Closed Bug 1299520 Opened 9 years ago Closed 9 years ago

Crash in NDXGI::CDevice::SignalSynchronizationObjectCB

Categories

(Core :: Audio/Video, defect)

Product:
Core
Component:
Audio/Video
Version:
49 Branch
Platform:
Unspecified
Windows 10
Type:
defect
Priority:
Not set
Severity:
critical

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla52
Tracking Flags:
Tracking Status
firefox49 --- affected
firefox-esr45 --- affected
firefox50 + fixed
firefox51 --- fixed
firefox52 --- fixed

People

(Reporter: mwolf, Assigned: philipp)

References

Details

(Keywords: regression, Whiteboard: [gfx-noted])

Crash Data

Attachments

(1 file)

bug1299520.patch
4.44 KB, patch
mozbugz
: review+
ritu
: approval-mozilla-aurora+
ritu
: approval-mozilla-beta+
Details | Diff | Splinter Review
This bug was filed from the Socorro interface and is report bp-09e74f17-b67c-41af-a017-27ad42160831. ============================================================= the crashes occur randomly while browsing or watching videos e.g. on youtube after i installed my new rx 480. i use the most recent drivers aka 16.8.2 on windows 10 1607 build. also i will link some of the other crashes here: https://crash-stats.mozilla.com/report/index/bp-c6a3a601-4630-4c87-af16-a8aae2160831 https://crash-stats.mozilla.com/report/index/bp-655b2856-a8c3-4a54-b24a-f4d5e2160831 https://crash-stats.mozilla.com/report/index/bp-3c528ba3-4873-4eda-90fc-79e712160830 https://crash-stats.mozilla.com/report/index/bp-4e1a01a2-f245-4a11-abc4-0031f2160830 https://crash-stats.mozilla.com/report/index/bp-7c011fed-e219-4296-bd82-167d62160830 https://crash-stats.mozilla.com/report/index/bp-dfdba574-0985-4705-b351-2bbef2160830 https://crash-stats.mozilla.com/report/index/bp-0510f391-8e63-46db-a807-4b9312160830 https://crash-stats.mozilla.com/report/index/bp-4679bf01-20f6-4ccf-b7b3-4da0b2160828 https://crash-stats.mozilla.com/report/index/bp-f8072738-ab6f-49bf-9273-1a06d2160828 https://crash-stats.mozilla.com/report/index/bp-a15d325e-5095-47bc-afde-36f5d2160828 https://crash-stats.mozilla.com/report/index/bp-d377057d-9309-4f97-8974-621ee2160828 https://crash-stats.mozilla.com/report/index/bp-ef8ccd84-e8a9-4a08-a55c-c6fb32160827 https://crash-stats.mozilla.com/report/index/bp-bf9de491-cf19-4d43-b8b3-860052160827 PS: this bugreport shows always up in the crashlogs: https://bugzilla.mozilla.org/show_bug.cgi?id=1255732
Version: unspecified → 49 Branch
{ "application": { "name": "Firefox", "osVersion": "Windows_NT 10.0", "version": "49.0b8", "buildID": "20160829102229", "userAgent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:49.0) Gecko/20100101 Firefox/49.0", "safeMode": false, "updateChannel": "beta", "supportURL": "https://support.mozilla.org/1/firefox/49.0/WINNT/de/", "numTotalWindows": 1, "numRemoteWindows": 0, "remoteAutoStart": false, "autoStartStatus": 2 }, "modifiedPreferences": { "accessibility.typeaheadfind.flashBar": 0, "browser.cache.disk.smart_size.first_run": false, "browser.cache.disk.hashstats_reported": 1, "browser.cache.disk.filesystem_reported": 1, "browser.cache.disk.capacity": 358400, "browser.cache.frecency_experiment": 1, "browser.cache.disk.smart_size.use_old_max": false, "browser.cache.disk.smart_size_cached_value": 358400, "browser.download.importedFromSqlite": true, "browser.download.manager.alertOnEXEOpen": false, "browser.places.smartBookmarksVersion": 8, "browser.sessionstore.upgradeBackup.latestBuildID": "20160825132718", "browser.startup.homepage_override.mstone": "49.0", "browser.startup.homepage_override.buildID": "20160829102229", "browser.tabs.warnOnClose": false, "browser.urlbar.userMadeSearchSuggestionsChoice": true, "dom.apps.lastUpdate.buildID": "20160829102229", "dom.apps.reset-permissions": true, "dom.push.userAgentID": "f69f301234e2453f8ed1395607686a37", "dom.apps.lastUpdate.mstone": "49.0", "dom.w3c_touch_events.expose": false, "dom.mozApps.used": true, "extensions.lastAppVersion": "49.0", "font.internaluseonly.changed": true, "gfx.crash-guard.status.": 2, "gfx.crash-guard.glcontext.gfx.driver-init.webgl-angle-force-d3d11": false, "gfx.crash-guard.d3d11layers.appVersion": "49.0", "gfx.driver-init.appVersion": "42.0", "gfx.crash-guard.glcontext.driverVersion": "16.150.0.0", "gfx.crash-guard.glcontext.gfx.driver-init.webgl-angle-try-d3d11": true, "gfx.crash-guard.glcontext.gfx.driver-init.direct3d11-angle": true, "gfx.crash-guard.glcontext.deviceID": "0x6818", "gfx.crash-guard.d3d11layers.driverVersion": "16.300.2511.1001", "gfx.crash-guard.glcontext.appVersion": "45.0", "gfx.direct3d.prefer_10_1": true, "gfx.crash-guard.d3d11layers.deviceID": "0x67df", "gfx.driver-init.deviceID": "0x6818", "gfx.driver-init.driverVersion": "15.201.1151.0", "gfx.driver-init.status": 2, "gfx.crash-guard.glcontext.gfx.driver-init.webgl-angle": true, "gfx.crash-guard.d3d11layers.feature-d3d11": true, "gfx.crash-guard.status.d3d11video": 2, "gfx.crash-guard.status.glcontext": 2, "gfx.crash-guard.d3d11layers.feature-d2d": true, "gfx.crash-guard.status.d3d11layers": 2, "gfx.driver-init.feature-d3d11": true, "gfx.crash-guard.glcontext.gfx.driver-init.webgl-angle-force-warp": false, "gfx.driver-init.feature-d2d": true, "gfx.direct3d.last_used_feature_level_idx": 0, "gfx.crash-guard.status.d3d9video": 2, "media.youtube-ua.override": false, "media.gmp-eme-adobe.lastUpdate": 1458897688, "media.webrtc.debug.aec_log_dir": "C:\\Users\\MARTIN~1\\AppData\\Local\\Temp", "media.webrtc.debug.log_file": "C:\\Users\\MARTIN~1\\AppData\\Local\\Temp\\WebRTC.log", "media.gmp-widevinecdm.lastUpdate": 1470945618, "media.gmp-gmpopenh264.lastUpdate": 1471345764, "media.benchmark.vp9.versioncheck": 1, "media.gmp-eme-adobe.abi": "x86_64-msvc-x64", "media.hardware-video-decoding.failed": false, "media.gmp-widevinecdm.abi": "x86_64-msvc-x64", "media.gmp-gmpopenh264.version": "1.6", "media.gmp-manager.lastCheck": 1472646359, "media.gmp-widevinecdm.version": "1.4.8.903", "media.gmp-eme-adobe.version": "17", "media.gmp.storage.version.observed": 1, "media.benchmark.vp9.fps": 243, "media.gmp-gmpopenh264.abi": "x86_64-msvc-x64", "media.gmp-manager.buildID": "20160829102229", "network.predictor.cleaned-up": true, "network.cookie.prefsMigrated": true, "places.history.expiration.transient_current_max_pages": 104858, "places.database.lastMaintenance": 1472359296, "plugin.importedState": true, "plugin.disable_full_page_plugin_for_types": "application/pdf", "print.printer_5350DN.print_shrink_to_fit": true, "print.printer_Brother_HL-5350DN_series_Printer.print_footerleft": "&PT", "print.printer_Brother_HL-5350DN_series.print_unwriteable_margin_left": 0, "print.printer_Brother_HL-5350DN_series.print_edge_left": 0, "print.printer_Brother_HL-5350DN_series.print_margin_top": "0.5", "print.printer_Brother_HL-5350DN_series_Printer.print_paper_name": "", "print.printer_Brother_HL-5350DN_series_Printer.print_shrink_to_fit": true, "print.printer_5350DN.print_edge_left": 0, "print.printer_Brother_HL-5350DN_series_Printer.print_paper_data": 9, "print.printer_Brother_HL-5350DN_series.print_paper_size_type": 0, "print.printer_Brother_HL-5350DN_series.print_bgimages": false, "print.printer_Brother_HL-5350DN_series_Printer.print_edge_left": 0, "print.printer_Brother_HL-5350DN_series_Printer.print_page_delay": 50, "print.printer_Brother_HL-5350DN_series_Printer.print_reversed": false, "print.printer_5350DN.print_orientation": 1, "print.printer_Brother_HL-5350DN_series.print_downloadfonts": false, "print.printer_Brother_HL-5350DN_series_Printer.print_footerright": "&D", "print.printer_Brother_HL-5350DN_series_Printer.print_margin_right": "0.5", "print.printer_5350DN.print_bgcolor": false, "print.printer_Brother_HL-5350DN_series_Printer.print_unwriteable_margin_bottom": 