK_2FM;
break;
default:
break;
}
}
/*
* Now validate that some info we got makes sense.
*/
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
if ((softc->max_blk < softc->media_blksize) ||
(softc->min_blk > softc->media_blksize &&
softc->media_blksize)) {
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,
"BLOCK LIMITS (%d..%d) could not match current "
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
"block settings (%d)- adjusting\n", softc->min_blk,
softc->max_blk, softc->media_blksize);
softc->max_blk = softc->min_blk =
softc->media_blksize;
}
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
/*
* Now put ourselves into the right frame of mind based
* upon quirks...
*/
tryagain:
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
/*
* If we want to be in FIXED mode and our current blocksize
* is not equal to our last blocksize (if nonzero), try and
* set ourselves to this last blocksize (as the 'preferred'
* block size). The initial quirkmatch at registry sets the
* initial 'last' blocksize. If, for whatever reason, this
* 'last' blocksize is zero, set the blocksize to 512,
* or min_blk if that's larger.
*/
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
if ((softc->quirks & SA_QUIRK_FIXED) &&
Some prettifying for the quirk comments. Add a SA_QUIRK_NO_MODESEL type and use it for the OnStream real SCSI device (not the broken one). This one is still broken in that it can't be set to the same fixed block size it reports [ unflattering comments about this company elided ]. If we're unable to set buffered mode on, complain, but drive on. It's not a fatal error to not be in buffered mode.
2000-05-09 04:54:10 +00:00
(softc->quirks & SA_QUIRK_NO_MODESEL) == 0 &&
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
(softc->media_blksize != softc->last_media_blksize)) {
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
softc->media_blksize = softc->last_media_blksize;
if (softc->media_blksize == 0) {
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
softc->media_blksize = 512;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
if (softc->media_blksize < softc->min_blk) {
softc->media_blksize = softc->min_blk;
}
}
error = sasetparams(periph, SA_PARAM_BLOCKSIZE,
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
softc->media_blksize, 0, 0, SF_NO_PRINT);
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
if (error) {
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,
"unable to set fixed blocksize to %d\n",
softc->media_blksize);
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
goto exit;
}
}
if ((softc->quirks & SA_QUIRK_VARIABLE) &&
(softc->media_blksize != 0)) {
softc->last_media_blksize = softc->media_blksize;
softc->media_blksize = 0;
error = sasetparams(periph, SA_PARAM_BLOCKSIZE,
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
0, 0, 0, SF_NO_PRINT);
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
if (error) {
/*
* If this fails and we were guessing, just
* assume that we got it wrong and go try
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
* fixed block mode. Don't even check against
* density code at this point.
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
*/
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
if (guessing) {
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
softc->quirks &= ~SA_QUIRK_VARIABLE;
softc->quirks |= SA_QUIRK_FIXED;
if (softc->last_media_blksize == 0)
softc->last_media_blksize = 512;
goto tryagain;
}
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,
"unable to set variable blocksize\n");
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
goto exit;
}
}
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
/*
* Now that we have the current block size,
* set up some parameters for sastart's usage.
*/
if (softc->media_blksize) {
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 |= SA_FLAG_FIXED;
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
if (powerof2(softc->media_blksize)) {
softc->blk_shift =
ffs(softc->media_blksize) - 1;
softc->blk_mask = softc->media_blksize - 1;
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 {
softc->blk_mask = ~0;
softc->blk_shift = 0;
}
} else {
/*
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
* The SCSI-3 spec allows 0 to mean "unspecified".
* The SCSI-1 spec allows 0 to mean 'infinite'.
*
* Either works here.
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->max_blk == 0) {
softc->max_blk = ~0;
}
softc->blk_shift = 0;
if (softc->blk_gran != 0) {
softc->blk_mask = softc->blk_gran - 1;
} else {
softc->blk_mask = 0;
}
}
if (write_protect)
softc->flags |= SA_FLAG_TAPE_WP;
if (comp_supported) {
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
if (softc->saved_comp_algorithm == 0)
softc->saved_comp_algorithm =
softc->comp_algorithm;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->flags |= SA_FLAG_COMP_SUPP;
if (comp_enabled)
softc->flags |= SA_FLAG_COMP_ENABLED;
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
softc->flags |= SA_FLAG_COMP_UNSUPP;
Some prettifying for the quirk comments. Add a SA_QUIRK_NO_MODESEL type and use it for the OnStream real SCSI device (not the broken one). This one is still broken in that it can't be set to the same fixed block size it reports [ unflattering comments about this company elided ]. If we're unable to set buffered mode on, complain, but drive on. It's not a fatal error to not be in buffered mode.
2000-05-09 04:54:10 +00:00
if ((softc->buffer_mode == SMH_SA_BUF_MODE_NOBUF) &&
(softc->quirks & SA_QUIRK_NO_MODESEL) == 0) {
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
error = sasetparams(periph, SA_PARAM_BUFF_MODE, 0,
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
0, 0, SF_NO_PRINT);
The code that sees a drive (at mount time) not in buffered mode and attempts to set buffered mode was printing out "unable to set buffered mode" no matter what. Oops. Spotted by: Joerg Wunsch <joerg_wunsch@uriah.heep.sax.de> MFC after: 3 weeks
2001-09-14 19:00:51 +00:00
if (error == 0) {
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
softc->buffer_mode = SMH_SA_BUF_MODE_SIBUF;
The code that sees a drive (at mount time) not in buffered mode and attempts to set buffered mode was printing out "unable to set buffered mode" no matter what. Oops. Spotted by: Joerg Wunsch <joerg_wunsch@uriah.heep.sax.de> MFC after: 3 weeks
2001-09-14 19:00:51 +00:00
} 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,
"unable to set buffered mode\n");
The code that sees a drive (at mount time) not in buffered mode and attempts to set buffered mode was printing out "unable to set buffered mode" no matter what. Oops. Spotted by: Joerg Wunsch <joerg_wunsch@uriah.heep.sax.de> MFC after: 3 weeks
2001-09-14 19:00:51 +00:00
}
Some prettifying for the quirk comments. Add a SA_QUIRK_NO_MODESEL type and use it for the OnStream real SCSI device (not the broken one). This one is still broken in that it can't be set to the same fixed block size it reports [ unflattering comments about this company elided ]. If we're unable to set buffered mode on, complain, but drive on. It's not a fatal error to not be in buffered mode.
2000-05-09 04:54:10 +00:00
error = 0; /* not an error */
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +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
+Match against T4000* for HP QIC quirks (not T4000S* where it doesn't then catch a T4000s) + Set *some* kind of error at EOM if we're in fixed mode and have pending errs. Do not clear the ERR_PENDING bit if more buffers are queued. + Release the start_ccb in this case also, else we hang forever on rewinding. + Any kind of error for load to BOT in samount should then cause an attempt to use REWIND to come back to BOT. Do the initial load command quietly. + In samount, if we succeed, set the relative position markers.
1999-03-01 01:07:47 +00:00
if (error == 0) {
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
softc->flags |= SA_FLAG_TAPE_MOUNTED;
+Match against T4000* for HP QIC quirks (not T4000S* where it doesn't then catch a T4000s) + Set *some* kind of error at EOM if we're in fixed mode and have pending errs. Do not clear the ERR_PENDING bit if more buffers are queued. + Release the start_ccb in this case also, else we hang forever on rewinding. + Any kind of error for load to BOT in samount should then cause an attempt to use REWIND to come back to BOT. Do the initial load command quietly. + In samount, if we succeed, set the relative position markers.
1999-03-01 01:07:47 +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
exit:
if (rblim != NULL)
Eliminate the use of M_TEMP.
2007-05-14 21:48:53 +00:00
free(rblim, M_SCSISA);
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
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
if (error != 0) {
softc->dsreg = MTIO_DSREG_NIL;
+Match against T4000* for HP QIC quirks (not T4000S* where it doesn't then catch a T4000s) + Set *some* kind of error at EOM if we're in fixed mode and have pending errs. Do not clear the ERR_PENDING bit if more buffers are queued. + Release the start_ccb in this case also, else we hang forever on rewinding. + Any kind of error for load to BOT in samount should then cause an attempt to use REWIND to come back to BOT. Do the initial load command quietly. + In samount, if we succeed, set the relative position markers.
1999-03-01 01:07:47 +00:00
} else {
softc->fileno = softc->blkno = 0;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_REST;
+Match against T4000* for HP QIC quirks (not T4000S* where it doesn't then catch a T4000s) + Set *some* kind of error at EOM if we're in fixed mode and have pending errs. Do not clear the ERR_PENDING bit if more buffers are queued. + Release the start_ccb in this case also, else we hang forever on rewinding. + Any kind of error for load to BOT in samount should then cause an attempt to use REWIND to come back to BOT. Do the initial load command quietly. + In samount, if we succeed, set the relative position markers.
1999-03-01 01:07:47 +00:00
}
Another 1FM@EOD quirk (from mike@sentex.net for a Seagate STT20000), and add a config option that allows one to default to 1FM@EOD for tapes otherwise unquirked or unknown as to which to prefer. Note that tcopy will be broken for these tapes until tcopy is fixed.
1999-10-02 20:17:16 +00:00
#ifdef SA_1FM_AT_EOD
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if ((softc->quirks & SA_QUIRK_2FM) == 0)
softc->quirks |= SA_QUIRK_1FM;
Another 1FM@EOD quirk (from mike@sentex.net for a Seagate STT20000), and add a config option that allows one to default to 1FM@EOD for tapes otherwise unquirked or unknown as to which to prefer. Note that tcopy will be broken for these tapes until tcopy is fixed.
1999-10-02 20:17:16 +00:00
#else
if ((softc->quirks & SA_QUIRK_1FM) == 0)
softc->quirks |= SA_QUIRK_2FM;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
#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
} else
xpt_release_ccb(ccb);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
/*
* If we return an error, we're not mounted any more,
* so release any device reservation.
*/
if (error != 0) {
(void) sareservereleaseunit(periph, FALSE);
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
} else {
/*
* Clear I/O residual.
