Closed Bug 1194654 Opened 9 years ago Closed 9 years ago

Firefox win64 startup crash [40+]

Categories

(Core :: Graphics, defect)

Product:
Core
Component:
Graphics
Version:
40 Branch
Type:
defect
Priority:
Not set
Severity:
critical

Tracking

()

Status:
RESOLVED WORKSFORME

People

(Reporter: toadyshadow101, Unassigned)

References

Details

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

Attachments

(1 file)

Vsync Tests
19.06 KB, application/x-zip-compressed
Details
User Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:40.0) Gecko/20100101 Firefox/40.0 Build ID: 20150813140032 Steps to reproduce: Used Firefox 40 beta x64, Firefox nightly x64 on windows 10 Actual results: Previously working, From 40.0 release up to the latest nightly x64bit Firefox ceases to run. Firefox will not start and have to terminate the process in task manager. Expected results: Firefox should have run. On a clean profile its initial start Firefox x64 will run until you close Firefox then wait 1 minute then start it again the process will just hang. If you navigate to the profile and delete extensions.json and the lock file after terminating the process in task manager, Firefox will run again until next time having to repeat the process. When you first start windows Firefox x64 will let you run until to close and try to open it again. Firefox used to work until possible windows updates, Resetting the computer back to factory has no effect. Running moz regression: Last working: app_name: firefox build_date: 2015-03-24 build_path: c:\users\DevPC-A03\appdata\local\temp\tmpaur39o\2015-03-24--mozilla-central--firefox-39.0a1.en-US.win64.zip build_txt_url: http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/2015/03/2015-03-24-03-02-07-mozilla-central/firefox-39.0a1.en-US.win64.txt build_type: nightly build_url: http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/2015/03/2015-03-24-03-02-07-mozilla-central/firefox-39.0a1.en-US.win64.zip changeset: 840cfd5bc971 pushlog_url: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=82ae3b4e2215&tochange=37d3dcbf23a9 repo: mozilla-central repository: https://hg.mozilla.org/mozilla-central Issue is present to current nightly: app_name: firefox build_date: 2015-03-25 build_path: c:\users\DevPC-A03\appdata\local\temp\tmpaur39o\2015-03-25--mozilla-central--firefox-39.0a1.en-US.win64.zip build_txt_url: http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/2015/03/2015-03-25-08-30-46-mozilla-central/firefox-39.0a1.en-US.win64.txt build_type: nightly build_url: http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/2015/03/2015-03-25-08-30-46-mozilla-central/firefox-39.0a1.en-US.win64.zip changeset: db0409de517a pushlog_url: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=840cfd5bc971&tochange=db0409de517a repo: mozilla-central repository: https://hg.mozilla.org/mozilla-central About:support information: { "application": { "name": "Firefox", "version": "43.0a1", "buildID": "20150813030208", "userAgent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:43.0) Gecko/20100101 Firefox/43.0", "safeMode": false, "updateChannel": "nightly", "supportURL": "https://support.mozilla.org/1/firefox/43.0a1/WINNT/en-US/", "numTotalWindows": 1, "numRemoteWindows": 1, "remoteAutoStart": true }, "modifiedPreferences": { "browser.cache.disk.smart_size.first_run": false, "browser.cache.disk.capacity": 1048576, "browser.cache.disk.filesystem_reported": 1, "browser.cache.frecency_experiment": 1, "browser.download.importedFromSqlite": true, "browser.places.smartBookmarksVersion": 7, "browser.startup.homepage_override.mstone": "43.0a1", "browser.startup.homepage_override.buildID": "20150813030208", "dom.apps.reset-permissions": true, "dom.mozApps.used": true, "extensions.lastAppVersion": "43.0a1", "gfx.crash-guard.d3d11layers.appVersion": "43.0a1", "gfx.direct3d.last_used_feature_level_idx": 0, "gfx.crash-guard.d3d11layers.feature-d3d11": true, "gfx.crash-guard.d3d11layers.feature-d2d": true, "gfx.crash-guard.d3d11layers.driverVersion": "10.18.15.4256", "gfx.crash-guard.d3d11layers.deviceID": "0x0a16", "gfx.crash-guard.status.d3d9video": 2, "gfx.crash-guard.status.d3d11layers": 2, "media.hardware-video-decoding.failed": false, "network.cookie.prefsMigrated": true, "places.history.expiration.transient_current_max_pages": 104858, "plugin.disable_full_page_plugin_for_types": "application/pdf", "privacy.sanitize.migrateFx3Prefs": true, "security.sandbox.content.tempDirSuffix": "{ecf09f9c-63b4-4e42-8c15-5b36a0acd5be}" }, "lockedPreferences": {}, "graphics": { "numTotalWindows": 1, "numAcceleratedWindows": 1, "windowLayerManagerType": "Direct3D 