CHECK() failed in ui::HardwareDisplayController::SchedulePageFlip() in Youtube |
||||||||||||||||||||||
Issue descriptionThis regression happens between commit c16e7e464ad15925ee652fffd9eed8f6c355b1be Author: Dmitry Gozman <dgozman@chromium.org> Date: Wed Oct 17 00:52:25 2018 +0000 and commit b430c9d0c83a00fef1ecdefd8c0d83f935c68b24 Author: Tom Finegan <tomfinegan@google.com> Date: Thu Oct 18 00:59:19 2018 +0000 After regression, Youtube hardware overlay is broken with following error logs. Users cannot watch Youtube. I'm testing it on octopus. [3518:3518:1018/170622.848220:ERROR:gpu_interface_provider.cc(87)] Not implemented reached in virtual void content::GpuInterfaceProvider::RegisterOzoneGpuInterfaces(service_manager::BinderRegistry *) [3518:3540:1018/170622.926980:ERROR:service_manager_connection_impl.cc(318)] Can't create service multidevice_setup. No handler found. [3561:3573:1018/170628.924107:ERROR:hardware_display_plane_manager_atomic.cc(134)] Failed to commit properties for page flip.: No space left on device (28) [3561:3573:1018/170628.924211:FATAL:hardware_display_controller.cc(121)] Check failed: status. SchedulePageFlip failed #0 0x59883e3f038f base::debug::StackTrace::StackTrace() #1 0x59883e36fd70 logging::LogMessage::~LogMessage() #2 0x59883c26d566 ui::HardwareDisplayController::SchedulePageFlip() #3 0x59883c2682e4 ui::DrmWindow::SchedulePageFlip() #4 0x59883c25f94b ui::DrmThread::OnPlanesReadyForPageFlip() #5 0x59883c260c97 _ZN4base8internal13FunctorTraitsIMN2ui9DrmThreadEFviNS_12OnceCallbackIFvN3gfx10SwapResultENSt3__110unique_ptrINS5_8GpuFenceENS7_14default_deleteIS9_EEEEEEENS4_IFvRKNS5_20PresentationFeedbackEEEENS7_6vectorINS2_15DrmOverlayPlaneENS7_9allocatorISL_EEEEEvE6InvokeISQ_NS_7WeakPtrIS3_EEJiSE_SJ_SO_EEEvT_OT0_DpOT1_ #6 0x59883c260bd7 _ZN4base8internal7InvokerINS0_9BindStateIMN2ui9DrmThreadEFviNS_12OnceCallbackIFvN3gfx10SwapResultENSt3__110unique_ptrINS6_8GpuFenceENS8_14default_deleteISA_EEEEEEENS5_IFvRKNS6_20PresentationFeedbackEEEENS8_6vectorINS3_15DrmOverlayPlaneENS8_9allocatorISM_EEEEEJNS_7WeakPtrIS4_EEiSF_SK_EEEFvSP_EE7RunOnceEPNS0_13BindStateBaseEOSP_ #7 0x59883c2749d1 _ZN4base8internal7InvokerINS0_9BindStateINS_12OnceCallbackIFvNSt3__16vectorIN2ui15DrmOverlayPlaneENS4_9allocatorIS7_EEEEEEEJSA_EEEFvvEE7RunOnceEPNS0_13BindStateBaseE #8 0x59883e406d91 base::debug::TaskAnnotator::RunTask() #9 0x59883e37590f base::MessageLoop::RunTask() #10 0x59883e375cb2 base::MessageLoop::DoWork() #11 0x59883e403999 base::MessagePumpLibevent::Run() #12 0x59883e395e45 base::RunLoop::Run() #13 0x59883e3ca643 base::Thread::ThreadMain() #14 0x59883e3fdb4d base::(anonymous namespace)::ThreadFunc() #15 0x7f1d6b61a2b8 <unknown> #16 0x7f1d6aadbfad clone Received signal 6 #0 0x59883e3f038f base::debug::StackTrace::StackTrace() #1 0x59883e3eff01 base::debug::(anonymous namespace)::StackDumpSignalHandler() #2 0x7f1d6b6242e0 <unknown> #3 0x7f1d6aa18dd2 gsignal #4 0x7f1d6aa1abf6 abort #5 0x59883e3eecf5 base::debug::BreakDebugger() #6 0x59883e3701f1 logging::LogMessage::~LogMessage() #7 0x59883c26d566 ui::HardwareDisplayController::SchedulePageFlip() #8 0x59883c2682e4 ui::DrmWindow::SchedulePageFlip() #9 0x59883c25f94b ui::DrmThread::OnPlanesReadyForPageFlip() #10 0x59883c260c97 _ZN4base8internal13FunctorTraitsIMN2ui9DrmThreadEFviNS_12OnceCallbackIFvN3gfx10SwapResultENSt3__110unique_ptrINS5_8GpuFenceENS7_14default_deleteIS9_EEEEEEENS4_IFvRKNS5_20PresentationFeedbackEEEENS7_6vectorINS2_15DrmOverlayPlaneENS7_9allocatorISL_EEEEEvE6InvokeISQ_NS_7WeakPtrIS3_EEJiSE_SJ_SO_EEEvT_OT0_DpOT1_ #11 0x59883c260bd7 _ZN4base8internal7InvokerINS0_9BindStateIMN2ui9DrmThreadEFviNS_12OnceCallbackIFvN3gfx10SwapResultENSt3__110unique_ptrINS6_8GpuFenceENS8_14default_deleteISA_EEEEEEENS5_IFvRKNS6_20PresentationFeedbackEEEENS8_6vectorINS3_15DrmOverlayPlaneENS8_9allocatorISM_EEEEEJNS_7WeakPtrIS4_EEiSF_SK_EEEFvSP_EE7RunOnceEPNS0_13BindStateBaseEOSP_ #12 0x59883c2749d1 _ZN4base8internal7InvokerINS0_9BindStateINS_12OnceCallbackIFvNSt3__16vectorIN2ui15DrmOverlayPlaneENS4_9allocatorIS7_EEEEEEEJSA_EEEFvvEE7RunOnceEPNS0_13BindStateBaseE #13 0x59883e406d91 base::debug::TaskAnnotator::RunTask() #14 0x59883e37590f base::MessageLoop::RunTask() #15 0x59883e375cb2 base::MessageLoop::DoWork() #16 0x59883e403999 base::MessagePumpLibevent::Run() #17 0x59883e395e45 base::RunLoop::Run() #18 0x59883e3ca643 base::Thread::ThreadMain() #19 0x59883e3fdb4d base::(anonymous namespace)::ThreadFunc() #20 0x7f1d6b61a2b8 <unknown> #21 0x7f1d6aadbfad clone r8: 00007f1d6ad8b8e0 r9: 00007f1d67f75700 r10: 0000000000000008 r11: 0000000000000202 r12: 00007f1d67f74698 r13: 0000000000000076 r14: 00007f1d67f746a0 r15: 00007f1d67f746a8 di: 0000000000000de9 si: 0000000000000df5 bp: 00007f1d67f73c80 bx: 00007f1d67f73e20 dx: 0000000000000006 ax: 0000000000000000 cx: 00007f1d6aa18dd2 sp: 00007f1d67f73b58 ip: 00007f1d6aa18dd2 efl: 0000000000000202 cgf: 002b000000000033 erf: 0000000000000000 trp: 0000000000000000 msk: 0000000000000000 cr2: 0000000000000000 [end of stack trace] Calling _exit(1). Core file will not be generated. device-enumerator: scan all dirs device-enumerator: scanning /sys/bus device-enumerator: scanning /sys/class [1:14:1018/170630.821855:ERROR:render_media_log.cc(30)] MediaEvent: MEDIA_ERROR_LOG_ENTRY {"error":"video decode error"} [1:1:1018/170630.839765:ERROR:render_media_log.cc(30)] MediaEvent: PIPELINE_ERROR PIPELINE_ERROR_DECODE [3686:3686:1018/170630.916607:ERROR:gles2_cmd_decoder.cc(17615)] [.BrowserCompositor-0x359c15507000]GL ERROR :GL_INVALID_OPERATION : glCreateAndConsumeTextureCHROMIUM: invalid mailbox name [3686:3686:1018/170630.917400:ERROR:gles2_cmd_decoder.cc(10223)] [.BrowserCompositor-0x359c15507000]RENDER WARNING: texture bound to texture unit 0 is not renderable. It maybe non-power-of-2 and have incompatible texture filtering. [3686:3686:1018/170630.920187:ERROR:gles2_cmd_decoder.cc(17615)] [.BrowserCompositor-0x359c15507000]GL ERROR :GL_INVALID_OPERATION : glCreateAndConsumeTextureCHROMIUM: invalid mailbox name [3686:3686:1018/170630.920505:ERROR:gles2_cmd_decoder.cc(10223)] [.BrowserCompositor-0x359c15507000]RENDER WARNING: texture bound to texture unit 0 is not renderable. It maybe non-power-of-2 and have incompatible texture filtering. [3686:3686:1018/170630.923517:ERROR:gles2_cmd_decoder.cc(17615)] [.BrowserCompositor-0x359c15507000]GL ERROR :GL_INVALID_OPERATION : glCreateAndConsumeTextureCHROMIUM: invalid mailbox name [3686:3686:1018/170630.923852:ERROR:gles2_cmd_decoder.cc(10223)] [.BrowserCompositor-0x359c15507000]RENDER WARNING: texture bound to texture unit 0 is not renderable. It maybe non-power-of-2 and have