Bug#883615: Acknowledgement ([CRITICAL] Stretch p-u 9.3 breaks NVidia driver and X.org)

Julien Aubin julien.aubin at gmail.com
Wed Dec 6 06:24:24 UTC 2017 

2017-12-06 7:07 GMT+01:00 Julien Aubin <julien.aubin at gmail.com>:

>
>
> 2017-12-06 3:34 GMT+01:00 Julien Aubin <julien.aubin at gmail.com>:
>
>>
>>
>> Le 5 déc. 2017 23:08, "Luca Boccassi" <bluca at debian.org> a écrit :
>>
>> On Tue, 2017-12-05 at 22:23 +0100, Julien Aubin wrote:
>> > To help debugging, could you provide me please some link to a newer
>> > NVidia
>> > driver release built for stretch please (and some notice to install
>> > it) ?
>>
>> Unfortunately I don't think we have anything that can be used in
>> Stretch - IIRC both 384 and 387 dependencies are structured for the
>> glvnd transition, so they can't be installed on Stretch.
>>
>> > Thus :
>> > - If I disable from a custom xorg.conf file module glx everything
>> > loads
>> > fine (as long as I use lightdm as a login manager)
>> >
>> > Thanks a lot.
>>
>> Could you clarify what this means? You are manually blacklisting glx?
>>
>>
>>
> The fact driver 38x cannot be installed on Stretch is very bad as it
> prevents newer GPUs from working w/ Debian stable. Newer low-end NVidia
> MX1** are already impacted.
>
> FYI here's what I get in kern.log when enabling glx in xorg :
>
> ec  6 06:59:44 pccorei7-4770 kernel: [ 3682.820990] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 06:59:53 pccorei7-4770 kernel: [ 3691.431679] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 06:59:53 pccorei7-4770 kernel: [ 3691.753169] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 06:59:54 pccorei7-4770 kernel: [ 3692.545921] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 06:59:54 pccorei7-4770 kernel: [ 3692.866820] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 06:59:55 pccorei7-4770 kernel: [ 3693.542921] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 06:59:55 pccorei7-4770 kernel: [ 3693.864460] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 06:59:56 pccorei7-4770 kernel: [ 3694.539028] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 06:59:56 pccorei7-4770 kernel: [ 3694.860824] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 06:59:57 pccorei7-4770 kernel: [ 3695.536803] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 06:59:57 pccorei7-4770 kernel: [ 3695.876240] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 07:02:55 pccorei7-4770 kernel: [ 3873.365928] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 07:02:55 pccorei7-4770 kernel: [ 3873.706339] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 07:02:56 pccorei7-4770 kernel: [ 3874.563884] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 07:02:56 pccorei7-4770 kernel: [ 3874.885761] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 07:02:57 pccorei7-4770 kernel: [ 3875.562088] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 07:02:57 pccorei7-4770 kernel: [ 3875.901835] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 07:02:58 pccorei7-4770 kernel: [ 3876.809216] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 07:02:58 pccorei7-4770 kernel: [ 3877.130778] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 07:02:59 pccorei7-4770 kernel: [ 3877.807965] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 07:02:59 pccorei7-4770 kernel: [ 3878.144767] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 07:03:37 pccorei7-4770 kernel: [ 3915.290650] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 07:03:37 pccorei7-4770 kernel: [ 3915.611177] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 07:03:38 pccorei7-4770 kernel: [ 3916.320719] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 07:03:38 pccorei7-4770 kernel: [ 3916.641984] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 07:03:39 pccorei7-4770 kernel: [ 3917.300914] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 07:03:39 pccorei7-4770 kernel: [ 3917.621818] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 07:03:40 pccorei7-4770 kernel: [ 3918.315331] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 07:03:40 pccorei7-4770 kernel: [ 3918.652675] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
> Dec  6 07:03:41 pccorei7-4770 kernel: [ 3919.563312] nvidia-modeset:
> Allocated GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @
> PCI:0000:01:00.0
> Dec  6 07:03:41 pccorei7-4770 kernel: [ 3919.882492] nvidia-modeset: Freed
> GPU:0 (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
>
>
>
Okay after further debugging I found that the following X.org option avoids
the crash :
Option "AllowGLXWithComposite" "False"



