* 2.5.25 regression from 2.6.24: smartctl fails to access drive
@ 2008-04-02 6:00 Gordon Farquharson
2008-04-02 6:06 ` Gordon Farquharson
0 siblings, 1 reply; 3+ messages in thread
From: Gordon Farquharson @ 2008-04-02 6:00 UTC (permalink / raw)
To: linux-ide; +Cc: bzolnier
On a GLAN Tank (arm/mach-iop32x) smartctl fails with kernel versions
2.6.25-rc1 and later with the following error messages:
hda: task_no_data_intr: status=0x51 { DriveReady SeekComplete Error }
hda: task_no_data_intr: error=0x04 { DriveStatusError }
ide: failed opcode was: 0xb0
smartctl does not fail with kernel version 2.6.24 on this platform.
The version of smartctl is the same for both kernel tests.
This issue has been discussed in
http://bugzilla.kernel.org/show_bug.cgi?id=10086 and based on the
tests conducted, Bartlomiej Zolnierkiewicz wrote,
<--- quote --->
Thanks, from looking at the log:
| hda: tf: feat 0xd8 nsect 0x00 lbal 0x01 lbam 0x4f lbah 0xc2 dev 0x00 cmd 0xb0
| hda: hob: nsect 0x00 lbal 0x00 lbam 0x00 lbah 0x00
| hda: tf: feat 0x00 nsect 0x00 lbal 0x00 lbam 0x00 lbah 0x00 dev 0x00 cmd 0xb0
| hda: hob: nsect 0x00 lbal 0x00 lbam 0x00 lbah 0x00
We got SMART command (cmd 0xb0) with all other registers set to 0x00 and it is
(theoretically) impossible to get it through HDIO_DRIVE_CMD (since
ide_cmd_ioctl() always sets lbam to 0x4f and lbah to 0xc2). Maybe it was
passed through HDIO_DRIVE_TASK or HDIO_DRIVE_TASKFILE or we have the bug
somewhere (N.B. the behavior of erroing it is correct if the user-space
actually requested such command)...
It seems to be a distinct problem from the one that was hit by Anders and later
fixed by Linus. Could you try git-bisecting it, please?
<--- end quote --->
Since then, I manually found that the problem commit occurred between
2.6.24 and 2.6.25-rc1. At that point, I started using git-bisect, but
all kernels I have tested so far fail to boot on the GLAN Tank.
Unfortunately, I won't have access to the hardware for the next two
and half weeks, and I have run out of time to track down the commit
that is causing the kernels to fail to boot and therefore also the
commit that is causing smartctl to fail. I wanted to post this message
to the mailing list as a record of the work done to date and so that I
can pick up this problem when I have access to the hardware again.
Sorry for not being able to track this problem down before 2.6.25 is released.
Gordon
--
Gordon Farquharson
GnuPG Key ID: 32D6D676
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: 2.5.25 regression from 2.6.24: smartctl fails to access drive
2008-04-02 6:00 2.5.25 regression from 2.6.24: smartctl fails to access drive Gordon Farquharson
@ 2008-04-02 6:06 ` Gordon Farquharson
2008-04-03 2:07 ` Gordon Farquharson
0 siblings, 1 reply; 3+ messages in thread
From: Gordon Farquharson @ 2008-04-02 6:06 UTC (permalink / raw)
To: linux-ide; +Cc: bzolnier
Hi All
On Wed, Apr 2, 2008 at 12:00 AM, Gordon Farquharson
<gordonfarquharson@gmail.com> wrote:
> Since then, I manually found that the problem commit occurred between
> 2.6.24 and 2.6.25-rc1. At that point, I started using git-bisect, but
> all kernels I have tested so far fail to boot on the GLAN Tank.
Update: commit 4383f18b7f94a4d668c5eec68645c75d44556235 does boot, but
smartctl fails with this kernel. I may have time to do a couple of
more git-bisects.
Gordon
--
Gordon Farquharson
GnuPG Key ID: 32D6D676
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: 2.5.25 regression from 2.6.24: smartctl fails to access drive
2008-04-02 6:06 ` Gordon Farquharson
@ 2008-04-03 2:07 ` Gordon Farquharson
0 siblings, 0 replies; 3+ messages in thread
From: Gordon Farquharson @ 2008-04-03 2:07 UTC (permalink / raw)
To: linux-ide; +Cc: bzolnier
On Wed, Apr 2, 2008 at 12:06 AM, Gordon Farquharson
<gordonfarquharson@gmail.com> wrote:
> On Wed, Apr 2, 2008 at 12:00 AM, Gordon Farquharson
> <gordonfarquharson@gmail.com> wrote:
>
> > Since then, I manually found that the problem commit occurred between
> > 2.6.24 and 2.6.25-rc1. At that point, I started using git-bisect, but
> > all kernels I have tested so far fail to boot on the GLAN Tank.
The earliest commit between 2.6.24 and 2.6.25-rc1 that boots on the
GLAN Tank is 66a2c077c36ee5c56e77ba3757702c4464011da6. Unfortunately,
smartctl does not work with this kernel, so the commit that is causing
the problem must have been made before this commit.
I guess the next step will be to revert the IDE commits before
66a2c077c36ee5c56e77ba3757702c4464011da6 one by one to see when
smartctl starts working. Is there an easy way to list all the commits
to drivers/ide that occurred before
66a2c077c36ee5c56e77ba3757702c4464011da6? I'll try this when I have
access to the hardware again (in around two weeks).
Gordon
--
Gordon Farquharson
GnuPG Key ID: 32D6D676
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2008-04-03 2:07 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2008-04-02 6:00 2.5.25 regression from 2.6.24: smartctl fails to access drive Gordon Farquharson
2008-04-02 6:06 ` Gordon Farquharson
2008-04-03 2:07 ` Gordon Farquharson
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox