Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.debian.kernel > #64452 > unrolled thread

Bug#932250: linux-image-amd64: random segfaults without taking something down

Back to article view | Back to linux.debian.kernel

Contents

  Bug#932250: linux-image-amd64: random segfaults without taking something down m.alfaeko@gmail.com - 2019-07-17 01:30 +0200
    Bug#932250: linux-image-amd64: adding segfault data m.alfaeko@gmail.com - 2019-07-17 01:50 +0200
      Bug#932250: linux-image-amd64: adding segfault data Ben Hutchings <ben@decadent.org.uk> - 2019-09-25 00:40 +0200
    Bug#932250: linux-image-amd64: random segfaults without taking something down Bernhard Übelacker <bernhardu@mailbox.org> - 2019-07-17 11:10 +0200
    Processed: Re: Bug#932250: linux-image-amd64: random segfaults  without taking something down "Debian Bug Tracking System" <owner@bugs.debian.org> - 2019-07-18 04:10 +0200
    Bug#932250: linux-image-amd64: random segfaults without taking something down Ben Hutchings <ben@decadent.org.uk> - 2019-07-18 04:10 +0200
    Bug#932250: Fwd: Bug#932250: linux-image-amd64: random segfaults without taking something down a a <m.alfaeko@gmail.com> - 2019-07-18 13:20 +0200
    Bug#932250: [Fwd: Re: Bug#932250: linux-image-amd64: random segfaults without taking something down] Ben Hutchings <ben@decadent.org.uk> - 2019-07-18 13:20 +0200
    Processed: Re: linux-image-amd64: adding segfault data "Debian Bug Tracking System" <owner@bugs.debian.org> - 2019-09-25 00:40 +0200
    Bug#932250: 932250 Paul Kerry <p.kerry@sheffield.ac.uk> - 2019-10-28 15:20 +0100
    Bug#932250: 932250 Paul Kerry <p.kerry@sheffield.ac.uk> - 2019-11-04 12:30 +0100
    Bug#932250: 932250 Paul Kerry <p.kerry@sheffield.ac.uk> - 2019-11-04 13:40 +0100
      Bug#932250: random segfaults Bernhard Übelacker <bernhardu@mailbox.org> - 2020-11-10 11:10 +0100
        Bug#932250: random segfaults Paul Kerry <p.kerry@sheffield.ac.uk> - 2020-11-10 11:40 +0100
          Bug#932250: random segfaults Salvatore Bonaccorso <carnil@debian.org> - 2020-11-13 10:30 +0100

#64452 — Bug#932250: linux-image-amd64: random segfaults without taking something down

Package: linux-image-amd64
Version: 4.19+105
Severity: important

Hi.

Im currently running debian buster kernel and i experienced some segfaults.
Segfaults are random, which prevents me from pinpointing the trigger.
First i tried disabling overcommit on memory, the behavior stoped.
Today it started during installing updates for libreoffice. Two first
segfaults were caused by xfdesktop, the two more with gmain.
Another odd thing is that the error is in libglib and the code is completly
identical. The RAM was tested with memtest86+, memtester. Previous release
with backport kernel was working perfectly even with overcommit. I
guess there is a small chance of defective hw.
My next suspect is apparmor. Im going to use the system with disabled
apparmor to see if this problem persist.

Thank you

I hope this helps make debian more stable and reliable.

Have a nice day

-- System Information:
  APT prefers stable
  APT policy: (500, 'stable')
Architecture: amd64 (x86_64)

Kernel: Linux 4.19.0-5-amd64 (SMP w/4 CPU cores)
Kernel taint flags: TAINT_OOT_MODULE, TAINT_UNSIGNED_MODULE
Locale: LANG=en_US.utf8, LC_CTYPE=en_US.utf8 (charmap=UTF-8), LANGUAGE=en_US.utf8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages linux-image-amd64 depends on:
ii  linux-image-4.19.0-5-amd64  4.19.37-5

linux-image-amd64 recommends no packages.

linux-image-amd64 suggests no packages.

