GitHub/mt8127/android_kernel_alcatel_ttab.git
17 years ago[SCSI] aacraid: Correct SMC products in aacraid.txt
Salyzyn, Mark [Wed, 4 Apr 2007 19:49:54 +0000 (15:49 -0400)]
[SCSI] aacraid: Correct SMC products in aacraid.txt

Correct a spelling mistake for the SMC product names (replace 'B' with
'R') in the Documentation/scsi/aacraid.txt file. This is a follow-up to
a documentation patch '[PATCH] aacraid: Add SMC and SUN products to
README' submitted and accepted to scsi-misc-2.6 on March 27 2007.

Signed-off-by: Mark Salyzyn <aacraid@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] scsi_error.c: Add EH Start Unit retry
Brian King [Thu, 29 Mar 2007 20:25:52 +0000 (15:25 -0500)]
[SCSI] scsi_error.c: Add EH Start Unit retry

Currently, the scsi error handler will issue a START_UNIT
command if the drive indicates it needs its motor started
and the allow_restart flag is set in the scsi_device. If,
after the scsi error handler invokes a host adapter reset
due to error recovery, a device is in a unit attention
state AND also needs a START_UNIT, that device will be placed
offline. The disk array devices on an ipr RAID adapter
will do exactly this when in a dual initiator configuration.
This patch adds a single retry to the EH initiated
START_UNIT.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Patch modified and
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: [Fastboot] Panics for AACRAID driver during 'insmod' for kexec test.
Salyzyn, Mark [Fri, 30 Mar 2007 14:30:48 +0000 (10:30 -0400)]
[SCSI] aacraid: [Fastboot] Panics for AACRAID driver during 'insmod' for kexec test.

Attached is the patch I feel will address this issue. As an added
'perk' I have also added the code to detect if the controller was
previously initialized for interrupted operations by ANY operating
system should the reset_devices kernel parameter not be set and we are
dealing with a naïve kexec without the addition of this kernel
parameter. The reset handler is also improved. Related to reset
operations, but not pertinent specifically to this issue, I have also
altered the handling somewhat so that we reset the adapter if we feel
it is taking too long (three minutes) to start up.

We have not unit tested the reset_devices flag propagation to this
driver code, nor have we unit tested the check for the interrupted
operations under the conditions of a naively issued kexec. We are
submitting this modified driver to our Q/A department for integration
testing in our current programs. I would appreciate an ACK to this
patch should it resolve the issue described in this thread...

Signed-off-by: Mark Salyzyn <aacraid@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ipr: Driver version to 2.3.2
Brian King [Thu, 29 Mar 2007 17:43:57 +0000 (12:43 -0500)]
[SCSI] ipr: Driver version to 2.3.2

Bump driver version.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ipr: Faster sg list fetch
Brian King [Thu, 29 Mar 2007 17:43:50 +0000 (12:43 -0500)]
[SCSI] ipr: Faster sg list fetch

Improve overall command performance by embedding the scatterlist
in the command block used by the adapter. This decreases
the overall number of DMAs required for a single command.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ipr: Return better qc_issue errors
Brian King [Thu, 29 Mar 2007 17:43:43 +0000 (12:43 -0500)]
[SCSI] ipr: Return better qc_issue errors

If qc_issue fails for some reason, return a better error
to libata.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ipr: Disrupt device error
Brian King [Thu, 29 Mar 2007 17:43:37 +0000 (12:43 -0500)]
[SCSI] ipr: Disrupt device error

Add entry in ipr error translation table for an error
received when a device is forced into the failed state
by the user.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ipr: Improve async error logging level control
Brian King [Thu, 29 Mar 2007 17:43:30 +0000 (12:43 -0500)]
[SCSI] ipr: Improve async error logging level control

Add the ability to control how much error data gets logged
on a per error basis.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ipr: PCI unblock config access fix
Brian King [Thu, 29 Mar 2007 17:43:23 +0000 (12:43 -0500)]
[SCSI] ipr: PCI unblock config access fix

Fix to make sure user config accesses get re-enabled if the
PCI config write to start BIST fails.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ipr: Fix for oops following SATA request sense
Brian King [Thu, 29 Mar 2007 17:43:17 +0000 (12:43 -0500)]
[SCSI] ipr: Fix for oops following SATA request sense

This patch fixes a problem discovered on a system with some
bad SATA devices attached. If a command to a SATAPI device
times out and the device gets reset as part of error recovery,
its possible that ipr will set err_mask to indicate a device
error has occurred. If this happens, a request sense will get
issued by libata. Since scsi core thinks this scsi command is
now completed, because the device reset handler completed successfully,
scsi core will free the associated scsi command, which may
cause an oops when that request sense is completed later
by ipr. This patch ensures that any commands that get aborted
as a result of a device reset set err_mask appropriately so
that the request sense does not get sent.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ipr: Log error for SAS dual path switch
Brian King [Thu, 29 Mar 2007 17:43:09 +0000 (12:43 -0500)]
[SCSI] ipr: Log error for SAS dual path switch

For ipr SAS adapters that support dual pathing, this
patch modifies ipr to log an error when a path fails.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ipr: Enable logging of debug error data for all devices
Brian King [Thu, 29 Mar 2007 17:43:01 +0000 (12:43 -0500)]
[SCSI] ipr: Enable logging of debug error data for all devices

The ipr driver has a sysfs attribute which can be used to
adjust the logging level of the driver for error events.
The error response data for commands can be dumped by
increasing the logging level of the ipr driver. This
currently only works for JBOD passthrough devices.
This patch enables this function for all devices, including
RAID devices, to aid in debugging problems.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ipr: Add new PCI-E IDs to device table
Brian King [Thu, 29 Mar 2007 17:42:53 +0000 (12:42 -0500)]
[SCSI] ipr: Add new PCI-E IDs to device table

Adds support for some new PCI-E ipr adapters.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ipr: Remove auto RAID create module parameter
Brian King [Thu, 29 Mar 2007 17:42:47 +0000 (12:42 -0500)]
[SCSI] ipr: Remove auto RAID create module parameter

Remove the auto RAID 0 array creation module parameter, since
support for this function has been removed from the firmware.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ipr: Make adapter operational timeout be per adapter type
Brian King [Thu, 29 Mar 2007 17:42:40 +0000 (12:42 -0500)]
[SCSI] ipr: Make adapter operational timeout be per adapter type

Some ipr adapters may take longer than others to come operational.
This patch makes this timeout different for different adapters,
while still preserving the module parameter which can be used
to globally override the default.