0, "print.printer_5350DN.print_oddpages": true, "print.printer_5350DN.print_unwriteable_margin_right": 0, "print.printer_Brother_HL-5350DN_series.print_paper_height": " 11,00", "print.printer_Brother_HL-5350DN_series_Printer.print_bgimages": false, "print.printer_Brother_HL-5350DN_series_Printer.print_in_color": true, "print.printer_Brother_HL-5350DN_series.print_headerright": "&U", "print.printer_Brother_HL-5350DN_series_Printer.print_unwriteable_margin_right": 0, "print.printer_Brother_HL-5350DN_series_Printer.print_margin_left": "0.5", "print.printer_Brother_HL-5350DN_series_Printer.print_oddpages": true, "print.printer_5350DN.print_reversed": false, "print.printer_Brother_HL-5350DN_series_Printer.print_paper_size_type": 0, "print.printer_5350DN.print_footercenter": "", "print.printer_Brother_HL-5350DN_series.print_shrink_to_fit": false, "print.printer_5350DN.print_in_color": true, "print.printer_Brother_HL-5350DN_series.print_headercenter": "", "print.printer_Brother_HL-5350DN_series_Printer.print_unwriteable_margin_left": 0, "print.printer_Brother_HL-5350DN_series.print_edge_top": 0, "print.printer_Brother_HL-5350DN_series.print_paper_name": "", "print.printer_Brother_HL-5350DN_series_Printer.print_margin_bottom": "0.5", "print.printer_Brother_HL-5350DN_series.print_duplex": 500944964, "print.printer_Brother_HL-5350DN_series.print_margin_right": "0.5", "print.printer_Brother_HL-5350DN_series.print_unwriteable_margin_bottom": 0, "print.printer_Brother_HL-5350DN_series.print_edge_right": 0, "print.printer_Brother_HL-5350DN_series.print_footercenter": "", "print.printer_5350DN.print_paper_name": "", "print.printer_Brother_HL-5350DN_series.print_page_delay": 50, "print.printer_5350DN.print_margin_top": "0.5", "print.printer_5350DN.print_bgimages": false, "print.printer_Brother_HL-5350DN_series.print_orientation": 0, "print.printer_Brother_HL-5350DN_series_Printer.print_edge_top": 0, "print.printer_5350DN.print_duplex": -437918235, "print.printer_5350DN.print_paper_width": "210,00", "print.printer_Brother_HL-5350DN_series.print_evenpages": true, "print.printer_Brother_HL-5350DN_series.print_margin_bottom": "0.5", "print.printer_Brother_HL-5350DN_series_Printer.print_footercenter": "", "print.printer_Brother_HL-5350DN_series_Printer.print_headercenter": "", "print.printer_Brother_HL-5350DN_series_Printer.print_resolution": 600, "print.printer_5350DN.print_headerleft": "&T", "print.printer_Brother_HL-5350DN_series.print_unwriteable_margin_top": 0, "print.printer_Brother_HL-5350DN_series_Printer.print_headerleft": "&T", "print.printer_Brother_HL-5350DN_series_Printer.print_headerright": "&U", "print.printer_5350DN.print_edge_top": 0, "print.printer_Brother_HL-5350DN_series.print_to_file": false, "print.printer_5350DN.print_unwriteable_margin_top": 0, "print.printer_5350DN.print_headerright": "&U", "print.printer_5350DN.print_margin_bottom": "0.5", "print.printer_5350DN.print_to_file": false, "print.printer_Brother_HL-5350DN_series.print_colorspace": "", "print.printer_5350DN.print_unwriteable_margin_bottom": 0, "print.printer_5350DN.print_scaling": " 1,00", "print.printer_Brother_HL-5350DN_series_Printer.print_bgcolor": false, "print.printer_5350DN.print_evenpages": true, "print.printer_Brother_HL-5350DN_series.print_bgcolor": false, "print.printer_5350DN.print_paper_data": 9, "print.printer_Brother_HL-5350DN_series_Printer.print_unwriteable_margin_top": 0, "print.printer_Brother_HL-5350DN_series.print_edge_bottom": 0, "print.printer_5350DN.print_headercenter": "", "print.printer_Brother_HL-5350DN_series.print_scaling": " 0,90", "print.printer_Brother_HL-5350DN_series.print_resolution_name": "", "print.printer_Brother_HL-5350DN_series_Printer.print_to_file": false, "print.printer_Brother_HL-5350DN_series.print_paper_data": 9, "print.printer_Brother_HL-5350DN_series.print_margin_left": "0.5", "print.printer_Brother_HL-5350DN_series_Printer.print_edge_right": 0, "print.printer_Brother_HL-5350DN_series.print_command": "", "print.printer_5350DN.print_footerright": "&D", "print.printer_Brother_HL-5350DN_series_Printer.print_paper_size_unit": 1, "print.printer_5350DN.print_edge_bottom": 0, "print.printer_5350DN.print_margin_right": "0.5", "print.printer_5350DN.print_resolution": 600, "print.printer_Brother_HL-5350DN_series_Printer.print_scaling": " 1,00", "print.printer_Brother_HL-5350DN_series.print_unwriteable_margin_right": 0, "print.printer_Brother_HL-5350DN_series_Printer.print_margin_top": "0.5", "print.printer_Brother_HL-5350DN_series_Printer.print_duplex": -437918235, "print.printer_5350DN.print_unwriteable_margin_left": 0, "print.printer_Brother_HL-5350DN_series.print_footerleft": "&PT", "print.printer_Brother_HL-5350DN_series_Printer.print_paper_height": "297,00", "print.printer_5350DN.print_paper_size_unit": 1, "print.printer_Brother_HL-5350DN_series.print_paper_size_unit": 1, "print.printer_5350DN.print_edge_right": 0, "print.printer_Brother_HL-5350DN_series.print_paper_width": " 8,50", "print.printer_Brother_HL-5350DN_series.print_resolution": 207233092, "print.printer_Brother_HL-5350DN_series.print_in_color": true, "print.printer_Brother_HL-5350DN_series_Printer.print_evenpages": true, "print.printer_5350DN.print_page_delay": 50, "print.printer_Brother_HL-5350DN_series.print_plex_name": "", "print.printer_Brother_HL-5350DN_series_Printer.print_edge_bottom": 0, "print.printer_Brother_HL-5350DN_series_Printer.print_orientation": 0, "print.printer_Brother_HL-5350DN_series_Printer.print_paper_width": "210,00", "print.printer_Brother_HL-5350DN_series.print_headerleft": "&T", "print.printer_5350DN.print_paper_height": "297,00", "print.printer_5350DN.print_footerleft": "&PT", "print.printer_5350DN.print_margin_left": "0.5", "print.printer_Brother_HL-5350DN_series.print_footerright": "&D", "print.printer_Brother_HL-5350DN_series.print_oddpages": true, "print.printer_Brother_HL-5350DN_series.print_reversed": false, "privacy.sanitize.migrateClearSavedPwdsOnExit": true, "privacy.donottrackheader.enabled": true, "privacy.sanitize.migrateFx3Prefs": true, "security.disable_button.openDeviceManager": false, "security.disable_button.openCertManager": false, "services.sync.declinedEngines": "", "services.sync.lastPing": 1472618683, "services.sync.lastSync": "Wed Aug 31 2016 18:10:30 GMT+0200", "services.sync.numClients": 2, "services.sync.engine.prefs.modified": false, "storage.vacuum.last.index": 1, "storage.vacuum.last.places.sqlite": 1470834141, "ui.osk.debug.keyboardDisplayReason": "IKPOS: Touch screen not found." }, "lockedPreferences": {}, "javaScript": { "incrementalGCEnabled": true }, "accessibility": { "isActive": false, "forceDisabled": 0 }, "libraryVersions": { "NSPR": { "minVersion": "4.12", "version": "4.12" }, "NSS": { "minVersion": "3.25", "version": "3.25" }, "NSSUTIL": { "minVersion": "3.25", "version": "3.25" }, "NSSSSL": { "minVersion": "3.25", "version": "3.25" }, "NSSSMIME": { "minVersion": "3.25", "version": "3.25" } }, "userJS": { "exists": false }, "crashes": { "submitted": [ { "id": "bp-09e74f17-b67c-41af-a017-27ad42160831", "date": 1472652674782, "pending": false }, { "id": "bp-c6a3a601-4630-4c87-af16-a8aae2160831", "date": 1472646300133, "pending": false }, { "id": "bp-655b2856-a8c3-4a54-b24a-f4d5e2160831", "date": 1472646030103, "pending": false }, { "id": "bp-3c528ba3-4873-4eda-90fc-79e712160830", "date": 1472573798009, "pending": false }, { "id": "bp-4e1a01a2-f245-4a11-abc4-0031f2160830", "date": 1472543213535, "pending": false }, { "id": "bp-7c011fed-e219-4296-bd82-167d62160830", "date": 1472543183860, "pending": false }, { "id": "bp-dfdba574-0985-4705-b351-2bbef2160830", "date": 