*/
softc->last_io_resid = 0;
softc->last_ctl_resid = 0;
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
}
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +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
}
scsi_sa.h
2000-10-31 22:34:51 +00:00
/*
* How many filemarks do we need to write if we were to terminate the
* tape session right now? Note that this can be a negative 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
static int
scsi_sa.h
2000-10-31 22:34:51 +00:00
samarkswanted(struct cam_periph *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
{
int markswanted;
struct sa_softc *softc;
softc = (struct sa_softc *)periph->softc;
markswanted = 0;
if ((softc->flags & SA_FLAG_TAPE_WRITTEN) != 0) {
markswanted++;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (softc->quirks & SA_QUIRK_2FM)
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
markswanted++;
}
scsi_sa.h
2000-10-31 22:34:51 +00:00
markswanted -= softc->filemarks;
return (markswanted);
}
static int
sacheckeod(struct cam_periph *periph)
{
int error;
int markswanted;
markswanted = samarkswanted(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
scsi_sa.h
2000-10-31 22:34:51 +00:00
if (markswanted > 0) {
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
error = sawritefilemarks(periph, markswanted, 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
} else {
error = 0;
}
return (error);
}
static int
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
saerror(union ccb *ccb, u_int32_t cflgs, u_int32_t sflgs)
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
static const char *toobig =
Correct spelling of 'supplied'. PR: misc/39528
2002-06-19 20:44:48 +00:00
"%d-byte tape record bigger than supplied buffer\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
struct cam_periph *periph;
struct sa_softc *softc;
struct ccb_scsiio *csio;
struct scsi_sense_data *sense;
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
uint64_t resid = 0;
int64_t info = 0;
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
cam_status status;
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
int error_code, sense_key, asc, ascq, error, aqvalid, stream_valid;
int sense_len;
uint8_t stream_bits;
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 sa_softc *)periph->softc;
csio = &ccb->csio;
sense = &csio->sense_data;
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
sense_len = csio->sense_len - csio->sense_resid;
scsi_extract_sense_len(sense, sense_len, &error_code, &sense_key,
&asc, &ascq, /*show_errors*/ 1);
if (asc != -1 && ascq != -1)
aqvalid = 1;
else
aqvalid = 0;
if (scsi_get_stream_info(sense, sense_len, NULL, &stream_bits) == 0)
stream_valid = 1;
else
stream_valid = 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
error = 0;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
status = csio->ccb_h.status & CAM_STATUS_MASK;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/*
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
* Calculate/latch up, any residuals... We do this in a funny 2-step
* so we can print stuff here if we have CAM_DEBUG enabled for this
* unit.
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
*/
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
if (status == CAM_SCSI_STATUS_ERROR) {
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
if (scsi_get_sense_info(sense, sense_len, SSD_DESC_INFO, &resid,
&info) == 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
if ((softc->flags & SA_FLAG_FIXED) != 0)
resid *= softc->media_blksize;
} else {
resid = csio->dxfer_len;
info = resid;
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
if ((softc->flags & SA_FLAG_FIXED) != 0) {
if (softc->media_blksize)
info /= softc->media_blksize;
}
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 breakage introduced by r256843: removing the SA_CCB_WAITING bit left some of the decisions based on its counterpart, SA_CCB_BUFFER_IO being random. As a result, propagation of the residual information for the SPACE command was broken, so the number of filemarks encountered during a SPACE operation was miscalculated. Consequently, systems relying on properly tracked filemark counters (like Bacula) fell apart. The change also removes a switch/case in sadone() which r256843 degraded to a single remaining case label. PR: 192285 Approved by: ken MFC after: 2 weeks
2014-07-31 22:09:50 +00:00
if (csio->cdb_io.cdb_bytes[0] == SA_READ ||
csio->cdb_io.cdb_bytes[0] == SA_WRITE) {
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
bcopy((caddr_t) sense, (caddr_t) &softc->last_io_sense,
sizeof (struct scsi_sense_data));
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
bcopy(csio->cdb_io.cdb_bytes, softc->last_io_cdb,
(int) csio->cdb_len);
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
softc->last_io_resid = resid;
Make some attempt to accomodate kern/24221- propate residual to mt_resid even if it might overflow. If we do a read or set of hardware or logical block position, we also clear Tape Frozen status.
2001-01-19 21:08:15 +00:00
softc->last_resid_was_io = 1;
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
} else {
bcopy((caddr_t) sense, (caddr_t) &softc->last_ctl_sense,
sizeof (struct scsi_sense_data));
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
bcopy(csio->cdb_io.cdb_bytes, softc->last_ctl_cdb,
(int) csio->cdb_len);
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
softc->last_ctl_resid = resid;
Make some attempt to accomodate kern/24221- propate residual to mt_resid even if it might overflow. If we do a read or set of hardware or logical block position, we also clear Tape Frozen status.
2001-01-19 21:08:15 +00:00
softc->last_resid_was_io = 0;
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
}
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_INFO, ("CDB[0]=0x%x Key 0x%x "
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
"ASC/ASCQ 0x%x/0x%x CAM STATUS 0x%x flags 0x%x resid %jd "
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
"dxfer_len %d\n", csio->cdb_io.cdb_bytes[0] & 0xff,
sense_key, asc, ascq, status,
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
(stream_valid) ? stream_bits : 0, (intmax_t)resid,
csio->dxfer_len));
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
} else {
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
CAM_DEBUG(periph->path, CAM_DEBUG_INFO,
("Cam Status 0x%x\n", status));
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
}
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
switch (status) {
case CAM_REQ_CMP:
return (0);
case CAM_SCSI_STATUS_ERROR:
/*
* If a read/write command, we handle it here.
*/
Merge CAM locking changes from the projects/camlock branch to radically reduce lock congestion and improve SMP scalability of the SCSI/ATA stack, preparing the ground for the coming next GEOM direct dispatch support. Replace big per-SIM locks with bunch of smaller ones: - per-LUN locks to protect device and peripheral drivers state; - per-target locks to protect list of LUNs on target; - per-bus locks to protect reference counting; - per-send queue locks to protect queue of CCBs to be sent; - per-done queue locks to protect queue of completed CCBs; - remaining per-SIM locks now protect only HBA driver internals. While holding LUN lock it is allowed (while not recommended for performance reasons) to take SIM lock. The opposite acquisition order is forbidden. All the other locks are leaf locks, that can be taken anywhere, but should not be cascaded. Many functions, such as: xpt_action(), xpt_done(), xpt_async(), xpt_create_path(), etc. are no longer require (but allow) SIM lock to be held. To keep compatibility and solve cases where SIM lock can't be dropped, all xpt_async() calls in addition to xpt_done() calls are queued to completion threads for async processing in clean environment without SIM lock held. Instead of single CAM SWI thread, used for commands completion processing before, use multiple (depending on number of CPUs) threads. Load balanced between them using "hash" of the device B:T:L address. HBA drivers that can drop SIM lock during completion processing and have sufficient number of completion threads to efficiently scale to multiple CPUs can use new function xpt_done_direct() to avoid extra context switch. Make ahci(4) driver to use this mechanism depending on hardware setup. Sponsored by: iXsystems, Inc. MFC after: 2 months
2013-10-21 12:00:26 +00:00
if (csio->cdb_io.cdb_bytes[0] == SA_READ ||
csio->cdb_io.cdb_bytes[0] == SA_WRITE) {
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
break;
}
/*
* If this was just EOM/EOP, Filemark, Setmark or ILI detected
* on a non read/write command, we assume it's not an error
* and propagate the residule and return.
*/
if ((aqvalid && asc == 0 && ascq > 0 && ascq <= 5) ||
(aqvalid == 0 && sense_key == SSD_KEY_NO_SENSE)) {
csio->resid = resid;
QFRLS(ccb);
return (0);
}
/*
* Otherwise, we let the common code handle this.
*/
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
return (cam_periph_error(ccb, cflgs, sflgs, &softc->saved_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
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/*
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
* XXX: To Be Fixed
* We cannot depend upon CAM honoring retry counts for these.
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
*/
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
case CAM_SCSI_BUS_RESET:
case CAM_BDR_SENT:
if (ccb->ccb_h.retry_count <= 0) {
return (EIO);
}
/* FALLTHROUGH */
default:
return (cam_periph_error(ccb, cflgs, sflgs, &softc->saved_ccb));
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
}
/*
* Handle filemark, end of tape, mismatched record sizes....
* From this point out, we're only handling read/write cases.
* Handle writes && reads differently.
*/
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (csio->cdb_io.cdb_bytes[0] == SA_WRITE) {
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
if (sense_key == SSD_KEY_VOLUME_OVERFLOW) {
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->resid = resid;
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
error = ENOSPC;
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
} else if ((stream_valid != 0) && (stream_bits & SSD_EOM)) {
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
softc->flags |= SA_FLAG_EOM_PENDING;
/*
* Grotesque as it seems, the few times
* I've actually seen a non-zero resid,
* the tape drive actually lied and had
Fix type in comment, 's/writtent/written/' PR: 56954 Submitted by: Dan Langille <dan@langille.org>
2004-01-18 00:17:14 +00:00
* written all the data!.
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
*/
csio->resid = 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 a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
} else {
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
csio->resid = resid;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (sense_key == SSD_KEY_BLANK_CHECK) {
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
if (softc->quirks & SA_QUIRK_1FM) {
error = 0;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->flags |= SA_FLAG_EOM_PENDING;
} else {
error = EIO;
}
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
} else if ((stream_valid != 0) && (stream_bits & SSD_FILEMARK)){
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
if (softc->flags & SA_FLAG_FIXED) {
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
error = -1;
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 |= SA_FLAG_EOF_PENDING;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
}
/*
* Unconditionally, if we detected a filemark on a read,
* mark that we've run moved a file ahead.
*/
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
if (softc->fileno != (daddr_t) -1) {
softc->fileno++;
softc->blkno = 0;
Fix PR 24220 by using the periph private field0 for a CCB for not only CCB type but also extra flags- one of which can be "position updated". In other changes: Add in a SA_QUIRK_NO_CPAGE quirk so that it's possible to avoid using a (broken) device's implementation of he DEVICE COMPRESSION page. Also do a couple of printout cleanups. As per some discussion on FreeBSD-scsi, skip doing tape flushing if we're reading tape logical block location (MTIOCRDSPOS).
2001-01-15 22:28:11 +00:00
csio->ccb_h.ccb_pflags |= SA_POSITION_UPDATED;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +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
}
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
}
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/*
* Incorrect Length usually applies to read, but can apply to writes.
*/
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
if (error == 0 && (stream_valid != 0) && (stream_bits & SSD_ILI)) {
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (info < 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(csio->ccb_h.path, toobig,
csio->dxfer_len - info);
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
csio->resid = csio->dxfer_len;
error = EIO;
} else {
csio->resid = resid;
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
if (softc->flags & SA_FLAG_FIXED) {
softc->flags |= SA_FLAG_EIO_PENDING;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
}
/*
* Bump the block number if we hadn't seen a filemark.
* Do this independent of errors (we've moved anyway).
