227805 2021-07-08 11:11:02 -0700 [Mac wk2] fast/canvas/canvas-composite-image.html is a flaky crash 2021-07-12 19:51:27 -0700 1 1 1 Unclassified WebKit WebKit2 WebKit Nightly Build Unspecified Unspecified NEW https://bugs.webkit.org/show_bug.cgi?id=227874 https://bugs.webkit.org/show_bug.cgi?id=227881 InRadar P2 Normal --- 1 ayumi_kojima ayumi_kojima ap cdumez dino kkinnunen tsavell webkit-bot-watchers-bugzilla webkit-bug-importer oldest_to_newest 1775727 0 ayumi_kojima 2021-07-08 11:11:02 -0700 fast/canvas/canvas-composite-image.html History: https://results.webkit.org/?suite=layout-tests&test=fast%2Fcanvas%2Fcanvas-composite-image.html I was able to reproduce the crash with run-webkit-tests fast/canvas/canvas-composite-image.html --iterations 50 --debug -f —child-process 20 No crash log found for com.apple.WebKit.WebContent.Development:30562. stdout: stderr: com.apple.WebKit.WebContent.Development terminated (pid 30562) because the process crashed LEAK: 1 WebPageProxy 1775728 1 webkit-bug-importer 2021-07-08 11:11:55 -0700 <rdar://problem/80333731> 1775731 2 433144 ayumi_kojima 2021-07-08 11:24:36 -0700 Created attachment 433144 TestExpectations update 1775739 3 433144 tsavell 2021-07-08 11:38:17 -0700 Comment on attachment 433144 TestExpectations update Clearing flags on attachment: 433144 Committed r279738 (239520@main): <https://commits.webkit.org/239520@main> 1775861 4 ayumi_kojima 2021-07-08 16:36:53 -0700 Test appears to have passed on my terminal at r278544 with run-webkit-tests fast/canvas/canvas-composite-image.html --iterations 50 --debug -f --no-timeout —child-process 20 but also got a crash log: ============ No crash log found for com.apple.WebKit.WebContent.Development:56305. stdout: stderr: com.apple.WebKit.WebContent.Development terminated (pid 56305) because the process crashed LEAK: 1 WebPageProxy ============ I tried r277000, r278000, and r278300 and it seemed to have passed at all revisions. 1775890 5 ap 2021-07-08 18:13:35 -0700 Also happening on release builds on internal bots. This output is produced by TestController::webProcessDidTerminate() when getting kWKProcessTerminationReasonCrash. I looked at this test run: https://build.webkit.org/#/builders/23/builds/2729/steps/11/logs/stdio 01:18:49.635 29797 Using port 'mac-bigsur-wk2' 01:18:49.639 29797 "uname -m" took 0.00s 01:18:49.639 29797 Test configuration: <bigsur, x86_64, debug> 01:18:49.639 29797 Placing test results in /Volumes/Data/worker/bigsur-debug-tests-wk2/build/layout-test-results 01:18:49.639 29797 Using Debug build ... 01:25:14.269 29797 worker/8 fast/canvas/canvas-color-space-display-p3-ImageData.html passed 01:25:14.539 29797 worker/8 fast/canvas/canvas-color-space-display-p3.html passed 01:25:18.010 29797 worker/8 fast/canvas/canvas-composite-alpha.html passed 01:25:19.120 29797 worker/8 fast/canvas/canvas-composite-canvas.html passed 01:25:19.267 29797 worker/8 fast/canvas/canvas-composite-fill-repaint.html passed 01:25:19.465 29797 worker/8 com.apple.WebKit.WebContent.Development crash, pid = 30562 01:25:19.485 29797 worker/8 