1472529130592, "pending": false }, { "id": "bp-0510f391-8e63-46db-a807-4b9312160830", "date": 1472520645740, "pending": false }, { "id": "bp-4679bf01-20f6-4ccf-b7b3-4da0b2160828", "date": 1472414841807, "pending": false }, { "id": "bp-f8072738-ab6f-49bf-9273-1a06d2160828", "date": 1472406877353, "pending": false }, { "id": "bp-a15d325e-5095-47bc-afde-36f5d2160828", "date": 1472403038905, "pending": false } ], "pending": 0 }, "extensions": [ { "name": "Adblock Plus", "version": "2.7.3", "isActive": true, "id": "{d10d0bf8-f5b5-c8b4-a8b2-2b9879e08c5d}" }, { "name": "Disconnect", "version": "3.15.3.1-signed.1-signed", "isActive": true, "id": "2.0@disconnect.me" }, { "name": "DownThemAll!", "version": "2.0.19", "isActive": true, "id": "{DDC359D1-844A-42a7-9AA1-88A850A938A8}" }, { "name": "Multi-process staged rollout", "version": "1.0", "isActive": true, "id": "e10srollout@mozilla.org" }, { "name": "Password Exporter", "version": "1.3.2", "isActive": true, "id": "{B17C1C5A-04B1-11DB-9804-B622A1EF5492}" }, { "name": "Pocket", "version": "1.0.4", "isActive": true, "id": "firefox@getpocket.com" }, { "name": "ProxTube", "version": "2.2.2", "isActive": true, "id": "ich@maltegoetz.de" }, { "name": "Tab Auto Reload", "version": "1.0.17", "isActive": true, "id": "TabAutoReload@schuzak.jp" }, { "name": "Web Compat", "version": "1.0", "isActive": true, "id": "webcompat@mozilla.org" }, { "name": "YouTube High Definition", "version": "48.1", "isActive": true, "id": "{7b1bf0b6-a1b9-42b0-b75d-252036438bdc}" }, { "name": "Hide Add-on-Bar-Close-Button", "version": "0.3.0.1-signed.1-signed", "isActive": false, "id": "hide-add-on-bar-close-button@mauricioid.com" }, { "name": "NoScript", "version": "2.9.0.14", "isActive": false, "id": "{73a6fe31-595d-460b-a920-fcc0f8843232}" }, { "name": "Speed Dial", "version": "0.9.6.18", "isActive": false, "id": "{64161300-e22b-11db-8314-0800200c9a66}" }, { "name": "Tab Mix Plus", "version": "0.5.0.0", "isActive": false, "id": "{dc572301-7619-498c-a57d-39143191b318}" }, { "name": "Video DownloadHelper", "version": "6.0.0", "isActive": false, "id": "{b9db16a4-6edc-47ec-a1f4-b86292ed211d}" }, { "name": "WOT", "version": "20151208", "isActive": false, "id": "{a0d7ccb3-214d-498b-b4aa-0e8fda9a7bf7}" } ], "experiments": [], "graphics": { "numTotalWindows": 1, "numAcceleratedWindows": 1, "windowLayerManagerType": "Direct3D 11", "windowLayerManagerRemote": true, "supportsHardwareH264": "Yes; Using D3D11 API", "adapterDescription": "Radeon (TM) RX 480 Graphics", "adapterVendorID": "0x1002", "adapterDeviceID": "0x67df", "adapterSubsysID": "2372148c", "adapterRAM": "4095", "adapterDrivers": "aticfx64 aticfx64 aticfx64 amdxc64 aticfx32 aticfx32 aticfx32 amdxc32 atiumd64 atidxx64 atidxx64 atiumdag atidxx32 atidxx32 atiumdva atiumd6a atitmm64", "driverVersion": "16.300.2511.1001", "driverDate": "8-11-2016", "adapterDescription2": "", "adapterVendorID2": "", "adapterDeviceID2": "", "adapterSubsysID2": "", "adapterRAM2": "", "adapterDrivers2": "", "driverVersion2": "", "driverDate2": "", "isGPU2Active": false, "direct2DEnabled": true, "directWriteEnabled": true, "directWriteVersion": "10.0.14393.0", "webglRenderer": "Google Inc. -- ANGLE (Radeon (TM) RX 480 Graphics Direct3D11 vs_5_0 ps_5_0)", "info": { "AzureCanvasBackend": "direct2d 1.1", "AzureCanvasAccelerated": 0, "AzureFallbackCanvasBackend": "cairo", "AzureContentBackend": "direct2d 1.1" }, "featureLog": { "features": [ { "name": "HW_COMPOSITING", "description": "Compositing", "status": "available", "log": [ { "type": "default", "status": "available" } ] }, { "name": "D3D11_COMPOSITING", "description": "Direct3D11 Compositing", "status": "available", "log": [ { "type": "default", "status": "available" } ] }, { "name": "D3D9_COMPOSITING", "description": "Direct3D9 Compositing", "status": "available", "log": [ { "type": "default", "status": "available" } ] }, { "name": "DIRECT2D", "description": "Direct2D", "status": "available", "log": [ { "type": "default", "status": "available" } ] }, { "name": "D3D11_HW_ANGLE", "description": "Direct3D11 hardware ANGLE", "status": "available", "log": [ { "type": "default", "status": "available" } ] } ], "fallbacks": [] }, "crashGuards": [] } }
Component: General → Graphics
Product: Firefox → Core
Status: UNCONFIRMED → NEW
Ever confirmed: true
it also crashes sometimes while i just move my mouse over some buttons on a website. sometimes the browser hangs while scrolling but does not crash.
this crash: https://crash-stats.mozilla.com/report/index/bp-d475fee0-00ea-4279-a630-52f362160831 occured just while scrolling
updated the videocard driver to 16.9.1 beta but i still get these crashes: https://crash-stats.mozilla.com/report/index/36e7fb99-fb7a-460c-b0b5-44b142160908
this crash signature seems to have risen over the whole 49 beta cycle and nearly exclusively occurs on amd graphics hardware on windows 8+10.
status-firefox49: --- → affected
status-firefox50: --- → affected
status-firefox51: --- → affected
Flags: needinfo?(milan)
It seems that Bug 1255732 got the same Crash Signature.
Whiteboard: [gfx-noted]
(In reply to Martin from comment #3) > this crash: > https://crash-stats.mozilla.com/report/index/bp-d475fee0-00ea-4279-a630- > 52f362160831 occured just while scrolling Was there any video playing in the background? E.g. when you are on the Facebook timeline the videos start playing automatically. Could you try setting the 'media.wmf.enabled' preference to false and report back if you still see the crashes?
Flags: needinfo?(mwolf)
i will try
Paul, these are all AMD crashes - any insight from reports linked in comment 3 and comment 4?
Flags: needinfo?(milan) → needinfo?(paul.blinzer)
(In reply to Milan Sreckovic [:milan] from comment #9) > Paul, these are all AMD crashes - any insight from reports linked in comment > 3 and comment 4? We have an upcoming driver release that should have a fix included, based on the crash signature seen.
Flags: needinfo?(paul.blinzer)
(In reply to Paul Blinzer from comment #10) > (In reply to Milan Sreckovic [:milan] from comment #9) > > Paul, these are all AMD crashes - any insight from reports linked in comment > > 3 and comment 4? > > We have an upcoming driver release that should have a fix included, based on > the crash signature seen. Is the crash that you fixed related to DXVA (or video playing in general) in any way? Maybe we can blocklist the affected versions.
since i disabled 'media.wmf.enabled' the crashes seem gone, im looking forward to the fix then. ty guys
Bulk move of gfx-noted bugs without priority to P3 for tracking.
Priority: -- → P3
https://crash-stats.mozilla.com/report/index/2851a3a0-f4d7-49df-b732-7b72b2160922 this is my latest crash with 16.9.2 hotfix and 'media.wmf.enabled' = true because i thought it was fixed by the driver...
Needinfoing Paul and Milan regarding comment 11. Looks like disabling WMF fixes the crash for Martin. Should we consider blocklisting the affected driver versions? What is the version of the first driver which is fixed? Removing needinfo from Martin, since he already replied to all our questions. Martin, please use 'media.wmf.enabled' = false as a workaround for now.
Flags: needinfo?(paul.blinzer)
Flags: needinfo?(mwolf)
Flags: needinfo?(milan)
http://support.amd.com/en-us/kb-articles/Pages/Radeon-Software-Crimson-Edition-16.9.2-Release-Notes.aspx i think paul mentions this driver (Intermittent crash may be experienced in Mozilla Firefox while doing video playback on some Radeon RX 400 Series graphics products.) but my last crash was with this driver. so of course i will set 'media.wmf.enabled' = false till amd has a true fix and/or paul responds to it.