> Hi
>>
>> Yes I used nvidia-xconfig to generate an xorg.conf file and in it I
>> disabled the glx module
>>
>>
>> > 2017-12-05 22:14 GMT+01:00 Julien Aubin <julien.aubin at gmail.com>:
>> >
>> > > GLX crashes here :
>> > >
>> > > #0  0x00007ffff588ad01 in __GI_____strtol_l_internal
>> > > (nptr=0x7fffffffe2b1
>> > > "001 GLX", endptr=0x7fffffffe2a8, base=10, group=<optimized out>,
>> > > loc=0x555555ad3620) at ../stdlib/strtol_l.c:293
>> > > #1  0x00005555555cd0cb in ?? ()
>> > > #2  0x00005555555bbeb0 in AddExtension ()
>> > > #3  0x00007ffff381d7b2 in ?? () from
>> > > /usr/lib/xorg/modules/linux/libglx.so
>> > >
>> > > #4  0x00005555555bc040 in ?? ()
>> > > #5  0x000000000000001d in ?? ()
>> > > #6  0x0000000000000200 in ?? ()
>> > > #7  0x0000000000000000 in ?? ()
>> > >
>> > >
>> > > 2017-12-05 21:47 GMT+01:00 Julien Aubin <julien.aubin at gmail.com>:
>> > >
>> > > > Hi,
>> > > >
>> > > > I reinstalled NVidia driver several times and the problem is
>> > > > still the
>> > > > same. Trying your method still results in the same issue.
>> > > >
>> > > > Now if I fully blacklist the nvidia driver I get a "better" black
>> > > > screen
>> > > > w/ Nouveau, in the sense that keyboard answers. Thus Xorg.0.log
>> > > > does not
>> > > > hang on the GLX stuff.
>> > > >
>> > > > Looks like this issue targets GeForce 10xx GPUs (as my 970 does
>> > > > seem
>> > > > unaffected as well).
>> > > >
>> > > > Rgds
>> > > >
>> > > >
>> > > >
>> > > > 2017-12-05 21:13 GMT+01:00 Luca Boccassi <bluca at debian.org>:
>> > > >
>> > > > > On Tue, 2017-12-05 at 20:55 +0100, Julien Aubin wrote:
>> > > > > > Looks like the culprit is NVidia modeset. Thus the issue does
>> > > > > > not
>> > > > > > happen w/
>> > > > > > a Maxwell GPU.
>> > > > > >
>> > > > > > [  542.167979] nvidia-modeset: Allocated GPU:0
>> > > > > > (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
>> > > > > >
>> > > > > >
>> > > > > > [  542.490521] nvidia-modeset: Freed GPU:0
>> > > > > > (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
>> > > > > > [  544.081034] nvidia-modeset: Allocated GPU:0
>> > > > > > (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
>> > > > > > [  544.385426] nvidia-modeset: Freed GPU:0
>> > > > > > (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
>> > > > > > [  546.079744] nvidia-modeset: Allocated GPU:0
>> > > > > > (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
>> > > > > > [  546.399641] nvidia-modeset: Freed GPU:0
>> > > > > > (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
>> > > > > > [  548.091421] nvidia-modeset: Allocated GPU:0
>> > > > > > (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
>> > > > > > [  548.395724] nvidia-modeset: Freed GPU:0
>> > > > > > (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
>> > > > > > [  550.088731] nvidia-modeset: Allocated GPU:0
>> > > > > > (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
>> > > > > > [  550.392898] nvidia-modeset: Freed GPU:0
>> > > > > > (GPU-c9899079-b539-164e-1242-191c547d3276) @ PCI:0000:01:00.0
>> > > > >
>> > > > > Didn't see any issue with my 780 (Kepler?).
>> > > > >
>> > > > > Are you sure the kernel modules is correctly rebuilt? I've seen
>> > > > > that
>> > > > > there was a kernel point release, and not all ABI changes are
>> > > > > reverted.
>> > > > > Which means DKMS won't rebuild the modules automatically.
>> > > > >
>> > > > > Try to remove it and rebuild it with:
>> > > > >
>> > > > > sudo dkms uninstall nvidia-current/375.82 -k 4.9.0-4-amd64
>> > > > > sudo dkms install nvidia-current/375.82 -k 4.9.0-4-amd64
>> > > > > > 2017-12-05 20:51 GMT+01:00 Julien Aubin <julien.aubin at gmail.c
>> > > > > > om>:
>> > > > > >
>> > > > > > > Driver hangs just here :
>> > > > > > > --------------------------->
>> > > > > > > [    32.680] (II) Initializing extension GLX
>> > > > > > > [    32.680] (II) Indirect GLX disabled.
>> > > > > > > ---------------->
>> > > > > > >
>> > > > > > > Normally I should get then [    34.659] (II) config/udev:
>> > > > > > > Adding
>> > > > > > > input
>> > > > > > > device Power Button (/dev/input/event5)
>> > > > > > >
>> > > > > > > But here it crashes without log.
>> > > > > > >
>> > > > > > > 2017-12-05 20:09 GMT+01:00 Debian Bug Tracking System <
>> > > > > > > owner at bugs.debian.org>:
>> > > > > > >
>> > > > > > > > Thank you for filing a new Bug report with Debian.
>> > > > > > > >
>> > > > > > > > You can follow progress on this Bug here: 883615:
>> > > > > > > > https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=883615.
>> > > > > > > >
>> > > > > > > > This is an automatically generated reply to let you know
>> > > > > > > > your
>> > > > > > > > message
>> > > > > > > > has been received.
>> > > > > > > >
>> > > > > > > > Your message is being forwarded to the package
>> > > > > > > > maintainers and
>> > > > > > > > other
>> > > > > > > > interested parties for their attention; they will reply
>> > > > > > > > in due
>> > > > > > > > course.
>> > > > > > > >
>> > > > > > > > Your message has been sent to the package maintainer(s):
>> > > > > > > >  Debian NVIDIA Maintainers <pkg-nvidia-devel at lists.alioth
>> > > > > > > > .debian.
>> > > > > > > > org>
>> > > > > > > >
>> > > > > > > > If you wish to submit further information on this
>> > > > > > > > problem, please
>> > > > > > > > send it to 883615 at bugs.debian.org.
>> > > > > > > >
>> > > > > > > > Please do not send mail to owner at bugs.debian.org unless
>> > > > > > > > you wish
>> > > > > > > > to report a problem with the Bug-tracking system.
>> > > > > > > >
>> > > > > > > > --
>> > > > > > > > 883615: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug
>> > > > > > > > =883615
>> > > > > > > > Debian Bug Tracking System
>> > > > > > > > Contact owner at bugs.debian.org with problems
>> > > > > > > >
>> > > > > > >
>> > > > > > >
>> > > > > >
>> > > > > > _______________________________________________
>> > > > > > pkg-nvidia-devel mailing list
>> > > > > > pkg-nvidia-devel at lists.alioth.debian.org
>> > > > > > http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/pkg-n
>> > > > > > vidia-de
>> > > > > > vel
>> > > >
>> > > >
>> > > >
>>
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pkg-nvidia-devel/attachments/20171206/cbb60729/attachment-0001.html>

More information about the pkg-nvidia-devel mailing list