11", "windowLayerManagerRemote": true, "supportsHardwareH264": true, "adapterDescription": "Intel(R) HD Graphics Family", "adapterVendorID": "0x8086", "adapterDeviceID": "0x0a16", "adapterSubsysID": "08661025", "adapterRAM": "Unknown", "adapterDrivers": "igdumdim64 igd10iumd64 igd10iumd64 igd12umd64 igdumdim32 igd10iumd32 igd10iumd32 igd12umd32", "driverVersion": "10.18.15.4256", "driverDate": "7-17-2015", "adapterDescription2": "", "adapterVendorID2": "", "adapterDeviceID2": "", "adapterSubsysID2": "", "adapterRAM2": "", "adapterDrivers2": "", "driverVersion2": "", "driverDate2": "", "isGPU2Active": false, "direct2DEnabled": true, "directWriteEnabled": true, "directWriteVersion": "10.0.10240.16430", "webglRenderer": "Google Inc. -- ANGLE (Intel(R) HD Graphics Family Direct3D11 vs_5_0 ps_5_0)", "info": { "AzureCanvasBackend": "direct2d 1.1", "AzureSkiaAccelerated": 0, "AzureFallbackCanvasBackend": "cairo", "AzureContentBackend": "direct2d 1.1", "ApzWheelInput": 1 } }, "javaScript": { "incrementalGCEnabled": true }, "accessibility": { "isActive": false, "forceDisabled": 0 }, "libraryVersions": { "NSPR": { "minVersion": "4.10.9 Beta", "version": "4.10.9 Beta" }, "NSS": { "minVersion": "3.19.3 Basic ECC", "version": "3.19.3 Basic ECC" }, "NSSUTIL": { "minVersion": "3.19.3", "version": "3.19.3" }, "NSSSSL": { "minVersion": "3.19.3 Basic ECC", "version": "3.19.3 Basic ECC" }, "NSSSMIME": { "minVersion": "3.19.3 Basic ECC", "version": "3.19.3 Basic ECC" } }, "userJS": { "exists": false }, "crashes": { "submitted": [], "pending": 0 }, "extensions": [], "experiments": [] } Also occurs in windows 8|8.1 On another windows 10 install with the latest windows updates all x64bit work, Swapping out the HDD of the computer the issue occurs on and installing a SDD with a clean install of windows 10 and drivers produces the same issue. Is there any further information i can provide please let me know thanks.
Flags: needinfo?(jmuizelaar)
If you can a mozilla-inbound regression window that would be much more helpful.
Flags: needinfo?(jmuizelaar)
Hi Jeff As requested. bad app_name: firefox build_path: c:\users\peter\appdata\local\temp\tmpbhhgnc\43e18ef66807--mozilla-inbound--firefox-39.0a1.en-US.win64.zip build_txt_url: https://queue.taskcluster.net/v1/task/6E2E7hlBQEuQ7OdjJ9lFAg/artifacts/public/build/firefox-39.0a1.en-US.win64.txt build_type: inbound build_url: https://queue.taskcluster.net/v1/task/6E2E7hlBQEuQ7OdjJ9lFAg/artifacts/public/build/firefox-39.0a1.en-US.win64.zip changeset: 43e18ef66807 pushlog_url: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=f0a4844f0ccd&tochange=43e18ef66807 repo: mozilla-inbound repository: https://hg.mozilla.org/integration/mozilla-inbound revision: 43e18ef6 timestamp: 1427247698 Last good app_name: firefox build_path: c:\users\peter\appdata\local\temp\tmpbhhgnc\43ef69c1f6b5--mozilla-inbound--firefox-39.0a1.en-US.win64.zip build_txt_url: https://queue.taskcluster.net/v1/task/_ABRSXeORs2tSbz9yUwTyg/artifacts/public/build/firefox-39.0a1.en-US.win64.txt build_type: inbound build_url: https://queue.taskcluster.net/v1/task/_ABRSXeORs2tSbz9yUwTyg/artifacts/public/build/firefox-39.0a1.en-US.win64.zip changeset: 43ef69c1f6b5 pushlog_url: https://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=43ef69c1f6b5&tochange=43e18ef66807 repo: mozilla-inbound repository: https://hg.mozilla.org/integration/mozilla-inbound revision: 43ef69c1 timestamp: 1427247348
Flags: needinfo?(jmuizelaar)
Blocks: 1144317
Severity: normal → critical
Component: Untriaged → Graphics
Keywords: crash
Product: Firefox → Core
Can you go to about:crashes and link us to the crash report please?
Flags: needinfo?(toadyshadow101)
(In reply to Mason Chang [:mchang] from comment #3) > Can you go to about:crashes and link us to the crash report please? Hi Mason Unfortunately no reports are generated as crash reporter does no trigger, The process hangs and must be terminated though task manager. Toggling the preferences for vsync (Silk) and also hardware acceleration has no effect.
Flags: needinfo?(toadyshadow101)
