Bug 60942 – Reduce code size for inline cache

• Component: New Bugs
• Status: RESOLVED FIXED

[Oliver Hunt]

Reduce code size for inline cache

[Oliver Hunt]

Created attachment 93735 [details]
Patch

[WebKit Review Bot]

Attachment 93735 [details] did not pass style-queue:

Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'Source/JavaScriptCore/ChangeLog', u'Source..." exit_code: 1

Source/JavaScriptCore/assembler/CodeLocation.h:133:  Missing space inside { }.  [whitespace/braces] [5]
Source/JavaScriptCore/assembler/CodeLocation.h:135:  Missing space inside { }.  [whitespace/braces] [5]
Source/JavaScriptCore/assembler/CodeLocation.h:137:  Missing space inside { }.  [whitespace/braces] [5]
Source/JavaScriptCore/assembler/X86Assembler.h:1023:  movl_mr_disp8 is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Source/JavaScriptCore/assembler/X86Assembler.h:1098:  movq_mr_disp8 is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Source/JavaScriptCore/assembler/X86Assembler.h:1687:  oneByteOp_disp8 is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Source/JavaScriptCore/assembler/X86Assembler.h:1801:  oneByteOp64_disp8 is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Source/JavaScriptCore/assembler/X86Assembler.h:2040:  memoryModRM_disp8 is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Total errors found: 8 in 21 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Collabora GTK+ EWS bot]

Comment on attachment 93735 [details]
Patch

Attachment 93735 [details] did not pass gtk-ews (gtk):
Output: http://queues.webkit.org/results/8701997

[Oliver Hunt]

Created attachment 93808 [details]
Patch

[WebKit Review Bot]

Attachment 93808 [details] did not pass style-queue:

Failed to run "['Tools/Scripts/check-webkit-style', '--diff-files', u'Source/JavaScriptCore/ChangeLog', u'Source..." exit_code: 1

Source/JavaScriptCore/assembler/X86Assembler.h:1023:  movl_mr_disp8 is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Source/JavaScriptCore/assembler/X86Assembler.h:1098:  movq_mr_disp8 is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Source/JavaScriptCore/assembler/X86Assembler.h:1687:  oneByteOp_disp8 is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Source/JavaScriptCore/assembler/X86Assembler.h:1801:  oneByteOp64_disp8 is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Source/JavaScriptCore/assembler/X86Assembler.h:2041:  memoryModRM_disp8 is incorrectly named. Don't use underscores in your identifier names.  [readability/naming] [4]
Total errors found: 5 in 21 files


If any of these errors are false positives, please file a bug against check-webkit-style.

[Gavin Barraclough]

Comment on attachment 93808 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=93808&action=review

> Source/JavaScriptCore/assembler/AbstractMacroAssembler.h:68
> +    static const int MaximumCompactPtrAlignedAddressOffset = 0x7FFFFFFF;

I think defining this here could be confusing, since you have have implementations of load32WithAddressOffsetPatch in all macro assemblers.  To avoid programmer errors I suggest moving this down into the architecture specific MacroAssemblers.

> Source/JavaScriptCore/assembler/MacroAssemblerX86Common.h:44
> +    static const int MaximumCompactPtrAlignedAddressOffset = ((1 << 8) - 1) / 2;

I think just 127 would be clearer here.

[Oliver Hunt]

Committed r86699: <http://trac.webkit.org/changeset/86699>

[thouraya]

Created attachment 94436 [details]
fix issue

Hello,

Fix issues related to the patch:
added load32WithCompactAddressOffsetPatch function and removed the duplicated function load32WithAddressOffsetPatch.

Regards,
Thouraya.

[Oliver Hunt]

Comment on attachment 94436 [details]
fix issue

I think it would be better simply to implement the compact version properly -- it looked like sh4 should be able to do this in a single instruction saving memory + cpu time