(In reply to Martin from comment #16) > http://support.amd.com/en-us/kb-articles/Pages/Radeon-Software-Crimson- > Edition-16.9.2-Release-Notes.aspx > i think paul mentions this driver > (Intermittent crash may be experienced in Mozilla Firefox while doing video > playback on some Radeon RX 400 Series graphics products.) > > but my last crash was with this driver. so of course i will set > 'media.wmf.enabled' = false till amd has a true fix and/or paul responds to > it. That driver package was expected to have a fix but it may be only a partial one, if a similar issue is still observed. Would it be possible to get an updated crashdump with the 16.9.2 driver to see if there is any other condition running into this ?
Flags: needinfo?(paul.blinzer)
(In reply to Martin from comment #18) > https://crash-stats.mozilla.com/report/index/2851a3a0-f4d7-49df-b732- > 7b72b2160922 this one is with 16.9.2 Martin, would you be willing to share a crash dump with AMD? It might contain personal information (e.g. the tabs that you had open at the time, etc.), so I perfectly understand if you'd rather not share it. The information that people can see at the URL is actually only part (the non-sensitive part) of your crash report. We're not going to share the sensitive part with anyone unless we have your explicit permission to do so.
how can i verify what pages i was on, since i think i could have been logged into my bank account and one of my customers websites.
Hi I want to report some info I found with this issue. First This is my crash report page: https://crash-stats.mozilla.com/report/index/c6ccec3e-54bd-4cd8-b4d4-598132160926 RX460 with 16.9.2/win10 64bit I created a new profile and still can reproduce this issue , following is the steps: 1.Create a new profile with firefox 49.0.1 and enable Hardware Acceleration. 2.Install these two plug-in : h264ify/uBlock Origin 3.In uBlock Origin's filter rule control panel, enable as much rule as you can. 4.Open this youtube videos and change video quality setting to 1080P: https://www.youtube.com/watch?v=6Dakd7EIgBE&list=RD6Dakd7EIgBE#t=3 5.While the video is playing ,look at the right-menu and click this title "Cello Wars (Star Wars Parody) Lightsaber Duel". I got: 1.Sometimes it loads fine. 2.Sometimes it is a black window with a loading animate until I F5 the page 3.Sometime it will hang firefox for 3~5 seconds and load the video. 3.Sometimes it crash like no tomorrow. You need to test this for several times by keeping switch from this video to "Cello Wars", while one video is playing. It seems only happens on 1080P especially 1080P@60FPS videos. If you do not install uBlock, Firefox seems still hangs/crash. If you do not install h264ify, Firefox still has chance to crash/hangs on some origin h264 DXVA videos which do not need h264ify. I tested Chrome and Edge on the same PC , as far as I can see, only Firefox has this issue. I installed the same plug-in on Chrome 53.0.2785.116 m (64-bit) and it works fine with H/W enabled.
> Would it be possible to get an updated crashdump with the 16.9.2 driver > to see if there is any other condition running into this ? Here are a few of the last days: https://crash-stats.mozilla.com/report/index/8315d35d-853f-420c-a2b1-5e0812160927 https://crash-stats.mozilla.com/report/index/b381e227-9fda-4758-b069-884e32160925 https://crash-stats.mozilla.com/report/index/027dd085-164c-4c50-86e8-9511b2160923 https://crash-stats.mozilla.com/report/index/2e3a8acc-378e-4c3e-a748-2809d2160923 https://crash-stats.mozilla.com/report/index/c599575f-f95a-46ed-9dc9-178c92160923 Graphics card is also rx 400 Series (470). If you need further Information, I'd be happy to provide them!
Alexander and/or Vonchen, would you be willing to share crash dumps with AMD? See comment 19 for an explanation of what this means.
Flags: needinfo?(vonchen2325)
Flags: needinfo?(alexander)
(In reply to Marco Castelluccio [:marco] from comment #23) > Alexander and/or Vonchen, would you be willing to share crash dumps with AMD? >> We're not going to share the sensitive part with anyone unless we have your explicit permission to do so. Yes, absolutely. No problem at all. How would I do that?
(In reply to Alexander Steinhöfer from comment #24) > Yes, absolutely. No problem at all. > > How would I do that? Thanks a lot! Don't worry, I'll take care of it.
Flags: needinfo?(vonchen2325)
Flags: needinfo?(alexander)
Paul, I've sent you the dumps, let us know if you're able to find the issue, if we can implement a workaround or if we need to wait for the next driver update.
Flags: needinfo?(paul.blinzer)
(In reply to Marco Castelluccio [:marco] from comment #23) > Alexander and/or Vonchen, would you be willing to share crash dumps with AMD? > See comment 19 for an explanation of what this means. Yes,it is fine, no problem. Hope it helps to find the cause of the issue. I have more reports if you needs them, but they are just from the same crash-loop, which means Firefox crashed-->checked send report-->click re-Lanuch Firefox->Firefox crashed-->endless loop. https://crash-stats.mozilla.com/report/index/49124e0a-9d81-4001-b23c-f951d2160928 https://crash-stats.mozilla.com/report/index/4c0123ee-d316-4f28-9cfb-28d2a2160926 https://crash-stats.mozilla.com/report/index/44f1458c-f7de-4cbf-b28a-3361a2160926
(In reply to vonchen from comment #27) > (In reply to Marco Castelluccio [:marco] from comment #23) > > Alexander and/or Vonchen, would you be willing to share crash dumps with AMD? > > See comment 19 for an explanation of what this means. > Yes,it is fine, no problem. > Hope it helps to find the cause of the issue. > > I have more reports if you needs them, but they are just from the same > crash-loop, which means Firefox crashed-->checked send report-->click > re-Lanuch Firefox->Firefox crashed-->endless loop. > > https://crash-stats.mozilla.com/report/index/49124e0a-9d81-4001-b23c- > f951d2160928 > https://crash-stats.mozilla.com/report/index/4c0123ee-d316-4f28-9cfb- > 28d2a2160926 > https://crash-stats.mozilla.com/report/index/44f1458c-f7de-4cbf-b28a- > 3361a2160926 Thank you for providing them. We are looking at the first set to root cause this issue right now.
Flags: needinfo?(paul.blinzer)
(In reply to Paul Blinzer from comment #28) > (In reply to vonchen from comment #27) > > (In reply to Marco Castelluccio [:marco] from comment #23) > > > Alexander and/or Vonchen, would you be willing to share crash dumps with AMD? > > > See comment 19 for an explanation of what this means. > > Yes,it is fine, no problem. > > Hope it helps to find the cause of the issue. > > > > I have more reports if you needs them, but they are just from the same > > crash-loop, which means Firefox crashed-->checked send report-->click > > re-Lanuch Firefox->Firefox crashed-->endless loop. > > > > https://crash-stats.mozilla.com/report/index/49124e0a-9d81-4001-b23c- > > f951d2160928 > > https://crash-stats.mozilla.com/report/index/4c0123ee-d316-4f28-9cfb- > > 28d2a2160926 > > https://crash-stats.mozilla.com/report/index/44f1458c-f7de-4cbf-b28a- > > 3361a2160926 > > Thank you for providing them. We are looking at the first set to root cause > this issue right now. Issue is located and will be addressed in an upcoming driver release.
Great, thanks Paul! Anything, short of blocking the driver versions, that we can avoid doing in the meantime?
Flags: needinfo?(milan)
Crash volume for signature 'NDXGI::CDevice::SignalSynchronizationObjectCB': - nightly (version 52): 34 crashes from 2016-09-19. - aurora (version 51): 86 crashes from 2016-09-19. - beta (version 50): 467 crashes from 2016-09-20. - release (version 49): 3916 crashes from 2016-09-05. - esr (version 45): 5 crashes from 2016-06-01. Crash volume on the last weeks (Week N is from 10-03 to 10-09): W. N-1 W. N-2 - nightly 22 12 - aurora 70 16 - beta 340 127 - release 3034 880 - esr 0 0 Affected platform: Windows Crash rank on the last 7 days: Browser Content Plugin - nightly #50 - aurora #81 #27 - beta #51 #39 #288 - release #14 #8 #329 - esr
status-firefox52: --- → affected
status-firefox-esr45: --- → affected
Hi Update to 16.10.1 still have this issue. https://crash-stats.mozilla.com/report/index/607cd3a3-c3e1-4aa6-9093-cd2502161006 Crash while watching: https://www.youtube.com/watch?v=xDMP3i36naA Plug-in: H264ify
[Tracking Requested - why for this release]: This crash spiked with the release of Beta 50.
tracking-firefox50: --- → ?
did you blacklist the drivers in ff50b5, since some videos dont play anymore andyoutube uses all of the sudden the flash player or is that a sideeffect of 'media.wmf.enabled'?
i meant 'media.wmf.enabled' = false
Hi Paul, could you please address Milan's question in comment 30? Thanks!
