[Bug 93652] Random crashes/freezing with amdgpu Fury X mesa 11.1

[bugzilla-daemon@freedesktop.org]

[-- Attachment #1.1: Type: text/plain, Size: 4448 bytes --]

https://bugs.freedesktop.org/show_bug.cgi?id=93652

--- Comment #19 from Adam Bolte <abolte@systemsaviour.com> ---
Updated to a stable Linux 4.11.0 kernel, with the LED patch that's still not
upstream (from https://bugs.freedesktop.org/show_bug.cgi?id=97590). Says
"Tainted" due to using the vhba (Virtual SCSI HBA, built via dkms) kernel
module from the CDEmu project. All config options are typical - built using
`make olddefconfig` on an up to date Debian Stretch install.

[    0.000000] Linux version 4.11.0+ (root@dragon) (gcc version 6.3.0 20170425
(Debian 6.3.0-16) ) #2 SMP Thu May 18 22:40:08 AEST 2017
[28034.309104] INFO: task Xorg:1278 blocked for more than 120 seconds.
[28034.309106]       Tainted: G           OE   4.11.0+ #2
[28034.309107] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this
message.
[28034.309107] Xorg            D    0  1278   1239 0x00000004
[28034.309108] Call Trace:
[28034.309111]  ? __schedule+0x3c2/0x8b0
[28034.309112]  ? __kfifo_in+0x2d/0x40
[28034.309113]  ? schedule+0x32/0x80
[28034.309124]  ? amd_sched_entity_push_job+0xb9/0x100 [amdgpu]
[28034.309125]  ? remove_wait_queue+0x60/0x60
[28034.309133]  ? amdgpu_job_submit+0x6e/0x90 [amdgpu]
[28034.309140]  ? amdgpu_vm_bo_split_mapping+0x510/0x6f0 [amdgpu]
[28034.309146]  ? amdgpu_vm_do_copy_ptes+0x90/0x90 [amdgpu]
[28034.309152]  ? amdgpu_vm_clear_freed+0x75/0xb0 [amdgpu]
[28034.309158]  ? amdgpu_gem_va_ioctl+0x3a6/0x400 [amdgpu]
[28034.309159]  ? __radix_tree_delete+0x30/0xa0
[28034.309160]  ? __check_object_size+0xfb/0x196
[28034.309164]  ? drm_ioctl+0x1ef/0x440 [drm]
[28034.309167]  ? drm_ioctl+0x1ef/0x440 [drm]
[28034.309173]  ? amdgpu_gem_metadata_ioctl+0x1c0/0x1c0 [amdgpu]
[28034.309174]  ? unmap_region+0xd9/0x120
[28034.309178]  ? amdgpu_drm_ioctl+0x49/0x80 [amdgpu]
[28034.309179]  ? do_vfs_ioctl+0x9f/0x600
[28034.309180]  ? do_munmap+0x356/0x470
[28034.309181]  ? SyS_ioctl+0x74/0x80
[28034.309181]  ? entry_SYSCALL_64_fastpath+0x1e/0xad
[28155.142288] INFO: task Xorg:1278 blocked for more than 120 seconds.
[28155.142290]       Tainted: G           OE   4.11.0+ #2
[28155.142290] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this
message.
[28155.142291] Xorg            D    0  1278   1239 0x00000004
[28155.142292] Call Trace:
[28155.142295]  ? __schedule+0x3c2/0x8b0
[28155.142296]  ? __kfifo_in+0x2d/0x40
[28155.142297]  ? schedule+0x32/0x80
[28155.142308]  ? amd_sched_entity_push_job+0xb9/0x100 [amdgpu]
[28155.142309]  ? remove_wait_queue+0x60/0x60
[28155.142317]  ? amdgpu_job_submit+0x6e/0x90 [amdgpu]
[28155.142324]  ? amdgpu_vm_bo_split_mapping+0x510/0x6f0 [amdgpu]
[28155.142330]  ? amdgpu_vm_do_copy_ptes+0x90/0x90 [amdgpu]
[28155.142336]  ? amdgpu_vm_clear_freed+0x75/0xb0 [amdgpu]
[28155.142342]  ? amdgpu_gem_va_ioctl+0x3a6/0x400 [amdgpu]
[28155.142343]  ? __radix_tree_delete+0x30/0xa0
[28155.142344]  ? __check_object_size+0xfb/0x196
[28155.142349]  ? drm_ioctl+0x1ef/0x440 [drm]
[28155.142352]  ? drm_ioctl+0x1ef/0x440 [drm]
[28155.142357]  ? amdgpu_gem_metadata_ioctl+0x1c0/0x1c0 [amdgpu]
[28155.142359]  ? unmap_region+0xd9/0x120
[28155.142363]  ? amdgpu_drm_ioctl+0x49/0x80 [amdgpu]
[28155.142364]  ? do_vfs_ioctl+0x9f/0x600
[28155.142365]  ? do_munmap+0x356/0x470
[28155.142366]  ? SyS_ioctl+0x74/0x80
[28155.142366]  ? entry_SYSCALL_64_fastpath+0x1e/0xad
(END)

This happened while playing The Darkness II (a d3d9 game) under a patched (for
Darkness II compatibility - https://bugs.winehq.org/attachment.cgi?id=54454)
Wine 2.8. This was *not* using Gallium on Nine. The screen froze up (I couldn't
alt-tab or switch virtual desktops) so I had to SSH in from a laptop to power
off safely.

Device: AMD FIJI (DRM 3.10.0 / 4.11.0+, LLVM 4.0.1) (0x7300)
Version: 17.2.0
OpenGL renderer string: Gallium 0.4 on AMD FIJI (DRM 3.10.0 / 4.11.0+, LLVM
4.0.1)
OpenGL core profile version string: 4.5 (Core Profile) Mesa 17.2.0-devel
(git-3db35a8bb9)
OpenGL core profile shading language version string: 4.50

Happened while running Mesa master (currently
f347bac30f4045a9583f95a5776484b1a2947183, Thu May 18 10:21:59 2017 +0300), with
the Dying Light patch set from
https://lists.freedesktop.org/archives/mesa-dev/2017-May/155428.html on top.

I only observed a crash just the one time during the entire ~6h play-through.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[-- Attachment #1.2: Type: text/html, Size: 5572 bytes --]

[-- Attachment #2: Type: text/plain, Size: 160 bytes --]

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel