Summary: | ARM/ARMv7: 4th argument register gets clobbered during storePtr call in JIT::updateTopCallFrame | ||
---|---|---|---|
Product: | WebKit | Reporter: | Mandeep Singh Baines <mandeep.baines> |
Component: | JavaScriptCore | Assignee: | Nobody <webkit-unassigned> |
Status: | RESOLVED FIXED | ||
Severity: | Normal | CC: | commit-queue, fpizlo, jbriance, kkristof, mark.lam, msaboff, ossy, rgabor, szledan, zherczeg |
Priority: | P2 | ||
Version: | 528+ (Nightly build) | ||
Hardware: | Unspecified | ||
OS: | Unspecified | ||
Bug Depends on: | |||
Bug Blocks: | 108645 | ||
Attachments: |
Description
Mandeep Singh Baines
2013-10-28 14:15:11 PDT
Hi, I have a fix for this, I'm just waiting that https://bugs.webkit.org/show_bug.cgi?id=123247 lands before Created attachment 215380 [details]
Fix 4th argument register trampling for ARM architecture.
This patch solves a lot of crashes for ARM_TRADITIONAL. Could you test it for ARMv7 and give me your feedback please ?
Created attachment 215381 [details]
Fix 4th argument register trampling for ARM architecture (with ChangeLog)
Better with the ChangeLog
Comment on attachment 215381 [details]
Fix 4th argument register trampling for ARM architecture (with ChangeLog)
I've seen a regression in the RegExpJIT for ARM, and after reading the code, it must be fixed too.
I'll submit a new patch soon
Created attachment 215403 [details]
Fix 4th argument register trampling for ARM architecture (with RegExpJIT fix)
Comment on attachment 215403 [details]
Fix 4th argument register trampling for ARM architecture (with RegExpJIT fix)
r=me. Thanks for cleaning this up.
Comment on attachment 215403 [details] Fix 4th argument register trampling for ARM architecture (with RegExpJIT fix) Clearing flags on attachment: 215403 Committed r158208: <http://trac.webkit.org/changeset/158208> All reviewed patches have been landed. Closing bug. |