I'm having the same problems described here, except I'm running Windows 7 x64. Forgive me if this is not the correct information to share - I'm illiterate to these things: { "application": { "name": "Firefox", "version": "39.0a1", "buildID": "20150324184446", "userAgent": "Mozilla/5.0 (Windows NT 6.1; Win64; x64; rv:39.0) Gecko/20100101 Firefox/39.0", "updateChannel": "default", "supportURL": "https://support.mozilla.org/1/firefox/39.0a1/WINNT/en-US/", "numTotalWindows": 1, "numRemoteWindows": 1, "remoteAutoStart": true }, "crashes": { "submitted": [], "pending": 0 }, "modifiedPreferences": { "accessibility.typeaheadfind.flashBar": 0, "browser.cache.disk.smart_size.first_run": false, "browser.cache.disk.smart_size.use_old_max": false, "browser.cache.disk.capacity": 358400, "browser.cache.disk.filesystem_reported": 1, "browser.cache.memory.capacity": 65536, "browser.cache.disk.smart_size_cached_value": 358400, "browser.cache.frecency_experiment": 4, "browser.display.show_image_placeholders": true, "browser.download.folderList": 2, "browser.download.importedFromSqlite": true, "browser.download.manager.alertOnEXEOpen": true, "browser.download.manager.showAlertOnComplete": false, "browser.download.manager.showWhenStarting": false, "browser.link.open_newwindow.restriction": 0, "browser.places.smartBookmarksVersion": 7, "browser.search.openintab": true, "browser.sessionstore.upgradeBackup.latestBuildID": "20150324184446", "browser.sessionstore.resume_from_crash": false, "browser.startup.homepage_override.mstone": "39.0a1", "browser.startup.homepage_override.buildID": "20150324184446", "browser.tabs.onTop": false, "browser.tabs.insertRelatedAfterCurrent.backup": true, "browser.tabs.warnOnClose": false, "browser.tabs.insertRelatedAfterCurrent": false, "browser.tabs.animate": false, "browser.urlbar.default.behavior": 32, "browser.urlbar.autocomplete.enabled": true, "browser.urlbar.suggest.history.onlyTyped": true, "dom.apps.reset-permissions": true, "dom.mozApps.used": true, "extensions.lastAppVersion": "39.0a1", "font.internaluseonly.changed": false, "gfx.direct3d.last_used_feature_level_idx": 0, "gfx.direct2d.disabled": true, "layers.acceleration.disabled": true, "media.gmp-gmpopenh264.lastUpdate": 1431623228, "media.gmp-gmpopenh264.version": "1.4", "media.hardware-video-decoding.failed": true, "media.gmp-manager.lastCheck": 1439742275, "media.gmp-manager.buildID": "20150816030217", "network.cookie.prefsMigrated": true, "network.http.max-persistent-connections-per-server": 8, "network.http.request.max-start-delay": 0, "network.http.max-connections": 48, "network.predictor.cleaned-up": true, "network.http.pipelining.maxrequests": 8, "network.http.max-connections-per-server": 16, "network.http.max-persistent-connections-per-proxy": 16, "places.database.lastMaintenance": 1439330480, "places.history.expiration.transient_current_max_pages": 104858, "plugin.importedState": true, "plugin.state.npdeployjava": 0, "plugin.disable_full_page_plugin_for_types": "application/pdf", "plugin.expose_full_path": true, "privacy.sanitize.migrateFx3Prefs": true, "privacy.donottrackheader.enabled": true, "security.sandbox.content.tempDirSuffix": "{68267743-711d-47a3-a4b2-240b168d5da2}", "storage.vacuum.last.places.sqlite": 1439330480, "storage.vacuum.last.index": 1 }, "lockedPreferences": {}, "graphics": { "numTotalWindows": 1, "numAcceleratedWindows": 0, "windowLayerManagerType": "Basic", "windowLayerManagerRemote": true, "numAcceleratedWindowsMessage": [ "" ], "adapterDescription": "AMD Radeon (TM) R9 200 Series", "adapterVendorID": "0x1002", "adapterDeviceID": "0x6811", "adapterSubsysID": "048d1043", "adapterRAM": "2048", "adapterDrivers": "aticfx64 aticfx64 aticfx64 aticfx32 aticfx32 aticfx32 atiumd64 atidxx64 atidxx64 atiumdag atidxx32 atidxx32 atiumdva atiumd6a atitmm64", "driverVersion": "15.200.1062.1004", "driverDate": "8-3-2015", "adapterDescription2": "", "adapterVendorID2": "", "adapterDeviceID2": "", "adapterSubsysID2": "", "adapterRAM2": "", "adapterDrivers2": "", "driverVersion2": "", "driverDate2": "", "isGPU2Active": false, "direct2DEnabled": false, "directWriteEnabled": false, "directWriteVersion": "6.2.9200.17461", "direct2DEnabledMessage": [ "" ], "webglRenderer": "Google Inc. -- ANGLE (AMD Radeon (TM) R9 200 Series Direct3D11 vs_5_0 ps_5_0)", "info": { "AzureCanvasBackend": "skia", "AzureSkiaAccelerated": 0, "AzureFallbackCanvasBackend": "cairo", "AzureContentBackend": "cairo", "ApzWheelInput": 1 } }, "javaScript": { "incrementalGCEnabled": true }, "accessibility": { "isActive": false, "forceDisabled": 0 }, "libraryVersions": { "NSPR": { "minVersion": "4.10.8", "version": "4.10.8" }, "NSS": { "minVersion": "3.18 Basic ECC", "version": "3.18 Basic ECC" }, "NSSUTIL": { "minVersion": "3.18", "version": "3.18" }, "NSSSSL": { "minVersion": "3.18 Basic ECC", "version": "3.18 Basic ECC" }, "NSSSMIME": { "minVersion": "3.18 Basic ECC", "version": "3.18 Basic ECC" } }, "userJS": { "exists": true }, "extensions": [], "experiments": [] }
From comment 4, it looks like this is non-silk related. However, it is odd that the inbound regression shows silk...
