[Bug 110848] Everything using GPU gets stuck after running+killing parallel Media loads (after running 3D benchmarks)

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=110848

--- Comment #5 from Eero Tamminen <eero.t.tamminen at intel.com> ---
Created attachment 144476
  --> https://bugs.freedesktop.org/attachment.cgi?id=144476&action=edit
Kernel stacks for all tasks when the problem happens

# for i in /proc/[0-9]*; do
    egrep '^(Name|State|Pid)' $i/status; cat $i/stack;
  done | grep -B1 -A4 "disk sleep"

Name:   khugepaged
State:  D (disk sleep)
Pid:    38
[<0>] kthread+0x113/0x130
[<0>] ret_from_fork+0x35/0x40
--
Name:   kworker/u8:0+i915
State:  D (disk sleep)
Pid:    7274
[<0>] __i915_gem_free_work+0x5f/0x90 [i915]
[<0>] process_one_work+0x1e9/0x410
[<0>] worker_thread+0x2d/0x3d0
--
Name:   ffmpeg
State:  D (disk sleep)
Pid:    7293
[<0>] __fput+0xae/0x200
[<0>] task_work_run+0x84/0xa0
[<0>] do_exit+0x308/0xba0

--- lots of similar ffmpeg tasks ---

Name:   ffmpeg
State:  D (disk sleep)
Pid:    8606
[<0>] chrdev_open+0xa3/0x1b0
[<0>] do_dentry_open+0x1c4/0x380
[<0>] path_openat+0x564/0x11f0

--- lots of similar ffmpeg tasks ---

Name:   ps
State:  D (disk sleep)
Pid:    8751
[<0>] proc_pid_cmdline_read+0x1e3/0x330
[<0>] vfs_read+0x91/0x140
[<0>] ksys_read+0x91/0xe0


I guess the ffmpeg processes in exit stage are ones from previous test, and the
ones trying to open device file ffmpeg processes from the next test.

To me it looks like kworker doing __i915_gem_free_work() stuff would be
blocking khugepaged task and everything else.


PS. Freeze when "ps" reads PID command lines seems to happen for running
processes:
$ for i in /proc/[0-9]*; do
    egrep '^(Name|State|Pid)' $i/status; cat $i/cmdline;
  done
...
Name:   ffmpeg
State:  Z (zombie)
Pid:    7937
Name:   ffmpeg
State:  Z (zombie)
Pid:    7977
Name:   ffmpeg
State:  R (running)
Pid:    8008
<freeze>

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20190607/7556ae82/attachment-0001.html>