killed pid 29896 01:26:04.494 29797 worker/8 worker/8 fast/canvas/canvas-composite-image.html crashed, (stderr lines): 01:26:04.494 29797 worker/8 com.apple.WebKit.WebContent.Development terminated (pid 30562) because the process crashed 01:26:04.494 29797 worker/8 LEAK: 1 WebPageProxy 01:26:04.495 29797 worker/8 killing driver But looking at system logs from this time, this process was entirely idle, and in fact it was running long before run-webkit-tests started?! % log show --start "2021-07-08 00:00:00" --end "2021-07-08 01:27:00" --info --debug | grep 30562 2021-07-08 00:06:39.305627-0700 0x354 Error 0x0 138 7 WindowServer: (SkyLight) [com.apple.SkyLight:default] _compareTimesAndApps: error getting owner of window 779303 2021-07-08 00:26:19.305628-0700 0x48d8a89 Default 0x7f1de65 354 2 pkd: (LaunchServices) [com.apple.launchservices:default] Waiting on thread <private> until Launch Services database seeding is complete. 2021-07-08 00:30:05.330562-0700 0x48dcc33 Error 0x0 19070 7 com.apple.WebKit.WebContent.Development: (SkyLight) [com.apple.SkyLight:default] 0 is not a valid connection ID. 2021-07-08 00:30:05.330562-0700 0x48dcc33 Error 0x0 19070 7 com.apple.WebKit.WebContent.Development: (SkyLight) [com.apple.SkyLight:default] 0 is not a valid connection ID. 2021-07-08 00:30:05.330562-0700 0x48dcc33 Error 0x0 19070 7 com.apple.WebKit.WebContent.Development: (SkyLight) [com.apple.SkyLight:default] 0 is not a valid connection ID. 2021-07-08 01:08:14.305620-0700 0x39a Default 0x0 143 5 loginwindow: [com.apple.loginwindow.logging:Standard] -[ApplicationManager checkInAppContext:eventData:] | ApplicationManager: App checkin failed 2021-07-08 01:08:17.305629-0700 0x354 Default 0x0 138 7 WindowServer: (SkyLight) [com.apple.SkyLight:default] [ ConnectionDebug ] Closing conn 0x1aabd3b, PID 25937 in session 257 on console 2021-07-08 01:12:28.230562-0700 0x4910d10 Default 0x0 105 2 launchservicesd: [com.apple.launchservices:cas] CHECKIN:0x0-0x6c043ed8 26982 com.apple.WebKit.WebContent 2021-07-08 01:24:46.298055-0700 0x491b7ed Default 0x0 105 2 launchservicesd: [com.apple.launchservices:cas] CHECKIN:0x0-0x6c3b724b 30562 com.apple.WebKit.WebContent 2021-07-08 01:24:46.315013-0700 0x491b9b7 Default 0x0 30562 7 com.apple.WebKit.WebContent.Development: (SkyLight) [com.apple.SkyLight:default] Set a breakpoint at SLSLogBreak to catch errors/faults as they are logged. 2021-07-08 01:24:46.315018-0700 0x491b9b7 Error 0x0 30562 7 com.apple.WebKit.WebContent.Development: (SkyLight) [com.apple.SkyLight:default] 0 is not a valid connection ID. 2021-07-08 01:24:46.315018-0700 0x491b9b7 Error 0x0 30562 7 com.apple.WebKit.WebContent.Development: (SkyLight) [com.apple.SkyLight:default] 0 is not a valid connection ID. 2021-07-08 01:24:46.315018-0700 0x491b9b7 Error 0x0 30562 7 com.apple.WebKit.WebContent.Development: (SkyLight) [com.apple.SkyLight:default] 0 is not a valid connection ID. 2021-07-08 01:24:46.315019-0700 0x491b9b7 Error 0x0 30562 7 com.apple.WebKit.WebContent.Development: (SkyLight) [com.apple.SkyLight:default] 0 is not a valid connection