From comment 2, the last good was 43ef69c1f6b5, which includes the refresh driver. I'm not quite sure how to reconcile that with the first bad being 43e18ef66807, which includes a merge plus the refresh driver? Does that mean something else in that merge broke it?
(In reply to Mason Chang [:mchang] from comment #7) > From comment 2, the last good was 43ef69c1f6b5, which includes the refresh > driver. I'm not quite sure how to reconcile that with the first bad being > 43e18ef66807, which includes a merge plus the refresh driver? Does that mean > something else in that merge broke it? After much testing using a user.js with user_pref("gfx.vsync.refreshdriver", false); user_pref("gfx.vsync.hw-vsync.enabled", false); user_pref("gfx.vsync.compositor", false); Has resolved it on most except for the latest nightly i will have to run mozregression to find where it regresses further. I know of other people who have encountered this issue and toggling off pref("gfx.vsync.refreshdriver", false); pref("gfx.vsync.hw-vsync.enabled", false); pref("gfx.vsync.compositor", false); appears to work for them, Is there any way to get a scale on how many this has affected since crash reporter does not trigger ? @Mason Chang Are there any test builds you would like me to try on a afflicted machine?
Flags: needinfo?(jmuizelaar)
(In reply to Toady from comment #8) > (In reply to Mason Chang [:mchang] from comment #7) > > From comment 2, the last good was 43ef69c1f6b5, which includes the refresh > > driver. I'm not quite sure how to reconcile that with the first bad being > > 43e18ef66807, which includes a merge plus the refresh driver? Does that mean > > something else in that merge broke it? > > After much testing using a user.js with > user_pref("gfx.vsync.refreshdriver", false); > user_pref("gfx.vsync.hw-vsync.enabled", false); > user_pref("gfx.vsync.compositor", false); > > Has resolved it on most except for the latest nightly i will have to run > mozregression to find where > it regresses further. > > I know of other people who have encountered this issue and toggling off > pref("gfx.vsync.refreshdriver", false); > pref("gfx.vsync.hw-vsync.enabled", false); > pref("gfx.vsync.compositor", false); > appears to work for them, Is there any way to get a scale on how many this > has affected since > crash reporter does not trigger ? > > @Mason Chang Are there any test builds you would like me to try on a > afflicted machine? Is there anyway you can revert to Firefox 39? You can download an old version here: https://ftp.mozilla.org/pub/mozilla.org/firefox/releases/39.0.3/ The vsync preferences should be disabled then. Can you go to about:config on Firefox 39, set the gfx.vsync.* preferences to true and restart Firefox 39? Are you still crashing then? Thanks!
Flags: needinfo?(toadyshadow101)
(In reply to Mason Chang [:mchang] from comment #9) > (In reply to Toady from comment #8) > > (In reply to Mason Chang [:mchang] from comment #7) > > > From comment 2, the last good was 43ef69c1f6b5, which includes the refresh > > > driver. I'm not quite sure how to reconcile that with the first bad being > > > 43e18ef66807, which includes a merge plus the refresh driver? Does that mean > > > something else in that merge broke it? > > > > After much testing using a user.js with > > user_pref("gfx.vsync.refreshdriver", false); > > user_pref("gfx.vsync.hw-vsync.enabled", false); > > user_pref("gfx.vsync.compositor", false); > > > > Has resolved it on most except for the latest nightly i will have to run > > mozregression to find where > > it regresses further. > > > > I know of other people who have encountered this issue and toggling off > > pref("gfx.vsync.refreshdriver", false); > > pref("gfx.vsync.hw-vsync.enabled", false); > > pref("gfx.vsync.compositor", false); > > appears to work for them, Is there any way to get a scale on how many this > > has affected since > > crash reporter does not trigger ? > > > > @Mason Chang Are there any test builds you would like me to try on a > > afflicted machine? > > Is there anyway you can revert to Firefox 39? You can download an old > version here: > > https://ftp.mozilla.org/pub/mozilla.org/firefox/releases/39.0.3/ > > The vsync preferences should be disabled then. Can you go to about:config on > Firefox 39, set the gfx.vsync.* preferences to true and restart Firefox 39? > Are you still crashing then? > > Thanks! When enabling gfx.vsync.* preferences and restarting Firefox build https://ftp.mozilla.org/pub/mozilla.org/firefox/releases/39.0.3/win32/en-US/Firefox%20Setup%2039.0.3.exe The application interface lagged and page content took 30+ seconds to render if it rendered, Firefox did not crash or hang but was unusable. When doing the same with https://ftp.mozilla.org/pub/mozilla.org/firefox/releases/39.0b7/win64/en-US/Firefox%20Setup%2039.0b7.exe The application has the symptoms described in the above posts. Would you like me to make a video, Would this help by giving you a visual ?
