Summary: | Cache flush problem on ARMv7 JSC | ||||||
---|---|---|---|---|---|---|---|
Product: | WebKit | Reporter: | Gabor Rapcsanyi <rgabor> | ||||
Component: | JavaScriptCore | Assignee: | Nobody <webkit-unassigned> | ||||
Status: | RESOLVED FIXED | ||||||
Severity: | Normal | CC: | barraclough, benjamin, cmarcelo, fpizlo, hausmann, laszlo.gombos, ojan.autocc, oliver, ossy, psychon, webkit.review.bot, zherczeg | ||||
Priority: | P2 | ||||||
Version: | 528+ (Nightly build) | ||||||
Hardware: | Unspecified | ||||||
OS: | Linux | ||||||
Bug Depends on: | |||||||
Bug Blocks: | 108645 | ||||||
Attachments: |
|
Description
Gabor Rapcsanyi
2013-03-05 07:50:09 PST
*** Bug 108632 has been marked as a duplicate of this bug. *** Could you revert Simon's workaround ( http://trac.webkit.org/changeset/144170 ) after or with this patch? Created attachment 191737 [details]
proposed fix
In another patch we should make it more sophisticated but for now it solves the problem.
I'm also reverting Simon's patch with this.
Yay, with this patch there are only 171 remaining inspector crashes on Thumb2, but they are unrelated to this bug - http://build.webkit.sed.hu/builders/ARMv7%20Linux%20Qt5%20Release%20%28Test%29/builds/7987 Comment on attachment 191737 [details]
proposed fix
r=me.
But I think the current flush mechanism invalidates a large amount of unnecessary, so we should improve with it.
Comment on attachment 191737 [details] proposed fix Clearing flags on attachment: 191737 Committed r145194: <http://trac.webkit.org/changeset/145194> All reviewed patches have been landed. Closing bug. |