Bug 88946 - [NVC1] black screen after resuming from hibernate
Summary: [NVC1] black screen after resuming from hibernate
Status: RESOLVED INVALID
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/nouveau (show other bugs)
Version: 7.7 (2012.06)
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Nouveau Project
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-02-03 17:07 UTC by bernie
Modified: 2016-02-23 07:26 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
syslog messages standy and resume - working (1.25 MB, text/plain)
2015-02-03 17:07 UTC, bernie
no flags Details
Xorg log standy and resume - working (63.99 KB, text/plain)
2015-02-03 17:09 UTC, bernie
no flags Details
syslog messages hibernate and resume - NOT working (681.38 KB, text/plain)
2015-02-03 17:10 UTC, bernie
no flags Details
Xorg log hibernate and resume - NOT working (65.30 KB, text/plain)
2015-02-03 17:10 UTC, bernie
no flags Details
lspci nvidia graphics card (13.89 KB, text/plain)
2015-02-03 17:11 UTC, bernie
no flags Details

Description bernie 2015-02-03 17:07:40 UTC
Created attachment 113124 [details]
syslog messages standy and resume - working

Hello,

I use Kubuntu 14.04.x together with a nvidia graphics card NVC1 (GF108) GT 430 passive and driver nouveau. 

Standby and resume works with the nouveau driver, but hibernate and resume don't.

There is a black screen after resume, then the monitor goes off (no signal) and no keyboard input is possible => Only a hard reset helps.

*** Package versions
Kubuntu 14.04.1 64bit
nouveau version 1.0.11-1-ubuntu2~trusty
Mesa 10.3.2, Xserver 1.16.0 (Utopic (14.10) Xstack)
kernel 3.16.x (also tested kernel 3.18.5 and 3.19rc7)

I have attached the following logs; with kernel parameter nouveau.debug=trace

--- syslog and xorg.log for the working standby 
syslog_standby_resume.txt	
standby_resume_Xorg.0.log.old_working.txt

--- syslog and xorg.log for the (hibernate) and resume problem 
syslog_hibernate_resume_NOT_working.txt
hibernate_resume_Xorg.0.log.old_NOT_working.txt

--- lspci.txt

There is a crash at the xserver;

[   101.332] nouveau_exa_upload_to_screen:380 - falling back to memcpy ignores tiling
[   101.348] nouveau_exa_download_from_screen:295 - falling back to memcpy ignores tiling
[   101.348] (EE) 
[   101.348] (EE) Backtrace:
[   101.348] (EE) 0: /usr/bin/X (xorg_backtrace+0x48) [0x7fbef984ae18]
[   101.349] (EE) 1: /usr/bin/X (0x7fbef96a6000+0x1a8e59) [0x7fbef984ee59]
[   101.349] (EE) 2: /lib/x86_64-linux-gnu/libc.so.6 (0x7fbef7826000+0x36d40) [0x7fbef785cd40]
[   101.349] (EE) 3: /lib/x86_64-linux-gnu/libc.so.6 (0x7fbef7826000+0x98d66) [0x7fbef78bed66]
[   101.349] (EE) 4: /usr/lib/xorg/modules/libexa.so (0x7fbef334d000+0x5356) [0x7fbef3352356]
[   101.349] (EE) 5: /usr/lib/xorg/modules/libexa.so (0x7fbef334d000+0x57fb) [0x7fbef33527fb]
[   101.349] (EE) 6: /usr/lib/xorg/modules/libexa.so (0x7fbef334d000+0x7e22) [0x7fbef3354e22]
[   101.349] (EE) 7: /usr/lib/xorg/modules/libexa.so (0x7fbef334d000+0x100c2) [0x7fbef335d0c2]
[   101.349] (EE) 8: /usr/lib/xorg/modules/libexa.so (0x7fbef334d000+0x8499) [0x7fbef3355499]
[   101.349] (EE) 9: /usr/bin/X (0x7fbef96a6000+0x131af9) [0x7fbef97d7af9]
[   101.349] (EE) 10: /usr/bin/X (0x7fbef96a6000+0x50281) [0x7fbef96f6281]
[   101.349] (EE) 11: /usr/bin/X (0x7fbef96a6000+0x5339e) [0x7fbef96f939e]
[   101.349] (EE) 12: /usr/bin/X (0x7fbef96a6000+0x571ea) [0x7fbef96fd1ea]
[   101.349] (EE) 13: /lib/x86_64-linux-gnu/libc.so.6 (__libc_start_main+0xf5) [0x7fbef7847ec5]
[   101.349] (EE) 14: /usr/bin/X (0x7fbef96a6000+0x4276e) [0x7fbef96e876e]
[   101.349] (EE) 
[   101.349] (EE) Segmentation fault at address 0x0
[   101.349] (EE) 
Fatal server error:
[   101.349] (EE) Caught signal 11 (Segmentation fault). Server aborting

Thank you for your help!

