Xorg crashes with SIGABRT

Bug #2059033 reported by Michael Schanne
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Mesa
New
Unknown
mesa (Ubuntu)
New
Undecided
Unassigned

Bug Description

I am running an Intel Kabylake CPU with Ubuntu 22.04, and i915 graphics driver. I am seeing random Xorg crashes due to SIGABRT, which kicks me back to the login screen.

Ubuntu version: 22.04
Kernel version: 5.19.0-50-generic
libgl1-mesa-dri version: 23.0.4-0ubuntu1~22.04.1

I am seeing this message in kern.log:

i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:85dffffb, in Xorg [1617]

I have also gotten a core dump; here is the track trace:

#0 __pthread_kill_implementation (no_tid=0, signo=6, threadid=140131874372224) at ./nptl/pthread_kill.c:44
#1 __pthread_kill_internal (signo=6, threadid=140131874372224) at ./nptl/pthread_kill.c:78
#2 __GI___pthread_kill (threadid=140131874372224, signo=signo@entry=6) at ./nptl/pthread_kill.c:89
#3 0x00007f72fec42476 in __GI_raise (sig=sig@entry=6) at ../sysdeps/posix/raise.c:26
#4 0x00007f72fec287f3 in __GI_abort () at ./stdlib/abort.c:79
#5 0x0000560c5cc756a0 in OsAbort () at ../../../../os/utils.c:1352
#6 0x0000560c5cc7aea9 in AbortServer () at ../../../../os/log.c:879
#7 0x0000560c5cc7be9a in FatalError (f=f@entry=0x560c5ccadc10 "Caught signal %d (%s). Server aborting\n") at ../../../../os/log.c:1017
#8 0x0000560c5cc72abd in OsSigHandler (unused=<optimized out>, sip=0x7ffff13ae7b0, signo=6) at ../../../../os/osinit.c:156
#9 OsSigHandler (signo=6, sip=0x7ffff13ae7b0, unused=<optimized out>) at ../../../../os/osinit.c:110
#10 <signal handler called>
#11 __pthread_kill_implementation (no_tid=0, signo=6, threadid=140131874372224) at ./nptl/pthread_kill.c:44
#12 __pthread_kill_internal (signo=6, threadid=140131874372224) at ./nptl/pthread_kill.c:78
#13 __GI___pthread_kill (threadid=140131874372224, signo=signo@entry=6) at ./nptl/pthread_kill.c:89
#14 0x00007f72fec42476 in __GI_raise (sig=sig@entry=6) at ../sysdeps/posix/raise.c:26
#15 0x00007f72fec287f3 in __GI_abort () at ./stdlib/abort.c:79
#16 0x00007f72fc49d86d in _iris_batch_flush (batch=0x560c5e916118, file=<optimized out>, line=<optimized out>) at ../src/gallium/drivers/iris/iris_batch.c:1116
#17 0x00007f72fd1da6ff in iris_fence_flush (ctx=0x560c5e915be0, out_fence=0x0, flags=<optimized out>) at ../src/gallium/drivers/iris/iris_fence.c:267
#18 0x00007f72fc55f127 in st_flush (flags=0, fence=0x0, st=0x560c5e93bbc0) at ../src/mesa/state_tracker/st_cb_flush.c:60
#19 st_glFlush (ctx=<optimized out>, gallium_flush_flags=0) at ../src/mesa/state_tracker/st_cb_flush.c:94
#20 0x00007f72fe56b604 in _glamor_block_handler (screen=0x560c5e8655a0, timeout=0x7ffff13af054) at ../../../../glamor/glamor.c:285
#21 0x00007f72fe5a6157 in msBlockHandler (pScreen=0x560c5e8655a0, timeout=0x7ffff13af054) at ../../../../../../../hw/xfree86/drivers/modesetting/driver.c:701
#22 0x0000560c5cb01095 in BlockHandler (pTimeout=pTimeout@entry=0x7ffff13af054) at ../../../../dix/dixutils.c:394
#23 0x0000560c5cc6c0c9 in WaitForSomething (are_ready=0) at ../../../../os/WaitFor.c:201
#24 0x0000560c5cafc3e7 in Dispatch () at ../../../../dix/dispatch.c:492
#25 0x0000560c5cb006b4 in dix_main (argc=11, argv=0x7ffff13af268, envp=<optimized out>) at ../../../../dix/main.c:272
#26 0x00007f72fec29d90 in __libc_start_call_main (main=main@entry=0x560c5cae95d0 <main>, argc=argc@entry=11, argv=argv@entry=0x7ffff13af268) at ../sysdeps/nptl/libc_start_call_main.h:58
#27 0x00007f72fec29e40 in __libc_start_main_impl (main=0x560c5cae95d0 <main>, argc=11, argv=0x7ffff13af268, init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>, stack_end=0x7ffff13af258) at ../csu/libc-start.c:392
#28 0x0000560c5cae9605 in _start ()

This post describes a similar issue:
https://gitlab.freedesktop.org/drm/intel/-/issues/4858

I tried using a workaround from the above link, to set GPU to max frequency using command "intel_gpu_frequency -m", this seemed to reduce the rate of occurrence from once per day to once every 3 days, but this is not a permanent solution.

Changed in mesa:
status: Unknown → New
Revision history for this message
Michael Schanne (mschan01) wrote :

Tried kernel parameter i915.enable_dc=0 as suggested by https://wiki.archlinux.org/title/intel_graphics#Crash.2Ffreeze_on_low_power_Intel_CPUs; still crashes

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.