incompatible texture filtering. [3686:3686:1018/170630.924043:ERROR:gles2_cmd_decoder.cc(17615)] [.BrowserCompositor-0x359c15507000]GL ERROR :GL_INVALID_OPERATION : glCreateAndConsumeTextureCHROMIUM: invalid mailbox name [3686:3686:1018/170630.924208:ERROR:gles2_cmd_decoder.cc(10223)] [.BrowserCompositor-0x359c15507000]RENDER WARNING: texture bound to texture unit 0 is not renderable. It maybe non-power-of-2 and have incompatible texture filtering. [3686:3686:1018/170630.924260:ERROR:gles2_cmd_decoder.cc(17615)] [.BrowserCompositor-0x359c15507000]GL ERROR :GL_INVALID_OPERATION : glCreateAndConsumeTextureCHROMIUM: invalid mailbox name [3686:3686:1018/170630.924303:ERROR:gles2_cmd_decoder.cc(10223)] [.BrowserCompositor-0x359c15507000]RENDER WARNING: texture bound to texture unit 0 is not renderable. It maybe non-power-of-2 and have incompatible texture filtering. [3686:3686:1018/170630.924365:ERROR:gles2_cmd_decoder.cc(17615)] [.BrowserCompositor-0x359c15507000]GL ERROR :GL_INVALID_OPERATION : glCreateAndConsumeTextureCHROMIUM: invalid mailbox name [3686:3686:1018/170630.924415:ERROR:gles2_cmd_decoder.cc(10223)] [.BrowserCompositor-0x359c15507000]RENDER WARNING: texture bound to texture unit 0 is not renderable. It maybe non-power-of-2 and have incompatible texture filtering. [3686:3686:1018/170630.926453:ERROR:gles2_cmd_decoder.cc(17615)] [.BrowserCompositor-0x359c15507000]GL ERROR :GL_INVALID_OPERATION : glCreateAndConsumeTextureCHROMIUM: invalid mailbox name [3686:3686:1018/170630.934670:ERROR:gles2_cmd_decoder.cc(17615)] [.BrowserCompositor-0x359c15507000]GL ERROR :GL_INVALID_OPERATION : glCreateAndConsumeTextureCHROMIUM: invalid mailbox name [3686:3686:1018/170630.934825:ERROR:gles2_cmd_decoder.cc(10223)] [.BrowserCompositor-0x359c15507000]RENDER WARNING: texture bound to texture unit 0 is not renderable. It maybe non-power-of-2 and have incompatible texture filtering. [3686:3686:1018/170630.934915:ERROR:gles2_cmd_decoder.cc(17615)] [.BrowserCompositor-0x359c15507000]GL ERROR :GL_INVALID_OPERATION : glCreateAndConsumeTextureCHROMIUM: invalid mailbox name [3686:3686:1018/170630.934958:ERROR:gles2_cmd_decoder.cc(10223)] [.BrowserCompositor-0x359c15507000]RENDER WARNING: texture bound to texture unit 0 is not renderable. It maybe non-power-of-2 and have incompatible texture filtering. [3686:3686:1018/170630.939091:ERROR:gles2_cmd_decoder.cc(17615)] [.BrowserCompositor-0x359c15507000]GL ERROR :GL_INVALID_OPERATION : glCreateAndConsumeTextureCHROMIUM: invalid mailbox name [3686:3686:1018/170630.939225:ERROR:gles2_cmd_decoder.cc(10223)] [.BrowserCompositor-0x359c15507000]RENDER WARNING: texture bound to texture unit 0 is not renderable. It maybe non-power-of-2 and have incompatible texture filtering. [3686:3686:1018/170630.939305:ERROR:gles2_cmd_decoder.cc(17615)] [.BrowserCompositor-0x359c15507000]GL ERROR :GL_INVALID_OPERATION : glCreateAndConsumeTextureCHROMIUM: invalid mailbox name
,
Oct 19
,
Oct 19
I submit the fix. https://chromium-review.googlesource.com/c/chromium/src/+/1291590 dcastagna, hoegsberg, could you review?
,
Oct 19
rjkroege@ this looks like something you'd be familiar with.
,
Oct 19
Dongseong already submitted the right fix. We're landing it now, I don't think we need to bother rjkroege@.
,
Oct 20
Issue 896423 has been merged into this issue.
