Closed Bug 454942 Opened 16 years ago Closed 16 years ago

TM: After enabling the new JIT support and browsing an ext.js based site, Firefox 3.1 crashes [@ nanojit::LIns::isop]

Categories

(Core :: JavaScript Engine, defect)

x86
Linux
defect
Not set
critical

Tracking

()

RESOLVED WORKSFORME

People

(Reporter: mark, Assigned: jimb)

References

()

Details

(Keywords: crash)

Crash Data

User-Agent:       Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1b1pre) Gecko/20080911020347 Minefield/3.1b1pre
Build Identifier: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1b1pre) Gecko/20080911020347 Minefield/3.1b1pre

After enabling the new JIT support from about:config, and browsing to an ext.js based application/site (such as the extjs.com example mentioned in this ticket) Firefox 3.1 immediately crashes.

Reproducible: Always

Steps to Reproduce:
1. Enabled JIT from about:config
2. Browse to http://extjs.com/deploy/dev/examples/feed-viewer/view.html
3. Cry like a baby, and submit a bug report.

Actual Results:  
Firefox crashes.

Expected Results:  
The application should be rendered.
Status: UNCONFIRMED → NEW
Ever confirmed: true
Summary: After enabling the new JIT support and browsing an ext.js based site, Firefox 3.1 crashes → TM: After enabling the new JIT support and browsing an ext.js based site, Firefox 3.1 crashes
Please read http://mversen.de/crash/
Keywords: crash, stackwanted
I've been able to reproduce this with M-C changeset f76419292bfd, although it takes a bit of fiddling (visiting the page again; opening the history side pane, selecting the page, closing the history pane).
Just loading the page using build Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1b1pre) Gecko/20080915032512 Minefield/3.1b1pre

http://crash-stats.mozilla.com/report/index/5f438a68-833f-11dd-a41f-001cc45a2c28?p=1
Signature	nanojit::LIns::isop(nanojit::LOpcode)
UUID	5f438a68-833f-11dd-a41f-001cc45a2c28
Time	2008-09-15 09:00:04-07
Uptime	715
Product	Firefox
Version	3.1b1pre
Build ID	20080915032512
OS	Windows NT
OS Version	5.1.2600 Service Pack 2
CPU	x86
CPU Info	GenuineIntel family 15 model 2 stepping 9
Crash Reason	EXCEPTION_ACCESS_VIOLATION
Crash Address	0x0
Comments	
Crashing Thread
Frame 	Module 	Signature 	Source
0 	js3250.dll 	nanojit::LIns::isop 	js/src/nanojit/LIR.h:343
1 	js3250.dll 	isi2f 	js/src/jstracer.cpp:290
2 	js3250.dll 	isPromoteInt 	js/src/jstracer.cpp:353
3 	js3250.dll 	TraceRecorder::writeBack 	js/src/jstracer.cpp:1381
4 	js3250.dll 	TraceRecorder::set 	js/src/jstracer.cpp:1411
5 	js3250.dll 	TraceRecorder::record_JSOP_CALL 	js/src/jstracer.cpp:4811
6 	js3250.dll 	js3250.dll@0x66992 	
7 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
8 	js3250.dll 	js_fun_apply 	js/src/jsfun.cpp:1731
9 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
10 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
11 	js3250.dll 	js_fun_apply 	js/src/jsfun.cpp:1731
12 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
13 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
14 	js3250.dll 	fun_call 	js/src/jsfun.cpp:1642
15 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
16 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
17 	js3250.dll 	js_fun_apply 	js/src/jsfun.cpp:1731
18 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
19 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
20 	js3250.dll 	js_fun_apply 	js/src/jsfun.cpp:1731
21 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
22 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
23 	js3250.dll 	js_fun_apply 	js/src/jsfun.cpp:1731
24 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
25 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
26 	js3250.dll 	js_fun_apply 	js/src/jsfun.cpp:1731
27 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
28 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
29 	js3250.dll 	js_fun_apply 	js/src/jsfun.cpp:1731
30 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
31 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
32 	js3250.dll 	js_fun_apply 	js/src/jsfun.cpp:1731
33 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
34 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
35 	js3250.dll 	js_fun_apply 	js/src/jsfun.cpp:1731
36 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
37 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
38 	js3250.dll 	js_fun_apply 	js/src/jsfun.cpp:1731
39 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
40 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
41 	js3250.dll 	js_fun_apply 	js/src/jsfun.cpp:1731
42 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
43 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
44 	js3250.dll 	js_fun_apply 	js/src/jsfun.cpp:1731
45 	js3250.dll 	js_Interpret 	js/src/jsinterp.cpp:4960
46 	js3250.dll 	js_Invoke 	js/src/jsinterp.cpp:1324
47 	xul.dll 	nsXPCWrappedJSClass::CallMethod 	js/src/xpconnect/src/xpcwrappedjsclass.cpp:1523
48 	xul.dll 	nsXPCWrappedJS::CallMethod 	js/src/xpconnect/src/xpcwrappedjs.cpp:565
49 	xul.dll 	PrepareAndDispatch 	xpcom/reflect/xptcall/src/md/win32/xptcstubs.cpp:114
50 	xul.dll 	SharedStub 	xpcom/reflect/xptcall/src/md/win32/xptcstubs.cpp:141
51 	xul.dll 	nsEventListenerManager::HandleEventSubType 	content/events/src/nsEventListenerManager.cpp:1080
Component: General → JavaScript Engine
Keywords: stackwanted
Product: Firefox → Core
QA Contact: general → general
Summary: TM: After enabling the new JIT support and browsing an ext.js based site, Firefox 3.1 crashes → TM: After enabling the new JIT support and browsing an ext.js based site, Firefox 3.1 crashes [@ nanojit::LIns::isop]
Version: unspecified → Trunk
As of Sep 22 10:58 -0500 (427a44cf7ccd), I'm unable to make this crash any more.  Could the original reporter verify and close if he agrees?
This probably argc > nargs which David fixed recently.
Using the latest available nightly (Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1b1pre) Gecko/20080922020442 Minefield/3.1b1pre)

