[webkit-reviews] review granted: [Bug 193993] [JSC] Shrink size of VM by lazily allocating IsoSubspaces for non-common types : [Attachment 361096] Patch

[bugzilla-daemon at webkit.org]

Keith Miller <keith_miller at apple.com> has granted Yusuke Suzuki
<ysuzuki at apple.com>'s request for review:
Bug 193993: [JSC] Shrink size of VM by lazily allocating IsoSubspaces for
non-common types
https://bugs.webkit.org/show_bug.cgi?id=193993

Attachment 361096: Patch

https://bugs.webkit.org/attachment.cgi?id=361096&action=review




--- Comment #34 from Keith Miller <keith_miller at apple.com> ---
Comment on attachment 361096
  --> https://bugs.webkit.org/attachment.cgi?id=361096
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=361096&action=review

r=me with some nits. Looks good!

> Source/JavaScriptCore/ChangeLog:31
> +	   can run this IsoSubspace code, the collector is never EndPhase. So
this is safe.

Is there anyway to assert that the GC isn't in EndPhase when creating an
IsoSubspace? Even if it's racy that might still be helpful. I would only worry
about this if it's like a one line change though.

> Source/JavaScriptCore/runtime/VM.cpp:1248
> +

Can we undef DYNAMIC_ISO_SUBSPACE_DEFINE_MEMBER_SLOW here?

> Source/JavaScriptCore/runtime/VM.cpp:1263
> +

Ditto on undef.

> Source/JavaScriptCore/runtime/VM.h:406
>      

Ditto on undef.

> Source/JavaScriptCore/runtime/VM.h:458
> +    DYNAMIC_SPACE_AND_SET_DEFINE_MEMBER(moduleProgramExecutableSpace)

Ditto on undef.