*/
Add descriptor sense support to CAM, and honor sense residuals properly in CAM. Desriptor sense is a new sense data format that originated in SPC-3. Among other things, it allows for an 8-byte info field, which is necessary to pass back block numbers larger than 4 bytes. This change adds a number of new functions to scsi_all.c (and therefore libcam) that abstract out most access to sense data. This includes a bump of CAM_VERSION, because the CCB ABI has changed. Userland programs that use the CAM pass(4) driver will need to be recompiled. camcontrol.c: Change uses of scsi_extract_sense() to use scsi_extract_sense_len(). Use scsi_get_sks() instead of accessing sense key specific data directly. scsi_modes: Update the control mode page to the latest version (SPC-4). scsi_cmds.c, scsi_target.c: Change references to struct scsi_sense_data to struct scsi_sense_data_fixed. This should be changed to allow the user to specify fixed or descriptor sense, and then use scsi_set_sense_data() to build the sense data. ps3cdrom.c: Use scsi_set_sense_data() instead of setting sense data manually. cam_periph.c: Use scsi_extract_sense_len() instead of using scsi_extract_sense() or accessing sense data directly. cam_ccb.h: Bump the CAM_VERSION from 0x15 to 0x16. The change of struct scsi_sense_data from 32 to 252 bytes changes the size of struct ccb_scsiio, but not the size of union ccb. So the version must be bumped to prevent structure mis-matches. scsi_all.h: Lots of updated SCSI sense data and other structures. Add function prototypes for the new sense data functions. Take out the inline implementation of scsi_extract_sense(). It is now too large to put in a header file. Add macros to calculate whether fields are present and filled in fixed and descriptor sense data scsi_all.c: In scsi_op_desc(), allow the user to pass in NULL inquiry data, and we'll assume a direct access device in that case. Changed the SCSI RESERVED sense key name and description to COMPLETED, as it is now defined in the spec. Change the error recovery action for a number of read errors to prevent lots of retries when the drive has said that the block isn't accessible. This speeds up reconstruction of the block by any RAID software running on top of the drive (e.g. ZFS). In scsi_sense_desc(), allow for invalid sense key numbers. This allows calling this routine without checking the input values first. Change scsi_error_action() to use scsi_extract_sense_len(), and handle things when invalid asc/ascq values are encountered. Add a new routine, scsi_desc_iterate(), that will call the supplied function for every descriptor in descriptor format sense data. Add scsi_set_sense_data(), and scsi_set_sense_data_va(), which build descriptor and fixed format sense data. They currently default to fixed format sense data. Add a number of scsi_get_*() functions, which get different types of sense data fields from either fixed or descriptor format sense data, if the data is present. Add a number of scsi_*_sbuf() functions, which print formatted versions of various sense data fields. These functions work for either fixed or descriptor sense. Add a number of scsi_sense_*_sbuf() functions, which have a standard calling interface and print the indicated field. These functions take descriptors only. Add scsi_sense_desc_sbuf(), which will print a formatted version of the given sense descriptor. Pull out a majority of the scsi_sense_sbuf() function and put it into scsi_sense_only_sbuf(). This allows callers that don't use struct ccb_scsiio to easily utilize the printing routines. Revamp that function to handle descriptor sense and use the new sense fetching and printing routines. Move scsi_extract_sense() into scsi_all.c, and implement it in terms of the new function, scsi_extract_sense_len(). The _len() version takes a length (which should be the sense length - residual) and can indicate which fields are present and valid in the sense data. Add a couple of new scsi_get_*() routines to get the sense key, asc, and ascq only. mly.c: Rename struct scsi_sense_data to struct scsi_sense_data_fixed. sbp_targ.c: Use the new sense fetching routines to get sense data instead of accessing it directly. sbp.c: Change the firewire/SCSI sense data transformation code to use struct scsi_sense_data_fixed instead of struct scsi_sense_data. This should be changed later to use scsi_set_sense_data(). ciss.c: Calculate the sense residual properly. Use scsi_get_sense_key() to fetch the sense key. mps_sas.c, mpt_cam.c: Set the sense residual properly. iir.c: Use scsi_set_sense_data() instead of building sense data by hand. iscsi_subr.c: Use scsi_extract_sense_len() instead of grabbing sense data directly. umass.c: Use scsi_set_sense_data() to build sense data. Grab the sense key using scsi_get_sense_key(). Calculate the sense residual properly. isp_freebsd.h: Use scsi_get_*() routines to grab asc, ascq, and sense key values. Calculate and set the sense residual. MFC after: 3 days Sponsored by: Spectra Logic Corporation
2011-10-03 20:32:55 +00:00
if ((stream_valid == 0) ||
(stream_bits & SSD_FILEMARK) == 0) {
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (softc->blkno != (daddr_t) -1) {
softc->blkno++;
Fix PR 24220 by using the periph private field0 for a CCB for not only CCB type but also extra flags- one of which can be "position updated". In other changes: Add in a SA_QUIRK_NO_CPAGE quirk so that it's possible to avoid using a (broken) device's implementation of he DEVICE COMPRESSION page. Also do a couple of printout cleanups. As per some discussion on FreeBSD-scsi, skip doing tape flushing if we're reading tape logical block location (MTIOCRDSPOS).
2001-01-15 22:28:11 +00:00
csio->ccb_h.ccb_pflags |=
SA_POSITION_UPDATED;
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
}
}
}
}
backout last commit- inadvertant
2001-07-30 00:22:57 +00:00
Clear SA_FLAG_ERR_PENDING for MTREW, MTERASE and MTRETENS ioctl cases. Clear residual counts after a successful samount (the user doesn't care that we got an N-kbyte residual on our test read). Change a lot of error handling code. 1. If we end up in saerror, check more carefully about the kind of error. If it is a CAM_SCSI_STATUS_ERROR and it is a read/write command, we'll be handling this in saerror. If it isn't a read/write command, check to see whether this is just an EOM/EOP check condition- if it is, just set residual and return normally. A residual and then a NO SENSE check condiftion with the ASC of 0 and ASCQ of between 1 and 4 are normal 'signifying' events, not errors per se, and we shouldn't give the command to cam_periph_error to do something relatively unpredictable with. 2. If we get a Bus Reset, had a BDR sent, or get the cam status of CAM_REQUEUE_REQ, check the retry count on the command. The default error handler, cam_periph_error, doesn't honor retry count in these cases. This may change in the future, but for now, make sure we set EIO and return without calling cam_periph_error if the retry count for the command with an error is zero. 3. Clean up the pending error case goop and handle cases more sensibly. The rules are: If command was a Write: If we got a SSD_KEY_VOLUME_OVERFLOW, the resid is propagated and we set ENOSPC as the error. Else if we got an EOM condition- just mark EOM pending. And set a residual of zero. For the longest time I was just propagating residual from the sense data- but my tape comparison tests were always failing because all drives I tested with actually *do* write the data anyway- the EOM (early warning) condition occurred *prior* to all of the data going out to media- that is, it was still buffered by the drive. This case is described in SCSI-2, 10.2.14, paragraph #d for the meaning of 'information field'. A better fix for this would be to issue a WFM command of zero to cause the drive to flush any buffered data, but this would require a fairly extensive rewrite. Else if the command was a READ: If we got a SSD_KEY_BLANK_CHECK- If we have a One Filemark EOT model- mark EOM as pending, otherwise set EIO as the erorr. Else if we found a Filemark- If we're in Fixed Block mode- mark EOF pending. If we had an ILI (Incorrect Length Indicator)- If the residual is less than zero, whine about tape record being too big for user's buffer, otherwise if we were in Fixed Block mode, mark EIO as pending. All 'pending' conditions mean that the command in question completes without error indication. It had succeeded, but a signifying event occurred during its execution which will apply to the *next* command that would be exexcuted. Except for the one EOM case above, we always propagate residual. Now, way back in sastart- if we notice any of the PENDING bits set, we don't run the command we've just pulled off the wait queue. Instead, we then figure out it's disposition based upon a previous command's association with a signifying event. If SA_FLAG_EOM_PENDING is set, we don't set an error. We just complete the command with residual set to the request count (not data moved, but no error). We continue on. If SA_FLAG_EOF_PENDING- if we have this, it's only because we're in Fixed Block mode- in which case we traverse all waiting buffers (which we can get in fixed block mode because physio has split things up) and mark them all as no error, but no data moved and complete them. If SA_FLAG_EIO_PENDING, just mark the buffer with an EIO error and complete it. Then we clear all of the pending state bits- we're done. MFC after: 4 weeks
2001-08-30 16:25:24 +00:00
if (error <= 0) {
/*
* Unfreeze the queue if frozen as we're not returning anything
* to our waiters that would indicate an I/O error has occurred
* (yet).
*/
QFRLS(ccb);
error = 0;
}
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
}
static int
sagetparams(struct cam_periph *periph, sa_params params_to_get,
u_int32_t *blocksize, u_int8_t *density, u_int32_t *numblocks,
int *buff_mode, u_int8_t *write_protect, u_int8_t *speed,
int *comp_supported, int *comp_enabled, u_int32_t *comp_algorithm,
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
sa_comp_t *tcs)
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 *ccb;
void *mode_buffer;
struct scsi_mode_header_6 *mode_hdr;
struct scsi_mode_blk_desc *mode_blk;
int mode_buffer_len;
struct sa_softc *softc;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
u_int8_t cpage;
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 error;
cam_status status;
softc = (struct sa_softc *)periph->softc;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
ccb = cam_periph_getccb(periph, 1);
Fix PR 24220 by using the periph private field0 for a CCB for not only CCB type but also extra flags- one of which can be "position updated". In other changes: Add in a SA_QUIRK_NO_CPAGE quirk so that it's possible to avoid using a (broken) device's implementation of he DEVICE COMPRESSION page. Also do a couple of printout cleanups. As per some discussion on FreeBSD-scsi, skip doing tape flushing if we're reading tape logical block location (MTIOCRDSPOS).
2001-01-15 22:28:11 +00:00
if (softc->quirks & SA_QUIRK_NO_CPAGE)
cpage = SA_DEVICE_CONFIGURATION_PAGE;
else
cpage = SA_DATA_COMPRESSION_PAGE;
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:
mode_buffer_len = sizeof(*mode_hdr) + sizeof(*mode_blk);
if (params_to_get & SA_PARAM_COMPRESSION) {
if (softc->quirks & SA_QUIRK_NOCOMP) {
*comp_supported = FALSE;
params_to_get &= ~SA_PARAM_COMPRESSION;
} else
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
mode_buffer_len += sizeof (sa_comp_t);
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
}
Work around a malloc locking problem.
2007-06-16 18:13:26 +00:00
/* XXX Fix M_NOWAIT */
mode_buffer = malloc(mode_buffer_len, M_SCSISA, M_NOWAIT | M_ZERO);
if (mode_buffer == NULL) {
xpt_release_ccb(ccb);
return (ENOMEM);
}
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
mode_hdr = (struct scsi_mode_header_6 *)mode_buffer;
mode_blk = (struct scsi_mode_blk_desc *)&mode_hdr[1];
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
/* it is safe to retry this */
scsi_mode_sense(&ccb->csio, 5, sadone, MSG_SIMPLE_Q_TAG, FALSE,
SMS_PAGE_CTRL_CURRENT, (params_to_get & SA_PARAM_COMPRESSION) ?
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
cpage : SMS_VENDOR_SPECIFIC_PAGE, mode_buffer, mode_buffer_len,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
SSD_FULL_SIZE, SCSIOP_TIMEOUT);
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
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
error = cam_periph_runccb(ccb, saerror, 0, SF_NO_PRINT,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
softc->device_stats);
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
status = ccb->ccb_h.status & CAM_STATUS_MASK;
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
if (error == EINVAL && (params_to_get & SA_PARAM_COMPRESSION) != 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 a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
* Hmm. Let's see if we can try another page...
* If we've already done that, give up on compression
* for this device and remember this for the future
* and attempt the request without asking for compression
* info.
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (cpage == SA_DATA_COMPRESSION_PAGE) {
cpage = SA_DEVICE_CONFIGURATION_PAGE;
goto retry;
}
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->quirks |= SA_QUIRK_NOCOMP;
Eliminate the use of M_TEMP.
2007-05-14 21:48:53 +00:00
free(mode_buffer, M_SCSISA);
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
goto retry;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
} else if (status == CAM_SCSI_STATUS_ERROR) {
/* Tell the user about the fatal error. */
scsi_sense_print(&ccb->csio);
goto sagetparamsexit;
}
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/*
* If the user only wants the compression information, and
* the device doesn't send back the block descriptor, it's
* no big deal. If the user wants more than just
* compression, though, and the device doesn't pass back the
* block descriptor, we need to send another mode sense to
* get the block descriptor.