Flags: needinfo?(toadyshadow101)
> When enabling gfx.vsync.* preferences and restarting Firefox build > https://ftp.mozilla.org/pub/mozilla.org/firefox/releases/39.0.3/win32/en-US/ > Firefox%20Setup%2039.0.3.exe > > The application interface lagged and page content took 30+ seconds to render > if it rendered, Firefox did not crash or hang but was unusable. Ahh this looks like bug 1147953, which wasn't uplifted to Gecko 39. Sorry about that! If you install a clean Firefox Beta (https://www.mozilla.org/en-US/firefox/channel/) or developer edition, does the crash still happen with the gfx.vsync.* preferences set to true? Also, can you make sure you use a clean profile? Can you also try launching firefox in safe mode? http://kb.mozillazine.org/Safe_Mode#Windows Does that resolve the issue for you?
Flags: needinfo?(toadyshadow101)
(In reply to Mason Chang [:mchang] from comment #11) > > When enabling gfx.vsync.* preferences and restarting Firefox build > > https://ftp.mozilla.org/pub/mozilla.org/firefox/releases/39.0.3/win32/en-US/ > > Firefox%20Setup%2039.0.3.exe > > > > The application interface lagged and page content took 30+ seconds to render > > if it rendered, Firefox did not crash or hang but was unusable. > > Ahh this looks like bug 1147953, which wasn't uplifted to Gecko 39. Sorry > about that! If you install a clean Firefox Beta > (https://www.mozilla.org/en-US/firefox/channel/) or developer edition, does > the crash still happen with the gfx.vsync.* preferences set to true? Also, > can you make sure you use a clean profile? > > Can you also try launching firefox in safe mode? > http://kb.mozillazine.org/Safe_Mode#Windows > Does that resolve the issue for you? Hi Mason Chang Unfortunately the x64bit still has the start hang. 40.0.2 x64 toggling gfx.vsync.* preferences set to false, The no longer exists. 41.0b4 x64 toggling gfx.vsync.* preferences set to false, Hangs regardless 42.0a2 x64 toggling gfx.vsync.* preferences set to false, Hangs regardless 43.0a1 x64 toggling gfx.vsync. preference set to false, Hangs regardless For me the x64 versions worked as intended. The builds affected by the bug, Won't start in safe mode as it hang during start up likely before safe mode is initialized. After 40.0.2 x64 in 41 and up there has been a further regression in that the toggling of the preference no longer work. Are you able to block list the driver for vsync? like you do with HWA
Flags: needinfo?(toadyshadow101) → needinfo?(mchang)
> > 40.0.2 x64 toggling gfx.vsync.* preferences set to false, The no longer > exists. > 41.0b4 x64 toggling gfx.vsync.* preferences set to false, Hangs regardless > 42.0a2 x64 toggling gfx.vsync.* preferences set to false, Hangs regardless > 43.0a1 x64 toggling gfx.vsync. preference set to false, Hangs regardless > For me the x64 versions worked as intended. What do you mean by the x64 versions worked as intended? What do you mean by hangs regardless? As in, nothing paints or firefox crashes on startup? Do the 32 bit builds work for you? > > The builds affected by the bug, Won't start in safe mode as it hang during > start up likely before safe mode is initialized. > > After 40.0.2 x64 in 41 and up there has been a further regression in that > the toggling of the preference no longer work. > > Are you able to block list the driver for vsync? like you do with HWA If the preferences are off, it won't matter if we block the driver for vsync. In this case, we don't even use vsync, so something else is broken.
Flags: needinfo?(mchang) → needinfo?(toadyshadow101)
(In reply to Mason Chang [:mchang] from comment #13) > > > > 40.0.2 x64 toggling gfx.vsync.* preferences set to false, The no longer > > exists. > > 41.0b4 x64 toggling gfx.vsync.* preferences set to false, Hangs regardless > > 42.0a2 x64 toggling gfx.vsync.* preferences set to false, Hangs regardless > > 43.0a1 x64 toggling gfx.vsync. preference set to false, Hangs regardless > > For me the x64 versions worked as intended. > > What do you mean by the x64 versions worked as intended? What do you mean by > hangs regardless? As in, nothing paints or firefox crashes on startup? Do > the 32 bit builds work for you? > > > > > The builds affected by the bug, Won't start in safe mode as it hang during > > start up likely before safe mode is initialized. > > > > After 40.0.2 x64 in 41 and up there has been a further regression in that > > the toggling of the preference no longer work. > > > > Are you able to block list the driver for vsync? like you do with HWA > > If the preferences are off, it won't matter if we block the driver for > vsync. In this case, we don't even use vsync, so something else is broken. Sorry {For me the x64 versions worked as intended.