,
Oct 20
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/4bf9d01dbf34f17c899095a937dabbb884ec87fe commit 4bf9d01dbf34f17c899095a937dabbb884ec87fe Author: Dongseong Hwang <dongseong.hwang@intel.com> Date: Sat Oct 20 02:26:49 2018 viz: Fix wrong uv_rect calculation CL:1045256 calculates |uv_rect| of the primary plane in wrong way. It causes drmModeAtomicAddProperty failure. This CL fixes it. BUG=896945 Change-Id: I6e914707cb961e8294e8ddd160b75e48ae5a2b64 Reviewed-on: https://chromium-review.googlesource.com/c/1291590 Commit-Queue: Dongseong Hwang <dongseong.hwang@intel.com> Reviewed-by: Daniele Castagna <dcastagna@chromium.org> Reviewed-by: Kristian H. Kristensen <hoegsberg@chromium.org> Reviewed-by: Dongseong Hwang <dongseong.hwang@intel.com> Cr-Commit-Position: refs/heads/master@{#601392} [modify] https://crrev.com/4bf9d01dbf34f17c899095a937dabbb884ec87fe/components/viz/service/display/overlay_processor.cc [modify] https://crrev.com/4bf9d01dbf34f17c899095a937dabbb884ec87fe/components/viz/service/display/overlay_unittest.cc
,
Oct 25
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/f8dfbe74c1f5c79051461cdf448abf7f8188336f commit f8dfbe74c1f5c79051461cdf448abf7f8188336f Author: Dongseong Hwang <dongseong.hwang@intel.com> Date: Thu Oct 25 00:58:37 2018 viz: add unittests for CL:1291590 BUG=896945 TEST=SingleOverlayOnTopTest.AllowVideoNormalTransformWithOutputSurfaceOverlay Change-Id: Iaa46909657f3e11f1d982f94aaff102676a74b4a Reviewed-on: https://chromium-review.googlesource.com/c/1292822 Commit-Queue: Daniele Castagna <dcastagna@chromium.org> Reviewed-by: Daniele Castagna <dcastagna@chromium.org> Reviewed-by: Dongseong Hwang <dongseong.hwang@intel.com> Cr-Commit-Position: refs/heads/master@{#602548} [modify] https://crrev.com/f8dfbe74c1f5c79051461cdf448abf7f8188336f/components/viz/service/display/overlay_unittest.cc
,
Nov 5
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/043bdd225bcd415dbf81ba078b7981b51eef1408 commit 043bdd225bcd415dbf81ba078b7981b51eef1408 Author: Wenzhao (Colin) Zang <wzang@chromium.org> Date: Mon Nov 05 23:36:45 2018 Revert "viz: add unittests for CL:1291590" This reverts commit f8dfbe74c1f5c79051461cdf448abf7f8188336f. Reason for revert: crbug.com/900373 Original change's description: > viz: add unittests for CL:1291590 > > BUG=896945 > TEST=SingleOverlayOnTopTest.AllowVideoNormalTransformWithOutputSurfaceOverlay > > Change-Id: Iaa46909657f3e11f1d982f94aaff102676a74b4a > Reviewed-on: https://chromium-review.googlesource.com/c/1292822 > Commit-Queue: Daniele Castagna <dcastagna@chromium.org> > Reviewed-by: Daniele Castagna <dcastagna@chromium.org> > Reviewed-by: Dongseong Hwang <dongseong.hwang@intel.com> > Cr-Commit-Position: refs/heads/master@{#602548} TBR=dongseong.hwang@intel.com,dcastagna@chromium.org # Not skipping CQ checks because original CL landed > 1 day ago. Bug: 896945, 900373 Change-Id: I57a2c9770df54447e207e7a1d1079006087829ae Reviewed-on: https://chromium-review.googlesource.com/c/1318359 Reviewed-by: Wenzhao (Colin) Zang <wzang@chromium.org> Commit-Queue: Wenzhao (Colin) Zang <wzang@chromium.org> Cr-Commit-Position: refs/heads/master@{#605517} [modify] https://crrev.com/043bdd225bcd415dbf81ba078b7981b51eef1408/components/viz/service/display/overlay_unittest.cc
,
Nov 6
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/ec93e1108c6b20b6df8b54a88e48b043f00b1f59 commit ec93e1108c6b20b6df8b54a88e48b043f00b1f59 Author: Wenzhao (Colin) Zang <wzang@chromium.org> Date: Tue Nov 06 01:33:30 2018 Revert "viz: Fix wrong uv_rect calculation" This reverts commit 4bf9d01dbf34f17c899095a937dabbb884ec87fe. Reason for revert: crbug.com/900373 Original change's description: > viz: Fix wrong uv_rect calculation > > CL:1045256 calculates |uv_rect| of the primary plane in wrong way. It causes > drmModeAtomicAddProperty failure. This CL fixes it. > > BUG=896945 > > Change-Id: I6e914707cb961e8294e8ddd160b75e48ae5a2b64 > Reviewed-on: https://chromium-review.googlesource.com/c/1291590 > Commit-Queue: Dongseong Hwang <dongseong.hwang@intel.com> > Reviewed-by: Daniele Castagna <dcastagna@chromium.org> > Reviewed-by: Kristian H. Kristensen <hoegsberg@chromium.org> > Reviewed-by: Dongseong Hwang <dongseong.hwang@intel.com> > Cr-Commit-Position: refs/heads/master@{#601392} TBR=dongseong.hwang@intel.com,dcastagna@chromium.org,hoegsberg@chromium.org # Not skipping CQ checks because original CL landed > 1 day ago. Bug: 896945, 900373 Change-Id: Ibab22a4fafb7be1ea0921f0cf774b5973d92b0b2 Reviewed-on: https://chromium-review.googlesource.com/c/1318796 Reviewed-by: Wenzhao (Colin) Zang <wzang@chromium.org> Commit-Queue: Wenzhao (Colin) Zang <wzang@chromium.org> Cr-Commit-Position: refs/heads/master@{#605559} [modify] https://crrev.com/ec93e1108c6b20b6df8b54a88e48b043f00b1f59/components/viz/service/display/overlay_processor.cc [modify] https://crrev.com/ec93e1108c6b20b6df8b54a88e48b043f00b1f59/components/viz/service/display/overlay_unittest.cc
