[webkit-reviews] review denied: [Bug 227204] Add a new pattern to instruction selector to utilize UBFIZ supported by ARM64 : [Attachment 431939] Patch

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Tue Jun 22 08:56:51 PDT 2021


Filip Pizlo <fpizlo at apple.com> has denied Yijia Huang <yijia_huang at apple.com>'s
request for review:
Bug 227204: Add a new pattern to instruction selector to utilize UBFIZ
supported by ARM64
https://bugs.webkit.org/show_bug.cgi?id=227204

Attachment 431939: Patch

https://bugs.webkit.org/attachment.cgi?id=431939&action=review




--- Comment #5 from Filip Pizlo <fpizlo at apple.com> ---
Comment on attachment 431939
  --> https://bugs.webkit.org/attachment.cgi?id=431939
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=431939&action=review

> Source/JavaScriptCore/ChangeLog:19
> +	   d = (n << lsb) & mask

Why not instead select (n & mask) << lsb?

Then, you can canonicalize (n << lsb) & mask to (n & mask) << lsb in
reduceStrength.

The benefit of (n & mask) << lsb being the canonical form is it's easier to
recognize. So, if some other part of the compiler needed to recognize this
pattern, they'd only have to recognize the easier one of them.


More information about the webkit-reviews mailing list