Signed-off-by: Brian King <brking@linux.vnet.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ibmvscsi: add slave_configure to allow device restart
Robert Jennings [Thu, 29 Mar 2007 17:30:40 +0000 (12:30 -0500)]
[SCSI] ibmvscsi: add slave_configure to allow device restart

Fixed the kernel-doc comment for ibmvscsi_slave_configure.  Thanks to
Randy Dunlap for pointing this out.

Adding a slave_configure function for the driver. Now the disks can be
restarted by the scsi mid-layer when the are disconnected and reconnected.

Signed-off-by: "Robert Jennings" <rcj@linux.vnet.ibm.com>
Signed-off-by: "Santiago Leon" <santil@us.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] ibmvscsi: allow for dynamic adjustment of server request_limit
Robert Jennings [Wed, 28 Mar 2007 17:45:46 +0000 (12:45 -0500)]
[SCSI] ibmvscsi: allow for dynamic adjustment of server request_limit

The request limit calculations used previously on the client failed to
mirror the state of the server.  Additionally, when a value < 3 was provided
there could be problems setting can_queue and handling abort and reset
commands.

Signed-off-by: "Robert Jennings" <rcj@linux.vnet.ibm.com>
Signed-off-by: Santiago Leon <santil@us.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: fix print of Firmware Build Date and add TSID
Salyzyn, Mark [Wed, 28 Mar 2007 17:44:54 +0000 (13:44 -0400)]
[SCSI] aacraid: fix print of Firmware Build Date and add TSID

The Adapter build date that is to be printed on instantiation was not
displayed as a result of the supplemental adapter information structure
not being in sync with the Firmware; the driver took an early test cycle
version that had a miss-sized padded region at the head and the
structure was not re-checked at the end of qualification. The Build Date
was not a priority and is merely a cosmetic enhancement, and the wrong
location for the start of the structure member would not induce any
side-effect problems. We updated the structure to match the actual
format, and added the TSID (Tech Support Identification) value print,
should it be present, to the adapter instantiation announcements during
driver load.

This later enhancement should improve the relationship between Service
folk & Tech Support if the printed value of the TSID found it's way into
the circular file labeled G...

Neither of these values show in sysfs (yet).

Signed-off-by: Mark Salyzyn <aacraid@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: remove unused or deprecated firmware constants
Salyzyn, Mark [Wed, 28 Mar 2007 13:22:39 +0000 (09:22 -0400)]
[SCSI] aacraid: remove unused or deprecated firmware constants

Just sweeping the floor clean in one spot. Some of these constants have
never been used in the driver or in the firmware (and thus are
meaningless). Triggered this patch because I discovered one of the
unused constants was actually incorrect and figured it was better to
clean them out than correct and update. There are no side effects at all
regarding this patch, it is purely cosmetic.

Signed-off-by: Mark Salyzyn <aacraid@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] zfcp: fix initialization of FSF timer
Christof Schmitt [Wed, 28 Mar 2007 12:20:40 +0000 (14:20 +0200)]
[SCSI] zfcp: fix initialization of FSF timer

Correctly initialize the timer for FSF requests with jiffies + timeout.

Cc: Swen Schillig <swen@vnet.ibm.com>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Christof Schmitt <christof.schmitt@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] qla2xxx: Remove duplicate pci_disable_device() call
Bernhard Walle [Tue, 27 Mar 2007 22:49:49 +0000 (00:49 +0200)]
[SCSI] qla2xxx: Remove duplicate pci_disable_device() call

On the path qla2x00_probe_one() -> probe_failed -> qla2x00_free_device(),
pci_disable_device() is executed twice, once in qla2x00_free_device()
and once in qla2x00_probe_one().

This patch removes the unnecessary call.

Signed-off-by: Bernhard Walle <bwalle@suse.de>
Acked-by: Andrew Vasquez <andrew.vasquez@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: resolve compiler warnings using ptrdiff_t
Salyzyn, Mark [Tue, 27 Mar 2007 19:07:28 +0000 (15:07 -0400)]
[SCSI] aacraid: resolve compiler warnings using ptrdiff_t

Unsigned long is not always the same size as a pointer, namely on 32 bit
systems with 64 bit address space. Ptrdiff_t is the same size as a
pointer in all configurations. By using ptrdiff_t we can mitigate the
warning messages on these configurations. There should be no side
effects of this patch.

Signed-off-by: Mark Salyzyn <aacraid@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: Add SMC and SUN products to README
Salyzyn, Mark [Tue, 27 Mar 2007 15:51:34 +0000 (11:51 -0400)]
[SCSI] aacraid: Add SMC and SUN products to README

Add SMC and SUN products to aacraid documentation

Signed-off-by: Mark Salyzyn <aacraid@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aic7xxx: make functions static
Adrian Bunk [Tue, 27 Mar 2007 05:59:29 +0000 (21:59 -0800)]
[SCSI] aic7xxx: make functions static

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: cleanups
Adrian Bunk [Tue, 27 Mar 2007 05:59:35 +0000 (21:59 -0800)]
[SCSI] aacraid: cleanups

- proper prototypes for global code in aacraid.h
- aac_rx_start_adapter() can now become static

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: "Salyzyn, Mark" <mark_salyzyn@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] dpt_i2o: remove dead code
Adrian Bunk [Tue, 27 Mar 2007 05:59:25 +0000 (21:59 -0800)]
[SCSI] dpt_i2o: remove dead code

The Coverity checker spotted this dead code introduced by
commit a07f353701acae77e023f6270e8af353b37af7c4.

