[GIT PATCH] SCSI fixes for 2.6.25-rc8

[GIT PATCH] SCSI fixes for 2.6.25-rc8

[James Bottomley]

This patch constitutes three bug fixes (the reversion of the BSG queue
fix is the most serious; without it we're getting eventual corruption in
systems).  The other two fix error legs or misleading reporting.  The
remaining 8 patches are all updates to the mvsas driver which was a late
addition at the -rc1 phase, so while they're not truly bug fixes, they
are enhancements to make that unreleased driver more roubust for 2.6.25.

The patch is here:

master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6.git

The short changelog is:

James Bottomley (2):
      libsas: Warn if ATA device detected but CONFIG_SCSI_SAS_ATA not set
      Revert " fix bsg queue oops with iscsi logout"

James Smart (1):
      hosts.c: fixes for "no error" reported after error scenarios

Ke Wei (8):
      mvsas: check subsystem id
      mvsas: get phy info.
      mvsas: fix the buffer of rx DMA overflow bug
      mvsas: retry aborting task.
      mvsas: check hd whether unplugged
      mvsas : interrupt handling
      mvsas: a tag handler implementation
      mvsas: fill in error info record and phy mode6 bits.

and the diffstat:

 drivers/scsi/hosts.c               |   16 
 drivers/scsi/libsas/sas_discover.c |    5 
 drivers/scsi/mvsas.c               |  711 +++++++++++++++++++++++++------------
 drivers/scsi/scsi_sysfs.c          |    2 
 4 files changed, 497 insertions(+), 237 deletions(-)

James

Re: [GIT PATCH] SCSI fixes for 2.6.25-rc8

[Boaz Harrosh]

On Tue, Apr 01 2008 at 23:36 +0300, James Bottomley <James.Bottomley@HansenPartnership.com> wrote:
> This patch constitutes three bug fixes (the reversion of the BSG queue
> fix is the most serious; without it we're getting eventual corruption in
> systems).  The other two fix error legs or misleading reporting.  The
> remaining 8 patches are all updates to the mvsas driver which was a late
> addition at the -rc1 phase, so while they're not truly bug fixes, they
> are enhancements to make that unreleased driver more roubust for 2.6.25.
> 
> The patch is here:
> 
> master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6.git
> 
> The short changelog is:
> 
> James Bottomley (2):
>       libsas: Warn if ATA device detected but CONFIG_SCSI_SAS_ATA not set
>       Revert " fix bsg queue oops with iscsi logout"
> 
<snip>

Does this mean that we still have the iscsi logout bug if used from bsg
in 2.6.25 kernel? I just want to know what are the out-of-tree patches 
I will need. Will the proper fix be candidate for the stable releases
of 2.6.25?

Thanks
Boaz

Re: [GIT PATCH] SCSI fixes for 2.6.25-rc8

[James Bottomley]

On Wed, 2008-04-02 at 21:05 +0300, Boaz Harrosh wrote:
> On Tue, Apr 01 2008 at 23:36 +0300, James Bottomley <James.Bottomley@HansenPartnership.com> wrote:
> > This patch constitutes three bug fixes (the reversion of the BSG queue
> > fix is the most serious; without it we're getting eventual corruption in
> > systems).  The other two fix error legs or misleading reporting.  The
> > remaining 8 patches are all updates to the mvsas driver which was a late
> > addition at the -rc1 phase, so while they're not truly bug fixes, they
> > are enhancements to make that unreleased driver more roubust for 2.6.25.
> > 
> > The patch is here:
> > 
> > master.kernel.org:/pub/scm/linux/kernel/git/jejb/scsi-rc-fixes-2.6.git
> > 
> > The short changelog is:
> > 
> > James Bottomley (2):
> >       libsas: Warn if ATA device detected but CONFIG_SCSI_SAS_ATA not set
> >       Revert " fix bsg queue oops with iscsi logout"
> > 
> <snip>
> 
> Does this mean that we still have the iscsi logout bug if used from bsg
> in 2.6.25 kernel? I just want to know what are the out-of-tree patches 
> I will need. Will the proper fix be candidate for the stable releases
> of 2.6.25?

I don't know yet.  The discussion is here:

http://marc.info/?t=120620205200001

James