Best regards, Bernhard
Comment 1 bernie 2015-02-03 17:09:05 UTC
Created attachment 113125 [details]
Xorg log standy and resume - working
Comment 2 bernie 2015-02-03 17:10:02 UTC
Created attachment 113126 [details]
syslog messages hibernate and resume - NOT working
Comment 3 bernie 2015-02-03 17:10:35 UTC
Created attachment 113127 [details]
Xorg log hibernate and resume - NOT working
Comment 4 bernie 2015-02-03 17:11:07 UTC
Created attachment 113128 [details]
lspci nvidia graphics card
Comment 5 Ilia Mirkin 2015-02-03 18:07:25 UTC
Looks like on resume you're getting a ton of

[   75.967312] nouveau E[    PBUS][0000:01:00.0] MMIO read of 0x00000000 FAULT at 0x610260 [ IBUS ]
[   75.967343] nouveau E[    PBUS][0000:01:00.0] MMIO read of 0x00000000 FAULT at 0x610260 [ IBUS ]

followed by

[   75.967345] nouveau E[   PDISP][0000:01:00.0][0xc000857b][ffff8803f3954100] timeout1: 0x00000000
[   75.967346] nouveau E[   PDISP][0000:01:00.0][0xc000857b][ffff8803f3954100] init failed, -16

and some other init failures. When the card fails to come up, X isn't particularly happy about that either, hence the black screen and general state of sadness. So the question is why does the card not come up properly...

By the way, when you say "standby" I assume you mean "suspend to ram", while when you say "hibernate", I assume you mean "suspend to disk" right?
Comment 6 bernie 2015-02-03 18:29:50 UTC
Yes, you are correct;

"standby" = "suspend to ram"
"hibernate" = "suspend to disk"
Comment 7 bernie 2015-02-08 19:36:14 UTC
Are there any new patches to test or any hints to solve this problem?
Comment 8 bernie 2015-02-11 15:20:56 UTC
In the final kernel 3.19 the problem persists. 

Shall I provide further debugging information?
Comment 9 bernie 2015-03-05 17:32:44 UTC
Same problem in 4.0rc2
Comment 10 bernie 2015-03-05 17:36:13 UTC
Is "suspend to disk"-support planned for driver nouveau?
Comment 11 Ilia Mirkin 2015-03-05 19:12:43 UTC
(In reply to B. A. from comment #10)
> Is "suspend to disk"-support planned for driver nouveau?

It's not broken on purpose... I believe it *ought* to work. It'll just require someone to debug this to figure out what's going wrong.
Comment 12 bernie 2015-03-05 19:26:07 UTC
What can I do to help you to debug this problem? 

NVC3 also seems to have this problem:
=> see bug
https://bugs.freedesktop.org/show_bug.cgi?id=67255

Is the NVC0 family (Fermi) in general broken with hibernate and resume, or is it dependent on the model of the graphics card?
Comment 13 Ilia Mirkin 2015-03-05 19:36:09 UTC
(In reply to B. A. from comment #12)
> What can I do to help you to debug this problem? 
> 
> NVC3 also seems to have this problem:
> => see bug
> https://bugs.freedesktop.org/show_bug.cgi?id=67255
> 
> Is the NVC0 family (Fermi) in general broken with hibernate and resume, or
> is it dependent on the model of the graphics card?

No clue. It could also be a new issue introduced with a recent kernel change (like all the fencing changes that have gone in). Or it could be an issue as old as time itself :)

If this were easy to figure out, someone would have already done it. I haven't a clue how to debug this... resume issues are notoriously difficult to investigate.
Comment 14 bernie 2015-03-05 19:55:22 UTC
Thank you for the answer!

I test this issue since kernel 3.2 and with other kernel versions (LTS kernel version from Ubuntu LTS point releases), but it never had worked.
Comment 15 Christopher M. Penalver 2016-02-23 07:26:15 UTC
B. A., given you are using a downstream version of nouveau, it will help immensely if you filed a new report with Ubuntu by ensuring you have the package xdiagnose installed, and that you click the Yes button for attaching additional debugging information running the following from a terminal:
ubuntu-bug xorg

Also, please feel free to subscribe me to it.

For more on why this is helpful, please see https://wiki.ubuntu.com/ReportingBugs.


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.