*/
if ((mode_hdr->blk_desc_len == 0) &&
(params_to_get & SA_PARAM_COMPRESSION) &&
(params_to_get & ~(SA_PARAM_COMPRESSION))) {
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
* Decrease the mode buffer length by the size of
* the compression page, to make sure the data
* there doesn't get overwritten.
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
mode_buffer_len -= sizeof (sa_comp_t);
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/*
* Now move the compression page that we presumably
* got back down the memory chunk a little bit so
* it doesn't get spammed.
*/
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
bcopy(&mode_hdr[0], &mode_hdr[1], sizeof (sa_comp_t));
bzero(&mode_hdr[0], sizeof (mode_hdr[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 a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/*
* Now, we issue another mode sense and just ask
* for the block descriptor, etc.
*/
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
scsi_mode_sense(&ccb->csio, 2, sadone, MSG_SIMPLE_Q_TAG, FALSE,
SMS_PAGE_CTRL_CURRENT, SMS_VENDOR_SPECIFIC_PAGE,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
mode_buffer, mode_buffer_len, SSD_FULL_SIZE,
SCSIOP_TIMEOUT);
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
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
error = cam_periph_runccb(ccb, saerror, 0, SF_NO_PRINT,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
softc->device_stats);
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (error != 0)
goto sagetparamsexit;
}
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (params_to_get & SA_PARAM_BLOCKSIZE)
*blocksize = scsi_3btoul(mode_blk->blklen);
if (params_to_get & SA_PARAM_NUMBLOCKS)
*numblocks = scsi_3btoul(mode_blk->nblocks);
if (params_to_get & SA_PARAM_BUFF_MODE)
*buff_mode = mode_hdr->dev_spec & SMH_SA_BUF_MODE_MASK;
if (params_to_get & SA_PARAM_DENSITY)
*density = mode_blk->density;
if (params_to_get & SA_PARAM_WP)
*write_protect = (mode_hdr->dev_spec & SMH_SA_WP)? TRUE : FALSE;
if (params_to_get & SA_PARAM_SPEED)
*speed = mode_hdr->dev_spec & SMH_SA_SPEED_MASK;
if (params_to_get & SA_PARAM_COMPRESSION) {
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
sa_comp_t *ntcs = (sa_comp_t *) &mode_blk[1];
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (cpage == SA_DATA_COMPRESSION_PAGE) {
struct scsi_data_compression_page *cp = &ntcs->dcomp;
*comp_supported =
(cp->dce_and_dcc & SA_DCP_DCC)? TRUE : FALSE;
*comp_enabled =
(cp->dce_and_dcc & SA_DCP_DCE)? TRUE : FALSE;
*comp_algorithm = scsi_4btoul(cp->comp_algorithm);
} else {
struct scsi_dev_conf_page *cp = &ntcs->dconf;
/*
* We don't really know whether this device supports
Fix typos - remove duplicate "the". PR: bin/154928 Submitted by: Eitan Adler <lists at eitanadler.com> MFC after: 3 days
2011-02-21 09:01:34 +00:00
* Data Compression if the algorithm field is
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
* zero. Just say we do.
*/
*comp_supported = TRUE;
*comp_enabled =
(cp->sel_comp_alg != SA_COMP_NONE)? TRUE : FALSE;
*comp_algorithm = cp->sel_comp_alg;
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
}
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (tcs != NULL)
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
bcopy(ntcs, tcs, sizeof (sa_comp_t));
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
}
if (CAM_DEBUGGED(periph->path, CAM_DEBUG_INFO)) {
int idx;
char *xyz = mode_buffer;
xpt_print_path(periph->path);
printf("Mode Sense Data=");
for (idx = 0; idx < mode_buffer_len; idx++)
printf(" 0x%02x", xyz[idx] & 0xff);
printf("\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
}
sagetparamsexit:
xpt_release_ccb(ccb);
Eliminate the use of M_TEMP.
2007-05-14 21:48:53 +00:00
free(mode_buffer, M_SCSISA);
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +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
}
/*
* The purpose of this function is to set one of four different parameters
* for a tape drive:
* - blocksize
* - density
* - compression / compression algorithm
* - buffering mode
*
* The assumption is that this will be called from saioctl(), and therefore
* from a process context. Thus the waiting malloc calls below. If that
* assumption ever changes, the malloc calls should be changed to be
* NOWAIT mallocs.
*
* Any or all of the four parameters may be set when this function is
* called. It should handle setting more than one parameter at once.
*/
static int
sasetparams(struct cam_periph *periph, sa_params params_to_set,
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
u_int32_t blocksize, u_int8_t density, u_int32_t calg,
Clean up and fix quirk table (was missing necessary wildcards) and add a couple from the old driver. Change format of quirk table to have a preferred block size for devices that need to be QUIRK_FIXED- this is loaded into the last_media_blocksize tag at saregister time and will be used in the first samount case. Change sasetparams to take a sense_flags argument so that probe time testing can be quieter (e.g. with SF_NO_PRINT). Fix a couple of silly bugs in the fixed/variable determination in samount- one was where there was a check against 'guessing' AND the density code being default density- *SMACK* - you're only guessing if you find the media code to be *other* than default density. Second bug was a test against current blocksize being zero- should be a test against whether current blocksize is not equal to the last blocksize if you had wanted to be fixed (suppose you came up in fixed, but not the preferred size?). And if you don't know what the fixed size should be, select 512 as the starting point, not BLKDEV_IOSIZE (reality wins). Finally, in doing the test set to variable mode, make sasetparams non-chatty.
1999-01-12 08:15:47 +00:00
u_int32_t sense_flags)
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 sa_softc *softc;
u_int32_t current_blocksize;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
u_int32_t current_calg;
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 current_density;
u_int8_t current_speed;
int comp_enabled, comp_supported;
void *mode_buffer;
int mode_buffer_len;
struct scsi_mode_header_6 *mode_hdr;
struct scsi_mode_blk_desc *mode_blk;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
sa_comp_t *ccomp, *cpage;
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 buff_mode;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
union ccb *ccb = 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
int error;
softc = (struct sa_softc *)periph->softc;
Work around the other uses of M_WAITOK.
2007-06-16 18:20:29 +00:00
ccomp = malloc(sizeof (sa_comp_t), M_SCSISA, M_NOWAIT);
if (ccomp == NULL)
return (ENOMEM);
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
/*
* Since it doesn't make sense to set the number of blocks, or
* write protection, we won't try to get the current value. We
* always want to get the blocksize, so we can set it back to the
* proper value.
*/
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
error = sagetparams(periph,
params_to_set | SA_PARAM_BLOCKSIZE | SA_PARAM_SPEED,
&current_blocksize, &current_density, NULL, &buff_mode, NULL,
&current_speed, &comp_supported, &comp_enabled,
&current_calg, ccomp);
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) {
Eliminate the use of M_TEMP.
2007-05-14 21:48:53 +00:00
free(ccomp, M_SCSISA);
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +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
}
mode_buffer_len = sizeof(*mode_hdr) + sizeof(*mode_blk);
if (params_to_set & SA_PARAM_COMPRESSION)
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
mode_buffer_len += sizeof (sa_comp_t);
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
Work around the other uses of M_WAITOK.
2007-06-16 18:20:29 +00:00
mode_buffer = malloc(mode_buffer_len, M_SCSISA, M_NOWAIT | M_ZERO);
if (mode_buffer == NULL) {
free(ccomp, M_SCSISA);
return (ENOMEM);
}
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
mode_hdr = (struct scsi_mode_header_6 *)mode_buffer;
mode_blk = (struct scsi_mode_blk_desc *)&mode_hdr[1];
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
ccb = cam_periph_getccb(periph, 1);
retry:
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 (params_to_set & SA_PARAM_COMPRESSION) {
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
if (mode_blk) {
cpage = (sa_comp_t *)&mode_blk[1];
} else {
cpage = (sa_comp_t *)&mode_hdr[1];
}
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
bcopy(ccomp, cpage, sizeof (sa_comp_t));
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
cpage->hdr.pagecode &= ~0x80;
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
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
cpage = 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 the caller wants us to set the blocksize, use the one they
* pass in. Otherwise, use the blocksize we got back from the
* mode select above.
*/
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
if (mode_blk) {
if (params_to_set & SA_PARAM_BLOCKSIZE)
scsi_ulto3b(blocksize, mode_blk->blklen);
else
scsi_ulto3b(current_blocksize, mode_blk->blklen);
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 for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
/*
* Set density if requested, else preserve old density.
* SCSI_SAME_DENSITY only applies to SCSI-2 or better
* devices, else density we've latched up in our softc.
*/
if (params_to_set & SA_PARAM_DENSITY) {
mode_blk->density = density;
} else if (softc->scsi_rev > SCSI_REV_CCS) {
mode_blk->density = SCSI_SAME_DENSITY;
} else {
mode_blk->density = softc->media_density;
}
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +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
/*
* For mode selects, these two fields must be zero.
*/
mode_hdr->data_length = 0;
mode_hdr->medium_type = 0;
/* set the speed to the current value */
mode_hdr->dev_spec = current_speed;
I forgot whom I got this from- only set single initiator buffered mode if we've recorded in our softc that we should set it.
2003-09-13 02:01:56 +00:00
/* if set, set single-initiator buffering mode */
if (softc->buffer_mode == SMH_SA_BUF_MODE_SIBUF) {
mode_hdr->dev_spec |= SMH_SA_BUF_MODE_SIBUF;
}
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 for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
if (mode_blk)
mode_hdr->blk_desc_len = sizeof(struct scsi_mode_blk_desc);
else
mode_hdr->blk_desc_len = 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
/*
* First, if the user wants us to set the compression algorithm or
* just turn compression on, check to make sure that this drive
* supports compression.
*/
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (params_to_set & SA_PARAM_COMPRESSION) {
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 compression algorithm is 0, disable compression.
* If the compression algorithm is non-zero, enable
* compression and set the compression type to the
* specified compression algorithm, unless the algorithm is
* MT_COMP_ENABLE. In that case, we look at the
* compression algorithm that is currently set and if it is
* non-zero, we leave it as-is. If it is zero, and we have
* saved a compression algorithm from a time when
* compression was enabled before, set the compression to
* the saved value.
*/
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
switch (ccomp->hdr.pagecode & ~0x80) {
Check in some months long pending minor debug output changes. Clarify that the implicit fallthrough was *not* intentional (thanks, Poul!) and reorganize the code so a correct fallthrough (with /* FALLTHROUGH */) occurs.
2003-06-02 00:37:42 +00:00
case SA_DEVICE_CONFIGURATION_PAGE:
{
struct scsi_dev_conf_page *dcp = &cpage->dconf;
if (calg == 0) {
dcp->sel_comp_alg = SA_COMP_NONE;
break;
}
if (calg != MT_COMP_ENABLE) {
dcp->sel_comp_alg = calg;
} else if (dcp->sel_comp_alg == SA_COMP_NONE &&
softc->saved_comp_algorithm != 0) {
dcp->sel_comp_alg = softc->saved_comp_algorithm;
}
break;
}
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
case SA_DATA_COMPRESSION_PAGE:
if (ccomp->dcomp.dce_and_dcc & SA_DCP_DCC) {
struct scsi_data_compression_page *dcp = &cpage->dcomp;
if (calg == 0) {
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
/*
* Disable compression, but leave the
* decompression and the capability bit
* alone.