,
Nov 7
,
Nov 28
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/1ceb07933f73defaa73c30153fcbe812ad1965e0 commit 1ceb07933f73defaa73c30153fcbe812ad1965e0 Author: Dongseong Hwang <dongseong.hwang@intel.com> Date: Wed Nov 28 01:15:08 2018 Reland "viz: Turn off or crop primary plane when possible" This is a reland of I7946d7acab334bffdf2bb49a1d9c424cf46fc610 and the following bug fixes as follows; "viz: Fix wrong uv_rect calculation" 4bf9d01dbf34f17c899095a937dabbb884ec87fe "viz: add unittests for CL:1291590" f8dfbe74c1f5c79051461cdf448abf7f8188336f Additional fix: - Fix the flickering issue because of primary |uv_rect| cropping even though an overlay attemption fails mainly due to limited pixel rate (e.g. 4k video). - Fix wrong |uv_rect| calculation which causes atomic page flip failures with "No space left on device (28)" error. Original change's description: > viz: Turn off or crop primary plane when possible > > For many fullscreen cases, the rest of the renderpass quads are solid > black quads. In that case, we can turn off the primary plane, and not > having to scan out an full screen of black pixels saves a good deal of > power. In other cases, we can crop down the primary plane down to the > content rectangle (youtube channel logo, for example) and save power. > > Averaging battery draw measurement over about a minute of play back > for a fullscreen youtube video, power savings for soraka is around > 500mW (6.5W to 6.0W) and 300mW for kevin (4.4W to 4.1W), when we > turn off the primary plane. > > Change-Id: I7946d7acab334bffdf2bb49a1d9c424cf46fc610 > Reviewed-on: https://chromium-review.googlesource.com/c/1045256 > Reviewed-by: Daniele Castagna <dcastagna@chromium.org> > Reviewed-by: Dongseong Hwang <dongseong.hwang@intel.com> > Commit-Queue: Kristian H. Kristensen <hoegsberg@chromium.org> > Cr-Commit-Position: refs/heads/master@{#600415} TEST=SingleOverlayOnTopTest.AllowVideoNormalTransformWithOutputSurfaceOverlay Bug: 896945, 900373 Change-Id: I88db118dbc3ca2fed8c8eaf1f9a4337cdae12990 Reviewed-on: https://chromium-review.googlesource.com/c/1346997 Commit-Queue: Dongseong Hwang <dongseong.hwang@intel.com> Reviewed-by: Dongseong Hwang <dongseong.hwang@intel.com> Reviewed-by: Kristian H. Kristensen <hoegsberg@chromium.org> Reviewed-by: Daniele Castagna <dcastagna@chromium.org> Cr-Commit-Position: refs/heads/master@{#611484} [modify] https://crrev.com/1ceb07933f73defaa73c30153fcbe812ad1965e0/components/viz/service/display/gl_renderer_unittest.cc [modify] https://crrev.com/1ceb07933f73defaa73c30153fcbe812ad1965e0/components/viz/service/display/overlay_processor.cc [modify] https://crrev.com/1ceb07933f73defaa73c30153fcbe812ad1965e0/components/viz/service/display/overlay_processor.h [modify] https://crrev.com/1ceb07933f73defaa73c30153fcbe812ad1965e0/components/viz/service/display/overlay_strategy_single_on_top.cc [modify] https://crrev.com/1ceb07933f73defaa73c30153fcbe812ad1965e0/components/viz/service/display/overlay_strategy_underlay.cc [modify] https://crrev.com/1ceb07933f73defaa73c30153fcbe812ad1965e0/components/viz/service/display/overlay_unittest.cc
,
Nov 29
,
Nov 29
Issue 896240 has been merged into this issue.
,
Nov 29
Issue 865739 has been merged into this issue.
,
Nov 29
Issue 901055 has been merged into this issue.
,
Dec 3
Since Chrome 72.0.3625.0 (which is the revision where the CL #12 landed),
Acer Chromebook Tab ("dru") started crashing a lot in the mentioned CHECK():
b/120393443 (internal bug)
Chrome log and the stack trace attached.
Can the crash be due to the reland?
,
Dec 3
#c17 - thank you for sharing logs. let me take a look
,
Dec 3
crash because of following line of the log [5561:5594:1202/185115.253442:ERROR:hardware_display_plane_manager_atomic.cc(140)] Failed to commit properties for page flip.: Invalid argument (22) I guess ARM device have some issue with partial crtc rectangle. kinaba@, Could you cc b/120393443 ot dongseong hwang <dongseong.hwang@intel.corp-partner.google.com> ?
,
Dec 4
#c17 - before we figure out ARM issue, there are 2 options for quick fix. 1. Disable crop optimization for ARM by #ifdef 2. Revert the patch Or we add more log and fix it as soon as we have enough information. 3. land https://chromium-review.googlesource.com/c/chromium/src/+/1359466 What is a better option?
