chrome crash after google search
Reported by
bau...@gmail.com,
Aug 5 2017
|
|||||||||||||
Issue descriptionUserAgent: Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/61.0.3163.31 Safari/537.36 Steps to reproduce the problem: 1. open new tab 2. make search 3. What is the expected behavior? work same as before display page and not crash What went wrong? crash, can't send comment (not work), can't sent problem with about page Did this work before? Yes before 61.0.3163.31 Chrome version: 61.0.3163.31 Channel: beta OS Version: 6.1 (Windows 7, Windows Server 2008 R2) Flash Version: very poor material design
,
Aug 8 2017
reproduce on my profil only since latest chrome update (version with very bad UI scheiss design forced for settings) I always reproduce crash when open google.fr But with this new very bad UI sheiss design it's very harsd to manage all settings.. clear selected cookies, enable or not option... (I swiched to Opera browser, to fast solve problem, google.fr work with all cookies imported) ID du rapport d'erreur importé : 8872d39268000000 (ID de plantage local : cfa4e8e1-ea3a-465c-9ea5-5328b71b8751) ID de plantage local : a64db96f-cfa8-4f87-b1a2-633e3e0c7bf0 ID de plantage local : 53e831a9-7e71-4748-98dd-86e4a915aeb1 ID du rapport d'erreur importé : 27e31dfc84000000 (ID de plantage local : 713f8f50-b336-4a32-a304-122193183603) ID de plantage local : a8d8ff03-000b-43ff-b5d7-af2843996983 ID de plantage local : 79f57c19-f388-4034-a54f-c99e55368498 ID du rapport d'erreur importé : 606b7a0010000000 (ID de plantage local : 7fd97bdc-7c69-4680-9abc-91cb3157e30b) ID du rapport d'erreur importé : 9526fff048000000 (ID de plantage local : 17e5a8a3-9c4b-43c4-84f1-ec1fb762ab77) ID du rapport d'erreur importé : 88e7d99268000000 (ID de plantage local : de9fce84-31f8-4c0d-83ff-ac05b0e8e593) ID de plantage local : e51a434e-abd4-4ae4-ab51-9f3832625515 ID de plantage local : 3d772011-1850-4840-9714-2109768bbcfe ID du rapport d'erreur importé : 3a37ec3f88000000 (ID de plantage local : 64851497-ad0c-4c39-8a9c-4f02e4dd87fd) ... more more... more test but not found why chrome crash.. tested without extension,same.
,
Aug 8 2017
Thank you for providing more feedback. Adding requester "mmanchala@chromium.org" to the cc list and removing "Needs-Feedback" label. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Aug 8 2017
more info: I found how to not reproduce problem with google.fr After more more search in setting; I found I autorize google.fr to access location. After set google.fr to default value for location (= request), I can open google.fr without crash, and page not request me to access location. When autorize google.fr to access location, chrome (tab) crash when open google.fr
,
Aug 8 2017
and linked to chrome://flags Disable 'Feature Policy' (Enables granting and removing access to features through the Feature-Policy HTTP header. #enable-feature-policy) can reproduce on canary 62.0.3179.0 and second problem linked to the same flags: menu/help/report an issue... not work when disable 'Feature Policy'
,
Aug 9 2017
Thanks for filing and updates!Checked with provided crash ids in crash server. Stack Trace: ------------ Thread 0 (id: 7792) CRASHED [EXCEPTION_ACCESS_VIOLATION_READ @ 0x00000000 ] MAGIC SIGNATURE THREAD Stack Quality83%Show frame trust levels 0x000007fed496c4c0 (chrome_child.dll -frame.cpp:193 ) blink::Frame::IsFeatureEnabled(blink::WebFeaturePolicyFeature) 0x000007fed497e093 (chrome_child.dll -deprecation.cpp:204 ) blink::Deprecation::CountDeprecationFeaturePolicy(blink::Document const &,blink::WebFeaturePolicyFeature) 0x000007fed59e9978 (chrome_child.dll -geolocation.cpp:171 ) blink::Geolocation::RecordOriginTypeAccess() 0x000007fed59e9edd (chrome_child.dll -geolocation.cpp:222 ) blink::Geolocation::StartRequest(blink::GeoNotifier *) 0x000007fed59ea4c7 (chrome_child.dll -geolocation.cpp:194 ) blink::Geolocation::getCurrentPosition(blink::PositionCallback *,blink::PositionErrorCallback *,blink::PositionOptions const &) 0x000007fed5aca494 (chrome_child.dll -v8geolocation.cpp:112 ) blink::GeolocationV8Internal::getCurrentPositionMethod 0x000007fed31216f3 (chrome_child.dll -api-arguments.cc:25 ) v8::internal::FunctionCallbackArguments::Call(void (*)(v8::FunctionCallbackInfo<v8::Value> const &)) 0x000007fed3121456 (chrome_child.dll -builtins-api.cc:112 ) v8::internal::`anonymous namespace'::HandleApiCallHelper<0> 0x000007fed3120e58 (chrome_child.dll -builtins-api.cc:142 ) v8::internal::Builtin_Impl_HandleApiCall 0x000007fed3120d7d (chrome_child.dll -builtins-api.cc:130 ) v8::internal::Builtin_HandleApiCall(int,v8::internal::Object * *,v8::internal::Isolate *) 0x00000185cc2047a0 0x000007fed32979ff (chrome_child.dll + 0x002d79ff ) 1)This crash is first started on 59.0.3071.115 and on latest Beta #61.0.3163.31 seeing 33 from 2 different clients. 2)This crash seen on Windows & chrome OS. 3)This crash is not seen on latest Canary,Dev & Stable. 62.0.3175.3 1.54% 1 62.0.3168.0 1.54% 1 61.0.3163.31 50.77% 33 -beta 61.0.3163.26 1.54% 1 61.0.3163.13 41.54% 27 61.0.3162.0 1.54% 1 Link to the list of builds: --------------------------- https://crash.corp.google.com/browse?q=custom_data.ChromeCrashProto.ptype%3D%27renderer%27%20AND%20custom_data.ChromeCrashProto.magic_signature_1.name%3D%27blink%3A%3AFrame%3A%3AIsFeatureEnabled%27&ignore_case=false&enable_rewrite=true&omit_field_name=&omit_field_value=&omit_field_opt=%3D#-property-selector,samplereports:5,+productversion Unable to find the suspect through code search. Hence untriaging the issue so that issue would get addressed. Thank You!
,
Aug 10 2017
Issue 753245 has been merged into this issue.
,
Aug 10 2017
Issue 752766 has been merged into this issue.
,
Aug 11 2017
,
Aug 16 2017
Unclear so far if this is WebFeaturePolicyFeature or Location related. Raymes, WDYT?
,
Aug 17 2017
I was able to reproduce this on ToT, running with FeaturePolicy disabled via command line, and doing a local search. (www.google.ca => search for "a thing near me") It appears that the code in Deprecation::CountDeprecationFeaturePolicy, which is tracking how many times the pre-feature-policy behaviour is being relied on, doesn't actually check to see whether FP is enabled before calling frame->IsFeatureEnabled(). I see this stack trace: [1:1:0817/123148.401295:FATAL:Frame.cpp(193)] Check failed: feature_policy. #0 0x7f8326c602bd base::debug::StackTrace::StackTrace() #1 0x7f8326c5e68c base::debug::StackTrace::StackTrace() #2 0x7f8326ceee0a logging::LogMessage::~LogMessage() #3 0x7f83155ed757 blink::Frame::IsFeatureEnabled() #4 0x7f83155e5900 blink::Deprecation::CountDeprecationFeaturePolicy() #5 0x7f831291f708 blink::Geolocation::RecordOriginTypeAccess() #6 0x7f831291fb00 blink::Geolocation::StartRequest() #7 0x7f831291f936 blink::Geolocation::getCurrentPosition() #8 0x7f8312276cf5 blink::GeolocationV8Internal::getCurrentPositionMethod() #9 0x7f8312275c25 blink::V8Geolocation::getCurrentPositionMethodCallback() #10 0x7f83178fa282 v8::internal::FunctionCallbackArguments::Call() #11 0x7f83179d9bf2 v8::internal::(anonymous namespace)::HandleApiCallHelper<>() #12 0x7f83179d8470 v8::internal::Builtin_Impl_HandleApiCall() #13 0x3889682843c4 <unknown>
,
Aug 18 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/fb86b10bda0326d2e05c3f5f4060120613f91a76 commit fb86b10bda0326d2e05c3f5f4060120613f91a76 Author: Ian Clelland <iclelland@google.com> Date: Fri Aug 18 14:25:20 2017 Fix crash in CountDeprecationWarning The code to count cross-origin uses of features which are going to be under the control of Feature Policy wasn't checking whether Feature Policy itself was enabled before asking the frame to check its policy. This caused a DCHECK to fire in Frame.cpp in that case. Without Feature Policy, we can't meaningfully check whether the usage of the feature would have been declined in every case, so this change causes CountDeprecationWarning to return early. This may introduce some undercounting, but since Feature Policy is now enabled by default, that should be minimal (and this undercounting is not new; it was just previously represented by crashes instead). Bug: 752762 Change-Id: Ia36a2f7c85de8cedc4a386803d428e9d351db012 Reviewed-on: https://chromium-review.googlesource.com/619366 Reviewed-by: Daniel Cheng <dcheng@chromium.org> Commit-Queue: Ian Clelland <iclelland@chromium.org> Cr-Commit-Position: refs/heads/master@{#495543} [modify] https://crrev.com/fb86b10bda0326d2e05c3f5f4060120613f91a76/third_party/WebKit/Source/core/frame/Deprecation.cpp
,
Aug 21 2017
,
Aug 21 2017
,
Aug 21 2017
This bug requires manual review: M61 has already been promoted to the beta branch, so this requires manual review Please contact the milestone owner if you have questions. Owners: amineer@(Android), cmasso@(iOS), ketakid@(ChromeOS), govind@(Desktop) For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Aug 21 2017
Before we approve merge to M61, please answer followings: * Is this M61 regression? Is it critical? * Is the change well baked/verified in Canary, having enough automation tests coverage and safe to merge to M61? * Any other important details to justify the merge. Please note We're only few weeks away from M61 Stable promotion, so merge bar is very high.
,
Aug 21 2017
1. Yes, this is a regression -- it causes a very quick renderer crash if Geolocation is requested when Feature Policy is disabled through flags. I don't know whether or not this counts as critical, since the flag is enabled by default, but it likely means that Chrome is unusable with the flag disabled. 2. I do *not* have an automated test for this, but I'll see if I can work one up quickly. 3. The change is quite small, just adding a safety check to ensure that FP is enabled before calling the function which will otherwise crash. Not calling that function has no effect on the renderer, since it just avoids logging one UMA data point.
,
Aug 21 2017
Please add automated test for this and then merge to M61. Approving merge to M61 branch 3163 based on comment #17.