*/
dcp->dce_and_dcc = SA_DCP_DCC;
dcp->dde_and_red |= SA_DCP_DDE;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
break;
}
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
/* enable compression && decompression */
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
dcp->dce_and_dcc = SA_DCP_DCE | SA_DCP_DCC;
dcp->dde_and_red |= SA_DCP_DDE;
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
/*
* If there, use compression algorithm from caller.
* Otherwise, if there's a saved compression algorithm
* and there is no current algorithm, use the saved
* algorithm. Else parrot back what we got and hope
* for the best.
*/
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (calg != MT_COMP_ENABLE) {
scsi_ulto4b(calg, dcp->comp_algorithm);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
scsi_ulto4b(calg, dcp->decomp_algorithm);
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
} else if (scsi_4btoul(dcp->comp_algorithm) == 0 &&
softc->saved_comp_algorithm != 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
scsi_ulto4b(softc->saved_comp_algorithm,
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
dcp->comp_algorithm);
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
scsi_ulto4b(softc->saved_comp_algorithm,
dcp->decomp_algorithm);
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +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
}
Check in some months long pending minor debug output changes. Clarify that the implicit fallthrough was *not* intentional (thanks, Poul!) and reorganize the code so a correct fallthrough (with /* FALLTHROUGH */) occurs.
2003-06-02 00:37:42 +00:00
/*
* Compression does not appear to be supported-
* at least via the DATA COMPRESSION page. It
* would be too much to ask us to believe that
* the page itself is supported, but incorrectly
* reports an ability to manipulate data compression,
* so we'll assume that this device doesn't support
* compression. We can just fall through for that.
*/
/* FALLTHROUGH */
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
default:
/*
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
* The drive doesn't seem to support compression,
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
* so turn off the set compression bit.
*/
params_to_set &= ~SA_PARAM_COMPRESSION;
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,
"device does not seem to support compression\n");
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/*
* If that was the only thing the user wanted us to set,
* clean up allocated resources and return with
* 'operation not supported'.
*/
if (params_to_set == SA_PARAM_NONE) {
Eliminate the use of M_TEMP.
2007-05-14 21:48:53 +00:00
free(mode_buffer, M_SCSISA);
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
xpt_release_ccb(ccb);
return (ENODEV);
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
}
/*
* That wasn't the only thing the user wanted us to set.
* So, decrease the stated mode buffer length by the
* size of the compression mode page.
*/
mode_buffer_len -= sizeof(sa_comp_t);
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/* It is safe to retry this operation */
scsi_mode_select(&ccb->csio, 5, sadone, MSG_SIMPLE_Q_TAG,
(params_to_set & SA_PARAM_COMPRESSION)? TRUE : FALSE,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
FALSE, mode_buffer, mode_buffer_len, SSD_FULL_SIZE, SCSIOP_TIMEOUT);
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
error = cam_periph_runccb(ccb, saerror, 0,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
sense_flags, softc->device_stats);
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (CAM_DEBUGGED(periph->path, CAM_DEBUG_INFO)) {
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
int idx;
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
char *xyz = mode_buffer;
xpt_print_path(periph->path);
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
printf("Err%d, Mode Select Data=", error);
Several changes having to do blocksize- mostly to force variable as the default. Attempt to determine (at mount time if not done so already) via density code whether a device should default to fixed mode or not. Attempts to set to variable that fail will cause fixed to be selected. Similarly, the '2 filemarks at EOM' quirk is now determined (or attempted to be determined) via density code. Some as yet not entirely tested code for coping with 2FM@EOD position is now also in place.
1998-12-17 18:56:23 +00:00
for (idx = 0; idx < mode_buffer_len; idx++)
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
printf(" 0x%02x", xyz[idx] & 0xff);
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
printf("\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 for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
if (error) {
/*
* If we can, try without setting density/blocksize.
*/
if (mode_blk) {
if ((params_to_set &
(SA_PARAM_DENSITY|SA_PARAM_BLOCKSIZE)) == 0) {
mode_blk = NULL;
goto retry;
}
} else {
mode_blk = (struct scsi_mode_blk_desc *)&mode_hdr[1];
cpage = (sa_comp_t *)&mode_blk[1];
}
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 were setting the blocksize, and that failed, we
* want to set it to its original value. If we weren't
* setting the blocksize, we don't want to change it.
*/
scsi_ulto3b(current_blocksize, mode_blk->blklen);
/*
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
* Set density if requested, else preserve old density.
* SCSI_SAME_DENSITY only applies to SCSI-2 or better
* devices, else density we've latched up in our 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
*/
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
if (params_to_set & SA_PARAM_DENSITY) {
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
mode_blk->density = current_density;
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
} else if (softc->scsi_rev > SCSI_REV_CCS) {
mode_blk->density = SCSI_SAME_DENSITY;
} else {
mode_blk->density = softc->media_density;
}
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 (params_to_set & SA_PARAM_COMPRESSION)
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
bcopy(ccomp, cpage, sizeof (sa_comp_t));
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
/*
* The retry count is the only CCB field that might have been
* changed that we care about, so reset it back to 1.
*/
ccb->ccb_h.retry_count = 1;
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
cam_periph_runccb(ccb, saerror, 0, sense_flags,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
softc->device_stats);
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 for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
xpt_release_ccb(ccb);
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
if (ccomp != NULL)
Eliminate the use of M_TEMP.
2007-05-14 21:48:53 +00:00
free(ccomp, M_SCSISA);
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
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
if (params_to_set & SA_PARAM_COMPRESSION) {
if (error) {
softc->flags &= ~SA_FLAG_COMP_ENABLED;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
/*
* Even if we get an error setting compression,
* do not say that we don't support it. We could
* have been wrong, or it may be media specific.
* softc->flags &= ~SA_FLAG_COMP_SUPP;
*/
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
softc->saved_comp_algorithm = softc->comp_algorithm;
softc->comp_algorithm = 0;
} else {
softc->flags |= SA_FLAG_COMP_ENABLED;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
softc->comp_algorithm = calg;
1) Fix some serious bugs (1 botch on my part which caused a filemark to be written even it the tape was opened readonly- 2 botches in deferred error handling for FIXED LENGTH mode which caused panic && hand resp.). Fixed a memory leak in sa_mount. 2) Fixed an annoying bug when turning of compression to actually reflect this for future status calls. 3) Implement the MTIOCERRSTAT call where latched control and I/O residuals and sense data are returned to the application asking for them.
1998-12-19 23:33:21 +00:00
}
}
Eliminate the use of M_TEMP.
2007-05-14 21:48:53 +00:00
free(mode_buffer, M_SCSISA);
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +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
}
static void
saprevent(struct cam_periph *periph, int action)
{
struct sa_softc *softc;
union ccb *ccb;
Sascha Blank <blank@uni-trier.de> convinced me I was an idiot about testing SA_QUIRK_2FM in samount. Fixed. Removed the NORRLS quirk (to save quirk space) and left the behaviour of being quiet about failed reserve/release (failed due Illegal Request) the same. Added a SF_QUIET_IR for prevent/allow for the same purposes.
1999-01-16 19:20:30 +00:00
int error, sf;
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 sa_softc *)periph->softc;
Sascha Blank <blank@uni-trier.de> convinced me I was an idiot about testing SA_QUIRK_2FM in samount. Fixed. Removed the NORRLS quirk (to save quirk space) and left the behaviour of being quiet about failed reserve/release (failed due Illegal Request) the same. Added a SF_QUIET_IR for prevent/allow for the same purposes.
1999-01-16 19:20:30 +00:00
if ((action == PR_ALLOW) && (softc->flags & SA_FLAG_TAPE_LOCKED) == 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
return;
Sascha Blank <blank@uni-trier.de> convinced me I was an idiot about testing SA_QUIRK_2FM in samount. Fixed. Removed the NORRLS quirk (to save quirk space) and left the behaviour of being quiet about failed reserve/release (failed due Illegal Request) the same. Added a SF_QUIET_IR for prevent/allow for the same purposes.
1999-01-16 19:20:30 +00:00
if ((action == PR_PREVENT) && (softc->flags & SA_FLAG_TAPE_LOCKED) != 0)
return;
Add a quirk type (and one for ARCHIVE Python's) that disables the dummy read at sa_mount that attempts to latch up density. This breaks on at least one drive and sends it into the weeds. Approved: jkh
2000-02-03 18:29:25 +00:00
/*
* We can be quiet about illegal requests.
*/
if (CAM_DEBUGGED(periph->path, CAM_DEBUG_INFO)) {
Sascha Blank <blank@uni-trier.de> convinced me I was an idiot about testing SA_QUIRK_2FM in samount. Fixed. Removed the NORRLS quirk (to save quirk space) and left the behaviour of being quiet about failed reserve/release (failed due Illegal Request) the same. Added a SF_QUIET_IR for prevent/allow for the same purposes.
1999-01-16 19:20:30 +00:00
sf = 0;
Add a quirk type (and one for ARCHIVE Python's) that disables the dummy read at sa_mount that attempts to latch up density. This breaks on at least one drive and sends it into the weeds. Approved: jkh
2000-02-03 18:29:25 +00:00
} else
Sascha Blank <blank@uni-trier.de> convinced me I was an idiot about testing SA_QUIRK_2FM in samount. Fixed. Removed the NORRLS quirk (to save quirk space) and left the behaviour of being quiet about failed reserve/release (failed due Illegal Request) the same. Added a SF_QUIET_IR for prevent/allow for the same purposes.
1999-01-16 19:20:30 +00:00
sf = SF_QUIET_IR;
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
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
ccb = cam_periph_getccb(periph, 1);
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
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
/* It is safe to retry this operation */
scsi_prevent(&ccb->csio, 5, sadone, MSG_SIMPLE_Q_TAG, action,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
SSD_FULL_SIZE, SCSIOP_TIMEOUT);
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
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
error = cam_periph_runccb(ccb, saerror, 0, sf, softc->device_stats);
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 &= ~SA_FLAG_TAPE_LOCKED;
else
softc->flags |= SA_FLAG_TAPE_LOCKED;
}
xpt_release_ccb(ccb);
}
static int
sarewind(struct cam_periph *periph)
{
union ccb *ccb;
struct sa_softc *softc;
int error;
softc = (struct sa_softc *)periph->softc;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
ccb = cam_periph_getccb(periph, 1);
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
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
/* It is safe to retry this operation */
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
scsi_rewind(&ccb->csio, 2, sadone, MSG_SIMPLE_Q_TAG, FALSE,
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
SSD_FULL_SIZE, REWIND_TIMEOUT);
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
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_REW;
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
error = cam_periph_runccb(ccb, saerror, 0, 0, softc->device_stats);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_REST;
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_release_ccb(ccb);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
if (error == 0)
softc->fileno = softc->blkno = (daddr_t) 0;
else
softc->fileno = softc->blkno = (daddr_t) -1;
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
saspace(struct cam_periph *periph, int count, scsi_space_code code)
{
union ccb *ccb;
struct sa_softc *softc;
int error;
softc = (struct sa_softc *)periph->softc;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
ccb = cam_periph_getccb(periph, 1);
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
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
/* This cannot be retried */
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
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
scsi_space(&ccb->csio, 0, sadone, MSG_SIMPLE_Q_TAG, code, count,
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
SSD_FULL_SIZE, SPACE_TIMEOUT);
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
Use the residual from a spacing operation to try and figure out where we *really* are. It should be noted that there is a degenerate case where soft tape location will be lost (not causing a frozen state- but causing the loss of reporting fileno/blockno)- that's where you backspace over a filemark- you stop backspacing as soon as you cross the filemark, but you have no idea what the record number now is because you have no idea how many records you are into the file you just backed into. Such is life. While I'm at it, also pick up residuals from writing filemarks. PR: 24222
2001-01-16 00:53:45 +00:00
/*
* Clear residual because we will be using it.