-- no debconf information

[toc] | [next] | [standalone]

#64453 — Bug#932250: linux-image-amd64: adding segfault data

[Multipart message — attachments visible in raw view] — view raw

Package: linux-image-amd64
Version: 4.19+105
Followup-For: Bug #932250

adding file with segfault data

-- System Information:
  APT prefers stable
  APT policy: (500, 'stable')
Architecture: amd64 (x86_64)

Kernel: Linux 4.19.0-5-amd64 (SMP w/4 CPU cores)
Kernel taint flags: TAINT_OOT_MODULE, TAINT_UNSIGNED_MODULE
Locale: LANG=en_US.utf8, LC_CTYPE=en_US.utf8 (charmap=UTF-8), LANGUAGE=en_US.utf8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)

Versions of packages linux-image-amd64 depends on:
ii  linux-image-4.19.0-5-amd64  4.19.37-5

linux-image-amd64 recommends no packages.

linux-image-amd64 suggests no packages.

-- no debconf information

[toc] | [prev] | [next] | [standalone]

#65206 — Bug#932250: linux-image-amd64: adding segfault data

[Multipart message — attachments visible in raw view] — view raw

Control: tag -1 - moreinfo
Control: tag -1 help

On Wed, 17 Jul 2019 01:41:16 +0200 m.alfaeko@gmail.com wrote:
> Package: linux-image-amd64
> Version: 4.19+105
> Followup-For: Bug #932250
> 
> adding file with segfault data

For some reason I didn't look at this until now.  The contents are:

xfdesktop[11344]: segfault at 16 ip 00007ff9fd0f29ea sp 00007ffffb934f00 error 4 in libglib-2.0.so.0.5800.3[7ff9fd0a2000+7e000]
2019-07-17T01:04:04.589422+02:00 debian kernel: [18604.040877] Code: 8d 48 ff 48 89 c3 49 89 ce 49 c1 e6 04 4c 01 f5 48 8b 45 00 48 85 c0 0f 84 c3 00 00 00 48 8b 58 08 48 85 db 0f 84 fe 01 00 00 <48> 8b 13 48 89 50 08 48 8b 45 08 48 85 c0 74 08 48 83 e8 01 48 89
xfdesktop[16727]: segfault at 16 ip 00007fe22e9aa9ea sp 00007ffd313bc000 error 4 in libglib-2.0.so.0.5800.3[7fe22e95a000+7e000]
2019-07-17T01:04:07.231175+02:00 debian kernel: [18606.685725] Code: 8d 48 ff 48 89 c3 49 89 ce 49 c1 e6 04 4c 01 f5 48 8b 45 00 48 85 c0 0f 84 c3 00 00 00 48 8b 58 08 48 85 db 0f 84 fe 01 00 00 <48> 8b 13 48 89 50 08 48 8b 45 08 48 85 c0 74 08 48 83 e8 01 48 89
gmain[16752]: segfault at 17 ip 00007fdada99a9ea sp 00007fdad7f7c7c0 error 4 in libglib-2.0.so.0.5800.3[7fdada94a000+7e000]
2019-07-17T01:04:21.214468+02:00 debian kernel: [18620.669154] Code: 8d 48 ff 48 89 c3 49 89 ce 49 c1 e6 04 4c 01 f5 48 8b 45 00 48 85 c0 0f 84 c3 00 00 00 48 8b 58 08 48 85 db 0f 84 fe 01 00 00 <48> 8b 13 48 89 50 08 48 8b 45 08 48 85 c0 74 08 48 83 e8 01 48 89
gmain[16974]: segfault at 1c ip 00007f3b5dcc29ea sp 00007f3b5b2a45f0 error 4 in libglib-2.0.so.0.5800.3[7f3b5dc72000+7e000]
2019-07-17T01:05:28.070894+02:00 debian kernel: [18687.524814] Code: 8d 48 ff 48 89 c3 49 89 ce 49 c1 e6 04 4c 01 f5 48 8b 45 00 48 85 c0 0f 84 c3 00 00 00 48 8b 58 08 48 85 db 0f 84 fe 01 00 00 <48> 8b 13 48 89 50 08 48 8b 45 08 48 85 c0 74 08 48 83 e8 01 48 89

