Happens on both of the skl CI machines we have: skl-i5k-2 and skl-i7k-2 For some strange reason only one testcase blows up though: igt@kms_pipe_crc_basic@read-crc-pipe-c-frame-sequence Full dmesg backtrace: [ 95.642367] ------------[ cut here ]------------ [ 95.642385] WARNING: CPU: 7 PID: 5661 at drivers/gpu/drm/i915/intel_display.c:13778 skl_max_scale.part.87+0x67/0x70 [i915]() [ 95.642386] WARN_ON_ONCE(!crtc_clock || cdclk < crtc_clock) [ 95.642387] Modules linked in: [ 95.642388] snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_codec_generic i915 x86_pkg_temp_thermal snd_hda_intel intel_powerclamp coretemp snd_hda_codec crct10dif_pclmul crc32_pclmul snd_hwdep snd_hda_core snd_pcm mei_me mei i2c_hid e1000e ptp pps_core [ 95.642401] CPU: 7 PID: 5661 Comm: kms_pipe_crc_ba Tainted: G U 4.4.0-rc4-gfxbench+ #1 [ 95.642402] Hardware name: System manufacturer System Product Name/Z170M-PLUS, BIOS 0408 08/28/2015 [ 95.642403] ffffffffa029aab0 ffff880243c3fa40 ffffffff813df80c ffff880243c3fa88 [ 95.642405] ffff880243c3fa78 ffffffff810747c1 0000000000010000 ffff8800a51f8548 [ 95.642408] ffff8800a3291148 ffff8800a51f8548 ffffffffa027c620 ffff880243c3fad8 [ 95.642410] Call Trace: [ 95.642413] [<ffffffff813df80c>] dump_stack+0x4e/0x82 [ 95.642416] [<ffffffff810747c1>] warn_slowpath_common+0x81/0xc0 [ 95.642417] [<ffffffff81074847>] warn_slowpath_fmt+0x47/0x50 [ 95.642430] [<ffffffffa02238a7>] skl_max_scale.part.87+0x67/0x70 [i915] [ 95.642441] [<ffffffffa0223958>] intel_check_primary_plane+0xa8/0xc0 [i915] [ 95.642452] [<ffffffffa0215ab9>] intel_plane_atomic_check+0x129/0x1f0 [i915] [ 95.642455] [<ffffffff814ea153>] drm_atomic_helper_check_planes+0x43/0x1d0 [ 95.642466] [<ffffffffa022f0d5>] intel_atomic_check+0x2f5/0x1090 [i915] [ 95.642469] [<ffffffff815001db>] ? drm_property_unreference_blob+0x4b/0xf0 [ 95.642471] [<ffffffff81500101>] ? drm_property_free_blob+0x21/0x60 [ 95.642473] [<ffffffff8150dc72>] drm_atomic_check_only+0x142/0x5e0 [ 95.642474] [<ffffffff8150d7c0>] ? drm_atomic_set_mode_prop_for_crtc+0x20/0x110 [ 95.642476] [<ffffffff8150e122>] drm_atomic_commit+0x12/0x50 [ 95.642478] [<ffffffff814ead55>] drm_atomic_helper_set_config+0x75/0xb0 [ 95.642479] [<ffffffff814fcc60>] drm_mode_set_config_internal+0x60/0x110 [ 95.642481] [<ffffffff81501876>] drm_mode_setcrtc+0x186/0x4e0 [ 95.642483] [<ffffffff814f3b5d>] drm_ioctl+0x13d/0x590 [ 95.642485] [<ffffffff815016f0>] ? drm_mode_setplane+0x1b0/0x1b0 [ 95.642487] [<ffffffff811d4eac>] do_vfs_ioctl+0x2fc/0x550 [ 95.642489] [<ffffffff8118d71a>] ? vm_munmap+0x4a/0x60 [ 95.642490] [<ffffffff811e091a>] ? __fget_light+0x6a/0x90 [ 95.642492] [<ffffffff811d513c>] SyS_ioctl+0x3c/0x70 [ 95.642494] [<ffffffff8179a5db>] entry_SYSCALL_64_fastpath+0x16/0x73 [ 95.642495] ---[ end trace 7f9751c10d528f7e ]---
Fixed by "[PATCH 1/4] drm/i915/skl: Do not allow scaling when crtc is disabled." from cdclk fixes series in case of crtc disable. For DPMS off additionally needed are "[PATCH 3/4] drm/i915: Do not acquire crtc state to check clock during modeset, v3." and "[PATCH 11/12] drm/i915: Keep track of the cdclk as if all crtc's were active." + previous patches to make it apply.
Since CI build CI_DRM_895 it jumped around, and now it's happening on pipe A/B and no longer on pipe C. Since we have patches for this review is the only blocker. Mika, can you please take a look at this?
Bug scrub: Set it to resolved as fixed.
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.