,
Dec 4
> kinaba@, Could you cc b/120393443 ot dongseong hwang <dongseong.hwang@intel.corp-partner.google.com> ? Oops sorry, I tried, but the bug is in a restricted component that didn't accept the partner account. That said, what's in the bug is the logs above and the reproduction step, and below is the repro. The crash is consistently observed while running a test for Android on ChromeOS: 1. Prepare the device in developer mode, sign-in, enable Google Play Store, enable its developer mode, connect adb https://developer.android.com/topic/arc/development-environment#debug_with_adb 2. Download and expand "Android 7.1 R23 Compatibility Test Suite (CTS) - ARM" from https://source.android.com/compatibility/cts/downloads on your workstation. 3. ./android-cts/tools/cts-tradefed run commandAndExit cts -m CtsThemeHostTestCases Re #21: I'm not in the position to assess the importance of #12, but if "2. revert the patch" is an acceptable option, I think that should be the one to take. The fix is needed in M72 which has already branched. I don't think we want to do complex things on the branch.
,
Dec 5
This is also causing breakage for full screen wayland apps in the Linux VM (Crostini). I would suggest reverting the breaking changes and then merging that into M72.
,
Dec 6
Revert CL prepared: https://chromium-review.googlesource.com/c/chromium/src/+/1363734
,
Dec 7
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/e1809a76301288b5b74afc6a785cc2c15da0f938 commit e1809a76301288b5b74afc6a785cc2c15da0f938 Author: Kazuhiro Inaba <kinaba@google.com> Date: Fri Dec 07 17:21:16 2018 Revert "Reland "viz: Turn off or crop primary plane when possible"" This reverts commit 1ceb07933f73defaa73c30153fcbe812ad1965e0. Reason for revert: ARC and Crostini on ARM broken (crbug.com/896945#c17) Original change's description: > Reland "viz: Turn off or crop primary plane when possible" > > This is a reland of I7946d7acab334bffdf2bb49a1d9c424cf46fc610 and > the following bug fixes as follows; > "viz: Fix wrong uv_rect calculation" 4bf9d01dbf34f17c899095a937dabbb884ec87fe > "viz: add unittests for CL:1291590" f8dfbe74c1f5c79051461cdf448abf7f8188336f > > Additional fix: > - Fix the flickering issue because of primary |uv_rect| cropping even though > an overlay attemption fails mainly due to limited pixel rate (e.g. 4k video). > - Fix wrong |uv_rect| calculation which causes atomic page flip failures with > "No space left on device (28)" error. > > Original change's description: > > viz: Turn off or crop primary plane when possible > > > > For many fullscreen cases, the rest of the renderpass quads are solid > > black quads. In that case, we can turn off the primary plane, and not > > having to scan out an full screen of black pixels saves a good deal of > > power. In other cases, we can crop down the primary plane down to the > > content rectangle (youtube channel logo, for example) and save power. > > > > Averaging battery draw measurement over about a minute of play back > > for a fullscreen youtube video, power savings for soraka is around > > 500mW (6.5W to 6.0W) and 300mW for kevin (4.4W to 4.1W), when we > > turn off the primary plane. > > > > Change-Id: I7946d7acab334bffdf2bb49a1d9c424cf46fc610 > > Reviewed-on: https://chromium-review.googlesource.com/c/1045256 > > Reviewed-by: Daniele Castagna <dcastagna@chromium.org> > > Reviewed-by: Dongseong Hwang <dongseong.hwang@intel.com> > > Commit-Queue: Kristian H. Kristensen <hoegsberg@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#600415} > > TEST=SingleOverlayOnTopTest.AllowVideoNormalTransformWithOutputSurfaceOverlay > > Bug: 896945, 900373 > Change-Id: I88db118dbc3ca2fed8c8eaf1f9a4337cdae12990 > Reviewed-on: https://chromium-review.googlesource.com/c/1346997 > Commit-Queue: Dongseong Hwang <dongseong.hwang@intel.com> > Reviewed-by: Dongseong Hwang <dongseong.hwang@intel.com> > Reviewed-by: Kristian H. Kristensen <hoegsberg@chromium.org> > Reviewed-by: Daniele Castagna <dcastagna@chromium.org> > Cr-Commit-Position: refs/heads/master@{#611484} TBR=dongseong.hwang@intel.com,dcastagna@chromium.org,hoegsberg@chromium.org Bug: 896945, 900373 Change-Id: Iec6d2812d94c286c495d90d8952b65ecffa4c010 Reviewed-on: https://chromium-review.googlesource.com/c/1363734 Reviewed-by: Kazuhiro Inaba <kinaba@chromium.org> Reviewed-by: Daniele Castagna <dcastagna@chromium.org> Commit-Queue: Daniele Castagna <dcastagna@chromium.org> Cr-Commit-Position: refs/heads/master@{#614732} [modify] https://crrev.com/e1809a76301288b5b74afc6a785cc2c15da0f938/components/viz/service/display/gl_renderer_unittest.cc [modify] https://crrev.com/e1809a76301288b5b74afc6a785cc2c15da0f938/components/viz/service/display/overlay_processor.cc [modify] https://crrev.com/e1809a76301288b5b74afc6a785cc2c15da0f938/components/viz/service/display/overlay_processor.h [modify] https://crrev.com/e1809a76301288b5b74afc6a785cc2c15da0f938/components/viz/service/display/overlay_strategy_single_on_top.cc [modify] https://crrev.com/e1809a76301288b5b74afc6a785cc2c15da0f938/components/viz/service/display/overlay_strategy_underlay.cc [modify] https://crrev.com/e1809a76301288b5b74afc6a785cc2c15da0f938/components/viz/service/display/overlay_unittest.cc
,
Dec 7
Adding M72 merge request for the revert #25.
