| Summary: | GPU hang resulting in Freeze(?) then unclean logout (possibly connected to LibreOffice) | ||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | DRI | Reporter: | wettererscheinung | ||||||||||||||
| Component: | DRM/Intel | Assignee: | Intel GFX Bugs mailing list <intel-gfx-bugs> | ||||||||||||||
| Status: | CLOSED DUPLICATE | QA Contact: | Intel GFX Bugs mailing list <intel-gfx-bugs> | ||||||||||||||
| Severity: | normal | ||||||||||||||||
| Priority: | medium | CC: | intel-gfx-bugs, wettererscheinung | ||||||||||||||
| Version: | unspecified | ||||||||||||||||
| Hardware: | x86-64 (AMD64) | ||||||||||||||||
| OS: | Linux (All) | ||||||||||||||||
| Whiteboard: | |||||||||||||||||
| i915 platform: | SKL | i915 features: | GPU hang | ||||||||||||||
| Attachments: |
|
||||||||||||||||
|
Description
wettererscheinung
2017-08-27 17:33:41 UTC
(In reply to wettererscheinung from comment #0) > What have I tried? > * I tried to purge xserver-xorg-video-intel. But it occured again. It's a gpu hang from using -modesetting. Please do attach the error state so that we can triage it. Dear Chris, thanks for your answer, is it possible to reconstruct/retrieve the error state, after I restarted? Because now it shows "No error state collected". Otherwise it will take one or two weeks until it occurs again. Sincerely yours Maria Created attachment 133911 [details]
syslog from 31.08.2017
This is the Syslog to the GPU Hang from today.
Created attachment 133912 [details]
Error state of the GPU Hang from 31.08.2017
This time the GPU hang occured earlier than the last times. My notebook was only one night on suspend since last reboot.
Added the requested info. Thanks alot for your time and help! Maria Hello Maria, Could you try to reproduce with intel_iommu=igfx_off on grub? If it works may be a dup of bug 89360 or bug 103076. Created attachment 135297 [details] sys_class_drm_card0_error_17-11-08.txt Dear Elizabeth, sorry for not answering so long and thanks for the hint. Since I didn't experience the bug for some time I thought/hoped it had vanished. Unluckily it happened again today (I added the GPU Hang error output to this mail). * Nonetheless, how do I apply this on grub? * I read that virtualization won't work anymore - is that true? (This would be a problem as I do use virtualbox regularly) * When I understood the info about this feature right you mean that possibly the GPU doesn't work correctly with the DMA Re-Mapping? Is that an hardware/guarantee issue? Bytheway I experience two kinds of occurances: 1 - freeze, then logout; 2 - total freeze, no change, only poweroff helps therefore no error report possible Yours Maria bugzilla-daemon@freedesktop.org: > Elizabeth <mailto:elizabethx.de.la.torre.mena@intel.com> changed bug > 102433 <https://bugs.freedesktop.org/show_bug.cgi?id=102433> > What Removed Added > Status NEW NEEDINFO > > *Comment # 6 <https://bugs.freedesktop.org/show_bug.cgi?id=102433#c6> on > bug 102433 <https://bugs.freedesktop.org/show_bug.cgi?id=102433> from > Elizabeth <mailto:elizabethx.de.la.torre.mena@intel.com> * > > Hello Maria, Could you try to reproduce with intel_iommu=igfx_off on grub? If > it works may be a dup of bug 89360 <show_bug.cgi?id=89360> or bug 103076 <show_bug.cgi?id=103076>. > > ------------------------------------------------------------------------ > You are receiving this mail because: > > * You are on the CC list for the bug. > * You reported the bug. > (In reply to wettererscheinung from comment #7) >... > * Nonetheless, how do I apply this on grub? Hello Maria, to apply this execute: $ sudo nano /etc/default/grub Add intel_iommu=igfx_off inside the "" after the grub command line, i.e.: GRUB_CMDLINE_LINUX_DEFAULT="quiet intel_iommu=igfx_off" Save and close. Then apply: $sudo update-grub And then reboot. > * I read that virtualization won't work anymore - is that true? > (This would be a problem as I do use virtualbox regularly) You can find more information over the internet: https://en.wikipedia.org/wiki/Input%E2%80%93output_memory_management_unit#Virtualization , Virtualization should keep working. > * When I understood the info about this feature right you mean that > possibly the GPU doesn't work correctly with the DMA Re-Mapping? > Is that an hardware/guarantee issue? You could do a memtest86 to be sure your memory is working correctly: On debian, do 'apt install memtest86'. You should see it in the grub options as a boot target that you can choose. There is no log. If memtest reports an error, you have to replace your memory. If it was a DMAR error, that should be follow on bug 89360. > Bytheway I experience two kinds of occurances: 1 - freeze, then logout; > 2 - total freeze, no change, only poweroff helps therefore no error > report possible Those could be different issues, though you would need to identify a patron to determine if they should be worked separately. > Yours > Maria From error state: ERROR: 0x00000000 FAULT_TLB_DATA: 0x0000001b 0xaacb0b2b Address 0x0000baacb0b2b000 GGTT DONE_REG: 0x07ffffff render command stream: START: 0x00011000 HEAD: 0xf9001d80 [0x00001d28] head = 0x00001d80, wraps = 1992 TAIL: 0x00001da8 [0x00001d80, 0x00001da8] CTL: 0x00003001 len=16384, enabled MODE: 0x00000000 HWS: 0xfffe8000 ACTHD: 0x00000000 f9001d80 at ring: 0x00000000 IPEIR: 0x00000000 IPEHR: 0x7a000004 INSTDONE: 0xffdfffff busy: CS SC_INSTDONE: 0xfffffbff SAMPLER_INSTDONE[0][0]: 0xffffffff SAMPLER_INSTDONE[0][1]: 0xffffffff SAMPLER_INSTDONE[0][2]: 0xffffffff ROW_INSTDONE[0][0]: 0xfffffffd ROW_INSTDONE[0][1]: 0xfffffffd ROW_INSTDONE[0][2]: 0xfffffffd batch: [0x00000000_044a6000, 0x00000000_044ae000] BBADDR: 0x00000000_044a631c BB_STATE: 0x00000020 INSTPS: 0x00008980 INSTPM: 0x00000000 FADDR: 0x00000000 00012da8 RC PSMI: 0x00000010 FAULT_REG: 0x00000000 SYNC_0: 0x00000000 SYNC_1: 0x00000000 SYNC_2: 0x00000000 GFX_MODE: 0x00008000 PDP0: 0x000000041915e000 PDP1: 0x0000000000000000 PDP2: 0x0000000000000000 PDP3: 0x0000000000000000 seqno: 0x002e45d4 last_seqno: 0x002e45d6 waiting: yes ring->head: 0x00001d00 ring->tail: 0x00001da8 hangcheck stall: yes hangcheck action: dead hangcheck action timestamp: 4331761496, 122744 ms ago ELSP[0]: pid 1042, ban score 0, seqno 2:002e45d5, emitted 123896ms ago, head 00001d28, tail 00001da8 ELSP[1]: pid 1904, ban score 0, seqno a:002e45d6, emitted 123896ms ago, head 00001c10, tail 00001c88 Active context: Xorg[1042] user_handle 1 hw_id 2, ban score 0 guilty 0 active 0 Created attachment 135397 [details]
sys_class_drm_card0_error_17-11-11.txt
Dear Elizabeth,
just now it happened again (Freeze for like 10 seconds, then sudden
logout), although "intel_iommu=igfx_off" was activated. It seems to only
happen, when I close LibreOffice before night, leave my computer active
and logged in over night and then the next day work for some time with
LibreOffice.
I attached the new report to this mail.
If you need any other logs or reports please tell me.
Thanks for your help!
Maria
Created attachment 135399 [details] sys_class_drm_card0_error_17-11-11_B.txt I have to correct myself, now it freezed and kicked shortly after a fresh reboot. This didn't used to happen before. It makes working pretty hard *sigh* I am typing the same text the third time ... Yours Maria Maria: > Dear Elizabeth, > > just now it happened again (Freeze for like 10 seconds, then sudden > logout), although "intel_iommu=igfx_off" was activated. It seems to only > happen, when I close LibreOffice before night, leave my computer active > and logged in over night and then the next day work for some time with > LibreOffice. > > I attached the new report to this mail. > > If you need any other logs or reports please tell me. > > Thanks for your help! > Maria > (In reply to wettererscheinung from comment #10) > Hello Maria, you can remove iommu parameter doing the same procedure, clearly it isn't related. I'm duplicating this bug to bug 101780 that is the same issue but reported earlier. Please keep track of the issue in that bug. *** This bug has been marked as a duplicate of bug 101780 *** |
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.