[webkit-reviews] review granted: [Bug 101084] Speed up TransformationMatrix::multiply() on modern ARM : [Attachment 172112] Patch

[bugzilla-daemon at webkit.org]

Gavin Barraclough <barraclough at apple.com> has granted Benjamin Poulain
<benjamin at webkit.org>'s request for review:
Bug 101084: Speed up TransformationMatrix::multiply() on modern ARM
https://bugs.webkit.org/show_bug.cgi?id=101084

Attachment 172112: Patch
https://bugs.webkit.org/attachment.cgi?id=172112&action=review

------- Additional Comments from Gavin Barraclough <barraclough at apple.com>
View in context: https://bugs.webkit.org/attachment.cgi?id=172112&action=review


R+ assuming that for loop is closed somewhere!

> Source/WebCore/platform/graphics/transforms/TransformationMatrix.cpp:978
> +	   "mov        r3, %[leftMatrix]\n\t"

rather than manually allocating r3, maybe it would be nicer to have a
leftMatrixIn, leftMatrixOut, let the compiler pick.

> Source/WebCore/platform/graphics/transforms/TransformationMatrix.cpp:1108
> +    for (unsigned i = 0; i < 4; ++i) {

I don't see this for loop being closed?