ID. 2021-07-08 01:24:46.315151-0700 0x491b9b7 Error 0x0 30562 7 com.apple.WebKit.WebContent.Development: (SkyLight) [com.apple.SkyLight:default] 0 is not a valid connection ID. 2021-07-08 01:24:46.315151-0700 0x491b9b7 Default 0x0 30562 7 com.apple.WebKit.WebContent.Development: (SkyLight) [com.apple.SkyLight:default] Invalid Connection ID 0 So this looks spurious? Also, there is nothing in the logs about crashing: $ log show --start "2021-07-08 01:20:00" --end "2021-07-08 01:27:00" --info --debug | grep crash 2021-07-08 01:26:20.280306-0700 0x491aa9c Fault 0x7f84b33 70 7 mediaremoted: (MediaServices) [com.apple.amp.MediaServices:Entitlements] Failed to retrieve application-identifier from SecTask. This typically indicates that the process is missing the application-identifier entitlement. It must match the process' bundle identifier. This is a client-issue. Check the crash report for the Coalition process. PID: 0 1775891 6 ap 2021-07-08 18:15:09 -0700 > and in fact it was running long before run-webkit-tests started?! Never mind that part, should have looked at grep output more carefully. Still, essentially no output from the process at all. 433144 2021-07-08 11:24:36 -0700 2021-07-08 17:50:24 -0700 TestExpectations update bug-227805-20210708112435.patch text/plain 1310 ayumi_kojima U3VidmVyc2lvbiBSZXZpc2lvbjogMjc5NzM2CmRpZmYgLS1naXQgYS9MYXlvdXRUZXN0cy9DaGFu Z2VMb2cgYi9MYXlvdXRUZXN0cy9DaGFuZ2VMb2cKaW5kZXggYjJjMGFkMWY1MDhiYjFkZjdlOTMx Mzc1NDNiZDA0NTZjODE1ZWQyYS4uODEzNmIzMzExMGI2ODc1MTcxMTEzOTk2ODBiODQ5ZTc5NmFi ZTRiZCAxMDA2NDQKLS0tIGEvTGF5b3V0VGVzdHMvQ2hhbmdlTG9nCisrKyBiL0xheW91dFRlc3Rz L0NoYW5nZUxvZwpAQCAtMSwzICsxLDEyIEBACisyMDIxLTA3LTA4ICBBeXVtaSBLb2ppbWEgIDxh eXVtaV9rb2ppbWFAYXBwbGUuY29tPgorCisgICAgICAgIFtNYWMgd2syIERlYnVnXSBmYXN0L2Nh bnZhcy9jYW52YXMtY29tcG9zaXRlLWltYWdlLmh0bWwgaXMgYSBmbGFreSBjcmFzaC4KKyAgICAg ICAgaHR0cHM6Ly9idWdzLndlYmtpdC5vcmcvc2hvd19idWcuY2dpP2lkPTIyNzgwNQorCisgICAg ICAgIFVucmV2aWV3ZWQgdGVzdCBnYXJkZW5pbmcuCisKKyAgICAgICAgKiBwbGF0Zm9ybS9tYWMt d2syL1Rlc3RFeHBlY3RhdGlvbnM6CisKIDIwMjEtMDctMDggIEF5dW1pIEtvamltYSAgPGF5dW1p X2tvamltYUBhcHBsZS5jb20+CiAKICAgICAgICAgWyBNYWMgRGVidWcgXSBpbnNwZWN0b3IvY2Fu dmFzL3NoYWRlclByb2dyYW0tYWRkLXJlbW92ZS13ZWJnbDIuaHRtbCBpcyBhIGZsYWt5IHRleHQg ZmFpbHVyZS4KZGlmZiAtLWdpdCBhL0xheW91dFRlc3RzL3BsYXRmb3JtL21hYy13azIvVGVzdEV4 cGVjdGF0aW9ucyBiL0xheW91dFRlc3RzL3BsYXRmb3JtL21hYy13azIvVGVzdEV4cGVjdGF0aW9u cwppbmRleCAyMDBmMDZlODQ0Zjc3OTUzMGU2ODlmYWE0NmM3OTA3ZDNkMDIzNDg0Li44NGIyZjhl ZDFjZjgzZmVhZTQ5ZGVlMGYxN2Y1ZTA4YTc5MzUwNjNmIDEwMDY0NAotLS0gYS9MYXlvdXRUZXN0 cy9wbGF0Zm9ybS9tYWMtd2syL1Rlc3RFeHBlY3RhdGlvbnMKKysrIGIvTGF5b3V0VGVzdHMvcGxh dGZvcm0vbWFjLXdrMi9UZXN0RXhwZWN0YXRpb25zCkBAIC0xMzg4LDMgKzEzODgsNSBAQCB3ZWJr aXQub3JnL2IvMjI3Nzc2IFsgQmlnU3VyIFJlbGVhc2UgYXJtNjQgXSBzY3JvbGxiYXJzL2Nvcm5l ci1yZXNpemVyLXdpbmRvdy1pbgogCiB3ZWJraXQub3JnL2IvMjI3Nzc2IFsgQ2F0YWxpbmEgUmVs ZWFzZSBdIHNjcm9sbGJhcnMvY29ybmVyLXJlc2l6ZXItd2luZG93LWluYWN0aXZlLmh0bWwgWyBQ YXNzIEltYWdlT25seUZhaWx1cmUgXQogCit3ZWJraXQub3JnL2IvMjI3ODA1IFsgRGVidWcgXSBm YXN0L2NhbnZhcy9jYW52YXMtY29tcG9zaXRlLWltYWdlLmh0bWwgWyBQYXNzIENyYXNoIF0KKwo=