Skip to content

[DG2] Running 'hdr' vulkan demo causes "RIP: dma_fence_array_create" trace in dmesg

  • Platform: Intel ARC A750 (DG2)
  • Kernel: 28f3fbf3ad9cc809521c7a308e29dfc27c445730 (drm-tip: 2022y-11m-09d-17h-31m-37s UTC integration manifest)
  • Mesa: 22.2.3
  • xorg-server: 21.1.4
  • GNOME Shell: 43.1 (either X or Wayland session, doesn't matter)
  • Monitor: 1920x1080

I built Sascha Willems' Vulkan demos from here: https://github.com/SaschaWillems/Vulkan

Running

$ ./bin/hdr -w 2560 -h 1440

causes this RIP error to appear in dmesg:

[  672.799397] RIP: 0010:dma_fence_array_create+0xdd/0x100
[  672.799404] Code: 45 85 e4 75 10 eb 2a 48 81 fa c0 cb 2e b6 74 1a 83 e8 01 72 1c 48 63 d0 48 8b 54 d5 00 48 8b 52 08 48 81 fa 60 cb 2e b6 75 dd <0f> 0b 83 e8 01 73 e4 48 89 d8 5b 5d 41 5c 41 5d 41 5e 41 5f c3 cc
[  672.799408] RSP: 0018:ffffb4b202bdfda8 EFLAGS: 00010246
[  672.799412] RAX: 0000000000000001 RBX: ffff9c0be7136780 RCX: 00000000000031e3
[  672.799415] RDX: ffffffffb62ecb60 RSI: ffff9c0be7136790 RDI: ffff9c0be7136780
[  672.799417] RBP: ffff9c0b026196c0 R08: 0000000000000001 R09: ffff9c0be7136780
[  672.799420] R10: ffffb4b202bdfe40 R11: 0000000000000001 R12: 0000000000000002
[  672.799422] R13: 0000000000000000 R14: 00000000000031e3 R15: 0000000000000001
[  672.799425] FS:  00007fe6ceae2340(0000) GS:ffff9c0eaf880000(0000) knlGS:0000000000000000
[  672.799428] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  672.799431] CR2: 00007fe6ce4ad00c CR3: 000000017572a001 CR4: 0000000000770ee0
[  672.799434] PKRU: 55555554
[  672.799435] Call Trace:
[  672.799438]  <TASK>
[  672.799443]  dma_resv_get_singleton+0x68/0x130
[  672.799453]  dma_buf_ioctl+0x123/0x370
[  672.799460]  __x64_sys_ioctl+0x8d/0xd0
[  672.799469]  do_syscall_64+0x58/0x80
[  672.799478]  ? syscall_exit_to_user_mode_prepare+0x18e/0x1c0
[  672.799485]  ? syscall_exit_to_user_mode+0x12/0x30
[  672.799488]  ? do_syscall_64+0x67/0x80
[  672.799494]  ? syscall_exit_to_user_mode_prepare+0x18e/0x1c0
[  672.799499]  ? syscall_exit_to_user_mode+0x12/0x30
[  672.799502]  ? do_syscall_64+0x67/0x80
[  672.799508]  ? syscall_exit_to_user_mode+0x12/0x30
[  672.799511]  ? do_syscall_64+0x67/0x80
[  672.799517]  ? do_syscall_64+0x67/0x80
[  672.799524]  entry_SYSCALL_64_after_hwframe+0x63/0xcd
[  672.799530] RIP: 0033:0x7fe6ce62dc0f
[  672.799535] Code: 00 48 89 44 24 18 31 c0 48 8d 44 24 60 c7 04 24 10 00 00 00 48 89 44 24 08 48 8d 44 24 20 48 89 44 24 10 b8 10 00 00 00 0f 05 <89> c2 3d 00 f0 ff ff 77 18 48 8b 44 24 18 64 48 2b 04 25 28 00 00
[  672.799538] RSP: 002b:00007ffc96a3ed80 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
[  672.799543] RAX: ffffffffffffffda RBX: 0000559cc72b66c0 RCX: 00007fe6ce62dc0f
[  672.799545] RDX: 00007ffc96a3ee10 RSI: 00000000c0086202 RDI: 000000000000000e
[  672.799547] RBP: 00007ffc96a3ee10 R08: 0000000000000000 R09: 0000000000000000
[  672.799550] R10: 0000559cc7549210 R11: 0000000000000246 R12: 00000000c0086202
[  672.799552] R13: 000000000000000e R14: 0000000000000000 R15: 00007ffc96a3ee10
[  672.799558]  </TASK>
[  672.799560] ---[ end trace 0000000000000000 ]---

Presumably that shouldn't be happening.

The window size seems to matter. -w 2560 -h 1440 is larger than my 1920x1080 monitor. Small sizes (-w 800 -h 600) don't seem to trigger this error.

To upload designs, you'll need to enable LFS and have an admin enable hashed storage. More information