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: "Salyzyn, Mark" <mark_salyzyn@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] fusion: make mptspi_target_destroy() static
Adrian Bunk [Tue, 27 Mar 2007 05:59:46 +0000 (21:59 -0800)]
[SCSI] fusion: make mptspi_target_destroy() static

Signed-off-by: Adrian Bunk <bunk@stusta.de>
Acked-by: Eric Moore <Eric.Moore@lsi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] Remove some unused SCSI-related kernel config variables.
Robert P. J. Day [Tue, 27 Mar 2007 05:59:33 +0000 (21:59 -0800)]
[SCSI] Remove some unused SCSI-related kernel config variables.

Remove the unused SCSI-related kernel config variables

SCSI_NCR53C8XX_PROFILE_SUPPORT
SCSI_NCR53C8XX_PROFILE
53C700_IO_MAPPED
AIC79XX_ENABLE_RD_STRM
AIC7XXX_PROBE_EISA_VL

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] dpt: whitespace cleanup
Henne [Tue, 27 Mar 2007 05:59:28 +0000 (21:59 -0800)]
[SCSI] dpt: whitespace cleanup

Remove some trailing whitespaces and some replace whitespaces with tabs.

Signed-off-by: Henrik Kretzschmar <henne@nachtwindheim.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: "Salyzyn, Mark" <mark_salyzyn@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] pcmcia: allow drivers to be built non-modular
Robert P. J. Day [Mon, 26 Mar 2007 20:06:45 +0000 (16:06 -0400)]
[SCSI] pcmcia: allow drivers to be built non-modular

Remove the Kconfig requirement that the PCMCIA SCSI drivers be built
only as modules, and allow them to be built into the kernel.

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Acked-by: Dominik Brodowski <linux@dominikbrodowski.net>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: Add likely() and unlikely()
Salyzyn, Mark [Mon, 26 Mar 2007 13:21:14 +0000 (09:21 -0400)]
[SCSI] aacraid: Add likely() and unlikely()

Add some likely() and unlikely() compiler hints in some of the aacraid
hardware interface layers. There should be no operational side effects
resulting from this patch and the changes should be mostly benign on x86
platforms.

Signed-off-by: Mark Salyzyn <aacraid@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] qla2xxx: fix for byteswap in fc_host fabric_name
Richard Lary [Thu, 22 Mar 2007 15:53:19 +0000 (10:53 -0500)]
[SCSI] qla2xxx: fix for byteswap in fc_host fabric_name

This patch fixes byte swap issue in qla2xxx driver
to fix corrupted fabric_name passed to
/sys/class/fc_host/host*/fabric_name.

Signed-off-by: Richard Lary <rlary@us.ibm.com>
Acked-by: Seokmann Ju <seokmann.ju@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] lpfc: fix oops when parsing dodgy VPD
Anton Blanchard [Wed, 21 Mar 2007 13:41:47 +0000 (08:41 -0500)]
[SCSI] lpfc: fix oops when parsing dodgy VPD

We have seen two cases where VPD on an emulex card has been incorrect
and we end up walking off the end of memory. It looks like someone made
an update (increased the length of a string) without increasing the
Length field. Then we do:

Length -= (3+i);

And since Length is unsigned it becomes very large and we loop forever
in the encapsulating:

while (Length > 0) {

If we make Length signed then we fall out of the loop and proceed on.

Its important to note we have only seen this in the lab and it may be
the only two cases of this in existence, but since the rest of the code
has been written to be resilient against bad VPD we may as well fix this
too.

Signed-off-by: Anton Blanchard <anton@samba.org>
Acked-by: James Smart <James.Smart@Emulex.Com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: cleanup and version stamp driver
Salyzyn, Mark [Wed, 21 Mar 2007 17:49:47 +0000 (13:49 -0400)]
[SCSI] aacraid: cleanup and version stamp driver

There is some residual cleanup of the last series of patches and the
need to bump the revision number to draw the line in the sand.

The cmd->SCp.phase is set in the aac_valid_context routine, then set
again to the same value following it's return. The cmd->scsi_done is set
twice in the aac_queuecommand routine. Free up the scsidev FILO in
aac_probe_container as it is not needed further down the function in any
case. Improve the efficiency of the abort handler kernel print
parameters. Bump revision number of driver to approximate the equivalent
in the Adaptec supplied version.

Signed-off-by: Mark Salyzyn <aacraid@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: check buffer address in aac_internal_transfer
Salyzyn, Mark [Wed, 21 Mar 2007 17:22:56 +0000 (13:22 -0400)]
[SCSI] aacraid: check buffer address in aac_internal_transfer

Captured a panic on an older kernel where an application issuing
commands via sg was sending requests that lacked a request_buffer, thus
the buffer pointer used in aac_internal_transer was NULL. The
application was fixed closing the issue, but felt it was advised to
immunize the driver against the eventuality.

Signed-off-by: Mark Salyzyn <aacraid@adaptec.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] fix scsi_wait_scan build problem
James Bottomley [Tue, 20 Mar 2007 19:44:04 +0000 (14:44 -0500)]
[SCSI] fix scsi_wait_scan build problem

The #ifdef MODULE around the export of scsi_complete_async_scans()
which is the API the scsi_wait_scan module uses is incorrect and
causes the symbol to be undefined in certain circumstances leading to
a build failure.  Remove the defines.

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] sd: fix up start/stop messages for new sd_printk() API
James Bottomley [Tue, 20 Mar 2007 17:26:03 +0000 (12:26 -0500)]
[SCSI] sd: fix up start/stop messages for new sd_printk() API

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] sd: implement START/STOP management
Tejun Heo [Tue, 20 Mar 2007 15:13:59 +0000 (00:13 +0900)]
[SCSI] sd: implement START/STOP management

Implement SBC START/STOP management.  sdev->mange_start_stop is added.
When it's set to one, sd STOPs the device on suspend and shutdown and
STARTs it on resume.  sdev->manage_start_stop defaults is in sdev
instead of scsi_disk cdev to allow ->slave_config() override the
default configuration but is exported under scsi_disk sysfs node as
sdev->allow_restart is.