These segfaults at least don't seem so random.  Every single one is in
the same code in libglib-2.0.so.0.5800.3.  (The ip addresses differ due
to ASLR, but the offsets are the same.)  The code bytes correspond to
part of the g_slice_alloc() function.

That seems more likely to be a bug in libglib, or perhaps some
intermediate library.

Ben.

-- 
Ben Hutchings
We get into the habit of living before acquiring the habit of thinking.
                                                         - Albert Camus

[toc] | [prev] | [next] | [standalone]

#64456

Hello,
the information might not yet be enough for
the maintainers to help.

Maybe you could install the package systemd-coredump.
In the output of 'journalctl --no-pager' should then
the segfaults appear too, followed by a backtrace
that you could forward to this bug.

Kind regards,
Bernhard

[toc] | [prev] | [next] | [standalone]

#64466 — Processed: Re: Bug#932250: linux-image-amd64: random segfaults without taking something down

Processing control commands:

> tag -1 moreinfo
Bug #932250 [linux-image-amd64] linux-image-amd64: random segfaults without taking something down
Added tag(s) moreinfo.

-- 
932250: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=932250
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems

[toc] | [prev] | [next] | [standalone]

#64467

[Multipart message — attachments visible in raw view] — view raw

Control: tag -1 moreinfo

On Wed, 2019-07-17 at 01:27 +0200, m.alfaeko@gmail.com wrote:
> Package: linux-image-amd64
> Version: 4.19+105
> Severity: important
> 
> Hi.
> 
> Im currently running debian buster kernel and i experienced some segfaults.
> Segfaults are random, which prevents me from pinpointing the trigger.
> First i tried disabling overcommit on memory, the behavior stoped.
> Today it started during installing updates for libreoffice. Two first
> segfaults were caused by xfdesktop, the two more with gmain.
> Another odd thing is that the error is in libglib and the code is completly
> identical. The RAM was tested with memtest86+, memtester. Previous release
> with backport kernel was working perfectly even with overcommit. I
> guess there is a small chance of defective hw.

Could it be overheating?

> My next suspect is apparmor. Im going to use the system with disabled
> apparmor to see if this problem persist.

What about the out-of-tree module you are using?  Can you run without
that?

> Thank you
> 
> I hope this helps make debian more stable and reliable.
[...]

Well, it isn't as helpful as it could be, because you chose to ignore
the request to report a bug against linux-image-4.19.0-5-amd64.

Ben.

-- 
Ben Hutchings
Any sufficiently advanced bug is indistinguishable from a feature.

[toc] | [prev] | [next] | [standalone]

#64471 — Bug#932250: Fwd: Bug#932250: linux-image-amd64: random segfaults without taking something down

[Multipart message — attachments visible in raw view] — view raw

It shouldnt be overheating, because i have custom setting on cpu heatsink
fan and the ram module has a own heatsink. Plus i have fan on case which
puls out the excees heat. The segfaults happend at low load, maybe it wasnt
above P2 power state. Also i have made sure that the cpu doenst exceed 90°C.

Out of tree modules are from virtual box.

After disabling overcommit and apparmor i didnt experience a segfault for
now. The segfaults started to dissapeer it seams.

I fear that this might be a bug tied to certain order of instructions. For
example if i enable debug it may change order of instruction and the bug
doesnt occur.
Reason that i reported this bug with package linux-image-amd64 is that i
didnt know about the existence of linux-mage-4.19.0-5-amd64 and i tried to
prevent creating bug file for nonexisting package. Should i file another
bug and put a link to this report ?

