[Bug 91577] [HD Graphics 4000] Segfault accessing rq->bo->handle

bugzilla-daemon at freedesktop.org bugzilla-daemon at freedesktop.org
Fri Aug 7 06:07:37 PDT 2015


https://bugs.freedesktop.org/show_bug.cgi?id=91577

--- Comment #27 from Dominic <fdo-bugs.aspseka at xoxy.net> ---
Was not able to trigger any crashes so far. Memory footprint drops as soon as I
stop selecting/deselecting text.

However, I was able to trigger oom condition by *not* stopping ( (de)selecting
every 2-3s or so on my machine is enough to keep Xorg busy on his CPU core and
eating more and more objects.) After about 65k objects, using a tad more than
4GiB, it starts freeing.

-----
65707 objects, 4687548416 bytes
24237 [15] objects, 1699127296 [8998912] bytes in gtt
  0 [0] active objects, 0 [0] bytes
  24237 [15] inactive objects, 1699127296 [8998912] bytes
0 unbound objects, 0 bytes
0 purgeable objects, 0 bytes
6 pinned mappable objects, 8962048 bytes
7 fault mappable objects, 40960 bytes
2147483648 [268435456] gtt total

batch pool: 1 objects, 8192 bytes (0 active, 0 inactive, 0 global, 0 shared,
8192 unbound)

systemd-logind: 83 objects, 37683200 bytes (0 active, 0 inactive, 0 global,
5242880 shared, 37601280 unbound)
Xwayland: 21 objects, 7675904 bytes (0 active, 0 inactive, 0 global, 0 shared,
7598080 unbound)
Xwayland: 0 objects, 0 bytes (0 active, 0 inactive, 0 global, 0 shared, 0
unbound)
systemd-logind: 65333 objects, 4518924288 bytes (0 active, 1677623296 inactive,
1677623296 global, 35459072 shared, 2835603456 unbound)
gnome-shell: 262 objects, 149680128 bytes (0 active, 29831168 inactive,
29831168 global, 25890816 shared, 100548608 unbound)
-----

I collected a huge log containing two "garbage collections", terminating xorg
via sysreq as soon as I saw the second one happening.
I'll try to attach the last 1M lines of it.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20150807/c30f928a/attachment.html>


More information about the intel-gfx-bugs mailing list