WebKit Bugzilla
New
Browse
Log In
×
Sign in with GitHub
or
Remember my login
Create Account
·
Forgot Password
Forgotten password account recovery
RESOLVED FIXED
144304
Fully replace ENABLE_LLINT_C_LOOP with ENABLE_JIT
https://bugs.webkit.org/show_bug.cgi?id=144304
Summary
Fully replace ENABLE_LLINT_C_LOOP with ENABLE_JIT
Michael Catanzaro
Reported
2015-04-27 18:32:03 PDT
In
bug #143998
we removed ENABLE_LLINT_C_LOOP from the CMake build; now let's remove it completely.
Attachments
Patch
(37.51 KB, patch)
2015-04-27 18:50 PDT
,
Michael Catanzaro
no flags
Details
Formatted Diff
Diff
Patch
(37.02 KB, patch)
2015-04-27 20:27 PDT
,
Michael Catanzaro
no flags
Details
Formatted Diff
Diff
Patch
(36.67 KB, patch)
2015-04-27 20:28 PDT
,
Michael Catanzaro
no flags
Details
Formatted Diff
Diff
Patch
(35.68 KB, patch)
2015-04-28 11:19 PDT
,
Michael Catanzaro
no flags
Details
Formatted Diff
Diff
Show Obsolete
(3)
View All
Add attachment
proposed patch, testcase, etc.
Michael Catanzaro
Comment 1
2015-04-27 18:50:35 PDT
Created
attachment 251805
[details]
Patch
Michael Catanzaro
Comment 2
2015-04-27 20:27:33 PDT
Created
attachment 251813
[details]
Patch
Michael Catanzaro
Comment 3
2015-04-27 20:28:42 PDT
Created
attachment 251814
[details]
Patch
Csaba Osztrogonác
Comment 4
2015-04-28 02:25:24 PDT
There are CLOOP buildbots on
https://build.webkit.org/waterfall
which still use --cloop option, they should be changed to --no-jit. But in this case the buildmaster should be restarted immediately after the patch landed. -
https://trac.webkit.org/browser/trunk/Tools/BuildSlaveSupport/build.webkit.org-config/master.cfg#L233
-
https://trac.webkit.org/browser/trunk/Tools/BuildSlaveSupport/build.webkit.org-config/master.cfg#L491
Maybe it would be less intrusive if we keep --cloop option which could set ENABLE_JIT to 0.
Michael Catanzaro
Comment 5
2015-04-28 11:18:07 PDT
OK, let's keep the --cloop argument to build-jsc then. I was hoping EWS would tell us if there is a problem with this patch, but I guess I'll need to keep an eye on that waterfall after landing.
Michael Catanzaro
Comment 6
2015-04-28 11:19:49 PDT
Created
attachment 251867
[details]
Patch
Geoffrey Garen
Comment 7
2015-04-28 11:25:10 PDT
Sounds like a good idea. Let's see if it builds.
Michael Catanzaro
Comment 8
2015-04-28 12:56:25 PDT
It's fine by EWS at least. If I get r+ then I will check the waterfall after this lands and roll it out if the two cloop bots don't like it, but I bet it'll be fine.
Geoffrey Garen
Comment 9
2015-04-28 14:42:14 PDT
Comment on
attachment 251867
[details]
Patch r=me
Michael Catanzaro
Comment 10
2015-04-28 15:51:43 PDT
Comment on
attachment 251867
[details]
Patch Clearing flags on attachment: 251867 Committed
r183514
: <
http://trac.webkit.org/changeset/183514
>
Michael Catanzaro
Comment 11
2015-04-28 15:51:50 PDT
All reviewed patches have been landed. Closing bug.
Michael Catanzaro
Comment 12
2015-04-28 16:00:04 PDT
It broke the cloop bots: "Cannot build FTL JIT without baseline JIT support!" I will fix it. Ossy, thanks for mentioning that these exist, otherwise I would have had no clue.
WebKit Commit Bot
Comment 13
2015-04-28 16:02:45 PDT
Re-opened since this is blocked by
bug 144359
Michael Catanzaro
Comment 14
2015-04-28 16:17:39 PDT
Committed
r183516
: <
http://trac.webkit.org/changeset/183516
>
Mark Lam
Comment 15
2015-06-01 10:00:43 PDT
(In reply to
comment #14
)
> Committed
r183516
: <
http://trac.webkit.org/changeset/183516
>
This broke the --cloop feature. The CLoop bot has not been building the CLoop ever since. Will fix in
https://bugs.webkit.org/show_bug.cgi?id=145516
.
Michael Catanzaro
Comment 16
2015-06-01 11:04:07 PDT
Oops, I was checking the cloop bots, but I guess "cloop bots are green and happy" is different from "cloop bots are actually testing cloop." :( Thanks for fixing.
Note
You need to
log in
before you can comment on or make changes to this bug.
Top of Page
Format For Printing
XML
Clone This Bug