Bug 125649 - cStack: Calling into FTL-generated code should be fast
Summary: cStack: Calling into FTL-generated code should be fast
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: JavaScriptCore (show other bugs)
Version: 528+ (Nightly build)
Hardware: All All
: P2 Normal
Assignee: Filip Pizlo
URL:
Keywords:
Depends on:
Blocks: 113621
  Show dependency tree / graph
 
Reported: 2013-12-12 13:00 PST by Filip Pizlo
Modified: 2014-01-30 13:23 PST (History)
8 users (show)

See Also:

Attachments
starting (10.09 KB, patch)
2013-12-12 13:56 PST, Filip Pizlo 		no flags Details | Formatted Diff | Diff
moar (37.05 KB, patch)
2013-12-13 14:29 PST, Filip Pizlo 		no flags Details | Formatted Diff | Diff
FTL-on-C-stack now knows that 1+1=2 (43.53 KB, patch)
2013-12-13 15:13 PST, Filip Pizlo 		no flags Details | Formatted Diff | Diff
it sort of runs crypto (51.40 KB, patch)
2013-12-13 21:24 PST, Filip Pizlo 		no flags Details | Formatted Diff | Diff
the patch (54.42 KB, patch)
2013-12-14 10:54 PST, Filip Pizlo 		ggaren: review+
Details | Formatted Diff | Diff
Show Obsolete (4) 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 Filip Pizlo 2013-12-12 13:00:52 PST 
Patch forthcoming.
Comment 1 Filip Pizlo 2013-12-12 13:56:53 PST 
Created attachment 219116 [details]
starting
Comment 2 Filip Pizlo 2013-12-13 14:29:05 PST 
Created attachment 219199 [details]
moar
Comment 3 Filip Pizlo 2013-12-13 15:13:55 PST 
Created attachment 219204 [details]
FTL-on-C-stack now knows that 1+1=2

It ran the first program.
Comment 4 Filip Pizlo 2013-12-13 21:24:58 PST 
Created attachment 219232 [details]
it sort of runs crypto

You have to disable the LLInt and GC, but it sort of works.
Comment 5 Filip Pizlo 2013-12-14 10:54:25 PST 
Created attachment 219254 [details]
the patch
Comment 6 Filip Pizlo 2013-12-14 13:35:08 PST 
Landed in http://trac.webkit.org/changeset/160600
Comment 7 Geoffrey Garen 2013-12-16 12:15:26 PST 
Comment on attachment 219254 [details]
the patch

View in context: https://bugs.webkit.org/attachment.cgi?id=219254&action=review

r=me

> Source/JavaScriptCore/runtime/Options.h:211
> +    v(bool, disableGC, false) \

I prefer for settings like this to be stated in the affirmative: "enableGC", defaulting to true. "Don't disable GC" is a hard phrase to parse.
Comment 8 Mark Lam 2014-01-13 18:44:21 PST 
Review status updated in r161938: <http://trac.webkit.org/r161938>.