*/
softc->last_ctl_resid = 0;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = (count < 0)? MTIO_DSREG_REV : MTIO_DSREG_FWD;
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
error = cam_periph_runccb(ccb, saerror, 0, 0, softc->device_stats);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_REST;
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_release_ccb(ccb);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
/*
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
* If a spacing operation has failed, we need to invalidate
* this mount.
*
* If the spacing operation was setmarks or to end of recorded data,
* we no longer know our relative position.
*
Use the residual from a spacing operation to try and figure out where we *really* are. It should be noted that there is a degenerate case where soft tape location will be lost (not causing a frozen state- but causing the loss of reporting fileno/blockno)- that's where you backspace over a filemark- you stop backspacing as soon as you cross the filemark, but you have no idea what the record number now is because you have no idea how many records you are into the file you just backed into. Such is life. While I'm at it, also pick up residuals from writing filemarks. PR: 24222
2001-01-16 00:53:45 +00:00
* If the spacing operations was spacing files in reverse, we
* take account of the residual, but still check against less
* than zero- if we've gone negative, we must have hit BOT.
*
* If the spacing operations was spacing records in reverse and
* we have a residual, we've either hit BOT or hit a filemark.
* In the former case, we know our new record number (0). In
* the latter case, we have absolutely no idea what the real
* record number is- we've stopped between the end of the last
* record in the previous file and the filemark that stopped
* our spacing backwards.
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
*/
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
if (error) {
softc->fileno = softc->blkno = (daddr_t) -1;
} else if (code == SS_SETMARKS || code == SS_EOD) {
softc->fileno = softc->blkno = (daddr_t) -1;
} else if (code == SS_FILEMARKS && softc->fileno != (daddr_t) -1) {
Use the residual from a spacing operation to try and figure out where we *really* are. It should be noted that there is a degenerate case where soft tape location will be lost (not causing a frozen state- but causing the loss of reporting fileno/blockno)- that's where you backspace over a filemark- you stop backspacing as soon as you cross the filemark, but you have no idea what the record number now is because you have no idea how many records you are into the file you just backed into. Such is life. While I'm at it, also pick up residuals from writing filemarks. PR: 24222
2001-01-16 00:53:45 +00:00
softc->fileno += (count - softc->last_ctl_resid);
if (softc->fileno < 0) /* we must of hit BOT */
softc->fileno = 0;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->blkno = 0;
} else if (code == SS_BLOCKS && softc->blkno != (daddr_t) -1) {
Use the residual from a spacing operation to try and figure out where we *really* are. It should be noted that there is a degenerate case where soft tape location will be lost (not causing a frozen state- but causing the loss of reporting fileno/blockno)- that's where you backspace over a filemark- you stop backspacing as soon as you cross the filemark, but you have no idea what the record number now is because you have no idea how many records you are into the file you just backed into. Such is life. While I'm at it, also pick up residuals from writing filemarks. PR: 24222
2001-01-16 00:53:45 +00:00
softc->blkno += (count - softc->last_ctl_resid);
if (count < 0) {
if (softc->last_ctl_resid || softc->blkno < 0) {
if (softc->fileno == 0) {
softc->blkno = 0;
} else {
softc->blkno = (daddr_t) -1;
}
}
}
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +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
return (error);
}
static int
sawritefilemarks(struct cam_periph *periph, int nmarks, int setmarks)
{
union ccb *ccb;
struct sa_softc *softc;
Use the residual from a spacing operation to try and figure out where we *really* are. It should be noted that there is a degenerate case where soft tape location will be lost (not causing a frozen state- but causing the loss of reporting fileno/blockno)- that's where you backspace over a filemark- you stop backspacing as soon as you cross the filemark, but you have no idea what the record number now is because you have no idea how many records you are into the file you just backed into. Such is life. While I'm at it, also pick up residuals from writing filemarks. PR: 24222
2001-01-16 00:53:45 +00:00
int error, nwm = 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 = (struct sa_softc *)periph->softc;
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
if (softc->open_rdonly)
return (EBADF);
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
ccb = cam_periph_getccb(periph, 1);
Use the residual from a spacing operation to try and figure out where we *really* are. It should be noted that there is a degenerate case where soft tape location will be lost (not causing a frozen state- but causing the loss of reporting fileno/blockno)- that's where you backspace over a filemark- you stop backspacing as soon as you cross the filemark, but you have no idea what the record number now is because you have no idea how many records you are into the file you just backed into. Such is life. While I'm at it, also pick up residuals from writing filemarks. PR: 24222
2001-01-16 00:53:45 +00:00
/*
* Clear residual because we will be using it.
*/
softc->last_ctl_resid = 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
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_FMK;
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
/* this *must* not be retried */
scsi_write_filemarks(&ccb->csio, 0, sadone, MSG_SIMPLE_Q_TAG,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
FALSE, setmarks, nmarks, SSD_FULL_SIZE, IO_TIMEOUT);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_REST;
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
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
error = cam_periph_runccb(ccb, saerror, 0, 0, softc->device_stats);
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 in block position/block locate functions.
1998-12-18 04:31:43 +00:00
if (error == 0 && nmarks) {
struct sa_softc *softc = (struct sa_softc *)periph->softc;
Use the residual from a spacing operation to try and figure out where we *really* are. It should be noted that there is a degenerate case where soft tape location will be lost (not causing a frozen state- but causing the loss of reporting fileno/blockno)- that's where you backspace over a filemark- you stop backspacing as soon as you cross the filemark, but you have no idea what the record number now is because you have no idea how many records you are into the file you just backed into. Such is life. While I'm at it, also pick up residuals from writing filemarks. PR: 24222
2001-01-16 00:53:45 +00:00
nwm = nmarks - softc->last_ctl_resid;
softc->filemarks += nwm;
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 the residual from a spacing operation to try and figure out where we *really* are. It should be noted that there is a degenerate case where soft tape location will be lost (not causing a frozen state- but causing the loss of reporting fileno/blockno)- that's where you backspace over a filemark- you stop backspacing as soon as you cross the filemark, but you have no idea what the record number now is because you have no idea how many records you are into the file you just backed into. Such is life. While I'm at it, also pick up residuals from writing filemarks. PR: 24222
2001-01-16 00:53:45 +00:00
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
xpt_release_ccb(ccb);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
/*
* Update relative positions (if we're doing that).
*/
if (error) {
softc->fileno = softc->blkno = (daddr_t) -1;
} else if (softc->fileno != (daddr_t) -1) {
Use the residual from a spacing operation to try and figure out where we *really* are. It should be noted that there is a degenerate case where soft tape location will be lost (not causing a frozen state- but causing the loss of reporting fileno/blockno)- that's where you backspace over a filemark- you stop backspacing as soon as you cross the filemark, but you have no idea what the record number now is because you have no idea how many records you are into the file you just backed into. Such is life. While I'm at it, also pick up residuals from writing filemarks. PR: 24222
2001-01-16 00:53:45 +00:00
softc->fileno += nwm;
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->blkno = 0;
}
Add in block position/block locate functions.
1998-12-18 04:31:43 +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
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
static int
sardpos(struct cam_periph *periph, int hard, u_int32_t *blkptr)
{
struct scsi_tape_position_data loc;
union ccb *ccb;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
struct sa_softc *softc = (struct sa_softc *)periph->softc;
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
int error;
/*
Fix PR 24220 by using the periph private field0 for a CCB for not only CCB type but also extra flags- one of which can be "position updated". In other changes: Add in a SA_QUIRK_NO_CPAGE quirk so that it's possible to avoid using a (broken) device's implementation of he DEVICE COMPRESSION page. Also do a couple of printout cleanups. As per some discussion on FreeBSD-scsi, skip doing tape flushing if we're reading tape logical block location (MTIOCRDSPOS).
2001-01-15 22:28:11 +00:00
* We try and flush any buffered writes here if we were writing
* and we're trying to get hardware block position. It eats
* up performance substantially, but I'm wary of drive firmware.
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
*
Fix PR 24220 by using the periph private field0 for a CCB for not only CCB type but also extra flags- one of which can be "position updated". In other changes: Add in a SA_QUIRK_NO_CPAGE quirk so that it's possible to avoid using a (broken) device's implementation of he DEVICE COMPRESSION page. Also do a couple of printout cleanups. As per some discussion on FreeBSD-scsi, skip doing tape flushing if we're reading tape logical block location (MTIOCRDSPOS).
2001-01-15 22:28:11 +00:00
* I think that *logical* block position is probably okay-
* but hardware block position might have to wait for data
* to hit media to be valid. Caveat Emptor.
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
*/
Fix PR 24220 by using the periph private field0 for a CCB for not only CCB type but also extra flags- one of which can be "position updated". In other changes: Add in a SA_QUIRK_NO_CPAGE quirk so that it's possible to avoid using a (broken) device's implementation of he DEVICE COMPRESSION page. Also do a couple of printout cleanups. As per some discussion on FreeBSD-scsi, skip doing tape flushing if we're reading tape logical block location (MTIOCRDSPOS).
2001-01-15 22:28:11 +00:00
if (hard && (softc->flags & SA_FLAG_TAPE_WRITTEN)) {
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
error = sawritefilemarks(periph, 0, 0);
if (error && error != EACCES)
return (error);
}
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
Grab the ccb *after* writing filemarks. PR: 21723 Submitted by: razuwaev@relex.ru
2000-10-05 17:02:20 +00:00
ccb = cam_periph_getccb(periph, 1);
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
scsi_read_position(&ccb->csio, 1, sadone, MSG_SIMPLE_Q_TAG,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
hard, &loc, SSD_FULL_SIZE, SCSIOP_TIMEOUT);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_RBSY;
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
error = cam_periph_runccb(ccb, saerror, 0, 0, softc->device_stats);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_REST;
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
if (error == 0) {
if (loc.flags & SA_RPOS_UNCERTAIN) {
error = EINVAL; /* nothing is certain */
} else {
*blkptr = scsi_4btoul(loc.firstblk);
}
}
xpt_release_ccb(ccb);
return (error);
}
static int
sasetpos(struct cam_periph *periph, int hard, u_int32_t *blkptr)
{
union ccb *ccb;
struct sa_softc *softc;
int error;
/*
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
* We used to try and flush any buffered writes here.