št 18. 7. 2019 o 4:04 Ben Hutchings <ben@decadent.org.uk> napísal(a):

> Control: tag -1 moreinfo
>
> On Wed, 2019-07-17 at 01:27 +0200, m.alfaeko@gmail.com wrote:
> > Package: linux-image-amd64
> > Version: 4.19+105
> > Severity: important
> >
> > Hi.
> >
> > Im currently running debian buster kernel and i experienced some
> segfaults.
> > Segfaults are random, which prevents me from pinpointing the trigger.
> > First i tried disabling overcommit on memory, the behavior stoped.
> > Today it started during installing updates for libreoffice. Two first
> > segfaults were caused by xfdesktop, the two more with gmain.
> > Another odd thing is that the error is in libglib and the code is
> completly
> > identical. The RAM was tested with memtest86+, memtester. Previous
> release
> > with backport kernel was working perfectly even with overcommit. I
> > guess there is a small chance of defective hw.
>
> Could it be overheating?
>
> > My next suspect is apparmor. Im going to use the system with disabled
> > apparmor to see if this problem persist.
>
> What about the out-of-tree module you are using?  Can you run without
> that?
>
> > Thank you
> >
> > I hope this helps make debian more stable and reliable.
> [...]
>
> Well, it isn't as helpful as it could be, because you chose to ignore
> the request to report a bug against linux-image-4.19.0-5-amd64.
>
> Ben.
>
> --
> Ben Hutchings
> Any sufficiently advanced bug is indistinguishable from a feature.
>
>
>

[toc] | [prev] | [next] | [standalone]

#64472 — Bug#932250: [Fwd: Re: Bug#932250: linux-image-amd64: random segfaults without taking something down]

-------- Forwarded Message --------
From: a a <m.alfaeko@gmail.com>
To: Ben Hutchings <ben@decadent.org.uk>
Subject: Re: Bug#932250: linux-image-amd64: random segfaults without 
taking something down
Date: Thu, 18 Jul 2019 13:09:53 +0200
Message-Id: <CANfEg5hbNfNpOXrHStG6b-
vUi3zbFFaw3ChtOx3+i2XRL4UbtQ@mail.gmail.com>

It shouldnt be overheating, because i have custom setting on cpu heatsink
fan and the ram module has a own heatsink. Plus i have fan on case which
puls out the excees heat. The segfaults happend at low load, maybe it wasnt
above P2 power state. Also i have made sure that the cpu doenst exceed 90°C.

Out of tree modules are from virtual box.

After disabling overcommit and apparmor i didnt experience a segfault for
now. The segfaults started to dissapeer it seams.

I fear that this might be a bug tied to certain order of instructions. For
example if i enable debug it may change order of instruction and the bug
doesnt occur.
Reason that i reported this bug with package linux-image-amd64 is that i
didnt know about the existence of linux-mage-4.19.0-5-amd64 and i tried to
prevent creating bug file for nonexisting package. Should i file another
bug and put a link to this report ?

št 18. 7. 2019 o 4:04 Ben Hutchings <ben@decadent.org.uk> napísal(a):