Flags: needinfo?(paul.blinzer)
Hi Milan, this was bulk-moved to a P3 but given that this crash is spiking, can we bump up the priority?
Flags: needinfo?(milan)
Priority: P3 → --
(In reply to Milan Sreckovic [:milan] from comment #30) > Great, thanks Paul! > > Anything, short of blocking the driver versions, that we can avoid doing in > the meantime? As far as workarounds, it is a similar multithread race condition as for a few other previous bugs (which really should not happen as only one application thread should be in the driver at a time). It seems that Firefox is prone to run into such conditions. And this is again a Win10 specific bug apparently. On some other news, the driver fix I referred to should have made it into the Crimson 16.10.1 hotfix drivers released last week. Could you please confirm that the issue is now addressed with that driver?
Flags: needinfo?(paul.blinzer)
Paul, vonchen in comment 32 said that the update to 16.10.1 still has the issue. Here's his crash report: https://crash-stats.mozilla.com/report/index/607cd3a3-c3e1-4aa6-9093-cd2502161006.
(In reply to Paul Blinzer from comment #38) > (In reply to Milan Sreckovic [:milan] from comment #30) > > Great, thanks Paul! > > > > Anything, short of blocking the driver versions, that we can avoid doing in > > the meantime? > As far as workarounds, it is a similar multithread race condition as for a > few other previous bugs (which really should not happen as only one > application thread should be in the driver at a time). It seems that Firefox > is prone to run into such conditions. And this is again a Win10 specific bug > apparently. > > On some other news, the driver fix I referred to should have made it into > the Crimson 16.10.1 hotfix drivers released last week. Could you please > confirm that the issue is now addressed with that driver? Hi As far as I know there are two version of 16.10.1 driver which one is non-whql released at early October and one is whql released last week(10/13).Which one do you mean? Or they are all the same? I tested non-whql 16.10.1 but the same issue occurred.I then roll-backed to 16.9.2 and I did not test whql 16.10.1. The reason is after 16.10.1 I have some weird crash not only firefox, but also 2d desktop works and gaming hard freezing that never happened from 16.7.3~16.9.2 on my RX460. I am not sure if there is something wrong on my pc or driver issue(It seems my OS hard driver have some lag issue I did some trick to make to back to normal which is backup and recover hdd partition). But anyway back to 16.9.2 everything is fine again.(firefox issue still), so I am waiting for next new driver. But if the whql 16.10.1 has the fix I maybe will try it , just want to make sure which version do you mean.
Hi Marco, or anyone knowing this issue. Is there really no way that fix this issue on Firefox side? Since Firefox 47/48 seems fine. I tried to completely disable Multi-Process feature by setting browser.tabs.remote.autostart.2 to false in Firefox 49.0.1, but the issue still occurred. (Although it is already disable because of add-on or accessibility tools, I whatever no way to use Firefox multi-process feature on my PC, even I forced enable this feature I still only can see one firefox.exe process in task manager.) I go back and tested old build, the following is result(with 16.9.2): 1.Firefox 47.01 : Works just fine. No see any lag hang freeze or crash. 2.Firefox 48.0.2: Has similar freezing behavior while loading video(click other tabs no response) ,only sometimes, but it will finally load the video and never crash.
@paul here is a crash from yesterday with 16.10.1 (13th of october version) https://crash-stats.mozilla.com/report/index/8e0aeb70-74d7-40e0-a102-fc5c42161017
Hey vonchen or Martin, if this bug was not present in 47 and is present in 48, perhaps we can fix it or understand it better if we manage to get a regression range. Could you try to get a regression range by using mozregression (you can download it, mozregression-gui.exe, from here: https://github.com/mozilla/mozregression/releases)? Thank you for your help!
sure i will try to bisect the issue with that tool. give me a couple of days please.
Hi Marco I am not really familiar with this tool so I do not know if this is what you needs, but anyway this is the result: http://i.imgur.com/AycogJd.jpg But,I am afraid this issue is randomly so even I mark it bad , but maybe next time it will runs fine. Crashed: app_name: firefox build_date: 2016-07-26 build_file: C:\Users\akato\.mozilla\mozregression\persist\2016-07-26--mozilla-central--firefox-50.0a1.en-US.win64.zip build_type: nightly build_url: https://archive.mozilla.org/pub/firefox/nightly/2016/07/2016-07-26-08-05-20-mozilla-central/firefox-50.0a1.en-US.win64.zip changeset: ceb63dec9267e9bb62f5e5e1f4c9d32d3ac1fbac pushlog_url: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=feaaf1af1065257b9178faca8b67eed9657b4a17&tochange=ceb63dec9267e9bb62f5e5e1f4c9d32d3ac1fbac repo_name: mozilla-central repo_url: https://hg.mozilla.org/mozilla-central hangs but finally load the video: app_name: firefox build_date: 2016-07-25 build_file: C:\Users\akato\.mozilla\mozregression\persist\2016-07-25--mozilla-central--firefox-50.0a1.en-US.win64.zip build_type: nightly build_url: https://archive.mozilla.org/pub/firefox/nightly/2016/07/2016-07-25-03-02-48-mozilla-central/firefox-50.0a1.en-US.win64.zip changeset: 7c669d5d63efceb12696cd65cfa72c296013dafb pushlog_url: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=e0bc88708ffed39aaab1fbc0ac461d93561195de&tochange=7c669d5d63efceb12696cd65cfa72c296013dafb repo_name: mozilla-central repo_url: https://hg.mozilla.org/mozilla-central I can reproduce the issue more often then other build by following(others is randomly and hard to find out) : https://ftp.mozilla.org/pub/firefox/releases/49.0.1/win64/en-US/Firefox%20Setup%2049.0.1.exe https://ftp.mozilla.org/pub/firefox/releases/50.0b8/win64/en-US/Firefox%20Setup%2050.0b8.exe New test video: https://www.youtube.com/watch?v=L1QY3uL1av0 New Test step: 1.launch fierfox with a clean profile 2.just past the video url and open the page This is what I got: http://i.imgur.com/PJip7QR.jpg But still randomly , sometime it loads. I maybe will keep testing but for now, the above are all. Hope it helps. I anyway now back to Firefox 47.0 for no issue watching DXVA video. If you needs info or I run into wrong testing process please tell me.
For some reason I re-upload the image at http://i.imgur.com/AycogJd.jpg New image is here: http://i.imgur.com/cU94o1J.jpg
(In reply to vonchen from comment #45) > Hi Marco > I am not really familiar with this tool so I do not know if this is what you > needs, but anyway this is the result: > http://i.imgur.com/AycogJd.jpg > > But,I am afraid this issue is randomly so even I mark it bad , but maybe > next time it will runs fine. To find the regression range, test a given build for a while. In theory the tool should tell you when the process is complete. The important information is when the failure started. So it's OK if you mark it bad when it works sometimes and doesn't work other times, the important thing is that you manage to mark as "bad" every build that fails (also if it doesn't *always* fail). Are you able to reproduce with or without the h264ify addon?
(In reply to Marco Castelluccio [:marco] from comment #47) > (In reply to vonchen from comment #45) > > Hi Marco > > I am not really familiar with this tool so I do not know if this is what you > > needs, but anyway this is the result: > > http://i.imgur.com/AycogJd.jpg > > > > But,I am afraid this issue is randomly so even I mark it bad , but maybe > > next time it will runs fine. > > To find the regression range, test a given build for a while. In theory the > tool should tell you when the process is complete. > The important information is when the failure started. So it's OK if you > mark it bad when it works sometimes and doesn't work other times, the > important thing is that you manage to mark as "bad" every build that fails > (also if it doesn't *always* fail). > > Are you able to reproduce with or without the h264ify addon? Hi Thanks for the answer. And yes, with this video: https://www.youtube.com/watch?v=L1QY3uL1av0 Firefox will just run it in DXVA mode which need not h264ify installed. And I reproduce the issue (only crashed page, not firefox crash) with the video. I reproduce it again in Fireofx 49.0.1 without any plugin(clean and new profile), this is the report: https://crash-stats.mozilla.com/report/index/fef52498-99a3-4119-934a-fecdc2161018#tab-details
Hi Marco Some new test result. I tested build 48 and marked all as bad. http://i.imgur.com/I79MRr5.jpg Test step: 1.Build a profile with 49.0.1 and install h264ify/Ublock/open some video page in it. 2.Use the profile while testing regression. 3.Play videos for about 1~3 min and then do switch to another video. The reason why I marked it as bad: At build 48 just like I mentioned before, it did not crash but freezes while switching form one dxva video to another video. If I see this behavior I mark it as bad. The last pushlog_url is: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=68d3781deda0d4d58ec9877862830db89669b3a5&tochange=7a6ad98de34e7374188231aec6a50eff87781913 Full detail: app_name: firefox build_date: 2016-03-08 07:40:24.188000 build_file: C:\Users\akato\.mozilla\mozregression\persist\7a6ad98de34e--mozilla-central--firefox-48.0a1.en-US.win64.zip build_type: inbound build_url: https://queue.taskcluster.net/v1/task/JnhKtphRRRKGofEWxfQYAw/runs/0/artifacts/public%2Fbuild%2Ffirefox-48.0a1.en-US.win64.zip changeset: 7a6ad98de34e7374188231aec6a50eff87781913 pushlog_url: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=68d3781deda0d4d58ec9877862830db89669b3a5&tochange=7a6ad98de34e7374188231aec6a50eff87781913 repo_name: mozilla-central repo_url: https://hg.mozilla.org/mozilla-central task_id: JnhKtphRRRKGofEWxfQYAw Hope this helps. I maybe will do another build test,if this did not help or you need some more info please tell me.