* Now we push this onto user applications to either
* flush the pending writes themselves (via a zero count
* WRITE FILEMARKS command) or they can trust their tape
* drive to do this correctly for them.
*/
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
softc = (struct sa_softc *)periph->softc;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
ccb = cam_periph_getccb(periph, 1);
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
scsi_set_position(&ccb->csio, 1, sadone, MSG_SIMPLE_Q_TAG,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
hard, *blkptr, SSD_FULL_SIZE, SPACE_TIMEOUT);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_POS;
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
error = cam_periph_runccb(ccb, saerror, 0, 0, softc->device_stats);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_REST;
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_release_ccb(ccb);
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
/*
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
* Note relative file && block number position as now unknown.
Add in block position/block locate functions.
1998-12-18 04:31:43 +00:00
*/
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->fileno = softc->blkno = (daddr_t) -1;
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
saretension(struct cam_periph *periph)
{
union ccb *ccb;
struct sa_softc *softc;
int error;
softc = (struct sa_softc *)periph->softc;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
ccb = cam_periph_getccb(periph, 1);
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
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
/* It is safe to retry this operation */
scsi_load_unload(&ccb->csio, 5, sadone, MSG_SIMPLE_Q_TAG, FALSE,
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
FALSE, TRUE, TRUE, SSD_FULL_SIZE, ERASE_TIMEOUT);
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
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_TEN;
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
error = cam_periph_runccb(ccb, saerror, 0, 0, softc->device_stats);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_REST;
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_release_ccb(ccb);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
if (error == 0)
softc->fileno = softc->blkno = (daddr_t) 0;
else
softc->fileno = softc->blkno = (daddr_t) -1;
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +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
}
static int
sareservereleaseunit(struct cam_periph *periph, int reserve)
{
union ccb *ccb;
struct sa_softc *softc;
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +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
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
softc = (struct sa_softc *)periph->softc;
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
ccb = cam_periph_getccb(periph, 1);
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
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
/* It is safe to retry this operation */
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
scsi_reserve_release_unit(&ccb->csio, 2, sadone, MSG_SIMPLE_Q_TAG,
A slightly more complete change to timeouts: 1. Add SA_IO_TIMEOUT as an option (4 minutes default) to cover reads, writes, wfm, test unit ready. 2. Add internal SCSIOP_TIMEOUT (e.g., for mode sense) at 1 minute. This should not require an option, but is cleaner to parameterize. MFC after: 1 week
2001-07-02 17:48:59 +00:00
FALSE, 0, SSD_FULL_SIZE, SCSIOP_TIMEOUT, reserve);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_RBSY;
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
error = cam_periph_runccb(ccb, saerror, 0,
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
SF_RETRY_UA | SF_NO_PRINT, softc->device_stats);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_REST;
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_release_ccb(ccb);
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
/*
* If the error was Illegal Request, then the device doesn't support
* RESERVE/RELEASE. This is not an error.
*/
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
if (error == EINVAL) {
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +00:00
error = 0;
Add a quirk NORRLS (no reserve/release) which can (and will) get set for the devices that don't actually support reserve/release (so we don't keep trying it). Add softc storage and manage storing last I/O and CTL commands that had errors (for correlative purposes). In saclose clear the 'MOUNTED' bit if we either rewind or unload (yes, this shouldn't be necessary since the next open should catch whether a tape change occurred, but I'm having some questions about that actually working so this is safer for the moment). Oh, forgot to mention in previous commit messages that some of the failures particularly at close time cause the tape to be ejected (for the sake of safety)- all this prior to redoing the state machine (which is in progress) which will try and handle this better. Complete the addition of the setmark support (from Martin.Birgmeier@aon.at).
1998-12-22 17:26:13 +00:00
}
Some fixes to handle fixed mode and variable mode more sensibly- and also incorporate some notion of which revision the device is. If it's < SCSI2, for example, READ BLOCK LIMITS is not a MANDATORY command. At any rate, the initial state is to try and read block limits to get a notion of the smallest and largest record size as well as the granularity. However, this doesn't mean that the device should actually *in* fixed block mode should the max && min be equal... *That* choice is (for now) determined by whether the device comes up with a blocksize of nonzero. If so, then it's a fixed block preferred device, otherwise not (this will change again soon). When actually doing I/O, and you're in fixed length mode, the block count is *not* the byte count divided by the minimum block size- it's the byte count divided by the current blocksize (or use shift/mask shortcuts if that worked out...). Then when you *change* the blocksize via an ioctl, make sure this actually propagates to the stored notion of blocksize (and update the shift/mask shortcuts). Misc Other: When doing a mode select, only use the SCSI_SAME_DENSITY (0x7f) code if the device is >= SCSI2- otherwise just use the saved density code. Recover from the ripple of ILLEGAL REQUEST not being 'retried' in that RESERVE/RELEASE is not a mandatory command for < SCSI2 (so ignore it if it fails).
1998-12-11 07:19:36 +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
saloadunload(struct cam_periph *periph, int load)
{
union ccb *ccb;
struct sa_softc *softc;
int error;
softc = (struct sa_softc *)periph->softc;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
ccb = cam_periph_getccb(periph, 1);
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
More bandaids. One important one from Sascha Blank (<blank@fox.uni-trier.de>) about quirks being set as arithmetic values, not as bitfields. Add HP, Kennedy and M4 1/2" reel quirk entries. Do a lot of gratuitous source changing. Audit all functions that build ccbs for the tape driver and decide whether each one can be retried or not. Still to do is some more state management post errors.
1999-01-16 04:02:31 +00:00
/* It is safe to retry this operation */
scsi_load_unload(&ccb->csio, 5, sadone, MSG_SIMPLE_Q_TAG, FALSE,
Correct some botched timeout defines. Allocate only 8KB for the test read in samount. Make things a lot quieter in samount (and other places). Fix ridiculous and not so ridiculous bugs in compression related code in both sagetparams and sasetparams.
1999-12-03 23:14:11 +00:00
FALSE, FALSE, load, SSD_FULL_SIZE, REWIND_TIMEOUT);
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
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = (load)? MTIO_DSREG_LD : MTIO_DSREG_UNL;
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
error = cam_periph_runccb(ccb, saerror, 0, 0, softc->device_stats);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_REST;
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_release_ccb(ccb);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
if (error || load == 0)
softc->fileno = softc->blkno = (daddr_t) -1;
else if (error == 0)
softc->fileno = softc->blkno = (daddr_t) 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
return (error);
}
static int
saerase(struct cam_periph *periph, int longerase)
{
union ccb *ccb;
struct sa_softc *softc;
int error;
softc = (struct sa_softc *)periph->softc;
Incorporate the O_NONBLOCK open semantics of Linux and Solaris. This allows an application to upon a tape (yea, even the non-control device) even if it cannot establish a mount session. If the open cannot establish a mount session and O_NONBLOCK was specified, the tape becomes 'open pending mount'. All I/O operations that would require access to a tape thereafter until a close attempt to initiate the mount session. If the mount session succeeds, the tape driver transitions to full open state, else returns an appropriate I/O error (ENXIO). At the same time, add a change that remembers whether tape is being opened read-only. If so, disallow 'write' operations like writing filemarks that bypass the normal read-only filtering operations that happen in the write(2) syscall. Reviewed by: ken, justin, grog MFC after: 2 weeks Suggested by: The Bacula Team
2006-01-14 14:32:41 +00:00
if (softc->open_rdonly)
return (EBADF);
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 SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
ccb = cam_periph_getccb(periph, 1);
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
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
scsi_erase(&ccb->csio, 1, sadone, MSG_SIMPLE_Q_TAG, FALSE, longerase,
Fix for 11815 (at mount time do a throwaway read of the front of the tape to force the drive to do a media access so it knows what media may be inserted). Also Ken's make_dev patches- relatively untested.
1999-11-17 06:05:09 +00:00
SSD_FULL_SIZE, ERASE_TIMEOUT);
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
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_ZER;
Allocate the devstat structure with devstat_new_entry().
2003-03-08 21:44:46 +00:00
error = cam_periph_runccb(ccb, saerror, 0, 0, softc->device_stats);
Extend unit numbers to a full 10 bits (split into sections of the minor). Establish and use a control mode open. Control mode opens may open the device without locking, but are prohibited from all but some ioctls. MTIOCGET always works. MTIOCERRSTAT works, but the clearing of latched error status is contingent upon whether another application has the device open, in which case an interruptible perip acquire is done. MTSETBSIZ, MTSETDNSTY and MTCOMP also require a periph aquire. Relative fileno and blkno are tracked. Note that just about any error will make these undefined, and if you space to EOD or use hardware block positioning, these are also lost until the next UNLOAD or REWIND. Driver state is also tracked and recorded in the unit softc to be passed back in mt_dsreg for a MTIOCGET call. Thanks to Dan Strick for suggesting this. Reintroduce 2 filemarks at EOD for all but QIC devices. I really think it's wrong, but there is a lot of 3rd party software that depends upon this (not the least of which is tcopy). Introduce a SA_QUIRK_1FM to ensure that some devices can be marked as only being able to do 1 FM at EOD. At samount time force a load to BOT if we aren't mounted. If the LOAD command fails, use the REWIND command (e.g., for the IBM 3590 which for some gawdawful reason doesn't support the LOAD (to BOT) command). Also at samount time, if you don't know fixed or variable, try to *set* to one of the known fixed (or variable, for special case) density codes. We only have to do this once per boot, so it's not that painful. This is another way to try and figure out the wierd QIC devices without having to quirk everything in the universe. A substantial amount of cleanup as to what operations can and what operations cannot be retried. Don't retry space operations if they fail- it'll just lead to lossage. Not yet done is invalidating mounts correctly after errors. ENOTIME.
1999-02-05 07:32:52 +00:00
softc->dsreg = MTIO_DSREG_REST;
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_release_ccb(ccb);
return (error);
}
Change #ifdef KERNEL to #ifdef _KERNEL in the public headers. "KERNEL" is an application space macro and the applications are supposed to be free to use it as they please (but cannot). This is consistant with the other BSD's who made this change quite some time ago. More commits to come.
1999-12-29 04:46:21 +00:00
#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
/*
* Read tape block limits command.
