Closed
Bug 1586662
Opened 5 years ago
Closed 4 years ago
MacroAssembler::move64 doesn't use xor instruction for zeroing register
Categories
(Core :: JavaScript Engine: JIT, enhancement, P3)
Core
JavaScript Engine: JIT
Tracking
()
RESOLVED
FIXED
87 Branch
People
(Reporter: anba, Assigned: jandem)
References
Details
Attachments
(1 file)
|
(deleted),
text/x-phabricator-request
|
Details |
x64 emits (for example):
[Codegen] movl $0x0, %edx
x86 emits (for example):
[Codegen] movl $0x0, %eax
[Codegen] movl $0x0, %ebp
when calling masm.move64(Imm64(0), dest);.
|
||
Comment 1•5 years ago
|
||
Be aware that xor changes the flags, and thus may not be usable in all situations.
|
||
Updated•5 years ago
|
Priority: -- → P3
|
Assignee | |
Comment 2•4 years ago
|
||
In theory xor could unexpectedly affect flags, but looking at the callers we don't
depend on flags being unaffected and this matches move32.
|
||
Updated•4 years ago
|
Assignee: nobody → jdemooij
Status: NEW → ASSIGNED
Pushed by jdemooij@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/d306abe67e3f Use mov instead of movq for move64 to benefit from xor optimization. r=iain
|
||
Comment 4•4 years ago
|
||
| bugherder | ||
Status: ASSIGNED → RESOLVED
Closed: 4 years ago
status-firefox87:
--- → fixed
Resolution: --- → FIXED
Target Milestone: --- → 87 Branch
|
||
Updated•4 years ago
|
You need to log in
before you can comment on or make changes to this bug.
Description
•