Thanks Vonchen, this means that Firefox 48 has always been affected. It might be difficult to find a regression range if it is an intermittent crash that just became more frequent.
Hi Marco This is new test result for Build47. I hope to find something even if it should be a good build which I know it is fine at release version. http://i.imgur.com/Vjfe2y2.jpg And I really found something weird. At 02-26/02-21/02-18 build I marked them bad because of the loading freeze behavior that the same as build48. The last pushlog_url for 0629918a is : https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=0629918a09ae87808efdda432d7852371ba37db6&tochange=1150ac4755c7bb35df4fc7504f6f1b6c257f400e I have no way to test 4d724d04 because of starting from 0629918a , the downloading speed is very slow. At build 4d724d04 it just stop downloading and I wait for a few minutes(5~10 min or more) still nothing happened. So I can not continue to test it. But since some build47 has bad mark, others not , so I hope this result can be a clue. I recorded a video to show how the loading freeze behaves: https://drive.google.com/open?id=0Bx6pDxKyikytQzRiWlB4WUVlRFU Hope this help.
Thanks, the regression range includes bug 1248496, which is expected given that we have these correlations with this signature: (99.62% in signature vs 06.05% overall) "DXVA2D3D11+" in app_notes = true (99.62% in signature vs 07.52% overall) "DXVA2D3D11?" in app_notes = true (87.90% in signature vs 01.83% overall) address = 0x4 (100.0% in signature vs 19.03% overall) adapter_vendor_id = Advanced Micro Devices, Inc. [AMD/ATI] (80.13% in signature vs 00.69% overall) adapter_device_id = 0x67df (99.87% in signature vs 25.39% overall) platform_pretty_version = Windows 10 (84.88% in signature vs 18.62% overall) platform_version = 10.0.14393 (100.0% in signature vs 36.83% overall) reason = EXCEPTION_ACCESS_VIOLATION_READ (91.55% in signature vs 36.67% overall) bios_manufacturer = American Megatrends Inc. (99.75% in signature vs 49.51% overall) "D2D1.1+" in app_notes = true (99.75% in signature vs 49.52% overall) "DWrite+" in app_notes = true (99.75% in signature vs 49.52% overall) "DWrite?" in app_notes = true (98.98% in signature vs 55.02% overall) os_arch = amd64 (00.30% in signature vs 40.59% overall) "D2D1.1-" in app_notes = true (42.12% in signature vs 00.65% overall) adapter_driver_version = 21.19.137.1 (00.64% in signature vs 39.54% overall) os_arch = x86 (40.64% in signature vs 00.47% overall) adapter_driver_version = 21.19.151.3 (46.92% in signature vs 19.47% overall) Addon "Adblock Plus" = true (25.39% in signature vs 00.94% overall) cpu_info = GenuineIntel family 6 model 94 stepping 3 | 4 (100.0% in signature vs 77.86% overall) shutdown_progress = null (100.0% in signature vs 78.66% overall) moz_crash_reason = null (48.24% in signature vs 25.89% overall) adapter_subsys_id = 00000000 (17.71% in signature vs 00.15% overall) adapter_subsys_id = 0b371002 Unfortunately, it's still unclear when the crash became more frequent, but I think it would be really hard to find that out with mozregression (and it's also possible that it is not a regression in Firefox, but in an AMD driver, or some other condition). Crashes with this signature became more frequent in August, which is when the Windows 10 Anniversary Update was released (and 84.88% of the crashes with this signature are with Windows 10 Anniversary Update). Perhaps this is the cause of the spike?
Hi Marco Thanks for the explanation. I would agree with your conclusion if I can reproduce the same issue on Chrome and Edge but I can not.But I am not saying I do not agree with you. As far as I know,for now AMD driver still has some issue related to DXVA(or video playing and Application) on RX400s and Still not fixed. Although they are maybe special conditions. DXVA corruption issue:(I did not really fact it on most of videos) https://community.amd.com/thread/205135 MadVR issue(I can avoid it by enable d3d11): https://community.amd.com/thread/206496 FAN issue: https://community.amd.com/thread/203282 You maybe wonder why GPU Fan going to 100% issue is DXVA related but anyway that's another story. And I have DXVA2 copy-back lag issue. The following is what I guessed. Maybe Firefox use DXVA in a way that will cause issue with AMD driver that they did not fix yet but other Browsers did not.Or they do some trick to avoid crash. I tested Chrome and Edge, they all also use D3D11 DXVA2, I find it out with a tool DXVAChecker: http://bluesky23.yukishigure.com/en/DXVAChecker.html So what's the different is what I want to figure out so maybe firefox can notice the point in the future. Since everybody use D3D11 DXVA2. And also, as far as I know , DXVA2 does not only mean DXVA2, it has two mode that is DXVA2-native and DXVA2-copy back which is LAV gives me the option. But I am not familiar with browser program coding and DXVA API, I am even not a programmer. So maybe browser's DXVA is not necessary the same as media player codec. Or they use different base engine. But from my view point they are all using the same standard which is HTML5 and DXVA API(I know there are different in detail depends on browser), but gives out different result in stability. Maybe next AMD driver will fix the issue(seriously I doubt) , but before that I wish to figure out the true reason why it crashed and other browsers not crash. That's why I am here. In fact I also reported this issue to AMD. I use firefox and it is hard for me to move to Chrome, although I use it too. Maybe it is a probability factor, but seriously whatever I have tested, Chrome and Edge is fine,at least I can not reproduce issue in a short time on them,and since I can reproduce the issue with clean firefox profile, I am afraid firefox add-on is not the factor. Just my two cents.
(In reply to vonchen from comment #40) > (In reply to Paul Blinzer from comment #38) > > (In reply to Milan Sreckovic [:milan] from comment #30) > > > Great, thanks Paul! > > > > > > Anything, short of blocking the driver versions, that we can avoid doing in > > > the meantime? > > As far as workarounds, it is a similar multithread race condition as for a > > few other previous bugs (which really should not happen as only one > > application thread should be in the driver at a time). It seems that Firefox > > is prone to run into such conditions. And this is again a Win10 specific bug > > apparently. > > > > On some other news, the driver fix I referred to should have made it into > > the Crimson 16.10.1 hotfix drivers released last week. Could you please > > confirm that the issue is now addressed with that driver? > Hi > As far as I know there are two version of 16.10.1 driver which one is > non-whql released at early October and one is whql released last > week(10/13).Which one do you mean? Or they are all the same? > > I tested non-whql 16.10.1 but the same issue occurred.I then roll-backed to > 16.9.2 and I did not test whql 16.10.1. > > The reason is after 16.10.1 I have some weird crash not only firefox, but > also 2d desktop works and gaming hard freezing that never happened from > 16.7.3~16.9.2 on my RX460. I am not sure if there is something wrong on my > pc or driver issue(It seems my OS hard driver have some lag issue I did some > trick to make to back to normal which is backup and recover hdd partition). > > But anyway back to 16.9.2 everything is fine again.(firefox issue still), so > I am waiting for next new driver. > > But if the whql 16.10.1 has the fix I maybe will try it , just want to make > sure which version do you mean. The two drivers are identical, main difference that in the time between these two posting it went through the formal WHQL certification process which is mostly a lot of bitwork (ex paperwork) but no change in the actual binary. For the repro or non-repro of this bug may not make much difference. But I am not certain based on your report if the fix made it into that release (need to track this through the process).
