Arch Linux

IMBJR

Member

From: INDCMPLX

Registered: 2014-03-16

Posts: 88

Screen lock during blender/minecraft sessions

For a few weeks now, I've been getting screen freezes (but with the mouse cursor unaffected) especially when either playing minecraft or using blender. Today, after setting up ssh, I've finally managed to get some evidence as to what is happening:

[19660.603798] INFO: task amdgpu_cs:0:529 blocked for more than 120 seconds.
[19660.603808]       Tainted: G           O    4.10.5-1-ARCH #1
[19660.603811] "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
[19660.603816] amdgpu_cs:0     D    0   529    503 0x00000000
[19660.603821] Call Trace:
[19660.603834]  __schedule+0x22f/0x700
[19660.603840]  schedule+0x3d/0x90
[19660.603846]  schedule_timeout+0x243/0x3d0
[19660.603851]  ? __wake_up_locked+0x13/0x20
[19660.603896]  ? amdgpu_sa_bo_free+0x5c/0x110 [amdgpu]
[19660.603902]  dma_fence_default_wait+0x1db/0x260
[19660.603906]  ? dma_fence_default_wait+0x260/0x260
[19660.603909]  dma_fence_wait_timeout+0x39/0x120
[19660.603941]  amdgpu_ctx_add_fence+0x66/0x1c0 [amdgpu]
[19660.603969]  amdgpu_cs_ioctl+0x106a/0x1480 [amdgpu]
[19660.603983]  drm_ioctl+0x21b/0x4c0 [drm]
[19660.604008]  ? amdgpu_cs_find_mapping+0xa0/0xa0 [amdgpu]
[19660.604012]  ? do_futex+0x32b/0xbd0
[19660.604032]  amdgpu_drm_ioctl+0x4f/0x90 [amdgpu]
[19660.604038]  do_vfs_ioctl+0xa3/0x5f0
[19660.604042]  ? __fget+0x77/0xb0
[19660.604046]  SyS_ioctl+0x79/0x90
[19660.604050]  entry_SYSCALL_64_fastpath+0x1a/0xa9
[19660.604053] RIP: 0033:0x7f17599080d7
[19660.604055] RSP: 002b:00007f17502df9f8 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
[19660.604058] RAX: ffffffffffffffda RBX: 0000000001e1f010 RCX: 00007f17599080d7
[19660.604060] RDX: 00007f17502dfbb0 RSI: 00000000c0186444 RDI: 000000000000000e
[19660.604062] RBP: 00007f17502dfbb0 R08: 0000000000ff9088 R09: 0000000000000000
[19660.604064] R10: 0000000000000000 R11: 0000000000000246 R12: 0000000040086409
[19660.604065] R13: 000000000000000e R14: 0000000000f80660 R15: 0000000000f80630

This all seems to have started after recent firmware updates, in which I was consistently getting ring/ib test failures at boot but which now are intermittant. When these occur, I typically reboot.

Any suggestions on tackling this? Searching around seems to reveal very little in the way of information.