When manage_start_stop is zero (the default value), this patch doesn't
introduce any behavior change.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Rejections fixed and
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] sd: fix return value of sd_sync_cache()
Tejun Heo [Tue, 20 Mar 2007 15:07:18 +0000 (00:07 +0900)]
[SCSI] sd: fix return value of sd_sync_cache()

sd_sync_cache() should return -errno on error, fix it.

Signed-off-by: Tejun Heo <htejun@gmail.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] fusion: remove VMWare guest OS remounted as read only work around
Eric Moore [Mon, 19 Mar 2007 16:31:51 +0000 (10:31 -0600)]
[SCSI] fusion: remove VMWare guest OS remounted as read only work around

This address the issue of VMWare guest OS being remounted as read-only
becuase the underlying device was held busy too long and at the
same time address Engenio MPP driver concerns over infinite retries.
This patch removes the code that snoops the SAM STATUS on busy, which
would be returning DID_BUS_BUSY, instead we return the status as is.
Retry hanlding seems to be properly handled in scsi_softirq_done,
where a busy sam status would only occurr for the time specified by
(cmd->allowed +1) * cmd->timeout_per_command.

Signed-off-by: Eric Moore <Eric.Moore@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] fusion: honour return value of pci_enable_device() in mpt_resume()
Horms [Mon, 19 Mar 2007 06:06:44 +0000 (15:06 +0900)]
[SCSI] fusion: honour return value of pci_enable_device() in mpt_resume()

Honour the return value of pci_enable_device(), which
seems to be a desirable thing to do:

  2.6.20-rc4
  gcc (GCC) 4.1.2 20061115 (prerelease) (Debian 4.1.1-21)

    CC [M]  drivers/message/fusion/mptbase.o
    drivers/message/fusion/mptbase.c: In function `mpt_resume':
    drivers/message/fusion/mptbase.c:1541: warning: ignoring return value
    of `pci_enable_device', declared with attribute warn_unused_result

It also in turn has mptscsih_resume() honour the return value of
mpt_resume()

I'm not sure about the handling of the other potential error cases
in mpt_resume(), of which there appear to be many. But this does
seem to be a good start.

Signed-off-by: Simon Horman <horms@verge.net.au>
Acked-by: "Moore, Eric" <Eric.Moore@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] fusion: remove unnecessary code in mptscsih_resume()
Horms [Fri, 16 Mar 2007 07:05:22 +0000 (16:05 +0900)]
[SCSI] fusion: remove unnecessary code in mptscsih_resume()

It seems that most of the code in mptscsih_resume() doesn't
do anything. This patch removes that code.

Signed-off-by: Simon Horman <horms@verge.net.au>
Acked-by: "Moore, Eric" <Eric.Moore@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] eata_pio: Remove FALSE/TRUE defines
Richard Knutsson [Sat, 17 Mar 2007 23:58:23 +0000 (00:58 +0100)]
[SCSI] eata_pio: Remove FALSE/TRUE defines

eata_generic.h is only included by eata_pio.c and it only uses
FALSE/TRUE in comments.

Signed-off-by: Richard Knutsson <ricknu-0@student.ltu.se>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] expose eh_timed_out to the host template
James Bottomley [Fri, 16 Mar 2007 22:44:41 +0000 (17:44 -0500)]
[SCSI] expose eh_timed_out to the host template

It looks like megaraid_sas at least needs this to throttle its commands
as they begin to time out.  The code keeps the existing transport
template use of eh_timed_out (and allows the transport to override the
host if they both have this callback).

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: Improved error handling
Mark Haverkamp [Thu, 15 Mar 2007 17:27:45 +0000 (10:27 -0700)]
[SCSI] aacraid: Improved error handling

Received from Mark Salyzyn,

This set of fixes improve error handling stability of the driver. A popular
manifestation of the problems is an NULL pointer reference in the interrupt
handler when referencing portions of the scsi command context, or in the
scsi_done handling when an offlined device is referenced.

The aacraid driver currently does not get notification of orphaned command
completions due to devices going offline. The driver also fails to handle the
commands that are finished by the error handler, and thus can complete again
later at the hands of the adapter causing situations of completion of an
invalid scsi command context. Test Unit Ready calls abort assuming that the
abort was successful, but are not, and thus when the interrupt from the adapter
occurs, they reference invalid command contexts. We add in a TIMED_OUT flag to
inform the aacraid FIB context that the interrupt service should merely release
the driver resources and not complete the command up. We take advantage of this
with the abort handler as well for select abortable commands. And we detect and
react if a command that can not be aborted is currently still outstanding to
the controller when reissued by the retry mechanism.

Signed-off-by: Mark Haverkamp <markh@linux-foundation.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: fix srb ioctl for 64 bits
Mark Haverkamp [Thu, 15 Mar 2007 17:27:32 +0000 (10:27 -0700)]
[SCSI] aacraid: fix srb ioctl for 64 bits

Received from Mark Salyzyn,

The raw srb ioctl is supposed to be able to take packets with 32 and 64 bit
virtual address SG elements, it did not handle the frames with 64 bit SG
elements well when communicating with 64 bit DMA capable adapters, and it did
not handle the 32 bit limited DMA adapters at all.  The enclosed patch now
handles all four quadrants (32 bit / 64 bit SG elements in SRB requests + 32
bit or 64 bit DMA capable adapters)

This fix is required before Java based management applications in a 64 bit user
space can submit raw srb requests to the array physical components via the
ioctl mechanism, the allocated user memory pool on 64 bit machines under this
environment forced the management software's hands to submit 64 bit user space
virtual address SG elements in via the ioctl.

Signed-off-by: Mark Haverkamp <markh@linux-foundation.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: remove un-needed references to container id (cid)
Mark Haverkamp [Thu, 15 Mar 2007 17:26:55 +0000 (10:26 -0700)]
[SCSI] aacraid: remove un-needed references to container id (cid)

Received from Mark Salyzyn,

This little patch removes the ',cid)' container identification argument
from some of the functions. The argument is used in some cases as merely
a debug helper and thus not used, and in others, the value can be
quickly acquired from the scsi command in their single solitary use in
the procedure rather than wasting resources on passing the argument in
from above.

Signed-off-by: Mark Haverkamp <markh@linux-foundation.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: Fix ioctl handling when adapter resets
Mark Haverkamp [Thu, 15 Mar 2007 17:27:21 +0000 (10:27 -0700)]
[SCSI] aacraid: Fix ioctl handling when adapter resets

Received from Mark Salyzyn,

Outstanding ioctl calls still have some problems with aborting cleanly
in the face of a reset iop recovery action should the adapter ever enter
into a Firmware Assert (BlinkLED) condition. The enclosed patch resolves
some uncovered flawed handling.

Signed-off-by: Mark Haverkamp <markh@linux-foundation.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: Fix blocking issue with container probing function (cast update)
Mark Haverkamp [Thu, 15 Mar 2007 19:55:07 +0000 (12:55 -0700)]
[SCSI] aacraid: Fix blocking issue with container probing function (cast update)

Received from Mark Salyzyn,

The aac_probe_container call blocks. This is an issue because it is called on
occasion in the context of the queuecommand handler. Once in a blue moon this
has resulted in a kernel panic sleeping during interrupt; or problems with some
embedded system versions of the kernel that depend on queuecommand to not
block. This ugly patch rewrites the aac_probe_container call into a new routine
_aac_probe_container that is an asynchronous state machine to complete the
series of operations. The legacy blocking aac_probe_container call used in
other areas of the driver (during initialization scanning for all targets and
in the separate hot-add/remove [aacraid] thread) merely issues
_aac_probe_container and then simple spins calling schedule() waiting for
completion.

Signed-off-by: Mark Haverkamp <markh@linux-foundation.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: Fix struct element name issue
Mark Haverkamp [Thu, 15 Mar 2007 17:26:22 +0000 (10:26 -0700)]
[SCSI] aacraid: Fix struct element name issue

Received from Mark Salyzyn,

This patch is to resolve a namespace issue that will result from a patch
expected in the future that adds a new interface; rationalized as
correcting a long term issue where hw_fib, instead of hw_fib_va, refers
to the virtual address space and hw_fib_pa refers to the physical
address space. A small fragment of this patch also cleans up an unused
variable that was close to the patch fragments.

Signed-off-by: Mark Haverkamp <markh@linux-foundation.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] aacraid: add restart adapter platform function
Mark Haverkamp [Thu, 15 Mar 2007 17:26:05 +0000 (10:26 -0700)]
[SCSI] aacraid: add restart adapter platform function

Received from Mark Salyzyn,

This patch updates the adapter restart function to deal with some
adapters that have specific IOP reset needs. Since the code for
restarting the adapter was in two places, changed over to utilizing a
platform function in one place.

Signed-off-by: Mark Haverkamp <markh@linux-foundation.org>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] set resid in scsi_io_completion() even for check condition
Pete Wyckoff [Tue, 13 Mar 2007 20:53:28 +0000 (16:53 -0400)]
[SCSI] set resid in scsi_io_completion() even for check condition

Some targets can return both valid data and sense information.
Always update the request data_len from the SCSI command residual.
Callers should interpret sense data to determine what parts of the
data are valid in case of a CHECK CONDITION status.

Signed-off-by: Pete Wyckoff <pw@osc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] tgt: remove the code to build sense
FUJITA Tomonori [Tue, 13 Mar 2007 01:07:15 +0000 (10:07 +0900)]
[SCSI] tgt: remove the code to build sense

tgt notifies a LLD of the failure with sense when it hits the
user-space daemon bugs. However, tgt doesn't know anything about SCSI
devices that initiators talks to. So it's impossible to send proper
sense buffer (format and contents).

This patch changes tgt not to notify a LLD of the failure with bogus
sense. Instead, tgt just re-queues the failure command to the internal
list so that it will be freed cleanly later on when the scsi_host is
removed.

Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] fc_transport: update potential link speeds
James Smart [Mon, 12 Mar 2007 19:16:35 +0000 (14:16 -0500)]
[SCSI] fc_transport: update potential link speeds

This patch updates the FC transport for all speeds identified in
SM-HBA.  Note: it does not sync the "bit" definitions, as that is
actually insulated from user-space via the sysfs text string. (I could
do it, but it does introduce a potential binary-incompatibility).

Signed-off-by: James Smart <James.Smart@emulex.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] qla2xxx: Update version number to 8.01.07-k6.
Andrew Vasquez [Mon, 12 Mar 2007 17:41:31 +0000 (10:41 -0700)]
[SCSI] qla2xxx: Update version number to 8.01.07-k6.

Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] qla2xxx: Allow the extended-error-logging flag to be dynamic.
Andrew Vasquez [Mon, 12 Mar 2007 17:41:30 +0000 (10:41 -0700)]
[SCSI] qla2xxx: Allow the extended-error-logging flag to be dynamic.

The module parameter, ql2xextended_error_logging, can now be
set dynamically by writing to the following sysfs entry:

/sys/module/qla2xxx/parameters/ql2xextended_error_logging

This alleviates the need for the driver to be unloaded and
reloaded in order to enable logging.

Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] qla2xxx: Drop acquisition of hardware_lock during flash manipulations.
Andrew Vasquez [Mon, 12 Mar 2007 17:41:29 +0000 (10:41 -0700)]
[SCSI] qla2xxx: Drop acquisition of hardware_lock during flash manipulations.

There's no need given, I/O has been quiesced, RISC
interrupts have been disabled, and finally the RISC has been
paused.  Flash manipulation on ISP21xx, ISP22xx, and ISP23xx
parts requires the RISC to go through a full reset to
recover.

Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] qla2xxx: Add cond_resched() calls during HBA flash manipulation.
Andrew Vasquez [Mon, 12 Mar 2007 17:41:28 +0000 (10:41 -0700)]
[SCSI] qla2xxx: Add cond_resched() calls during HBA flash manipulation.

We're observing soft lockups during HBA FLASH retrieval and
update.  Add cond_resched() each time around the tight-loops
during flash read()s/write()s.

Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] qla2xxx: Add scan_[start|finish]() callbacks for ISP24xx HBAs.
Andrew Vasquez [Mon, 12 Mar 2007 17:41:27 +0000 (10:41 -0700)]
[SCSI] qla2xxx: Add scan_[start|finish]() callbacks for ISP24xx HBAs.

Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] qla2xxx: fix RSCN handling on big-endian systems
Malahal Naineni [Mon, 12 Mar 2007 17:41:26 +0000 (10:41 -0700)]
[SCSI] qla2xxx: fix RSCN handling on big-endian systems

qla2xxx driver fails to handle RSCN events affecting area or domain due
to an endian issue on big endian systems.  This fixes the port_id_t
structure on big endian systems.

Signed-off-by: Malahal Naineni <malahal@us.ibm.com>
Acked-by: Seokmann Ju <seokmann.ju@qlogic.com>
Signed-off-by: Andrew Vasquez <andrew.vasquez@qlogic.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] make scsi_wait_scan always modular
James Bottomley [Sun, 11 Mar 2007 19:16:43 +0000 (14:16 -0500)]
[SCSI] make scsi_wait_scan always modular

Currently scsi_wait_scan is only built modular if SCSI is modular.
However, it's perfectly possible for a built in SCSI still to have
modular drivers and thus need scsi_wait_scan as a module.  Therefore,
scsi_wait_scan should always be built as a module (unless the kernel
doesn't support modules).

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] fusion: kernel-doc warning fixes
Randy Dunlap [Tue, 20 Feb 2007 19:17:03 +0000 (11:17 -0800)]
[SCSI] fusion: kernel-doc warning fixes

Fix kernel-doc warnings in fusion driver code.

Signed-off-by: Randy Dunlap <randy.dunlap@oracle.com>
Acked-by: "Moore, Eric" <Eric.Moore@lsi.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] sd: typo fix: sdkp_printk should be sd_printk
James Bottomley [Sun, 11 Mar 2007 17:25:33 +0000 (12:25 -0500)]
[SCSI] sd: typo fix: sdkp_printk should be sd_printk

Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] pci2000: Delete unused header file.
Robert P. J. Day [Sat, 10 Mar 2007 22:16:26 +0000 (17:16 -0500)]
[SCSI] pci2000: Delete unused header file.

This driver was removed a while ago by commit

099175c94a221fa2723b7273883c98cd32efe900

However, it seems that pci2000.h wasn't properly eliminated, so remove
it now.

Signed-off-by: Robert P. J. Day <rpjday@mindspring.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] tgt: fix scsi command leak
FUJITA Tomonori [Sat, 3 Mar 2007 00:55:55 +0000 (09:55 +0900)]
[SCSI] tgt: fix scsi command leak

The failure to map user-space pages leads to scsi command leak. It can
happens mostly because of user-space daemon bugs (or OOM). This patch
makes tgt just notify a LLD of the failure with sense when
blk_rq_map_user() fails.

Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] tgt: fix sesnse buffer problems
FUJITA Tomonori [Sat, 3 Mar 2007 00:55:54 +0000 (09:55 +0900)]
[SCSI] tgt: fix sesnse buffer problems

This patch simplify the way to notify LLDs of the command completion
and addresses the following sense buffer problems:

- can't handle both data and sense.
- forces user-space to use aligned sense buffer

tgt copies sense_data from userspace to cmnd->sense_buffer (if
necessary), maps user-space pages (if necessary) and then calls
host->transfer_response (host->transfer_data is removed).

Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] tgt: rm bio hacks in scsi tgt
Mike Christie [Sat, 3 Mar 2007 00:55:54 +0000 (09:55 +0900)]
[SCSI] tgt: rm bio hacks in scsi tgt

scsi tgt breaks up a command into multple scatterlists
if we cannot fit all the data in one. This was because
the block rq helpers did not support large requests and
because we can get a command of any old size so it is
hard to preallocate pages for scatterlist large enough
(we cannot really preallocate pages with the bio map
user path). In 2.6.20, we added large request support to
the block layer helper, blk_rq_map_user. And at LSF,
we talked about increasing SCSI_MAX_PHYS_SEGMENTS for
scsi tgt if we want to support really really :) large
(greater than 256 * PAGE_SIZE in the worst mapping case)
requests.

The only target currently implemented does not even support
the multiple scatterlists stuff and only supports smaller
requests, so this patch just coverts scsi tgt to use
blk_rq_map_user.

Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] iscsi_tcp: print useful error message when iscsi crc23c allocation fails
Mike Christie [Wed, 28 Feb 2007 23:32:21 +0000 (17:32 -0600)]
[SCSI] iscsi_tcp: print useful error message when iscsi crc23c allocation fails

People do not read the README and seem to like to
unselect the crc32c module even though iscsi_tcp selects
it for them. This patch spits a error that tells the user
that they really do need the module. Hopefully, we will
get fewer people asking about this now.

Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] iscsi_tcp: increase max_sectors
Mike Christie [Wed, 28 Feb 2007 23:32:20 +0000 (17:32 -0600)]
[SCSI] iscsi_tcp: increase max_sectors

For a while now, the block layer has seperated max sectors
and max hw sectors. Software iscsi has no limit so this patch
increases max hw sectors, so we can support large pass through
commands.

Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] libiscsi: use get_unaligned
Mike Christie [Wed, 28 Feb 2007 23:32:19 +0000 (17:32 -0600)]
[SCSI] libiscsi: use get_unaligned

Dave Miller meantioned that the data buffer in a past
sense fixup patch was not gauranteed to be aligned
properly for ia64. This patch has libiscsi use get_unalinged
to make sure. There are a couple more places in the
digest handling we may need to do this, but we are in the middle
of fixing that code for big endien systems so just the sense
access is fixed here.

Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] iscsi transport: use atomic for session_nr allocations
Mike Christie [Wed, 28 Feb 2007 23:32:18 +0000 (17:32 -0600)]
[SCSI] iscsi transport: use atomic for session_nr allocations

qla4xxx and iscsi_tcp or iser could be creating
sessions at the same time, so make session_nr id
allocation atomic.

Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] iscsi: rename DEFAULT_MAX_RECV_DATA_SEGMENT_LENGTH
Mike Christie [Wed, 28 Feb 2007 23:32:17 +0000 (17:32 -0600)]
[SCSI] iscsi: rename DEFAULT_MAX_RECV_DATA_SEGMENT_LENGTH

This patch renames DEFAULT_MAX_RECV_DATA_SEGMENT_LENGTH to avoid
confusion with the drivers default values (DEFAULT_MAX_RECV_DATA_SEGMENT_LENGTH
is the iscsi RFC specific default).

Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] libiscsi: clear mtask
Mike Christie [Wed, 28 Feb 2007 23:32:16 +0000 (17:32 -0600)]
[SCSI] libiscsi: clear mtask

Consolidate the mtask clearing code.

Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] libiscsi: flush work before freeing connection
Mike Christie [Wed, 28 Feb 2007 23:32:15 +0000 (17:32 -0600)]
[SCSI] libiscsi: flush work before freeing connection

It's possible that we call iscsi_xmitworker after iscsi_conn_release
which causes a oops. This patch flushes the workqueue.

Signed-off-by: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Signed-off-by: Mike Christie <michaelc@cs.wisc.edu>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] constants.c: Update ASC list and make it const
Martin K. Petersen [Wed, 28 Feb 2007 17:37:06 +0000 (12:37 -0500)]
[SCSI] constants.c: Update ASC list and make it const

Add missing additional sense code and provide pointer to upstream
reference (from Doug Gilbert).

Add missing const (from Michael Tokarev).

Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] sd: convert logging to new printk helpers
Martin K. Petersen [Wed, 28 Feb 2007 03:41:19 +0000 (22:41 -0500)]
[SCSI] sd: convert logging to new printk helpers

Convert the sd.c SCSI logging calls to scmd_printk()/sd_printk()
instead of plain printk().

Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] sd: make printing use a common prefix
Martin K. Petersen [Wed, 28 Feb 2007 03:40:55 +0000 (22:40 -0500)]
[SCSI] sd: make printing use a common prefix

Make SCSI disk printing more consistent:

 - Define sd_printk(), sd_print_sense_hdr() and sd_print_result()

 - Move relevant header bits into sd.h

 - Remove all the legacy disk_name passing and use scsi_disk pointers
   where possible

 - Switch printk() lines to the new sd_ functions so that output is
   consistent

Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] Make error printing more verbose
Martin K. Petersen [Wed, 28 Feb 2007 03:40:27 +0000 (22:40 -0500)]
[SCSI] Make error printing more verbose

This patch enhances SCSI error printing by:

 - Making use of scsi_print_result() in the completion functions.

 - Having scmd_printk() output the disk name (when applicable).

Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] constants.c: cleanup, verbose result printing
Martin K. Petersen [Wed, 28 Feb 2007 03:39:44 +0000 (22:39 -0500)]
[SCSI] constants.c: cleanup, verbose result printing

Clean up constants.c and make result printing more user friendly:

 - Refactor the command and sense functions so that the actual
   formatting can be called from the various helper functions with the
   correct prefix.

 - Replace scsi_print_hostbyte() and scsi_print_driverbyte() with
   scsi_print_result() which is verbose when CONFIG_SCSI_CONSTANTS is
   on.

Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years ago[SCSI] zfcp: fix likely/unlikely usage
Heiko Carstens [Wed, 21 Feb 2007 08:28:00 +0000 (09:28 +0100)]
[SCSI] zfcp: fix likely/unlikely usage

zfcp_fsf_protstatus_eval() takes always the 'wrong' branch.

Likely Profiling Results
 ---------------------------------------------------------
[+- ] Type | # True | # False | Function:Filename@Line
+unlikely  |   11042|        0  zfcp_fsf_protstatus_eval()

Acked-by: Swen Schillig <swen@vnet.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: James Bottomley <James.Bottomley@SteelEye.com>
17 years agoRevert "USB: pxa2xx_udc: fix hardcoded irq number"
Linus Torvalds [Sat, 10 Mar 2007 22:22:07 +0000 (14:22 -0800)]
Revert "USB: pxa2xx_udc: fix hardcoded irq number"

This reverts commit d2487cb4257dafb686f682285854fe7f02ca29d8.

Russell King points out that it's obviously bogus, and I have to agree.
Not only does "irq" not even exist in that scope, but we obviously need
to free the irq that we actually requested, and that's IRQ_USB.

Reported-by: Russell King <rmk@arm.linux.org.uk>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Greg KH <greg@kroah.com>
Cc: David Brownell <david-b@pacbell.net>,
Cc: Milan Svoboda <msvoboda@ra.rockwell.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6
Linus Torvalds [Sat, 10 Mar 2007 06:04:27 +0000 (22:04 -0800)]
Merge /pub/scm/linux/kernel/git/davem/net-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6:
  [IPV6] fix ipv6_getsockopt_sticky copy_to_user leak
  [IPV6]: Fix for ipv6_setsockopt NULL dereference
  [DCCP]: Initialise write_xmit_timer also on passive sockets
  [IPV4]: Fix rtm_to_ifaddr() error handling.

17 years agoMerge branch 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6
Linus Torvalds [Sat, 10 Mar 2007 06:00:51 +0000 (22:00 -0800)]
Merge branch 'release' of git://git./linux/kernel/git/aegl/linux-2.6

* 'release' of git://git.kernel.org/pub/scm/linux/kernel/git/aegl/linux-2.6:
  [IA64] refresh config files
  [IA64] put kdump_find_rsvd_region in __init
  [IA64] Remove sparse warning from unwind code
  [IA64] add missing syscall trace clear
  [IA64] Cleanup in crash.c
  [IA64] kexec: declare ia64_mca_pal_base in mca.h rather than kexec.h
  [IA64] pci_get_legacy_ide_irq should return irq (not GSI)
  [IA64] whitespace fixes for include/asm-ia64/sal.h
  [IA64] Cache error recovery
  [IA64] Proper handling of TLB errors from duplicate itr.d dropins

