Bug 93438

Summary: SEGV 0x00007fb7d53cdb0d in __kgem_retire_rq () from /usr/lib/xorg/modules/drivers/intel_drv.so
Product: DRI Reporter: Marc MERLIN <marc>
Component: DRM/IntelAssignee: Intel GFX Bugs mailing list <intel-gfx-bugs>
Status: CLOSED FIXED QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: major    
Priority: medium CC: intel-gfx-bugs
Version: unspecified   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features: GEM/Other
Attachments:
Description Flags
gdb traceback none

Description Marc MERLIN 2015-12-18 14:09:41 UTC
Created attachment 120585 [details]
gdb traceback

ii  xorg              1:7.7+12      amd64         X.Org X Window System
ii  xserver-xorg-vide 2:2.99.917-2  amd64         X.Org X server -- Intel i8xx, i9xx disp
kernel: 4.3.3-amd64

Program received signal SIGSEGV, Segmentation fault.
0x00007fb7d53cdb0d in __kgem_retire_rq () from /usr/lib/xorg/modules/drivers/intel_drv.so
#0  0x00007fb7d53cdb0d in __kgem_retire_rq () from /usr/lib/xorg/modules/drivers/intel_drv.so
No locals.
#1  0x00007fb7d53cecb7 in __kgem_retire_requests_upto () from /usr/lib/xorg/modules/drivers/intel_drv.so
No locals.
#2  0x00007fb7d53da393 in __kgem_bo_is_busy.part.69 () from /usr/lib/xorg/modules/drivers/intel_drv.so
No locals.
#3  0x00007fb7d53ff9e5 in sna_put_zpixmap_blt.isra.92 () from /usr/lib/xorg/modules/drivers/intel_drv.so
No locals.
#4  0x00007fb7d53ffc59 in sna_put_image () from /usr/lib/xorg/modules/drivers/intel_drv.so
Comment 1 Chris Wilson 2015-12-18 14:15:00 UTC
Please try again with http://cgit.freedesktop.org/xorg/driver/xf86-video-intel/ - it should be fixed.
Comment 2 Marc MERLIN 2015-12-25 18:15:37 UTC
Thanks.
The last one I found in debian experimental was:
xserver-xorg-video-intel_2%3a2.99.917+git20151217
Is that too old? Which date was the patch added?
(that driver didn't SEGV but I've found the laptop locked up when I came back the next morning, although since it's locked up, it's hard to say what at's fault)

For now, I've come back to 2:2.99.917-1 which doesn't seem to hang or crash, but I'm happy to try the 

Thanks, Marc
(Merry Xmas)
Comment 3 Marc MERLIN 2015-12-25 18:15:53 UTC
Thanks.
The last one I found in debian experimental was:
xserver-xorg-video-intel_2%3a2.99.917+git20151217
Is that too old? Which date was the patch added?
(that driver didn't SEGV but I've found the laptop locked up when I came back the next morning, although since it's locked up, it's hard to say what at's fault)

For now, I've come back to 2:2.99.917-1 which doesn't seem to hang or crash, but I'm happy to try the 

Thanks, Marc
(Merry Xmas)
Comment 4 Marc MERLIN 2015-12-25 18:17:06 UTC
Thanks.
The last one I found in debian experimental was:
xserver-xorg-video-intel_2%3a2.99.917+git20151217
Is that too old? Which date was the patch added?
(that driver didn't SEGV but I've found the laptop locked up when I came back the next morning, although since it's locked up, it's hard to say what at's fault)

For now, I've come back to 2:2.99.917-1 which doesn't seem to hang or crash, but I'm happy to try the 

Thanks, Marc
(Merry Xmas)
Comment 5 Marc MERLIN 2015-12-30 13:08:31 UTC
Doh, sorry for the triple post. Bugzilla kept telling me that my post hadn't gone through, when in fact it was lying :(

1) xserver-xorg-video-intel  2:2.99.917-1 hung on 4.3.3
2) 2:2.99.917+git20151217-1~exp1 gave a full hang too on 4.1.15 (not even the mouse moving, power cycle required)

I'm now trying 2:2.99.917+git20151217-1~exp1 on 4.3.3 and next I'll go back to 2:2.99.917-1 with 4.1.15.
If there is a newer git version I should be trying let me know, but at least this one just hangs completely, it does not crash (maybe not better, but not a SEGV)
Comment 6 Chris Wilson 2015-12-30 13:17:15 UTC
The crash is completely secondary to an underlying resource starvation bug, more than likely not from the ddx.
Comment 7 Marc MERLIN 2015-12-31 14:48:48 UTC
Thanks Chris, that makes sense.
I have a big screen (3K), that's why I bought that laptop.
The intel graphics has always felt underpowered for it, but now that I have it, I'll have to make do :)

Are there kernel / xorg options I can set to reduce performance and increase stability?
I could also run google-chrome with --disable-gpu, but that will kill some amount of sites that actually require hardware rendering.

Thank.
Comment 8 Marc MERLIN 2016-01-03 15:57:15 UTC
Ok, after more tests, more crashes, and more data lost:
The old xserver-xorg-video-intel 2:2.99.917-1 fails with 4.1.15 and 4.3.3 but when it does, my mouse keeps working, sound on an mplayer video keeps working (but no video), and I can switch back to a text console.
The X server does not come back to life though.

With 2:2.99.917+git20151217-1~exp1 , X server deadlocks, everything is lost, I need to power cycle (both with 4.1.15 and 4.3.3).

This is quite disappointing :(
I'm not running google-chrome-beta --disable-gpu to see if that helps.
Comment 9 Jani Nikula 2016-06-16 14:21:37 UTC
Does the problem persist with the latest kernel and userspace bits?
Comment 10 yann 2017-01-30 16:54:46 UTC
(In reply to Jani Nikula from comment #9)
> Does the problem persist with the latest kernel and userspace bits?

Timeout, assuming that this is fixed by now. If this is not the case, please re-test with latest kernel & xf86-video-intel to see if this issue is still occurring since there were improvements pushed in kernel and xf86-video-intel that will benefit to your system.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.