Bug 82741 - Wake up from Suspend to RAM crashes Toshiba Satellite laptop to black screen
Summary: Wake up from Suspend to RAM crashes Toshiba Satellite laptop to black screen
Status: NEEDINFO
Alias: None
Product: Drivers
Classification: Unclassified
Component: Network (show other bugs)
Hardware: x86-64 Linux
: P1 normal
Assignee: drivers_network@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-08-18 19:49 UTC by fireandfuel
Modified: 2014-11-29 16:22 UTC (History)
0 users

See Also:
Kernel Version: 3.16.1
Subsystem:
Regression: Yes
Bisected commit-id:


Attachments
dmesg, acpidump, lsmod & lspci output (392.08 KB, text/plain)
2014-08-18 19:49 UTC, fireandfuel
Details
lscpi output (24.60 KB, text/plain)
2014-08-20 22:30 UTC, fireandfuel
Details
Log file of git bisect (3.46 KB, text/x-log)
2014-08-31 21:04 UTC, fireandfuel
Details
Patch for 3.17.3 to fix the crash (486 bytes, patch)
2014-11-29 16:22 UTC, fireandfuel
Details | Diff

Description fireandfuel 2014-08-18 19:49:04 UTC
Created attachment 147151 [details]
dmesg, acpidump, lsmod & lspci output

With Kernel 3.16.1 my Toshiba Satellite C660D-10P laptop (PSC0UE-00D001GR)  crashes while wake up from suspend with a black screen. 
The crash is reproducible; every time I close the lid or suspend it from terminal or X11.

The crash happens on Kernel version 3.16, too.
The last known Kernel version that doesn't crash after suspend is 3.15.8.
Comment 1 Zhang Rui 2014-08-20 01:31:00 UTC
could you please check 3.17-rc1?
or could you please check if commit dee1592638ab7ea35a32179b73f9284dead49c03 fixes the crash for you or not.

Plus, please also attach the lspci output.
Comment 2 fireandfuel 2014-08-20 22:30:56 UTC
Created attachment 147581 [details]
lscpi output

Commit dee1592638ab7ea35a32179b73f9284dead49c03 applied as patch on 3.16.1 doesn't fix the crash. I will check 3.17-rc1 soon.

Sometimes the display doesn't initialize when I restart my laptop after the crash.
Comment 3 Rafael J. Wysocki 2014-08-21 03:54:56 UTC
Can you please run the "echo devices > /sys/power/pm_test" test (as described in Documentation/power/basic-pm-debugging.txt in the kernel source tree) and see if the machine survives that?
Comment 4 fireandfuel 2014-08-21 18:04:48 UTC
(In reply to Rafael J. Wysocki from comment #3)
> Can you please run the "echo devices > /sys/power/pm_test" test (as
> described in Documentation/power/basic-pm-debugging.txt in the kernel source
> tree) and see if the machine survives that?
The "echo devices > /sys/power/pm_test" test runs successfully on 3.16.1.
Comment 5 Zhang Rui 2014-08-24 07:17:47 UTC
then what about echo platform/core > /sys/power/pm_test?
Comment 6 fireandfuel 2014-08-24 11:02:47 UTC
(In reply to Zhang Rui from comment #5)
> then what about echo platform/core > /sys/power/pm_test?
Both tests run successfully.
Comment 7 Zhang Rui 2014-08-25 02:52:40 UTC
have you tried 3.17-rc1 yet?
can you please run git bisect to find out which commit introduce this regression?
Comment 8 Len Brown 2014-08-26 14:14:43 UTC
possibly a change related to Radeon?

01:05.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] RS880M [Mobility Radeon HD 4225/4250] (prog-if 00 [VGA controller])
Comment 9 fireandfuel 2014-08-26 18:24:47 UTC
(In reply to Zhang Rui from comment #7)
> have you tried 3.17-rc1 yet?
> can you please run git bisect to find out which commit introduce this
> regression?
3.17-rc2 produces the same crash. 
I hope that I'll have some time at weekend to run git bisect. I'm very busy this week.

(In reply to Len Brown from comment #8)
> possibly a change related to Radeon?
> 
> 01:05.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI]
> RS880M [Mobility Radeon HD 4225/4250] (prog-if 00 [VGA controller])
This was my first thought about this bug. 3.16 brought some changes to the radeon driver.
Comment 10 fireandfuel 2014-08-31 21:00:15 UTC
(In reply to Zhang Rui from comment #7)
> can you please run git bisect to find out which commit introduce this
> regression?
The first bad commit: b51ecea852b712618796d9eab8428a7d5f1f106f
Comment 11 fireandfuel 2014-08-31 21:04:02 UTC
Created attachment 148951 [details]
Log file of git bisect
Comment 12 Zhang Rui 2014-09-01 04:15:22 UTC
so does the problem still exist if you revert that patch or blacklist the r8169 driver?
Comment 13 fireandfuel 2014-09-01 06:46:12 UTC
(In reply to Zhang Rui from comment #12)
> so does the problem still exist if you revert that patch or blacklist the
> r8169 driver?
I blacklisted the r8169 driver on 3.17-rc2 and my system doesn't crash. After revert of the patch the system doesn't crash, too.
Comment 14 Zhang Rui 2014-09-01 07:51:25 UTC
Okay, reassign this to the driver experts.
Comment 15 fireandfuel 2014-11-29 16:22:27 UTC
Created attachment 159171 [details]
Patch for 3.17.3 to fix the crash

I partly reverted b51ecea852b712618796d9eab8428a7d5f1f106f to see which changes in b51ecea852b712618796d9eab8428a7d5f1f106f crashes my system. 

The bad change is in function "static void rtl_hw_start_8168g_1(struct rtl8169_private *tp)". I created a patch for 3.17.3 which fixes the crash on my system.

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