Bug 81907

Summary: Unreal Engine "Effects Cave" Demo crashes with SIGBUS in __memcpy_sse2_unaligned
Product: DRI Reporter: Christoph Haag <haagch>
Component: DRM/RadeonAssignee: Default DRI bug account <dri-devel>
Status: RESOLVED FIXED QA Contact:
Severity: normal    
Priority: medium    
Version: unspecified   
Hardware: Other   
OS: All   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
full backtrace none

Description Christoph Haag 2014-07-30 08:43:54 UTC 
Created attachment 103673 [details]
full backtrace

llvm-svn 214022
mesa git master from today without f4b0ab7afd83c811329211eae8167c9bf238870c
linux 3.16-rc7

00:02.0 VGA compatible controller: Intel Corporation 3rd Gen Core processor Graphics Controller (rev 09)
01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Wimbledon XT [Radeon HD 7970M] (rev ff)

intel with render nodes, tearfree and DRI3 and I run it with DRI_PRIME=1.

Because of bug 81834 I reverted f4b0ab7afd83c811329211eae8167c9bf238870c.

Now I get this crash in the "Effects Cave" demo of the unreal engine. More rarely (?) I get bug 80880 instead/first. In other demos (perhaps less complex) like TappyChicken I always get bug 80880 and never (?) this crash here:

0x00007ffff6a15859 in __memcpy_sse2_unaligned () from /usr/lib/libc.so.6


It seems to run for a few seconds, the music is playing, but the window contents are black (it seems nothing at all is rendered, sometimes there are white stripes that do not vanish until it crashes). Then it crashes with the attached backtrace.

The "binary" struct seem to be always one of the same few. Others are:

binary = {code = 0x7fffb9ced270 "~\n\376\276\t\003\374\276\002\003\002\310\002\002\006\310\001\001", code_size = 76, config = 0x7fffb9cf1a90 "(\260", config_size = 24, rodata = 0x0, rodata_size = 0, disassembled = 0}

and

binary = {code = 0x7fffb9f85a10 "~\n\376\276\t", <incomplete sequence \374\276>, code_size = 4160, config = 0x7fffb9e33840 "(\260", config_size = 24, rodata = 0x0, rodata_size = 0, disassembled = 0}
Comment 1 Christoph Haag 2014-08-01 10:24:25 UTC 
Some information in case someone tries to reproduce it.

I applied this patch additionally to the revert: https://bugs.freedesktop.org/show_bug.cgi?id=80880#c10

Many demos work now. But the Cave Demo does not work, it hangs the GPU.

I can not reproduce the original bug I posted here anymore with the patch, just the hang. It looks like this:

ring 0 stalled for more than 10146msec
GPU lockup (waiting for 0x000000000000a921 last fence id 0x000000000000a920 on ring 0)
failed to get a new IB (-35)
Saved 205 dwords of commands on ring 0.
GPU softreset: 0x0000004D
  GRBM_STATUS               = 0xE7DE5028
  GRBM_STATUS_SE0           = 0xEF800002
  GRBM_STATUS_SE1           = 0xEFC00002
  SRBM_STATUS               = 0x200000C0
  SRBM_STATUS2              = 0x00000000
  R_008674_CP_STALLED_STAT1 = 0x00000000
  R_008678_CP_STALLED_STAT2 = 0x00010800
  R_00867C_CP_BUSY_STAT     = 0x00408006
  R_008680_CP_STAT          = 0x84038647
  R_00D034_DMA_STATUS_REG   = 0x44C83146
  R_00D834_DMA_STATUS_REG   = 0x44C83D57
  VM_CONTEXT1_PROTECTION_FAULT_ADDR   0x00000000
  VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x00000000
GRBM_SOFT_RESET=0x0000DDFF
SRBM_SOFT_RESET=0x00100100
  GRBM_STATUS               = 0x00003028
  GRBM_STATUS_SE0           = 0x00000006
  GRBM_STATUS_SE1           = 0x00000006
  SRBM_STATUS               = 0x200000C0
  SRBM_STATUS2              = 0x00000000
  R_008674_CP_STALLED_STAT1 = 0x00000000
  R_008678_CP_STALLED_STAT2 = 0x00000000
  R_00867C_CP_BUSY_STAT     = 0x00000000
  R_008680_CP_STAT          = 0x00000000
  R_00D034_DMA_STATUS_REG   = 0x44C83D57
  R_00D834_DMA_STATUS_REG   = 0x44C83D57
GPU reset succeeded, trying to resume


I'll leave this bug report open, because I have no idea if this bug will reappear once the hang is fixed.
Comment 2 Michel Dänzer 2014-08-01 14:38:16 UTC 
(In reply to comment #1)
> Many demos work now. But the Cave Demo does not work, it hangs the GPU.

Hmm, that one works for me. Only the Elemental and Sci-Fi Hallway demos run into shader compiler failures and corresponding GPU VM faults for me, no hangs though.


> I'll leave this bug report open, because I have no idea if this bug will
> reappear once the hang is fixed.

Right, they're probably separate problems and need to be tracked separately.
Comment 3 Christoph Haag 2014-08-01 15:31:50 UTC 
Oh, the hang in comment 1 may have been my fault for running it with hyperz.

But without hyperz it hangs too, just less verbosely:

radeon 0000:01:00.0: ring 3 stalled for more than 10016msec
radeon 0000:01:00.0: GPU lockup (waiting for 0x0000000000011c4a last fence id 0x0000000000011b2a on ring 3)
[TTM] Buffer eviction failed
Comment 4 Christoph Haag 2014-08-13 10:25:25 UTC 
This here doesn't happen anymore with recent mesa git builds either.

Cave Effects demo now runs on my GPU.

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.