} typo, It was meant to be x86 Enabling vsync in 39 https://bugzilla.mozilla.org/show_bug.cgi?id=1194654#c10 did affect x86 41+ to latest nightly x86 works perfectly no issues. > > >What do you mean by > hangs regardless? As in, nothing paints or firefox crashes on startup? Yes, 41+ to latest nightly hangs regardless of vsync settings If you navigate to the profile and delete extensions.json and the lock file after terminating the process in task manager, Firefox will run again until next time having to repeat the process. 40.0.2 x64 toggling the vsync to off stops the startup hang.
Flags: needinfo?(toadyshadow101)
> Sorry {For me the x64 versions worked as intended.} typo, It was meant to be > x86 > Enabling vsync in 39 > https://bugzilla.mozilla.org/show_bug.cgi?id=1194654#c10 did affect x86 > > 41+ to latest nightly x86 works perfectly no issues. Does Gecko 40, 32 bit work for you? > > > > >What do you mean by > > hangs regardless? As in, nothing paints or firefox crashes on startup? > > Yes, 41+ to latest nightly hangs regardless of vsync settings If you > navigate to the profile and delete extensions.json and the lock file after > terminating the process in task manager, Firefox will run again until next > time having to repeat the process. > > 40.0.2 x64 toggling the vsync to off stops the startup hang. How are you getting 40, 41, and 42 bit 64 bit builds? I can't seem to download them anywhere from official Mozilla release channels. From googling a bit, it looks like we only have 64 bit builds for nightly. It's also quite odd that everything works fine for vsync with 32 bit builds...
Flags: needinfo?(toadyshadow101)
Hi Mason Chang Here is a list of builds with download links, Next to each build number it will say if the build is good or bad. 39.0.3 x86 (Good) 39.0.3 x86 with vsync enabled (Bad) URL: https://ftp.mozilla.org/pub/mozilla.org/firefox/releases/39.0.3/win32/en-US/Firefox%20Setup%2039.0.3.exe 39.0.3 x64 (Good) 39.0.3 x64 with vsync enabled (Bad) URL: https://ftp.mozilla.org/pub/mozilla.org/firefox/releases/39.0b7/win64/en-US/Firefox%20Setup%2039.0b7.exe 40.0.2 x86 (Good) 40.0.2 x86 with vsync enabled (Good) URL: http://ftp.mozilla.org/pub/firefox/releases/40.0.2/win32/en-US/Firefox%20Setup%2040.0.2.exe 40.0 x64 (Good) 40.0 x64 with vsync enabled (Bad) URL: http://ftp.mozilla.org/pub/firefox/candidates/40.0b9-candidates/build1/win64/en-US/Firefox%20Setup%2040.0b9.exe 40.0.2 x64 (Good)[Compiled from source] 40.0.2 x64 with vsync enabled (Bad)[Compiled from source] URL: http://ftp.mozilla.org/pub/firefox/releases/40.0.2/source/firefox-40.0.2.source.tar.bz2 41.0b4 x86 (Good) 41.0b4 x86 with vsync enabled (Good) URL: http://ftp.mozilla.org/pub/firefox/releases/41.0b4/win32/en-US/Firefox%20Setup%2041.0b4.exe 41.0b4 x64 (Bad) 41.0b4 x64 with vsync enabled (Bad) URL: http://ftp.mozilla.org/pub/firefox/releases/41.0b4/win64/en-US/Firefox%20Setup%2041.0b4.exe 42.0a2 x86 (Good) 42.0a2 x86 with vsync enabled (Good) URL: http://ftp.mozilla.org/pub/firefox/nightly/latest-mozilla-aurora/firefox-42.0a2.en-US.win32.installer.exe 42.0a2 x64 (Bad) 42.0a2 x64 with vsync enabled (Bad) URL: http://ftp.mozilla.org/pub/firefox/nightly/latest-mozilla-aurora/firefox-42.0a2.en-US.win64.installer.exe 43.0a1 x86 (Good) 43.0a1 x86 with vsync enabled (Good) URL: http://ftp.mozilla.org/pub/firefox/nightly/latest-mozilla-central/firefox-43.0a1.en-US.win32.installer.exe 43.0a1 x64 (Bad) 43.0a1 x64 with vsync enabled (Bad) URL: http://ftp.mozilla.org/pub/firefox/nightly/latest-mozilla-central/firefox-43.0a1.en-US.win64.installer.exe Yes, x86 builds work with 39.0.3 vsync enabled as the exception. The issue is mainly x64bit builds, Could this be a x64bit driver issue with vsync? In 40.0.2 x64 bit toggling vsync solved the start hang. In 41+ up to latest nightly toggling vsync has no affect has the start hang described above in first post where removing extensions.json from the user profile allows it to start, This is likely a different issue if it can be separated.
Flags: needinfo?(toadyshadow101)
Attached file Vsync Tests —