(In reply to Paul Blinzer from comment #54) > (In reply to vonchen from comment #40) > > (In reply to Paul Blinzer from comment #38) > > > (In reply to Milan Sreckovic [:milan] from comment #30) > > > > Great, thanks Paul! > > > > > > > > Anything, short of blocking the driver versions, that we can avoid doing in > > > > the meantime? > > > As far as workarounds, it is a similar multithread race condition as for a > > > few other previous bugs (which really should not happen as only one > > > application thread should be in the driver at a time). It seems that Firefox > > > is prone to run into such conditions. And this is again a Win10 specific bug > > > apparently. > > > > > > On some other news, the driver fix I referred to should have made it into > > > the Crimson 16.10.1 hotfix drivers released last week. Could you please > > > confirm that the issue is now addressed with that driver? > > Hi > > As far as I know there are two version of 16.10.1 driver which one is > > non-whql released at early October and one is whql released last > > week(10/13).Which one do you mean? Or they are all the same? > > > > I tested non-whql 16.10.1 but the same issue occurred.I then roll-backed to > > 16.9.2 and I did not test whql 16.10.1. > > > > The reason is after 16.10.1 I have some weird crash not only firefox, but > > also 2d desktop works and gaming hard freezing that never happened from > > 16.7.3~16.9.2 on my RX460. I am not sure if there is something wrong on my > > pc or driver issue(It seems my OS hard driver have some lag issue I did some > > trick to make to back to normal which is backup and recover hdd partition). > > > > But anyway back to 16.9.2 everything is fine again.(firefox issue still), so > > I am waiting for next new driver. > > > > But if the whql 16.10.1 has the fix I maybe will try it , just want to make > > sure which version do you mean. > > The two drivers are identical, main difference that in the time between > these two posting it went through the formal WHQL certification process > which is mostly a lot of bitwork (ex paperwork) but no change in the actual > binary. For the repro or non-repro of this bug may not make much difference. > But I am not certain based on your report if the fix made it into that > release (need to track this through the process). Hi Paul Thanks for the explanation and I will wait for next driver release. Martin also confirmed the issue still exists in 16.10.1 at comment 42: https://bugzilla.mozilla.org/show_bug.cgi?id=1299520#c42 And a reddit thread reply mentions he has less crash after 16.10.1: https://www.reddit.com/r/firefox/comments/55zpog/firefox_4901_constantly_crashing/ So this is my suppose, maybe a fix was included in 16.10.1 but not fully fix the issue, just low the frequency. The reason why I can find it out but others say it is better(only one time a day), maybe it is because of I always do some special stress testing it and use the browse very heavily. For example, while loading a video form youtube right-menu like this: https://www.youtube.com/watch?v=6Dakd7EIgBE&list=RD6Dakd7EIgBE#t=0 At the same time I quickly move to other webpage tab and click it.It has high chance to freeze Firefox after build 48.Try remove uBlock or adblock add-on can low the frequency but it will finally happened again. This issue does not happen every time, sometimes you want to test it but Firefox keeps doing fine. And when I though it is fine, next second it crashed. This behavior happens many time from my testing. But I can not reproduce it on Chrome and Edge. So if one only watches some video a day and then do other things they maybe never notice this issue. That's what I wonder.
Every amd driver release after 16.7.2 has firefox crashes from 48.x to 49.0.1. Since 16.9.x I have had glitches in chrome too. I had to roll back to 16.7.2 because of that and crashes in video playback with mpc-hc. Firefox and chrome has been rock-solid since installing 16.7.2 ever since. It's an amd bug for sure. 16.10.1 has not fixed it.
Hi Marco I finally can finish testing of Build47 today with regression tool. http://i.imgur.com/d8cw2lG.jpg The last bad pushlog_url: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=ec2ea269b3de08dfc9827971d1c62b70a39bf496&tochange=f2fa3f9f35445a0525ddceee2ebdba7278e01655 The last good pushlog_url: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=1911cab911de222bd1705fa752c845cfe2d02738&tochange=f2fa3f9f35445a0525ddceee2ebdba7278e01655 This testing only test the behavior of freeze firefox(no crash) while playing video and switch to another tab. Step: 1.Install ublock and h264ify 2.open a video in 1080P mode/open another webpage tab 3.while video is playing , click another video 4.before the video loading finish, quickly move mouse cursor to another tab and click it. If I see any freeze I mark it as bad. I have two video of release version of Firefox 47 and 49, to show how the different is: Firefox49: https://drive.google.com/open?id=0Bx6pDxKyikytUkp5bjJtYVBiR1U Firefox47: https://drive.google.com/open?id=0Bx6pDxKyikyteHZKczVaN0d2Wk0 Hope it helps.
(In reply to cookie1338 from comment #56) > Every amd driver release after 16.7.2 has firefox crashes from 48.x to > 49.0.1. Since 16.9.x I have had glitches in chrome too. I had to roll back > to 16.7.2 because of that and crashes in video playback with mpc-hc. Firefox > and chrome has been rock-solid since installing 16.7.2 ever since. It's an > amd bug for sure. 16.10.1 has not fixed it. Hi I have had display driver TDR issue at August with driver from 16.7.3~16.8.2 on watching youtube DXVA video, that means Chrome/Firefox/Edge all the same issue on Win10 64bit. On my RX460. So yes at that time it is AMD driver issue. For me from 16.9.1 it seems fixed the issue on RX460. So I am fine now on watching DXVA video with Chrome/Edge/Firefox. But after Firefox 48/49 update I have Firefox freeze / crash issue, back to Firefox 47.0.1 fix the issue. And I am using 16.9.2 now, I have system instability problem with 16.10.1. So I have crash issue with Firefox 49 and have freeze issue with Firefox 48. This is my condition.
Updated to 16.10.2 and Firefox 49.0.2 http://i.imgur.com/U3XwI6w.jpg This DXVA caused Firefox crash issue seems fixed. Firefox will load video finally. I will test it for more days to make sure. But (even not doing stressing test on it) I sometimes still have loading freeze issue I mentioned above whit 48/49. Sometimes I got a loading circle and video never load , I have to F5 the page to load it again. http://i.imgur.com/rr0qRQB.jpg So still Firefox 47 release version is still the smoothest version for my case on RX460.
Hey Vonchen, thank you for the info. According to your testing from comment 57, at some point Firefox 47 started failing, so it looks like Firefox 47 *was affected*. Could you try testing Firefox 47 downloading it from http://ftp.mozilla.org/pub/firefox/releases/47.0/? The testing you made in comment 57 was with the Nightly version of Firefox 47, which might have had some differences with the release version.
Hi Marco Ok I tested Firefox 47.0(release version) that downloaded from the URL you providing. It has no issue just like 47.0.1(release version). I record a video to show how it works: https://drive.google.com/open?id=0Bx6pDxKyikyteElYNVZaNXpteEk (This is with 16.10.2 driver, but I think since I still have freeze/lag issue with Firefox49.0.2 so I need not to go back 16.9.2 driver for double check.) Also I just found that 47(release version)/47.0.1(release version) use DXVA2 but not showing D3D11 in DXVAChecker's trace log: http://i.imgur.com/MY4LCer.jpg Starting from Firefox48.0(release version) trace log shows it is D3D11 and start the freeze or crash issue: http://i.imgur.com/SadcjPq.jpg Chrome(release version) also use D3D11 in trace log,Chrome is fine and never lag or freeze: http://i.imgur.com/WkGHwRR.jpg Microsoft Edge also use D3d11 and has no issue too: http://i.imgur.com/FXjSrvo.jpg The reason I can only test nightly build is mozregression-gui tool seems only download nightly build, I tried other options like this: http://i.imgur.com/SIE9pfY.jpg If I did not let the Repository be blank, it will just say unable to find enough data and do nothing. I am not sure if it is how it works or not. If there is any way I can test release build with mozregression-gui tool please tell me.
Hi Marco I finally find out this: media.windows-media-foundation.allow-d3d11-dxva The setting seems be set to true after build48 or build49. I just tested it. If I set media.windows-media-foundation.allow-d3d11-dxva to false, I can have the same smooth dxva playing as Build47 in Firefox 49(release version). Although disable it means no D3D11, only DXVA2. (Sorry I do not know how they are different, it seems DXVA2 does not mean D3D9 too. )
OK, everything seems to point to the enabling of D3D11 DXVA in bug 1248496. I suppose we have enabled it in Nightly 47 but it didn't ride the trains until Firefox 48. At least the crash should be fixed with the new driver version. Should we blocklist D3D11 DXVA on affected driver versions?
Flags: needinfo?(gsquelart)
(In reply to Marco Castelluccio [:marco] from comment #63) > OK, everything seems to point to the enabling of D3D11 DXVA in bug 1248496. > I suppose we have enabled it in Nightly 47 but it didn't ride the trains > until Firefox 48. > > At least the crash should be fixed with the new driver version. > > Should we blocklist D3D11 DXVA on affected driver versions? Marco, the smoothness issue when using D3D11 DXVA is surprising as the video pipeline should be more efficient and it is with other applications and AMD drivers using D3D11 Video. I assume that video path is used with other drivers too? Any similar issues reported? Could this be some maturity issue using D3D11 Video pipeline? Has anyone repro'd and profiled this case already?
See Also: → 1292273
Thanks for the pointer. Suggest to look at the D3D11 video path in Firefox, profile and optimize it thoroughly. But I guess that requires a new ticket and allows this one to be closed as 16.10.2 is now confirmed fixing the original problem.
Attached patch bug1299520.patch — — Splinter Review
(In reply to Marco Castelluccio [:marco] from comment #63) > OK, everything seems to point to the enabling of D3D11 DXVA in bug 1248496. > I suppose we have enabled it in Nightly 47 but it didn't ride the trains > until Firefox 48. > > At least the crash should be fixed with the new driver version. > > Should we blocklist D3D11 DXVA on affected driver versions? i have added the most affected driver versions to media.wmf.disable-d3d11-for-dlls in the attached patch if we want to go for that.