,
Aug 21 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/65302d8e4a97c073cbcc675da90c7a1ff1547d95 commit 65302d8e4a97c073cbcc675da90c7a1ff1547d95 Author: Ian Clelland <iclelland@google.com> Date: Mon Aug 21 23:04:28 2017 Add regression test for FP deprecation crash. This adds an automated test for the case where a feature-policy- deprecation warning is triggered when feature policy is disabled. It ensures that no usage is recorded, and that the renderer does not crash. Bug: 752762 Change-Id: I2da29263381c3078d2cb1bc83bcc3b05de564ee3 Reviewed-on: https://chromium-review.googlesource.com/624419 Commit-Queue: Ian Clelland <iclelland@chromium.org> Reviewed-by: Jeremy Roman <jbroman@chromium.org> Cr-Commit-Position: refs/heads/master@{#496105} [modify] https://crrev.com/65302d8e4a97c073cbcc675da90c7a1ff1547d95/third_party/WebKit/Source/core/frame/UseCounterTest.cpp
,
Aug 21 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/2132f13fb3702dadaf31c0785cecd10766634880 commit 2132f13fb3702dadaf31c0785cecd10766634880 Author: Ian Clelland <iclelland@google.com> Date: Mon Aug 21 23:24:52 2017 Fix crash in CountDeprecationWarning The code to count cross-origin uses of features which are going to be under the control of Feature Policy wasn't checking whether Feature Policy itself was enabled before asking the frame to check its policy. This caused a DCHECK to fire in Frame.cpp in that case. Without Feature Policy, we can't meaningfully check whether the usage of the feature would have been declined in every case, so this change causes CountDeprecationWarning to return early. This may introduce some undercounting, but since Feature Policy is now enabled by default, that should be minimal (and this undercounting is not new; it was just previously represented by crashes instead). TBR=iclelland@google.com (cherry picked from commit fb86b10bda0326d2e05c3f5f4060120613f91a76) Bug: 752762 Change-Id: Ia36a2f7c85de8cedc4a386803d428e9d351db012 Reviewed-on: https://chromium-review.googlesource.com/619366 Reviewed-by: Daniel Cheng <dcheng@chromium.org> Commit-Queue: Ian Clelland <iclelland@chromium.org> Cr-Original-Commit-Position: refs/heads/master@{#495543} Reviewed-on: https://chromium-review.googlesource.com/625016 Reviewed-by: Ian Clelland <iclelland@chromium.org> Cr-Commit-Position: refs/branch-heads/3163@{#739} Cr-Branched-From: ff259bab28b35d242e10186cd63af7ed404fae0d-refs/heads/master@{#488528} [modify] https://crrev.com/2132f13fb3702dadaf31c0785cecd10766634880/third_party/WebKit/Source/core/frame/Deprecation.cpp
,
Aug 21 2017
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/e30fc9ad343c20e2d67560f5ff057a2a28ee0945 commit e30fc9ad343c20e2d67560f5ff057a2a28ee0945 Author: Ian Clelland <iclelland@google.com> Date: Mon Aug 21 23:28:01 2017 Add regression test for FP deprecation crash. This adds an automated test for the case where a feature-policy- deprecation warning is triggered when feature policy is disabled. It ensures that no usage is recorded, and that the renderer does not crash. TBR=iclelland@google.com (cherry picked from commit 65302d8e4a97c073cbcc675da90c7a1ff1547d95) Bug: 752762 Change-Id: I2da29263381c3078d2cb1bc83bcc3b05de564ee3 Reviewed-on: https://chromium-review.googlesource.com/624419 Commit-Queue: Ian Clelland <iclelland@chromium.org> Reviewed-by: Jeremy Roman <jbroman@chromium.org> Cr-Original-Commit-Position: refs/heads/master@{#496105} Reviewed-on: https://chromium-review.googlesource.com/624956 Reviewed-by: Ian Clelland <iclelland@chromium.org> Cr-Commit-Position: refs/branch-heads/3163@{#740} Cr-Branched-From: ff259bab28b35d242e10186cd63af7ed404fae0d-refs/heads/master@{#488528} [modify] https://crrev.com/e30fc9ad343c20e2d67560f5ff057a2a28ee0945/third_party/WebKit/Source/core/frame/UseCounterTest.cpp
,
Sep 22 2017
,
Sep 22 2017
|
|||||||||||||
►
Sign in to add a comment |
|||||||||||||
Comment 1 by mmanchala@chromium.org
, Aug 8 2017Labels: Needs-Feedback Needs-Triage-M61