,
Dec 8
This bug requires manual review: Reverts referenced in bugdroid comments after merge request. Please contact the milestone owner if you have questions. Owners: govind@(Android), kariahda@(iOS), djmm@(ChromeOS), abdulsyed@(Desktop) For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Dec 10
+djmm@ for merge approval, of the revert #25.
,
Dec 10
,
Dec 10
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/a0ffbe6455fc7eb3cfa1a80d2429ffc064b8ece7 commit a0ffbe6455fc7eb3cfa1a80d2429ffc064b8ece7 Author: Kazuhiro Inaba <kinaba@google.com> Date: Mon Dec 10 12:14:50 2018 Revert "Reland "viz: Turn off or crop primary plane when possible"" This reverts commit 1ceb07933f73defaa73c30153fcbe812ad1965e0. Reason for revert: ARC and Crostini on ARM broken (crbug.com/896945#c17) Original change's description: > Reland "viz: Turn off or crop primary plane when possible" > > This is a reland of I7946d7acab334bffdf2bb49a1d9c424cf46fc610 and > the following bug fixes as follows; > "viz: Fix wrong uv_rect calculation" 4bf9d01dbf34f17c899095a937dabbb884ec87fe > "viz: add unittests for CL:1291590" f8dfbe74c1f5c79051461cdf448abf7f8188336f > > Additional fix: > - Fix the flickering issue because of primary |uv_rect| cropping even though > an overlay attemption fails mainly due to limited pixel rate (e.g. 4k video). > - Fix wrong |uv_rect| calculation which causes atomic page flip failures with > "No space left on device (28)" error. > > Original change's description: > > viz: Turn off or crop primary plane when possible > > > > For many fullscreen cases, the rest of the renderpass quads are solid > > black quads. In that case, we can turn off the primary plane, and not > > having to scan out an full screen of black pixels saves a good deal of > > power. In other cases, we can crop down the primary plane down to the > > content rectangle (youtube channel logo, for example) and save power. > > > > Averaging battery draw measurement over about a minute of play back > > for a fullscreen youtube video, power savings for soraka is around > > 500mW (6.5W to 6.0W) and 300mW for kevin (4.4W to 4.1W), when we > > turn off the primary plane. > > > > Change-Id: I7946d7acab334bffdf2bb49a1d9c424cf46fc610 > > Reviewed-on: https://chromium-review.googlesource.com/c/1045256 > > Reviewed-by: Daniele Castagna <dcastagna@chromium.org> > > Reviewed-by: Dongseong Hwang <dongseong.hwang@intel.com> > > Commit-Queue: Kristian H. Kristensen <hoegsberg@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#600415} > > TEST=SingleOverlayOnTopTest.AllowVideoNormalTransformWithOutputSurfaceOverlay > > Bug: 896945, 900373 > Change-Id: I88db118dbc3ca2fed8c8eaf1f9a4337cdae12990 > Reviewed-on: https://chromium-review.googlesource.com/c/1346997 > Commit-Queue: Dongseong Hwang <dongseong.hwang@intel.com> > Reviewed-by: Dongseong Hwang <dongseong.hwang@intel.com> > Reviewed-by: Kristian H. Kristensen <hoegsberg@chromium.org> > Reviewed-by: Daniele Castagna <dcastagna@chromium.org> > Cr-Commit-Position: refs/heads/master@{#611484} TBR=dongseong.hwang@intel.com,dcastagna@chromium.org,hoegsberg@chromium.org Bug: 896945, 900373 Change-Id: Iec6d2812d94c286c495d90d8952b65ecffa4c010 Reviewed-on: https://chromium-review.googlesource.com/c/1363734 Reviewed-by: Kazuhiro Inaba <kinaba@chromium.org> Reviewed-by: Daniele Castagna <dcastagna@chromium.org> Commit-Queue: Daniele Castagna <dcastagna@chromium.org> Cr-Original-Commit-Position: refs/heads/master@{#614732}(cherry picked from commit e1809a76301288b5b74afc6a785cc2c15da0f938) Reviewed-on: https://chromium-review.googlesource.com/c/1369760 Cr-Commit-Position: refs/branch-heads/3626@{#193} Cr-Branched-From: d897fb137fbaaa9355c0c93124cc048824eb1e65-refs/heads/master@{#612437} [modify] https://crrev.com/a0ffbe6455fc7eb3cfa1a80d2429ffc064b8ece7/components/viz/service/display/gl_renderer_unittest.cc [modify] https://crrev.com/a0ffbe6455fc7eb3cfa1a80d2429ffc064b8ece7/components/viz/service/display/overlay_processor.cc [modify] https://crrev.com/a0ffbe6455fc7eb3cfa1a80d2429ffc064b8ece7/components/viz/service/display/overlay_processor.h [modify] https://crrev.com/a0ffbe6455fc7eb3cfa1a80d2429ffc064b8ece7/components/viz/service/display/overlay_strategy_single_on_top.cc [modify] https://crrev.com/a0ffbe6455fc7eb3cfa1a80d2429ffc064b8ece7/components/viz/service/display/overlay_strategy_underlay.cc [modify] https://crrev.com/a0ffbe6455fc7eb3cfa1a80d2429ffc064b8ece7/components/viz/service/display/overlay_unittest.cc
