Nouveau - Sep 2015 - [Bug 91992] New: [NV4C] computer hangs after few minutes of use

https://bugs.freedesktop.org/show_bug.cgi?id=91992

            Bug ID: 91992
           Summary: [NV4C] computer hangs after few minutes of use
           Product: Mesa
           Version: unspecified
          Hardware: x86-64 (AMD64)
                OS: Linux (All)
            Status: NEW
          Severity: major
          Priority: medium
         Component: Drivers/DRI/nouveau
          Assignee: nouveau at lists.freedesktop.org
          Reporter: laurent.riffard+freedesktop at free.fr
        QA Contact: nouveau at lists.freedesktop.org

Created attachment 118238
  --> https://bugs.freedesktop.org/attachment.cgi?id=118238&action=edit
dmesg (ArchLinux)

Hello,

After few minutes (3 to 15) of use, the display freezes. 
1 or 2 seconds later, the screen is filled at once with repeated noise small
pattern. Then, display is frozen, no reaction at all to key presses or mouse.
NumLock etc. keyboard lights do not react. Ssh-connection dies. Machine does
not respond to ping. SysRq-keys cannot sync disks. Logs are not written to
disk. SysRq-keys (SysRq-reboot) do not work at all. Have to press reset button. 

The video card is NVIDIA Corporation C61 [GeForce 7025 / nForce 630a].

It does happen with 2 distros : ArchLinux and Ubuntu 15.04. The freeze is best
triggered by lauching LibreOffice or Firefox.

Adding nouveau.config=NvMSI=0 kernel parameter does not help.

Thanks
~~
laurent

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20150913/e57b91da/attachment-0001.html>

https://bugs.freedesktop.org/show_bug.cgi?id=91992

--- Comment #1 from Laurent Riffard <laurent.riffard+freedesktop at
free.fr> ---
Created attachment 118239
  --> https://bugs.freedesktop.org/attachment.cgi?id=118239&action=edit
Xorg.log (ArchLinux)

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20150913/a24de7d4/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=91992

--- Comment #2 from Laurent Riffard <laurent.riffard+freedesktop at
free.fr> ---
Created attachment 118240
  --> https://bugs.freedesktop.org/attachment.cgi?id=118240&action=edit
vbios.rom

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20150913/a0fc563b/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=91992

--- Comment #3 from Hans de Goede <jwrdegoede at fedoraproject.org> ---
Hi,

What desktop environment are you using ?

I've a feeling this may be caused by us running out of video-mem. Can you
see
if you can change the amount of memory assigned to the gpu in the BIOS, and if
making that larger helps ?

Regards,

Hans

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20150913/bf96154e/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=91992

--- Comment #4 from Laurent Riffard <laurent.riffard+freedesktop at
free.fr> ---
Hi,

I’m using Xfce or Cinnamon.

As you suggested, I increased the amount of RAM assigned to the video card,
from 32 Gb to 64 Gb, and I didn’t encounter any new problem.

Shouldn’t this short-memory condition be nicely handled by the driver ?

Regards,
Laurent

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20150914/28f4e778/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=91992

Hans de Goede <jwrdegoede at fedoraproject.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED

--- Comment #5 from Hans de Goede <jwrdegoede at fedoraproject.org> ---
Hi,

(In reply to Laurent Riffard from comment #4)
> Hi,
> 
> I’m using Xfce or Cinnamon.
> 
> As you suggested, I increased the amount of RAM assigned to the video card,
> from 32 Gb to 64 Gb, and I didn’t encounter any new problem.
I assume you mean MB not GB ? Yes 32MB is a very low setting, so this explains
the original problem quite nicely.
> Shouldn’t this short-memory condition be nicely handled by the driver ?
In an ideal world it would at least be handled better then crashing, but our
oom handling currently is rather poor (as you've noticed) . Fixing this
unfortunately is quite difficult and currently no one is working on this.

Since changing the setting in the BIOS fixes things, I'm closing this bug.

Regards,

Hans

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20150915/fa70e4c3/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=91992

--- Comment #6 from Laurent Riffard <laurent.riffard+freedesktop at
free.fr> ---
Hi,

(In reply to Hans de Goede from comment #5)
[...]
> > As you suggested, I increased the amount of RAM assigned to the video
card,
> > from 32 Gb to 64 Gb, and I didn’t encounter any new problem.
> 
> I assume you mean MB not GB ? Yes 32MB is a very low setting, so this
> explains the original problem quite nicely.
Yes, I mean Mb...

> > Shouldn’t this short-memory condition be nicely handled by the driver
?
> 
> In an ideal world it would at least be handled better then crashing, but
our
> oom handling currently is rather poor (as you've noticed) . Fixing this
> unfortunately is quite difficult and currently no one is working on this.
> 
> Since changing the setting in the BIOS fixes things, I'm closing this
bug.
Ok, thanks for your help

Regards,
Laurent

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://lists.freedesktop.org/archives/nouveau/attachments/20150915/f5a4e8fb/attachment.html>

https://bugs.freedesktop.org/show_bug.cgi?id=91992

Dave Odell <dmo2118 at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Blocks|                            |99968


Referenced Bugs:

https://bugs.freedesktop.org/show_bug.cgi?id=99968
[Bug 99968] [NV4C] System crashes reliably with glxinfo
-- 
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<https://lists.freedesktop.org/archives/nouveau/attachments/20170226/25b69eb4/attachment.html>