*/
void
scsi_read_block_limits(struct ccb_scsiio *csio, u_int32_t retries,
void (*cbfcnp)(struct cam_periph *, union ccb *),
u_int8_t tag_action,
struct scsi_read_block_limits_data *rlimit_buf,
u_int8_t sense_len, u_int32_t timeout)
{
struct scsi_read_block_limits *scsi_cmd;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
cam_fill_csio(csio, retries, cbfcnp, CAM_DIR_IN, tag_action,
(u_int8_t *)rlimit_buf, sizeof(*rlimit_buf), sense_len,
sizeof(*scsi_cmd), timeout);
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_cmd = (struct scsi_read_block_limits *)&csio->cdb_io.cdb_bytes;
bzero(scsi_cmd, sizeof(*scsi_cmd));
scsi_cmd->opcode = READ_BLOCK_LIMITS;
}
void
scsi_sa_read_write(struct ccb_scsiio *csio, u_int32_t retries,
void (*cbfcnp)(struct cam_periph *, union ccb *),
u_int8_t tag_action, int readop, int sli,
int fixed, u_int32_t length, u_int8_t *data_ptr,
u_int32_t dxfer_len, u_int8_t sense_len, u_int32_t timeout)
{
struct scsi_sa_rw *scsi_cmd;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
int read;
read = (readop & SCSI_RW_DIRMASK) == SCSI_RW_READ;
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_cmd = (struct scsi_sa_rw *)&csio->cdb_io.cdb_bytes;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
scsi_cmd->opcode = read ? SA_READ : SA_WRITE;
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_cmd->sli_fixed = 0;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
if (sli && read)
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_cmd->sli_fixed |= SAR_SLI;
if (fixed)
scsi_cmd->sli_fixed |= SARW_FIXED;
scsi_ulto3b(length, scsi_cmd->length);
scsi_cmd->control = 0;
Add unmapped I/O and larger I/O support to the sa(4) driver. We now pay attention to the maxio field in the XPT_PATH_INQ CCB, and if it is set, propagate it up to physio via the si_iosize_max field in the cdev structure. We also now pay attention to the PIM_UNMAPPED capability bit in the XPT_PATH_INQ CCB, and set the new SI_UNMAPPED cdev flag when the underlying SIM supports unmapped I/O. scsi_sa.c: Add unmapped I/O support and propagate the SIM's maximum I/O size up. Adjust scsi_tape_read_write() in the same way that scsi_read_write() was changed to support unmapped I/O. We overload the readop parameter with bits that tell us whether it's an unmapped I/O, and we need to set the CAM_DATA_BIO CCB flag. This change should be backwards compatible in source and binary forms. MFC after: 1 week Sponsored by: Spectra Logic
2013-08-16 16:14:32 +00:00
cam_fill_csio(csio, retries, cbfcnp, (read ? CAM_DIR_IN : CAM_DIR_OUT) |
((readop & SCSI_RW_BIO) != 0 ? CAM_DATA_BIO : 0),
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
tag_action, data_ptr, dxfer_len, sense_len,
sizeof(*scsi_cmd), timeout);
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
}
void
scsi_load_unload(struct ccb_scsiio *csio, u_int32_t retries,
void (*cbfcnp)(struct cam_periph *, union ccb *),
u_int8_t tag_action, int immediate, int eot,
int reten, int load, u_int8_t sense_len,
u_int32_t timeout)
{
struct scsi_load_unload *scsi_cmd;
scsi_cmd = (struct scsi_load_unload *)&csio->cdb_io.cdb_bytes;
bzero(scsi_cmd, sizeof(*scsi_cmd));
scsi_cmd->opcode = LOAD_UNLOAD;
if (immediate)
scsi_cmd->immediate = SLU_IMMED;
if (eot)
scsi_cmd->eot_reten_load |= SLU_EOT;
if (reten)
scsi_cmd->eot_reten_load |= SLU_RETEN;
if (load)
scsi_cmd->eot_reten_load |= SLU_LOAD;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
cam_fill_csio(csio, retries, cbfcnp, CAM_DIR_NONE, tag_action,
NULL, 0, sense_len, sizeof(*scsi_cmd), timeout);
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
}
void
scsi_rewind(struct ccb_scsiio *csio, u_int32_t retries,
void (*cbfcnp)(struct cam_periph *, union ccb *),
u_int8_t tag_action, int immediate, u_int8_t sense_len,
u_int32_t timeout)
{
struct scsi_rewind *scsi_cmd;
scsi_cmd = (struct scsi_rewind *)&csio->cdb_io.cdb_bytes;
bzero(scsi_cmd, sizeof(*scsi_cmd));
scsi_cmd->opcode = REWIND;
if (immediate)
scsi_cmd->immediate = SREW_IMMED;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
cam_fill_csio(csio, retries, cbfcnp, CAM_DIR_NONE, tag_action, NULL,
0, sense_len, sizeof(*scsi_cmd), timeout);
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
}
void
scsi_space(struct ccb_scsiio *csio, u_int32_t retries,
void (*cbfcnp)(struct cam_periph *, union ccb *),
u_int8_t tag_action, scsi_space_code code,
u_int32_t count, u_int8_t sense_len, u_int32_t timeout)
{
struct scsi_space *scsi_cmd;
scsi_cmd = (struct scsi_space *)&csio->cdb_io.cdb_bytes;
scsi_cmd->opcode = SPACE;
scsi_cmd->code = code;
scsi_ulto3b(count, scsi_cmd->count);
scsi_cmd->control = 0;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
cam_fill_csio(csio, retries, cbfcnp, CAM_DIR_NONE, tag_action, NULL,
0, sense_len, sizeof(*scsi_cmd), timeout);
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
}
void
scsi_write_filemarks(struct ccb_scsiio *csio, u_int32_t retries,
void (*cbfcnp)(struct cam_periph *, union ccb *),
u_int8_t tag_action, int immediate, int setmark,
u_int32_t num_marks, u_int8_t sense_len,
u_int32_t timeout)
{
struct scsi_write_filemarks *scsi_cmd;
scsi_cmd = (struct scsi_write_filemarks *)&csio->cdb_io.cdb_bytes;
bzero(scsi_cmd, sizeof(*scsi_cmd));
scsi_cmd->opcode = WRITE_FILEMARKS;
if (immediate)
scsi_cmd->byte2 |= SWFMRK_IMMED;
if (setmark)
scsi_cmd->byte2 |= SWFMRK_WSMK;
scsi_ulto3b(num_marks, scsi_cmd->num_marks);
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
cam_fill_csio(csio, retries, cbfcnp, CAM_DIR_NONE, tag_action, NULL,
0, sense_len, sizeof(*scsi_cmd), timeout);
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
}
/*
* The reserve and release unit commands differ only by their opcodes.
*/
void
scsi_reserve_release_unit(struct ccb_scsiio *csio, u_int32_t retries,
void (*cbfcnp)(struct cam_periph *, union ccb *),
u_int8_t tag_action, int third_party,
int third_party_id, u_int8_t sense_len,
u_int32_t timeout, int reserve)
{
struct scsi_reserve_release_unit *scsi_cmd;
scsi_cmd = (struct scsi_reserve_release_unit *)&csio->cdb_io.cdb_bytes;
bzero(scsi_cmd, sizeof(*scsi_cmd));
if (reserve)
scsi_cmd->opcode = RESERVE_UNIT;
else
scsi_cmd->opcode = RELEASE_UNIT;
if (third_party) {
scsi_cmd->lun_thirdparty |= SRRU_3RD_PARTY;
scsi_cmd->lun_thirdparty |=
((third_party_id << SRRU_3RD_SHAMT) & SRRU_3RD_MASK);
}
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
cam_fill_csio(csio, retries, cbfcnp, CAM_DIR_NONE, tag_action, NULL,
0, sense_len, sizeof(*scsi_cmd), timeout);
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
}
void
scsi_erase(struct ccb_scsiio *csio, u_int32_t retries,
void (*cbfcnp)(struct cam_periph *, union ccb *),
u_int8_t tag_action, int immediate, int long_erase,
u_int8_t sense_len, u_int32_t timeout)
{
struct scsi_erase *scsi_cmd;
scsi_cmd = (struct scsi_erase *)&csio->cdb_io.cdb_bytes;
bzero(scsi_cmd, sizeof(*scsi_cmd));
scsi_cmd->opcode = ERASE;
if (immediate)
scsi_cmd->lun_imm_long |= SE_IMMED;
if (long_erase)
scsi_cmd->lun_imm_long |= SE_LONG;
Add a SA_FLAG_COMP_SUPP flag (to quirk compression as SUPPORTED). Add a SA_FLAG_TAPE_FROZEN for (see below). Add a queue_count field to softc. Add HP T20* Travan-5 like tape device as a FIXED/512 type device. Works for me. Add TANDBERG SLR5 as a variable SA_QUIRK_1FM device. Change VIPER 2525 to 1024 byte blocksize. It's possible other drives should change too, but see below.. Change argument to sagetparams to be pointer to a sa_comp_t union- this can be either a DATA COMPRESSION or a DEVICE CONFIGURATION page. In general compression now tries to use the DATA COMPRESSION page and if that fails tries the DEVICE CONFIGURATION page. Change close routine to not rewind tape if there's a failure in either writing filemarks or in backing over one of two filemarks for a 2FM at EOT tape- instead mark the tape as 'frozen' and print a message saying that either an OFFLINE or REWIND or an MTEOM command is needed to clear this state (all bring certainty back to tape position). Fix sastrategy to not allow I/O to a frozen tape. Add MTIOCGETEOTMODEL/MTIOCSETEOTMODEL ioctls that get and set the EOT model for a tape drive (you can now dynamically change whether it's a 2 FM @ EOT or 1FM at EOT tape device). This ought to give folks something to handle the QIC drives we don't know about. Correctly propagate record of compression algorithm back. Clear FROZEN flag for EOM, REWIND and OFFLINE (and RETENSION and ERASE) cases. Fix an egregious bug in sadone that had left the device queue frozen for deferred (for fixed mode case) errors. Add comment in samount about how useless the test unit ready is for invalidating a mount (this has to be fixed later). Fix residual calculation (per Eivind) in saerror so that negative values for tape records being too large for the supplied buffer get caught. Do some other saerrro cleanup. Per Ken && Justin, add my name to copyright comment.
1999-05-11 04:01:35 +00:00
cam_fill_csio(csio, retries, cbfcnp, CAM_DIR_NONE, tag_action, NULL,
0, sense_len, sizeof(*scsi_cmd), timeout);
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 in block position/block locate functions.
1998-12-18 04:31:43 +00:00
/*
* Read Tape Position command.
*/
void
scsi_read_position(struct ccb_scsiio *csio, u_int32_t retries,
void (*cbfcnp)(struct cam_periph *, union ccb *),
u_int8_t tag_action, int hardsoft,
struct scsi_tape_position_data *sbp,
u_int8_t sense_len, u_int32_t timeout)
{
struct scsi_tape_read_position *scmd;
cam_fill_csio(csio, retries, cbfcnp, CAM_DIR_IN, tag_action,
(u_int8_t *)sbp, sizeof (*sbp), sense_len, sizeof(*scmd), timeout);
scmd = (struct scsi_tape_read_position *)&csio->cdb_io.cdb_bytes;
bzero(scmd, sizeof(*scmd));
scmd->opcode = READ_POSITION;
scmd->byte1 = hardsoft;
}
/*
* Set Tape Position command.
*/
void
scsi_set_position(struct ccb_scsiio *csio, u_int32_t retries,
void (*cbfcnp)(struct cam_periph *, union ccb *),
u_int8_t tag_action, int hardsoft, u_int32_t blkno,
u_int8_t sense_len, u_int32_t timeout)
{
struct scsi_tape_locate *scmd;
cam_fill_csio(csio, retries, cbfcnp, CAM_DIR_NONE, tag_action,
(u_int8_t *)NULL, 0, sense_len, sizeof(*scmd), timeout);
scmd = (struct scsi_tape_locate *)&csio->cdb_io.cdb_bytes;
bzero(scmd, sizeof(*scmd));
scmd->opcode = LOCATE;
if (hardsoft)
scmd->byte1 |= SA_SPOS_BT;
scsi_ulto4b(blkno, scmd->blkaddr);
}
Reference in New Issue Copy Permalink