Bug 164893 - Inlining should be disallowed when JSC_alwaysUseShadowChicken=true.
Summary: Inlining should be disallowed when JSC_alwaysUseShadowChicken=true.
Status: RESOLVED FIXED
Alias: None
Product: WebKit
Classification: Unclassified
Component: JavaScriptCore (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Mark Lam
URL:
Keywords: InRadar
Depends on:
Blocks:
 
Reported: 2016-11-17 15:32 PST by Mark Lam
Modified: 2016-11-17 16:24 PST (History)
5 users (show)

See Also:

Attachments
proposed patch. (1.36 KB, patch)
2016-11-17 15:35 PST, Mark Lam 		saam: review+
Details | Formatted Diff | Diff
proposed patch. (1.39 KB, patch)
2016-11-17 15:46 PST, Mark Lam 		saam: review+
Details | Formatted Diff | Diff
Show Obsolete (1) 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 Mark Lam 2016-11-17 15:32:38 PST 
The current shadow chicken implementation isn't written to work with inlining.
Comment 1 Mark Lam 2016-11-17 15:33:03 PST 
<rdar://problem/29146436>
Comment 2 Mark Lam 2016-11-17 15:35:18 PST 
Created attachment 295094 [details]
proposed patch.
Comment 3 Saam Barati 2016-11-17 15:37:43 PST 
Comment on attachment 295094 [details]
proposed patch.

View in context: https://bugs.webkit.org/attachment.cgi?id=295094&action=review

> Source/JavaScriptCore/dfg/DFGByteCodeParser.cpp:1412
> +    if (UNLIKELY(m_hasDebuggerEnabled || Options::alwaysUseShadowChicken())) {

Why not set the inlining depth to 1 inside recomputeDependentOptions?
Comment 4 Saam Barati 2016-11-17 15:38:15 PST 
Comment on attachment 295094 [details]
proposed patch.

Also, do you have a test?
Comment 5 Mark Lam 2016-11-17 15:41:40 PST 
(In reply to comment #3)
> Comment on attachment 295094 [details]
> proposed patch.
> 
> View in context:
> https://bugs.webkit.org/attachment.cgi?id=295094&action=review
> 
> > Source/JavaScriptCore/dfg/DFGByteCodeParser.cpp:1412
> > +    if (UNLIKELY(m_hasDebuggerEnabled || Options::alwaysUseShadowChicken())) {
> 
> Why not set the inlining depth to 1 inside recomputeDependentOptions?

I like this solution better.  I will change the patch.

(In reply to comment #4)
> Comment on attachment 295094 [details]
> proposed patch.
> 
> Also, do you have a test?

Sorry, don't have a reduced one.
Comment 6 Mark Lam 2016-11-17 15:46:25 PST 
Created attachment 295097 [details]
proposed patch.
Comment 7 Mark Lam 2016-11-17 16:24:15 PST 
Thanks for the review.  Landed in r208866: <http://trac.webkit.org/r208866>.