[webkit-reviews] review granted: [Bug 175772] [Win][Release] Crash when running testmasm executable. : [Attachment 318637] Patch

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Mon Aug 21 22:19:13 PDT 2017


Mark Lam <mark.lam at apple.com> has granted Per Arne Vollan <pvollan at apple.com>'s
request for review:
Bug 175772: [Win][Release] Crash when running testmasm executable.
https://bugs.webkit.org/show_bug.cgi?id=175772

Attachment 318637: Patch

https://bugs.webkit.org/attachment.cgi?id=318637&action=review




--- Comment #3 from Mark Lam <mark.lam at apple.com> ---
Comment on attachment 318637
  --> https://bugs.webkit.org/attachment.cgi?id=318637
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=318637&action=review

r=me

> Source/JavaScriptCore/ChangeLog:9
> +	   We need to save and restore the modified registers in case one or
more registers are callee saved
> +	   on the relevant platforms.

When I wrote the tests, I chose argument registers on purpose because I thought
they would not be callee saved registers.  Of course, I forgot that 32-bit x86
does not actually pass arguments in registers, and it turns out that
argumentGPR3 is ebx, which is a callee saved register.

I think it doesn't hurt to do what you're doing with pushing and popping the
argument registers for all platforms.  It makes the test agnostic of whether
the target platform treats any of the registers as callee saved or not, and
therefore, is more robust in a self-contained way.


More information about the webkit-reviews mailing list