Bug 19872 - phison unable to read partition table
Summary: phison unable to read partition table
Status: CLOSED CODE_FIX
Alias: None
Product: Drivers
Classification: Unclassified
Component: Staging (show other bugs)
Hardware: All Linux
: P1 high
Assignee: drivers_staging@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-10-07 21:11 UTC by Heinz Wiesinger
Modified: 2011-01-23 15:38 UTC (History)
2 users (show)

See Also:
Kernel Version: 2.6.36-rc7
Subsystem:
Regression: No
Bisected commit-id:


Attachments
dmesg log (30.24 KB, text/plain)
2010-10-07 21:11 UTC, Heinz Wiesinger
Details
proposed patch due to libata api change (486 bytes, patch)
2010-10-08 23:06 UTC, Greg Kroah-Hartman
Details | Diff

Description Heinz Wiesinger 2010-10-07 21:11:40 UTC
Created attachment 32712 [details]
dmesg log

I have a Verbatim ExpressCard SSD that uses the phison driver from staging.
On recent kernels when loading the module I see a lot of errors and the driver is eventually unable to read the partition table off the device, rendering the device useless on those kernels. (log attached)

I tested it with 2.6.35 and 2.6.36rc7, same behaviour. 2.6.32 works find. I haven't had the chance yet to test .33 and .34.
Comment 1 Greg Kroah-Hartman 2010-10-07 21:29:00 UTC
Can you run 'git bisect' to try to track down the patch that caused this
problem?  I can't seem to find my device to test this out myself, sorry.
Comment 2 Heinz Wiesinger 2010-10-08 22:48:28 UTC
Sure thing. Took a while, but according to my test the first bad commit is 1c5afdf7a629d2e77de8dd043b97a33dcd7e6dfa.

Hope that helps :)
Comment 3 Greg Kroah-Hartman 2010-10-08 23:06:10 UTC
Wonderful, it does, can you test out the patch attached here to see if it solves the problem or not for the driver?
Comment 4 Greg Kroah-Hartman 2010-10-08 23:06:42 UTC
Created attachment 32912 [details]
proposed patch due to libata api change
Comment 5 Heinz Wiesinger 2010-10-09 09:01:03 UTC
Which version is the patch for? It doesn't apply for me for either 2.6.36rc7 or 1c5afdf
Comment 6 Greg Kroah-Hartman 2010-10-09 18:31:04 UTC
On Sat, Oct 09, 2010 at 09:01:11AM +0000, bugzilla-daemon@bugzilla.kernel.org wrote:
> Which version is the patch for? It doesn't apply for me for either 2.6.36rc7
> or
> 1c5afdf

It should apply to 2.6.36-rc7.  It's just changing one function call to
another one, so if it somehow doesn't apply, can you do it by hand and
test it?
Comment 7 Heinz Wiesinger 2010-10-09 19:12:54 UTC
ok, done, and it seems to work fine here.
Thank you very much! :)
Comment 8 Greg Kroah-Hartman 2010-10-09 20:28:47 UTC
Wonderful, thanks for testing.  Sorry about the patch, now I see what I did wrong, that was my fault.

I've commited this to the staging-next tree and it will go into the .37 merge and backported to the proper kernels (.35 and .36)
Comment 9 Heinz Wiesinger 2010-10-09 20:44:54 UTC
No problem! Thank you for fixing this issue so fast!
Comment 10 Florian Mickler 2011-01-23 15:38:33 UTC
fixed in .37-rc1:
commit cf10700bf8047f0668dd874b607f89516612e6c7
Author: Greg Kroah-Hartman <gregkh@suse.de>
Date:   Sat Oct 9 13:26:12 2010 -0700

    Staging: phison: fix problem caused by libata change

Note You need to log in before you can comment on or make changes to this bug.