Bug 190283 – Disable Options::useWebAssemblyFastMemory() on linux if ASAN signal handling is not disabled.

• Status: RESOLVED
• Resolution: FIXED
• Component: JavaScriptCore

[Mark Lam]

Will also print a warning if we detect this condition.

[Radar WebKit Bug Importer]

<rdar://problem/45015752>

[Mark Lam]

Created attachment 351607 [details]
proposed patch.

[Keith Miller]

Comment on attachment 351607 [details]
proposed patch.

View in context: https://bugs.webkit.org/attachment.cgi?id=351607&action=review

r=me with comment.

> Source/JavaScriptCore/runtime/Options.cpp:581
> +                    && (strstr(asanOptions, "allow_user_segv_handler=1") || strstr(asanOptions, "handle_segv=0"));

Do we actually need both or does ASAN work with just allow_user_segv_handler=1?

[Mark Lam]

Comment on attachment 351607 [details]
proposed patch.

View in context: https://bugs.webkit.org/attachment.cgi?id=351607&action=review

>> Source/JavaScriptCore/runtime/Options.cpp:581
>> +                    && (strstr(asanOptions, "allow_user_segv_handler=1") || strstr(asanOptions, "handle_segv=0"));
> 
> Do we actually need both or does ASAN work with just allow_user_segv_handler=1?

We're checking for either here.

[Keith Miller]

Comment on attachment 351607 [details]
proposed patch.

View in context: https://bugs.webkit.org/attachment.cgi?id=351607&action=review

>>> Source/JavaScriptCore/runtime/Options.cpp:581
>>> +                    && (strstr(asanOptions, "allow_user_segv_handler=1") || strstr(asanOptions, "handle_segv=0"));
>> 
>> Do we actually need both or does ASAN work with just allow_user_segv_handler=1?
> 
> We're checking for either here.

Whoops, read that as &&. carry on!

[Mark Lam]

Thanks for the review.  Landed in r236839: <http://trac.webkit.org/r236839>.