,
Dec 19
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/a0ffbe6455fc7eb3cfa1a80d2429ffc064b8ece7 Commit: a0ffbe6455fc7eb3cfa1a80d2429ffc064b8ece7 Author: kinaba@google.com Commiter: kinaba@chromium.org Date: 2018-12-10 12:14:50 +0000 UTC Revert "Reland "viz: Turn off or crop primary plane when possible"" This reverts commit 1ceb07933f73defaa73c30153fcbe812ad1965e0. Reason for revert: ARC and Crostini on ARM broken (crbug.com/896945#c17) Original change's description: > Reland "viz: Turn off or crop primary plane when possible" > > This is a reland of I7946d7acab334bffdf2bb49a1d9c424cf46fc610 and > the following bug fixes as follows; > "viz: Fix wrong uv_rect calculation" 4bf9d01dbf34f17c899095a937dabbb884ec87fe > "viz: add unittests for CL:1291590" f8dfbe74c1f5c79051461cdf448abf7f8188336f > > Additional fix: > - Fix the flickering issue because of primary |uv_rect| cropping even though > an overlay attemption fails mainly due to limited pixel rate (e.g. 4k video). > - Fix wrong |uv_rect| calculation which causes atomic page flip failures with > "No space left on device (28)" error. > > Original change's description: > > viz: Turn off or crop primary plane when possible > > > > For many fullscreen cases, the rest of the renderpass quads are solid > > black quads. In that case, we can turn off the primary plane, and not > > having to scan out an full screen of black pixels saves a good deal of > > power. In other cases, we can crop down the primary plane down to the > > content rectangle (youtube channel logo, for example) and save power. > > > > Averaging battery draw measurement over about a minute of play back > > for a fullscreen youtube video, power savings for soraka is around > > 500mW (6.5W to 6.0W) and 300mW for kevin (4.4W to 4.1W), when we > > turn off the primary plane. > > > > Change-Id: I7946d7acab334bffdf2bb49a1d9c424cf46fc610 > > Reviewed-on: https://chromium-review.googlesource.com/c/1045256 > > Reviewed-by: Daniele Castagna <dcastagna@chromium.org> > > Reviewed-by: Dongseong Hwang <dongseong.hwang@intel.com> > > Commit-Queue: Kristian H. Kristensen <hoegsberg@chromium.org> > > Cr-Commit-Position: refs/heads/master@{#600415} > > TEST=SingleOverlayOnTopTest.AllowVideoNormalTransformWithOutputSurfaceOverlay > > Bug: 896945, 900373 > Change-Id: I88db118dbc3ca2fed8c8eaf1f9a4337cdae12990 > Reviewed-on: https://chromium-review.googlesource.com/c/1346997 > Commit-Queue: Dongseong Hwang <dongseong.hwang@intel.com> > Reviewed-by: Dongseong Hwang <dongseong.hwang@intel.com> > Reviewed-by: Kristian H. Kristensen <hoegsberg@chromium.org> > Reviewed-by: Daniele Castagna <dcastagna@chromium.org> > Cr-Commit-Position: refs/heads/master@{#611484} TBR=dongseong.hwang@intel.com,dcastagna@chromium.org,hoegsberg@chromium.org Bug: 896945, 900373 Change-Id: Iec6d2812d94c286c495d90d8952b65ecffa4c010 Reviewed-on: https://chromium-review.googlesource.com/c/1363734 Reviewed-by: Kazuhiro Inaba <kinaba@chromium.org> Reviewed-by: Daniele Castagna <dcastagna@chromium.org> Commit-Queue: Daniele Castagna <dcastagna@chromium.org> Cr-Original-Commit-Position: refs/heads/master@{#614732}(cherry picked from commit e1809a76301288b5b74afc6a785cc2c15da0f938) Reviewed-on: https://chromium-review.googlesource.com/c/1369760 Cr-Commit-Position: refs/branch-heads/3626@{#193} Cr-Branched-From: d897fb137fbaaa9355c0c93124cc048824eb1e65-refs/heads/master@{#612437} |
||||||||||||||||||||||
►
Sign in to add a comment |
||||||||||||||||||||||
Comment 1 by dongseon...@intel.com
, Oct 19