Maybe we should just disable it on all versions < 16.10.2. With release there might be driver versions that we don't see with Beta.
(In reply to Marco Castelluccio [:marco] from comment #67) > Maybe we should just disable it on all versions < 16.10.2. With release > there might be driver versions that we don't see with Beta. Marco, just to confirm: the "disable" would be a fallback to DX9 DVXVA, but not to SW render, correct?
yes, that would be through the mechanism added in bug 1273691
(In reply to Paul Blinzer from comment #68) > (In reply to Marco Castelluccio [:marco] from comment #67) > > Maybe we should just disable it on all versions < 16.10.2. With release > > there might be driver versions that we don't see with Beta. > > Marco, just to confirm: the "disable" would be a fallback to DX9 DVXVA, but > not to SW render, correct? Yes, at least until we can actually fix the bug with D3D11. (In reply to Marco Castelluccio [:marco] from comment #67) > Maybe we should just disable it on all versions < 16.10.2. With release > there might be driver versions that we don't see with Beta. I got confused with the other bugs that we're seeing on Beta with NVIDIA, forget it :)
(In reply to [:philipp] from comment #69) > yes, that would be through the mechanism added in bug 1273691 Good, so I think one can close out this ticket after vonchen's info on the freeze/lag has been transferred and is investigated separately.
Attachment #8803426 - Flags: review?(gsquelart)
Anthony, there are some suggestions as to blocking D3D11 DXVA (down to D3D9 DXVA) in the previous few comments, what is the right thing to do?
Component: Graphics → Audio/Video
Flags: needinfo?(milan) → needinfo?(ajones)
Keywords: regression
The patch looks fine, it's only adding a few driver versions to our D3D11 block-list (with fallback to D3D9). I'll wait for Anthony's response before giving a final r+.
Flags: needinfo?(gsquelart)
> Anthony, there are some suggestions as to blocking D3D11 DXVA (down to D3D9 > DXVA) in the previous few comments, what is the right thing to do? Sounds reasonable to me. There is a performance penalty because D3D9 uses a readback to synchronise but seems better than a crash.
Flags: needinfo?(ajones)
Assignee: nobody → madperson
Comment on attachment 8803426 [details] [diff] [review] bug1299520.patch Review of attachment 8803426 [details] [diff] [review]: ----------------------------------------------------------------- r+ it is, then. Thank you for doing this.
Attachment #8803426 - Flags: review?(gsquelart) → review+
Keywords: checkin-needed
Comment on attachment 8803426 [details] [diff] [review] bug1299520.patch Approval Request Comment [Feature/regressing bug #]: enabling D3D11 DXVA, bug 1248496 [User impact if declined]: users with an AMD gpu and driver versions below 16.10.2 might experience crashes during media playback (the crash signature is currently accounting for 0.64% of browser crashes on release) [Describe test coverage new/current, TreeHerder]: affected user vonchen confirmed in comment #62 that manually disabling media.windows-media-foundation.allow-d3d11-dxva fixes the freeze/crash during playback [Risks and why]: low, it is making use of the purpose-built preference to fall back to D3D9 DXVA playback, which was the default state pre-48 [String/UUID change made/needed]: none
Attachment #8803426 - Flags: approval-mozilla-beta?
Attachment #8803426 - Flags: approval-mozilla-aurora?
Comment on attachment 8803426 [details] [diff] [review] bug1299520.patch Fixes a top crash, Aurora51+, Beta50+
Attachment #8803426 - Flags: approval-mozilla-beta?
Attachment #8803426 - Flags: approval-mozilla-beta+
Attachment #8803426 - Flags: approval-mozilla-aurora?
Attachment #8803426 - Flags: approval-mozilla-aurora+
If all goes well, this should be in 50.0b10.
tracking-firefox50: ? → +
(In reply to Wes Kocher (:KWierso) from comment #78) > Does this still need to land on trunk? Yes.
Pushed by kwierso@gmail.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/853e2a69fd10 Add more affected AMD driver versions to d3d11 dxva blacklist. r=gerald
Keywords: checkin-needed
See Also: → 1312595
https://hg.mozilla.org/mozilla-central/rev/853e2a69fd10
Status: NEW → RESOLVED
Closed: 9 years ago
status-firefox52: affected → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla52
hi, although the patch has landed in 50.0b10 there are still crashes there with the driver versions that were supposed to fallback to dxva d3d9 through the pref change (also repeatedly for some users): https://crash-stats.mozilla.com/search/?signature=%3DNDXGI%3A%3ACDevice%3A%3ASignalSynchronizationObjectCB&product=Firefox&version=50.0b10&process_type=browser&_facets=signature&_facets=user_comments&_facets=platform_pretty_version&_facets=install_time&_facets=adapter_driver_version&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#facet-adapter_driver_version those reports also show "DXVA2D3D11+" in their app notes... do you see anything obvious why the preference doesn't seem to do its job there? thank you (i also manually confirmed on a local installation that the new expanded driver versions are present in media.wmf.disable-d3d11-for-dlls after the 50.0b10 update)
Flags: needinfo?(gsquelart)
I think the issue in this case is that we're checking for the existance of these files in C:\Windows\System32, but the file for the people that are submitting these crashes is at different paths (e.g. in one of the reports at C:\Windows\System32\DriverStore\FileRepository\c0306790.inf_amd64_b82087329794b4e0\atidxx32.dll). We could check the version of the file that is loaded in our process, instead of trying to find it in C:\Windows\System32.
Depends on: 1313339
Crash Signature: [@ NDXGI::CDevice::SignalSynchronizationObjectCB] → [@ NDXGI::CDevice::SignalSynchronizationObjectCB] [@ d3d11.dll@0xd3fb0 ]
why not reopening then?(In reply to Marco Castelluccio [:marco] from comment #84) > I think the issue in this case is that we're checking for the existance of > these files in C:\Windows\System32, but the file for the people that are > submitting these crashes is at different paths (e.g. in one of the reports > at > C:\Windows\System32\DriverStore\FileRepository\c0306790. > inf_amd64_b82087329794b4e0\atidxx32.dll). > > We could check the version of the file that is loaded in our process, > instead of trying to find it in C:\Windows\System32. Shouldn't this bug be reopened?
bug 1313339 is dealing with the follow-up.
This is still a top crasher in the "content" list, currently #4, maybe because of bug 1313339.
(In reply to Marco Castelluccio [:marco] from comment #87) > This is still a top crasher in the "content" list, currently #4, maybe > because of bug 1313339. Looks that all the crashes are with quite old driver versions, before the fix has been introduced (two major driver releases ago). I wonder if it would be possible to nudge people affected to download a newer driver either from Windows Update or from www.amd.com.
It looks like bug 1313339 fixed the problem. The last crash with Aurora is with build ID 20161230004019 [1] (the first Aurora build ID that contains the change from bug 1313339 is 20161231004004) and there are no crashes with 51.0b11 [2]. [1]: https://crash-stats.mozilla.com/search/?signature=%3DNDXGI%3A%3ACDevice%3A%3ASignalSynchronizationObjectCB&adapter_driver_version=%3D21.19.137.1&product=Firefox&version=52.0a2&date=%3E%3D2016-12-30T17%3A35%3A17.000Z&date=%3C2017-01-06T17%3A35%3A17.000Z&_sort=-date&_facets=signature&_facets=version&_facets=build_id&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=build_id&_columns=platform#facet-build_id [2]: https://crash-stats.mozilla.com/search/?signature=%3DNDXGI%3A%3ACDevice%3A%3ASignalSynchronizationObjectCB&adapter_driver_version=%3D21.19.137.1&product=Firefox&version=51.0b11&version=51.0b10&date=%3E%3D2016-12-30T17%3A35%3A11.000Z&date=%3C2017-01-06T17%3A35%3A11.000Z&_sort=-date&_facets=version&_columns=date&_columns=signature&_columns=product&_columns=version&_columns=platform#facet-version (In reply to Paul Blinzer from comment #88) > Looks that all the crashes are with quite old driver versions, before the > fix has been introduced (two major driver releases ago). I wonder if it > would be possible to nudge people affected to download a newer driver either > from Windows Update or from www.amd.com. This is a good idea, not only for this case but in general (e.g. we often have crashes due to malware or antivirus software, we could suggest the user what to do). I will file a bug to figure out how to do this with the best possible UI.
Flags: needinfo?(gsquelart)
Crash Signature: [@ NDXGI::CDevice::SignalSynchronizationObjectCB] [@ d3d11.dll@0xd3fb0 ] → [@ NDXGI::CDevice::SignalSynchronizationObjectCB] [@ d3d11.dll@0xd3fb0 ] [@ d3d11.dll@0xce99f ] [@ d3d11.dll@0xd474e ]
Blocks: 1334561
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: