From cooper.chiou at intel.com Tue Feb 9 03:21:58 2021 From: cooper.chiou at intel.com (Chiou, Cooper) Date: Tue, 9 Feb 2021 03:21:58 +0000 Subject: [I915-ci-infra] i915 upstream patch got unknown serious failure in i915-CI report Message-ID: Hi i915-CI team, This is Cooper from CCG/CCPE Chrome Graphics team, I?ve submitted i915 patch to resolve Gen9 Pentium/Celeron sku VP8 hardware encoding GPU hang issue in (https://patchwork.freedesktop.org/series/81764/), but Intel-gfx CI log mentioned there is serious unknown failure listed in https://www.mail-archive.com/intel-gfx at lists.freedesktop.org/msg240492.html. Series: drm/i915: Enable WaProgramMgsrForCorrectSliceSpecificMmioReads for Gen9 URL : https://patchwork.freedesktop.org/series/81764/ State : failure External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_18518/index.html I can?t access https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_18518/index.html to get more detail information to understand what failure is. Need your help to share more detail on this unknown failure. If it?s regarding to ?phoronix-test-suite? performance drop, Intel kernel validation team has clarified there was no any performance regression after refined test environment here (https://patchwork.freedesktop.org/patch/390869/?series=81764&rev=1#comment_754336) Since this patch is blocking Chrome VP8 HW encoding feature enabling on all KBL/GLK/CML Pentium/Celeron Chrome projects and did impact Chrome projects video call performance, now Google team is escalating to get more clear information on this patch CI failure reason and want to know why this patch isn?t merged in upstream as well. Please help on this urgent. Thanks, andrescj at google.com> #108Feb 9, 2021 03:33AM Cooper. Thanks for following up on that! Some CI failures (not performance regressions) in ?here. I'm unable to open the links with the claimed failures (access denied), so I can't tell if they're serious. Also, ff there are no regressions of any kind, is there any reason the patch hasn't been merged upstream? [Intel-gfx] ? Fi.CI.BAT: failure for drm/i915: Enable WaProgramMgsrForCorrectSliceSpecificMmioReads for Gen9 Patchwork Wed, 16 Sep 2020 13:00:10 -0700 == Series Details == Series: drm/i915: Enable WaProgramMgsrForCorrectSliceSpecificMmioReads for Gen9 URL : https://patchwork.freedesktop.org/series/81764/ State : failure == Summary == CI Bug Log - changes from CI_DRM_9019 -> Patchwork_18518 ==================================================== Summary ------- **FAILURE** Serious unknown changes coming with Patchwork_18518 absolutely need to be verified manually. If you think the reported changes have nothing to do with the changes introduced in Patchwork_18518, please notify your bug team to allow them to document this new failure mode, which will reduce false positives in CI. External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_18518/index.html --> Access Denied Best Regards, Cooper +886-2-6622-1166 -------------- next part -------------- An HTML attachment was scrubbed... URL: From petri.latvala at intel.com Tue Feb 9 06:45:31 2021 From: petri.latvala at intel.com (Petri Latvala) Date: Tue, 9 Feb 2021 08:45:31 +0200 Subject: [I915-ci-infra] i915 upstream patch got unknown serious failure in i915-CI report In-Reply-To: References: Message-ID: On Tue, Feb 09, 2021 at 03:21:58AM +0000, Chiou, Cooper wrote: > Hi i915-CI team, > > This is Cooper from CCG/CCPE Chrome Graphics team, I?ve submitted i915 patch to resolve Gen9 Pentium/Celeron sku VP8 hardware encoding GPU hang issue in (https://patchwork.freedesktop.org/series/81764/), but Intel-gfx CI log mentioned there is serious unknown failure listed in https://www.mail-archive.com/intel-gfx at lists.freedesktop.org/msg240492.html. > > > Series: drm/i915: Enable WaProgramMgsrForCorrectSliceSpecificMmioReads for Gen9 > > URL : https://patchwork.freedesktop.org/series/81764/ > > State : failure > > External URL: > > https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_18518/index.html > > > > I can?t access https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_18518/index.html to get more detail information to understand what failure is. Need your help to share more detail on this unknown failure. > Hi Cooper The regressions hit by the patch were these parts of the mail: Possible regressions igt at i915_pm_rpm@module-reload: fi-cfl-8109u: PASS -> DMESG-WARN +35 similar issues igt at i915_selftest@live at blt: fi-snb-2520m: PASS -> DMESG-FAIL igt at i915_selftest@live at gem_contexts: fi-skl-lmem: PASS -> DMESG-WARN The full logs and full results are no longer available, because we only keep premerge results around for a couple of weeks. I've queued the patch for a retest to get more up-to-date information. -- Petri Latvala