> Control: tag -1 moreinfo
> 
> On Wed, 2019-07-17 at 01:27 +0200, m.alfaeko@gmail.com wrote:
> > Package: linux-image-amd64
> > Version: 4.19+105
> > Severity: important
> > 
> > Hi.
> > 
> > Im currently running debian buster kernel and i experienced some
> segfaults.
> > Segfaults are random, which prevents me from pinpointing the trigger.
> > First i tried disabling overcommit on memory, the behavior stoped.
> > Today it started during installing updates for libreoffice. Two first
> > segfaults were caused by xfdesktop, the two more with gmain.
> > Another odd thing is that the error is in libglib and the code is
> completly
> > identical. The RAM was tested with memtest86+, memtester. Previous
> release
> > with backport kernel was working perfectly even with overcommit. I
> > guess there is a small chance of defective hw.
> 
> Could it be overheating?
> 
> > My next suspect is apparmor. Im going to use the system with disabled
> > apparmor to see if this problem persist.
> 
> What about the out-of-tree module you are using?  Can you run without
> that?
> 
> > Thank you
> > 
> > I hope this helps make debian more stable and reliable.
> [...]
> 
> Well, it isn't as helpful as it could be, because you chose to ignore
> the request to report a bug against linux-image-4.19.0-5-amd64.
> 
> Ben.
> 
> --
> Ben Hutchings
> Any sufficiently advanced bug is indistinguishable from a feature.
> 
> 
> 
-- 
Ben Hutchings
Any sufficiently advanced bug is indistinguishable from a feature.

[toc] | [prev] | [next] | [standalone]

#65205 — Processed: Re: linux-image-amd64: adding segfault data

Processing control commands:

> tag -1 - moreinfo
Bug #932250 [linux-image-amd64] linux-image-amd64: random segfaults without taking something down
Removed tag(s) moreinfo.
> tag -1 help
Bug #932250 [linux-image-amd64] linux-image-amd64: random segfaults without taking something down
Added tag(s) help.

-- 
932250: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=932250
Debian Bug Tracking System
Contact owner@bugs.debian.org with problems

[toc] | [prev] | [next] | [standalone]

#65528 — Bug#932250: 932250

Hi

We are experiencing the same issues as the o/p, with random segfaults 
like this with the same "Code" values...

[252199.319638] xfdesktop[7276]: segfault at b ip 00007fea64a219ea sp 
00007ffcf0678c00 error 4 in libglib-2.0.so.0.5800.3[7fea649d1000+7e000]
[252199.319662] Code: 8d 48 ff 48 89 c3 49 89 ce 49 c1 e6 04 4c 01 f5 48 
8b 45 00 48 85 c0 0f 84 c3 00 00 00 48 8b 58 08 48 85 db 0f 84 fe 01 00 
00 <48> 8b 13 48 89 50 08 48 8b 45 08 48 85 c0 74 08 48 83 e8 01 48 89

when using xfce 4.12.5 on debian 10 buster amd64.

"xfce4-panel" and "gmain" are sometimes shown as well as/instead of 
"xfdesktop" in the dmesg output.


This is happening on standalone debian 10 machines and also on 10x/30x 
systems in a network of debian 10 machines. The same segfault is shown 
at random times, but mainly after login. Some machines are using nvidia 
proprietary graphics, but the majority are using intel modesetting from 
the kernel. We are using a 4.19.80 custom kernel built from vanilla 
sources. This also happens with a brand new user profile.

When searching online, I came across several references to the segfault 
but with older versions of libglib2.0-0 like this for instance...

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=783865

and also these bugs look relevant...
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=933202
and...
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=934105

but I don't really want to build and upgrade to xfce 4.14 as mentioned 
in those bug reports.


I presume this would suggest this is a problem with the newer libglib2 
2.58.3-2+deb10u1 which is in debian 10?

I'm temporarily testing using libglib2.0-0:amd64 2.62.1-1 from the 
bullseye testing repo and haven't had a segfault as yet (fingers crossed).


Perhaps re-assign this bug to libglib2.0-0 ?


Cheers
Paul.
--

[toc] | [prev] | [next] | [standalone]

#65554 — Bug#932250: 932250

Hi all

Following up on from my previous post, further testing with 
libglib2.0-0:amd64 2.62.1-1 packages installed from testing eventually 
caused the same segfault but with the newer libglib2 version shown in 
dmesg...

[231681.228074] xfdesktop[31999]: segfault at a ip 00007f7788322b9d sp 
00007ffec3615cd0 error 4 in libglib-2.0.so.0.6200.1[7f77882d1000+80000]

so this would appear to be a red-herring as far as libglib2 is 
concerned. I have reverted back to the standard libglib2 debian 10 packages.