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/driver-2.6
Linus Torvalds [Sat, 10 Mar 2007 05:58:28 +0000 (21:58 -0800)]
Merge /pub/scm/linux/kernel/git/gregkh/driver-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/gregkh/driver-2.6:
  kobject: new_device->kref wasn't putted after error in kobject_move()
  driver core: export device_rename
  Remove devfs from MAINTAINERS
  Driver core: add device symlink back to sysfs
  Revert "driver core: refcounting fix"

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/pci-2.6
Linus Torvalds [Sat, 10 Mar 2007 05:58:10 +0000 (21:58 -0800)]
Merge /pub/scm/linux/kernel/git/gregkh/pci-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/gregkh/pci-2.6:
  pci: fix section mismatch warning
  PCI: aer: fix section mismatch warning
  pcie: fix section mismatch warning
  PCI: allow multiple calls to pcim_pin_device()

17 years agoMerge master.kernel.org:/pub/scm/linux/kernel/git/gregkh/usb-2.6
Linus Torvalds [Sat, 10 Mar 2007 05:57:46 +0000 (21:57 -0800)]
Merge /pub/scm/linux/kernel/git/gregkh/usb-2.6

* master.kernel.org:/pub/scm/linux/kernel/git/gregkh/usb-2.6: (23 commits)
  USB Elan FTDI: check for workqueue creation
  USB: fix spinlock recursion in cdc-acm.c
  USB: fix Unaligned access in EHCI driver
  USB: Product ID for FT232RL in ftdi_sio
  USBNET: DM9501: Add Corega FEther USB-TXC support.
  USB: ipaq.c: Additional devices
  USB: further fix for usb-serial
  USB: fix usb-serial device naming bug
  USB: RTS/DTR signal patch for airprime driver
  USB: ftdi_sio: use port_probe / port_remove thereby fixing access to the latency_timer
  usb-serial: fix shutdown / device_unregister order
  USB: add Additional PIDs in ftdi_sio
  USB: add QL355P power supply ids to fdti_sio
  USB: New device IDs for cp2101 driver
  USB: kill dead code from hub.c
  USB: ratelimit debounce error messages
  USB: pxa2xx_udc: fix hardcoded irq number
  UHCI: fix port resume problem
  USB: set the correct interval for interrupt URBs
  USB: goku_udc: Remove crude cache coherency code
  ...

17 years agoPull misc-for-upstream into release branch
Len Brown [Sat, 10 Mar 2007 04:19:50 +0000 (23:19 -0500)]
Pull misc-for-upstream into release branch

17 years agoPull bugzilla-8110 into release branch
Len Brown [Sat, 10 Mar 2007 04:19:25 +0000 (23:19 -0500)]
Pull bugzilla-8110 into release branch

17 years agoPull bugzilla-8066 into release branch
Len Brown [Sat, 10 Mar 2007 04:19:19 +0000 (23:19 -0500)]
Pull bugzilla-8066 into release branch

17 years agoPull bugzilla-7907 into release branch
Len Brown [Sat, 10 Mar 2007 04:18:53 +0000 (23:18 -0500)]
Pull bugzilla-7907 into release branch

17 years agoPull bugzilla-7570 into release branch
Len Brown [Sat, 10 Mar 2007 04:18:46 +0000 (23:18 -0500)]
Pull bugzilla-7570 into release branch

17 years agoPull bugzilla-6859 into release branch
Len Brown [Sat, 10 Mar 2007 04:18:35 +0000 (23:18 -0500)]
Pull bugzilla-6859 into release branch

17 years agoPull bugzilla-6316 into release branch
Len Brown [Sat, 10 Mar 2007 04:18:22 +0000 (23:18 -0500)]
Pull bugzilla-6316 into release branch

17 years agoPull bugzilla-5966 into release branch
Len Brown [Sat, 10 Mar 2007 04:18:05 +0000 (23:18 -0500)]
Pull bugzilla-5966 into release branch

17 years agoPull ibm into release branch
Len Brown [Sat, 10 Mar 2007 04:17:46 +0000 (23:17 -0500)]
Pull ibm into release branch

17 years agoPull altix into release branch
Len Brown [Sat, 10 Mar 2007 04:17:39 +0000 (23:17 -0500)]
Pull altix into release branch

17 years agoACPI: ec: fix race in status register access
Alexey Starikovskiy [Wed, 7 Mar 2007 23:29:35 +0000 (18:29 -0500)]
ACPI: ec: fix race in status register access

Delay the read of the EC status register until
after the event that caused it occurs -- otherwise
it is possible to read and act on stale status that was
associated with the previous event.

Do this with a perpetually incrementing "event_count" to detect
when a new event occurs and it is safe to read status.

There is no workaround for polling mode -- it is inherently
exposed to reading and acting on stale status, since it
doesn't have an interrupt to tell it the event completed.

http://bugzilla.kernel.org/show_bug.cgi?id=8110

Signed-off-by: Alexey Starikovskiy <alexey.y.starikovskiy@intel.com>
Signed-off-by: Len Brown <len.brown@intel.com>
17 years agoUSB Elan FTDI: check for workqueue creation
Cyrill Gorcunov [Tue, 6 Mar 2007 10:47:44 +0000 (02:47 -0800)]
USB Elan FTDI: check for workqueue creation

Avoid NULL pointer usage if workqueue creation failed.

Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: fix spinlock recursion in cdc-acm.c
Oliver Neukum [Tue, 6 Mar 2007 09:47:04 +0000 (10:47 +0100)]
USB: fix spinlock recursion in cdc-acm.c

this fixes the spinlock recursion issue. The older fix was incomplete.

Signed-off-by: Oliver Neukum <oneukum@suse.de>
Acked-by: Pete Zaitcev <zaitcev@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
17 years agoUSB: fix Unaligned access in EHCI driver
Max Dmitrichenko [Mon, 5 Mar 2007 23:45:01 +0000 (02:45 +0300)]
USB: fix Unaligned access in EHCI driver

I get following warnings on spar64:
Kernel unaligned access at TPC[1000c9e4] ehci_hub_control+0x54c/0x68c [ehci_hcd]

Despite of the comment in the patched code, the type cast used there
does make unaligned access. The fix was made as it's done in
ohci-hub.c.

Signed-off-by: Max Dmitrichenko <dmitrmax@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>