[webkit-reviews] review granted: [Bug 170351] JavaScriptCore: Do not use BLX for immediates (ARM-32) : [Attachment 305990] The patch

[bugzilla-daemon at webkit.org]

Mark Lam <mark.lam at apple.com> has granted  review:
Bug 170351: JavaScriptCore: Do not use BLX for immediates (ARM-32)
https://bugs.webkit.org/show_bug.cgi?id=170351

Attachment 305990: The patch

https://bugs.webkit.org/attachment.cgi?id=305990&action=review




--- Comment #2 from Mark Lam <mark.lam at apple.com> ---
Comment on attachment 305990
  --> https://bugs.webkit.org/attachment.cgi?id=305990
The patch

View in context: https://bugs.webkit.org/attachment.cgi?id=305990&action=review

I was wondering why this never posed a problem before for us before.  I did a
local build, and see that the compiled LowLevelInterpreter.o does emit blx
instructions.  However, in the llinked version of the code, the linker changed
the blx into bl instructions.  If the toolchain does not fix this for us, then
this bug would have manifested.

Hence, r=me

> Source/JavaScriptCore/ChangeLog:8
> +	   'call' meta-instruction (which may be found in
LowLevelInterperter.asm

typo: /LowLevelInterperter.asm/LowLevelInterpreter.asm/

> Source/JavaScriptCore/ChangeLog:16
> +	   This behaviour is unwanted, since the C++ code and the offlineasm
code

/offlineasm code/offlineasm generated code/