Having done more testing with the standard libglib2 debian 10 packages, 
it now appears that installing packages via apt which add or change xfce 
menu items via .desktop files while someone is logged in using xfce, can 
trigger the dmesg error, so it could be a bug somewhere in xfce 4.12. 
Maybe the issue has been fixed upstream in xfce 4.14 as mentioned in the 
other links?


Cheers
Paul.
--

[toc] | [prev] | [next] | [standalone]

#65555 — Bug#932250: 932250

see also this archived bug...

https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=934105

which mentions issues when upgrading google-chrome packages causing the 
same/similar issue, albeit with sid packages.


Cheers
Paul.
--

[toc] | [prev] | [next] | [standalone]

#68537 — Bug#932250: random segfaults

Hello Paul,
is it possible to install the package systemd-coredump on
the systems showing this crash?

Then after the next crash in the output of 'journalctl --no-pager'
should the segfault appear followed by a backtrace
that you could forward to this bug.

This should clarify which function calls lead to the crash.

More information on improving that information additionally
are here: https://wiki.debian.org/HowToGetABacktrace

Kind regards,
Bernhard

[toc] | [prev] | [next] | [standalone]

#68538 — Bug#932250: random segfaults

Hi Bernhard and thanks for the follow up.

I've not had anymore segfaults since reporting this issue last year.

I'd still go for the hypothesis of it being a xfce related issue which I 
linked too before:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=934105

This can be alleviated by installing packages without xfce running or by 
repackaging pre-built  packages and removing .desktop files within those 
packages.

For me, this is certainly not a linux-image-amd64 bug.

Cheers
Paul.

On Tue, 10 Nov 2020 11:04:13 +0100 =?UTF-8?Q?Bernhard_=c3=9cbelacker?= 
<bernhardu@mailbox.org> wrote:
> Hello Paul,
> is it possible to install the package systemd-coredump on
> the systems showing this crash?
> 
> Then after the next crash in the output of 'journalctl --no-pager'
> should the segfault appear followed by a backtrace
> that you could forward to this bug.
> 
> This should clarify which function calls lead to the crash.
> 
> More information on improving that information additionally
> are here: https://wiki.debian.org/HowToGetABacktrace
> 
> Kind regards,
> Bernhard
> 
> 

-- 

======================================================================
Mr. Paul Kerry
Astronomy Technical Support and Computer Manager,
Department of Physics and Astronomy, University of Sheffield,
Hicks Building, Hounsfield Road, Sheffield, S3 7RH, ENGLAND.
Email: p.kerry@sheffield.ac.uk
Web: http://paulkerry.staff.shef.ac.uk/
IM Chat: paulkerry@jabber.org (XMPP)
Skype: paulkerry2
Phone: +44 (0) 114 22 23551

I don't expect you to respond to my email outside your working hours.

At the University of Sheffield we value and encourage flexible working
patterns, so please be assured that I respect your working pattern and
I am looking forward to your response when you are next working.
======================================================================

[toc] | [prev] | [next] | [standalone]

#68552 — Bug#932250: random segfaults

Hi,

On Tue, Nov 10, 2020 at 10:37:24AM +0000, Paul Kerry wrote:
> 
> Hi Bernhard and thanks for the follow up.
> 
> I've not had anymore segfaults since reporting this issue last year.
> 
> I'd still go for the hypothesis of it being a xfce related issue which I
> linked too before:
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=934105
> 
> This can be alleviated by installing packages without xfce running or by
> repackaging pre-built  packages and removing .desktop files within those
> packages.
> 
> For me, this is certainly not a linux-image-amd64 bug.

So should we merge it together with #934105 ?

Regards,
Salvatore

[toc] | [prev] | [standalone]

Back to top | Article view | linux.debian.kernel

csiph-web