We tried reproducing this on a similar GPU but could not. I'm not quite convinced this is a vsync issue if it still hangs on those other versions. Also, from comment 0, it works on first start up, then hangs and works again once you delete extensions.json, which doesn't sound like a vsync issue. Just to confirm though, can you run these two programs and see if they hang? They just read vsync timestamps and print them out. One is a 32 bit build, the other is a 64 bit build. Can you just run them for a couple of seconds and attach their output here? Thanks!
Flags: needinfo?(toadyshadow101)
(In reply to Mason Chang [:mchang] from comment #17) > Created attachment 8653542 [details] > Vsync Tests > > We tried reproducing this on a similar GPU but could not. I'm not quite > convinced this is a vsync issue if it still hangs on those other versions. > Also, from comment 0, it works on first start up, then hangs and works again > once you delete extensions.json, which doesn't sound like a vsync issue. > Just to confirm though, can you run these two programs and see if they hang? > They just read vsync timestamps and print them out. One is a 32 bit build, > the other is a 64 bit build. Can you just run them for a couple of seconds > and attach their output here? Thanks! Hi Mason Chang The attached test is missing the the msvc2013 debug redist, But i was able to resolve that, Do they output any logs or do you just want me to copy the console screen output. With these builds below, Toggling vsync* to false did resolve the startup hang. 40.0 x64 (Good) 40.0 x64 with vsync enabled (Bad) URL: http://ftp.mozilla.org/pub/firefox/candidates/40.0b9-candidates/build1/win64/en-US/Firefox%20Setup%2040.0b9.exe 40.0.2 x64 (Good)[Compiled from source] 40.0.2 x64 with vsync enabled (Bad)[Compiled from source] URL: http://ftp.mozilla.org/pub/firefox/releases/40.0.2/source/firefox-40.0.2.source.tar.bz2 With these builds below, Toggling vsync* to true did create the startup hang in the x64 but made the x86 GUI all messed up. 39.0.3 x86 (Good) 39.0.3 x86 with vsync enabled (Bad) URL: https://ftp.mozilla.org/pub/mozilla.org/firefox/releases/39.0.3/win32/en-US/Firefox%20Setup%2039.0.3.exe 39.0.3 x64 (Good) 39.0.3 x64 with vsync enabled (Bad) URL: https://ftp.mozilla.org/pub/mozilla.org/firefox/releases/39.0b7/win64/en-US/Firefox%20Setup%2039.0b7.exe These below, Although the mozregression inbound lands on the same spot as the vsync enabled patch the issue is likely different Here is where both issues appear to start https://bugzilla.mozilla.org/show_bug.cgi?id=1194654#c2 41+ up to the latest nightly is likely not related to the vsync, How do i separate them, Running mozregression again inbound lands on the same spot am i missing something? 41.0b4 x64 (Bad) 41.0b4 x64 with vsync enabled (Bad) URL: http://ftp.mozilla.org/pub/firefox/releases/41.0b4/win64/en-US/Firefox%20Setup%2041.0b4.exe 42.0a2 x64 (Bad) 42.0a2 x64 with vsync enabled (Bad) URL: http://ftp.mozilla.org/pub/firefox/nightly/latest-mozilla-aurora/firefox-42.0a2.en-US.win64.installer.exe 43.0a1 x64 (Bad) 43.0a1 x64 with vsync enabled (Bad) URL: http://ftp.mozilla.org/pub/firefox/nightly/latest-mozilla-central/firefox-43.0a1.en-US.win64.installer.exe
Flags: needinfo?(toadyshadow101) → needinfo?(mchang)
(In reply to Toady from comment #18) > (In reply to Mason Chang [:mchang] from comment #17) > > Created attachment 8653542 [details] > Hi Mason Chang > > The attached test is missing the the msvc2013 debug redist, But i was able > to resolve that, Do they output any logs or do you just > want me to copy the console screen output. Just copy the console screen output please. > 41+ up to the latest nightly is likely not related to the vsync, How do i separate them, Running > mozregression again inbound lands on the same spot am i missing something? You installed visual studio right? It might just be better to get a profile with visual studio so we can see what's going on. I'm still not convinced even on Gecko 40, that vsync is an issue since the code hasn't changed from 40-43. Especially since you can load on first boot, but then the profile gets locked, and the crash only happens the second time. Also, it works fine on x32, which shouldn't be different with vsync at least. Can you please grab a profile with visual studio? Can you first disable e10s. Then you can do: 1) From the toolbar on top, Analyze -> Profiler -> New Performance Session 2) Analyze -> Profiler -> Attach/Detach 3) Choose Firefox.exe 4) Let it run a couple of seconds. 5) Push stop profiling. You should see a report on the left. Right click on that and click export and attach that to this bug. Thanks!
Flags: needinfo?(mchang) → needinfo?(toadyshadow101)
