Title: [243330] trunk/Source/_javascript_Core
Revision
243330
Author
[email protected]
Date
2019-03-21 14:36:26 -0700 (Thu, 21 Mar 2019)

Log Message

B3::Opcode can fit in a single byte, shrinking B3Value by 8 bytes
https://bugs.webkit.org/show_bug.cgi?id=196014

Reviewed by Keith Miller.

B3::Opcode has less than one hundred cases, so it can easily fit in one byte (from two currently)
This shrinks B3::Kind from 4 bytes to 2 (by removing the byte of padding at the end).
This in turns eliminate padding from B3::Value, shrinking it by 8 bytes (out of 80).

* b3/B3Opcode.h:

Modified Paths

Diff

Modified: trunk/Source/_javascript_Core/ChangeLog (243329 => 243330)


--- trunk/Source/_javascript_Core/ChangeLog	2019-03-21 21:16:37 UTC (rev 243329)
+++ trunk/Source/_javascript_Core/ChangeLog	2019-03-21 21:36:26 UTC (rev 243330)
@@ -1,3 +1,16 @@
+2019-03-21  Robin Morisset  <[email protected]>
+
+        B3::Opcode can fit in a single byte, shrinking B3Value by 8 bytes
+        https://bugs.webkit.org/show_bug.cgi?id=196014
+
+        Reviewed by Keith Miller.
+
+        B3::Opcode has less than one hundred cases, so it can easily fit in one byte (from two currently)
+        This shrinks B3::Kind from 4 bytes to 2 (by removing the byte of padding at the end).
+        This in turns eliminate padding from B3::Value, shrinking it by 8 bytes (out of 80).
+
+        * b3/B3Opcode.h:
+
 2019-03-21  Michael Catanzaro  <[email protected]>
 
         Unreviewed, more clang 3.8 build fixes

Modified: trunk/Source/_javascript_Core/b3/B3Opcode.h (243329 => 243330)


--- trunk/Source/_javascript_Core/b3/B3Opcode.h	2019-03-21 21:16:37 UTC (rev 243329)
+++ trunk/Source/_javascript_Core/b3/B3Opcode.h	2019-03-21 21:36:26 UTC (rev 243330)
@@ -37,7 +37,7 @@
 // Warning: In B3, an Opcode is just one part of a Kind. Kind is used the way that an opcode
 // would be used in simple IRs. See B3Kind.h.
 
-enum Opcode : int16_t {
+enum Opcode : uint8_t {
     // A no-op that returns Void, useful for when you want to remove a value.
     Nop,
     
_______________________________________________
webkit-changes mailing list
[email protected]
https://lists.webkit.org/mailman/listinfo/webkit-changes

Reply via email to