Bug 235388 - [JSC] Fix non-JIT Windows LLInt
Summary: [JSC] Fix non-JIT Windows LLInt
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: New Bugs (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Yusuke Suzuki
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2022-01-19 16:51 PST by Yusuke Suzuki
Modified: 2022-01-20 12:07 PST (History)
7 users (show)

See Also:

Attachments
Patch (8.03 KB, patch)
2022-01-19 16:53 PST, Yusuke Suzuki 		ews-feeder: commit-queue-
Details | Formatted Diff | Diff
Patch (8.77 KB, patch)
2022-01-19 16:57 PST, Yusuke Suzuki 		mark.lam: review+
Details | Formatted Diff | Diff
Patch (9.56 KB, patch)
2022-01-19 17:34 PST, Yusuke Suzuki 		no flags Details | Formatted Diff | Diff
Patch (9.98 KB, patch)
2022-01-19 17:36 PST, Yusuke Suzuki 		no flags Details | Formatted Diff | Diff
Show Obsolete (2) View All Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Yusuke Suzuki 2022-01-19 16:51:24 PST 
[JSC] Fix non-JIT Windows LLInt
Comment 1 Yusuke Suzuki 2022-01-19 16:53:03 PST 
Created attachment 449531 [details]
Patch
Comment 2 Yusuke Suzuki 2022-01-19 16:57:29 PST 
Created attachment 449532 [details]
Patch
Comment 3 Mark Lam 2022-01-19 17:29:20 PST 
Comment on attachment 449532 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=449532&action=review

r=me with fixes.

> Source/JavaScriptCore/llint/LowLevelInterpreter32_64.asm:131
> +    elsif C_LOOP or C_LOOP_WIN
> +        cloopCallSlowPath3 function, a0, a1, a2

This should come first because C_LOOP can have X86 or X86_WIN enabled too (or any other CPU arch).  I'm surprised that it works for cCall2 (don't know why).  Maybe should fix that too for consistency.

> Source/JavaScriptCore/llint/LowLevelInterpreter64.asm:165
> +        addp 64, sp

Maybe this `addp` should come after the register moves below.  r0 is still pointing to the reserved stack position, right?  They should remain protected by sp until after we move the results out into registers.
Comment 4 Yusuke Suzuki 2022-01-19 17:30:39 PST 
Comment on attachment 449532 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=449532&action=review

Thank you!

>> Source/JavaScriptCore/llint/LowLevelInterpreter32_64.asm:131
>> +        cloopCallSlowPath3 function, a0, a1, a2
> 
> This should come first because C_LOOP can have X86 or X86_WIN enabled too (or any other CPU arch).  I'm surprised that it works for cCall2 (don't know why).  Maybe should fix that too for consistency.

OK, done :)

>> Source/JavaScriptCore/llint/LowLevelInterpreter64.asm:165
>> +        addp 64, sp
> 
> Maybe this `addp` should come after the register moves below.  r0 is still pointing to the reserved stack position, right?  They should remain protected by sp until after we move the results out into registers.

Sounds good, done :)
Comment 5 Yusuke Suzuki 2022-01-19 17:34:10 PST 
Created attachment 449534 [details]
Patch
Comment 6 Yusuke Suzuki 2022-01-19 17:36:04 PST 
Created attachment 449535 [details]
Patch
Comment 7 EWS 2022-01-19 20:04:21 PST 
Committed r288268 (246210@main): <https://commits.webkit.org/246210@main>

All reviewed patches have been landed. Closing bug and clearing flags on attachment 449535 [details].
Comment 8 Radar WebKit Bug Importer 2022-01-20 12:07:43 PST 
<rdar://problem/87842831>