(In reply to Mason Chang [:mchang] from comment #19) > (In reply to Toady from comment #18) > > (In reply to Mason Chang [:mchang] from comment #17) > > > Created attachment 8653542 [details] > > Hi Mason Chang > > > > The attached test is missing the the msvc2013 debug redist, But i was able > > to resolve that, Do they output any logs or do you just > > want me to copy the console screen output. > > Just copy the console screen output please. > > > 41+ up to the latest nightly is likely not related to the vsync, How do i separate them, Running > > mozregression again inbound lands on the same spot am i missing something? > > You installed visual studio right? No i had copied the msvcp120d.dll & msvcr120d.dll from another computer with VS installed to the laptop pc with the issue to test. Unfortunately am no longer able to run further test as my friend wanted his laptop back, Since its been over 2 weeks and i said it should only take a few days to run tests and report the issue, On his brand new laptop he asked me to setup with a SSD for him. I wish i could be of more help sorry.
Flags: needinfo?(toadyshadow101)
Ahh, thanks for your help thus far. What kind of laptop was it if I may ask? Maybe we can get a cheap version of this laptop. Thanks!
Flags: needinfo?(toadyshadow101)
(In reply to Mason Chang [:mchang] from comment #21) > Ahh, thanks for your help thus far. What kind of laptop was it if I may ask? > Maybe we can get a cheap version of this laptop. Thanks! It was this model, He paid just over $700 retail http://www.acer.com.au/ac/en/AU/content/model/NX.ML8SA.004 Again sorry i couldn't be of more help.
Flags: needinfo?(toadyshadow101)
Whiteboard: [gfx-noted]
Seems, that i have a similar issue on my Acer vn7-571 laptop, but i've found a solution on http://www.camp-firefox.de/forum/viewtopic.php?f=3&t=115258&start=15#p985800. When ePowerEvent.exe is not running, firefox starts normally.
(In reply to Mikhail from comment #23) > Seems, that i have a similar issue on my Acer vn7-571 laptop, but i've found > a solution on > http://www.camp-firefox.de/forum/viewtopic.php?f=3&t=115258&start=15#p985800. > When ePowerEvent.exe is not running, firefox starts normally. Based on minimal research ePowerEvent.exe is Acer's proprietary power management software. I wonder if we have any contacts at Acer who may be able to assist?
This happens to me on an Acer as well. Problem happens on 45.0b9 but not on Developer Edition 46.0a2.
Toady, are you still crashing at startup with 64b Aurora 46 or Nightly 47?
Flags: needinfo?(toadyshadow101)
(In reply to Loic from comment #26) > Toady, are you still crashing at startup with 64b Aurora 46 or Nightly 47? Hi Loic As mentioned here https://bugzilla.mozilla.org/show_bug.cgi?id=1194654#c20 i lost the ability to test further when my friend got his laptop back, If i could be of more assistance i would be but without the means to reproduce it i am afraid that ill be of little use for this issue sorry. Mikhail in https://bugzilla.mozilla.org/show_bug.cgi?id=1194654#c23 maybe able to help since had the issue in question or jrx in https://bugzilla.mozilla.org/show_bug.cgi?id=1194654#c25 again sorry.
Flags: needinfo?(toadyshadow101)
Mikhail, jrx said he doesn't crash anymore with Firefox DE 46.0a2. Are you able to crash Firefox DE in the same way (maybe by enabling again ePowerEvent.exe)? https://www.mozilla.org/en-US/firefox/developer/all/
Flags: needinfo?(mikel1)
Loic(In reply to Loic from comment #28) > Mikhail, jrx said he doesn't crash anymore with Firefox DE 46.0a2. Are you > able to crash Firefox DE in the same way (maybe by enabling again > ePowerEvent.exe)? > https://www.mozilla.org/en-US/firefox/developer/all/ FWIW, 45.0b10 has the same issue.
(In reply to Loic from comment #28) > Mikhail, jrx said he doesn't crash anymore with Firefox DE 46.0a2. Are you > able to crash Firefox DE in the same way (maybe by enabling again > ePowerEvent.exe)? > https://www.mozilla.org/en-US/firefox/developer/all/ Yep, Firefox DE 46.0a2 works fine.
Flags: needinfo?(mikel1)
OK, good, guess whatever it is is coming down the pipe soon :)
(In reply to Mikhail from comment #30) > (In reply to Loic from comment #28) > > Mikhail, jrx said he doesn't crash anymore with Firefox DE 46.0a2. Are you > > able to crash Firefox DE in the same way (maybe by enabling again > > ePowerEvent.exe)? > > https://www.mozilla.org/en-US/firefox/developer/all/ > > Yep, Firefox DE 46.0a2 works fine. Installed 46.0b1 yesterday. So far it seems to be working.
I'm closing this bug report since recent comments indicate this was resolved. Please reopen if you can reproduce this crash on a current version of Firefox.
Status: UNCONFIRMED → RESOLVED
Closed: 9 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: