Created attachment 107916 [details] dmesg ==System Environment== -------------------------- Regression: Yes. It's a igt regression bug Good commit: 2faf5a8665bba9ab7ad64ac6aa8c13be8f176d7d Bud commit: 303fe742fcdd7da6f03f948a4e41b529ee39613a Non-working platforms: BYT ==kernel== -------------------------- origin/drm-intel-nightly: 2ea23cd593ba60ead60e2f796fae675aa4475b1a(fails) drm-intel-nightly: 2014y-10m-13d-09h-34m-29s UTC integration manifest origin/drm-intel-next-queued: fff95501f2095fa8a9dbd7e177bbf642d241b2fe(fails) drm/i915: intel_backlight scale() math WA origin/drm-intel-fixes: bfe01a5ba2490f299e1d2d5508cbbbadd897bbe9(fails) Linux 3.17 ==Bug detailed description== igt/gem_userptr_blits subcases cause "WARNING: CPU: 0 PID: 4681 at drivers/gpu/drm/i915/i915_gem_userptr.c:86 cancel_userptr+0xb6/0xf8 [i915]()" Case list: igt/gem_userptr_blits/forked-sync-swapping-multifd-mempressure-interruptible igt/gem_userptr_blits/forked-sync-swapping-multifd-mempressure-normal Output: "root@x-bytm02:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests# ./gem_userptr_blits --run-subtest forked-sync-swapping-multifd-mempressure-interruptible IGT-Version: 1.8-ga0b5c6d (x86_64) (Linux: 3.17.0_drm-intel-nightly_2ea23c_20141016+ x86_64) Aperture size is 2048 MiB Total RAM is 3837 MiB Testing unsynchronized mappings... Testing synchronized mappings... Subtest forked-sync-swapping-multifd-mempressure-interruptible: SUCCESS (97.534s) root@x-bytm02:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests# dmesg -r|egrep ""<[1-4]>""|grep drm <4>[ 346.029662] WARNING: CPU: 3 PID: 4365 at drivers/gpu/drm/i915/i915_gem_userptr.c:86 cancel_userptr+0xb6/0xf8 [i915]() <4>[ 346.029665] Modules linked in: dm_mod snd_hda_codec_hdmi snd_hda_codec_conexant snd_hda_codec_generic iTCO_wdt iTCO_vendor_support snd_hda_intel snd_hda_controller pcspkr snd_hda_codec snd_hwdep snd_pcm r8169 i2c_i801 snd_timer snd lpc_ich soundcore mfd_core iosf_mbi battery wmi ac acpi_cpufreq uvcvideo videobuf2_vmalloc videobuf2_memops videobuf2_core v4l2_common videodev i915 button video drm_kms_helper drm cfbfillrect cfbimgblt cfbcopyarea <4>[ 346.029704] CPU: 3 PID: 4365 Comm: gem_userptr_bli Not tainted 3.17.0_drm-intel-nightly_2ea23c_20141016+ #550" ==Reproduce steps== ---------------------------- 1. ./gem_userptr_blits --run-subtest forked-sync-swapping-multifd-mempressure-interruptible
Would you believe this is fixed by my s/seqno/requests/? It is! The failure here is that we fail to unbind because of the GPU reset.
(In reply to Chris Wilson from comment #1) > Would you believe this is fixed by my s/seqno/requests/? It is! > > The failure here is that we fail to unbind because of the GPU reset. Test passes on latest -nightly(0642a51748ff46a205cf2a3fc45ba18cc92c9bda) root@x-bytm02:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests# ./gem_userptr_blits --run-subtest forked-sync-swapping-multifd-mempressure-interruptible IGT-Version: 1.8-ga973aab (x86_64) (Linux: 3.18.0-rc3_drm-intel-nightly_0642a5_20141106+ x86_64) Aperture size is 2048 MiB Total RAM is 3837 MiB Testing unsynchronized mappings... Testing synchronized mappings... Subtest forked-sync-swapping-multifd-mempressure-interruptible: SUCCESS (44.070s) root@x-bytm02:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests# dmesg -r|egrep "<[1-4]>"|grep drm root@x-bytm02:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests#
The handling of the EIO was not fixed, you just stopped triggering the GPU hang.
Closing old verified+worksforme.
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.