Last seen: CI_DRM_3852: 2018-02-28 / 224 runs ago

Happened again:

https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_4108/fi-bsw-n3050/igt@gem_sync@basic-each.html

(gem_sync:3107) igt_aux-CRITICAL: Test assertion failure function sig_abort, file ../lib/igt_aux.c:481:
(gem_sync:3107) igt_aux-CRITICAL: Failed assertion: !"GPU hung"


https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_4108/fi-bsw-n3050/igt@kms_pipe_crc_basic@hang-read-crc-pipe-c.html

(kms_pipe_crc_basic:3373) igt_gt-CRITICAL: Test assertion failure function igt_force_gpu_reset, file ../lib/igt_gt.c:415:
(kms_pipe_crc_basic:3373) igt_gt-CRITICAL: Failed assertion: !wedged
Subtest hang-read-crc-pipe-C failed.


https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_4108/fi-bsw-n3050/igt@kms_pipe_crc_basic@suspend-read-crc-pipe-c.html

[  648.484962] ------------[ cut here ]------------
[  648.484973] WARN_ON(intel_gpu_reset(i915, (~0)))
[  648.485230] WARNING: CPU: 1 PID: 2921 at drivers/gpu/drm/i915/i915_gem.c:4978 i915_gem_sanitize+0x4d/0x80 [i915]
[  648.485242] Modules linked in: vgem btusb btrtl btbcm btintel bluetooth ecdh_generic intel_powerclamp coretemp crct10dif_pclmul crc32_pclmul ghash_clmulni_intel r8169 mii lpc_ich snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic i915 snd_hda_intel snd_hda_codec snd_hwdep snd_hda_core snd_pcm pinctrl_cherryview prime_numbers
[  648.485501] CPU: 1 PID: 2921 Comm: kworker/u4:31 Tainted: G     U  W         4.17.0-rc2-CI-CI_DRM_4108+ #1
[  648.485512] Hardware name:  /NUC5CPYB, BIOS PYBSWCEL.86A.0058.2016.1102.1842 11/02/2016
[  648.485528] Workqueue: events_unbound async_run_entry_fn
[  648.485716] RIP: 0010:i915_gem_sanitize+0x4d/0x80 [i915]
[  648.485728] RSP: 0018:ffffc90000aebcc8 EFLAGS: 00010286
[  648.485749] RAX: 0000000000000000 RBX: ffff88016c890000 RCX: 0000000000000001
[  648.485760] RDX: 0000000080000001 RSI: ffffffff82121491 RDI: 00000000ffffffff
[  648.485770] RBP: ffff88016c890070 R08: 0000000075a89f58 R09: 0000000000000000
[  648.485781] R10: 0000000000000000 R11: 0000000000000000 R12: ffff88016c8989b8
[  648.485791] R13: 0000000000000000 R14: 0000000000000000 R15: 0000000000000000
[  648.485803] FS:  0000000000000000(0000) GS:ffff88017fd00000(0000) knlGS:0000000000000000
[  648.485814] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  648.485826] CR2: 00007fec90dd73d8 CR3: 0000000005210000 CR4: 00000000001006e0
[  648.485836] Call Trace:
[  648.486029]  i915_gem_suspend+0xec/0x140 [i915]
[  648.486203]  i915_drm_suspend+0x5f/0x160 [i915]
[  648.486231]  pci_pm_suspend+0x7c/0x130
[  648.486252]  ? pci_pm_freeze+0xb0/0xb0
[  648.486267]  dpm_run_callback+0x5d/0x2f0
[  648.486295]  __device_suspend+0x11f/0x600
[  648.486322]  ? dpm_watchdog_set+0x60/0x60
[  648.486364]  async_suspend+0x15/0x90
[  648.486382]  async_run_entry_fn+0x34/0x160
[  648.486408]  process_one_work+0x229/0x6a0
[  648.486453]  worker_thread+0x35/0x380
[  648.486482]  ? process_one_work+0x6a0/0x6a0
[  648.486495]  kthread+0x119/0x130
[  648.486511]  ? _kthread_create_on_node+0x60/0x60
[  648.486537]  ret_from_fork+0x3a/0x50
[  648.486597] Code: dc 03 00 84 c0 74 f1 be ff ff ff ff 48 89 df e8 0a da 03 00 85 c0 74 e0 48 c7 c6 00 f8 20 a0 48 c7 c7 e2 76 1f a0 e8 a3 71 f9 e0 <0f> 0b eb c9 48 8d 6f 70 31 f6 48 89 ef e8 51 d1 83 e1 48 89 df 
[  648.487196] irq event stamp: 1160
[  648.487212] hardirqs last  enabled at (1159): [<ffffffff810f8827>] vprintk_emit+0x4b7/0x4d0
[  648.487226] hardirqs last disabled at (1160): [<ffffffff81a0111c>] error_entry+0x7c/0x100
[  648.487239] softirqs last  enabled at (1132): [<ffffffff81c0032b>] __do_softirq+0x32b/0x4e1
[  648.487254] softirqs last disabled at (1111): [<ffffffff8108b884>] irq_exit+0xa4/0xb0
[  648.487434] WARNING: CPU: 1 PID: 2921 at drivers/gpu/drm/i915/i915_gem.c:4978 i915_gem_sanitize+0x4d/0x80 [i915]
[  648.487445] ---[ end trace 46a07151b792d058 ]---
[  649.548697] [drm:vlv_suspend_complete [i915]] *ERROR* timeout waiting for GT wells to go off

Come back in 6 months to see if this is gone for good!

commit 5692251c254a3d561316c4e8e10c77e470b60658
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Sat Apr 28 12:15:32 2018 +0100

    drm/i915/lrc: Scrub the GPU state of the guilty hanging request
    
    Previously, we just reset the ring register in the context image such
    that we could skip over the broken batch and emit the closing
    breadcrumb. However, on resume the context image and GPU state would be
    reloaded, which may have been left in an inconsistent state by the
    reset. The presumption was that at worst it would just cause another
    reset and skip again until it recovered, however it seems just as likely
    to cause an unrecoverable hang. Instead of risking loading an incomplete
    context image, restore it back to the default state.
    
    v2: Fix up off-by-one from including the ppHSWP in with the register
    state.
    v3: Use a ring local to compact a few lines.
    v4: Beware setting the ring local before checking for a NULL request.

Closing, thanks.