New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 705324 link

Starred by 3 users

Issue metadata

Status: Assigned
Merged: issue 701287
Owner:
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Linux
Pri: 2
Type: Bug


Participants' hotlists:
Hotlist-AsmJsParser


Sign in to add a comment

V8 correctness failure in configs: x64,ignition:x64,ignition_turbo_opt

Project Member Reported by ClusterFuzz, Mar 27 2017

Issue description

Cc: clemensh@chromium.org bradnelson@chromium.org ahaas@chromium.org
Labels: -Pri-1 Pri-2
Status: Available (was: Untriaged)
PTAL
Mergedinto: 701287
Status: Duplicate (was: Available)
Project Member

Comment 3 by ClusterFuzz, May 16 2017

ClusterFuzz has detected this issue as fixed in range 45316:45317.

Detailed report: https://clusterfuzz.com/testcase?key=6223285387001856

Fuzzer: foozzie_js_mutation
Job Type: v8_foozzie
Platform Id: linux

Crash Type: V8 correctness failure
Crash Address: 
Crash State:
  configs: x64,ignition:x64,ignition_turbo_opt
  sources: 518
  
Sanitizer: address (ASAN)

Regressed: V8: 43718:43719
Fixed: V8: 45316:45317

Reproducer Testcase: https://clusterfuzz.com/download?testcase_id=6223285387001856


See https://dev.chromium.org/Home/chromium-security/bugs/reproducing-clusterfuzz-bugs for more information.

If you suspect that the result above is incorrect, try re-doing that job on the test case report page.
Cc: machenb...@chromium.org
Wait... How did the suppression of ".caller" fix this?
a bug :(
Cc: bmeu...@chromium.org
Is this a duplicate of 701287 then? 701287 is fixed, but this one is not...
Cc: mstarzinger@chromium.org
Status: Available (was: Duplicate)
Indeed. This case still repros. Keeping it available then.
Cc: marja@chromium.org
Possibly duplicate of  issue 720320 .
Cc: jarin@chromium.org
 Issue 732009  has been merged into this issue.
Owner: mstarzinger@chromium.org
Status: Assigned (was: Available)
This is due to missing deoptimization information when we use TurboFan to optimize asm.js code. We won't fix this for TurboFan. Unfortunately the asm.js-to-WebAssembly translator suffers a similar problem. I'll grab this and see if we can eventually fix this with the validator, definietely not a blocker of any kind though. The following is the stack-trace of normal execution versus asm.js validation.

======

$ ./out/x64.debug/d8 test/mjsunit/foo.js
v8-foozzie source: /v8/test/mjsunit/asm/math-fround.js
test/mjsunit/foo.js:61: TypeError: Cannot convert object to primitive value
    a = +a;
        ^
TypeError: Cannot convert object to primitive value
    at __f_3 (test/mjsunit/foo.js:61:9)
    at test/mjsunit/foo.js:66:1

======

$ ./out/x64.debug/d8 test/mjsunit/foo.js --validate-asm
v8-foozzie source: /v8/test/mjsunit/asm/math-fround.js
test/mjsunit/foo.js:66: TypeError: Cannot convert object to primitive value
__f_3(function() {});
^
TypeError: Cannot convert object to primitive value
    at test/mjsunit/foo.js:66:1

======

 Issue 708144  has been merged into this issue.

Sign in to add a comment