[Webkit-unassigned] [Bug 146440] Crash on xLarge memory allocation using bmalloc on 32bit systems

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=146440

--- Comment #16 from Carlos Garcia Campos <cgarcia at igalia.com> ---
Comment on attachment 256038
  --> https://bugs.webkit.org/attachment.cgi?id=256038
Patch proposal

View in context: https://bugs.webkit.org/attachment.cgi?id=256038&action=review

Do you know how important this optimization is? My main concern, like with all other patches that disable optimizations to work around compiler bugs, is that it's very difficult to know when we can stop disabling it, and we end up with the optimization disabled forever. Also, I think we should report these bugs to GCC as well.

>>> Source/WebCore/CMakeLists.txt:3581
>>> +if (CMAKE_COMPILER_IS_GNUCXX AND "${LOWERCASE_CMAKE_HOST_SYSTEM_PROCESSOR}" MATCHES "(i[3-6]86|x86)")
>> 
>> I think this should be: ... MATCHES "(i[3-6]86|x86)$")
>> 
>> Otherwise this becomes true also for AMD64 with GCC (where LOWERCASE_CMAKE_HOST_SYSTEM_PROCESSOR = x86_64 ). (yes, there is a bug on the previous code introduced on 128855 <http://trac.webkit.org/r128855>)
> 
> Also I don't think this will work for ARM 32-bit machines or $RANDOMARCH 32-bit machines.

Yes, we do a similar match to add the --no-keep-memory linker option in 32 bits. I think we could move it to a common place, detecting also non intel 32 bits processors and exposing a variable to CMake.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20150703/e58c639b/attachment.html>