I still see Firefox crashing when enabling the JIT, the crash log/report for this is:

http://crash-stats.mozilla.com/report/index/88897f94-88f7-11dd-9ecc-0013211cbf8a?p=1
Mark: sorry for confusion, Jim and Andreas are talking about tracemonkey builds, not mozilla-central builds. We will update mozilla-central today.

/be
After downloading the latest trace-monkey build from http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/latest-tracemonkey/firefox-3.1b1pre.en-US.linux-i686.tar.bz2 [ Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.1b1pre) Gecko/20080922020927 Minefield/3.1b1pre ] I still saw the crash after enabling the JIT:

http://crash-stats.mozilla.com/report/index/f4ba0e9e-88fa-11dd-84db-001cc4e2bf68?p=1
Assignee: nobody → brendan
Status: NEW → ASSIGNED
Jim, could you please work with Mark to reproduce and diagnose? Thanks,

/be
Assignee: brendan → jim
Confirmed. Just saw the crash.
Will do.
This might be the globalObj->dslots reallocation issue we are debugging at the moment in Zimbra. I will retest once we have that fixed.
I can't reproduce it any more with TM tip. If you still see the bug with the most recent tinderbox build of tracemonkey (ftp://ftp.mozilla.org/pub/firefox/tinderbox-builds/tracemonkey-linux), please re-open.

http://hg.mozilla.org/tracemonkey/rev/e8d1e82b64be
Status: ASSIGNED → RESOLVED
Closed: 16 years ago
Resolution: --- → WORKSFORME
I've not tried compiling from tip, but I just downloaded ftp://ftp.mozilla.org/pub/firefox/tinderbox-builds/tracemonkey-linux/1222201978/firefox-3.1b1pre.en-US.linux-i686.tar.bz2 and still crashed:

http://crash-stats.mozilla.com/report/index/1498da6b-89b7-11dd-a03c-001cc45a2ce4?p=1

I see there's now a newer build since I downloaded so will try that as well.
We just fixed this a few minutes ago so the tinderboxes should pick this up and produce builds "soon".
Lets reopen this until you can confirm that its fixed.
Status: RESOLVED → REOPENED
Resolution: WORKSFORME → ---
Reproduced on tm trunk 2fa92f95eaaf (Wed 13:12 -0700).  Helps to click on "Reading Pane"?
Is bug 455625 a dupe of this? If so I can reproduce it, the underlying crash is bug 456494.
For me - the crash occurs on initial load, so I don't even get a chance to
click on 'reading pane'.  bug 455625 sounds very dupe-like (from a user point of view at least).
Jim, we think this tracemonkey commit fixed this bug among others:

http://hg.mozilla.org/tracemonkey/rev/487cb5edf3c8

Could you verify and if so, close this bug? David sees another crash on the bug URL, which should be filed if it's not already, but we'd like to get past this one. Thanks,

/be
Not sure if this crash is related, but with the latest 3.1 nightly build, WITHOUT the JIT enabled, clicking through my ext.js based app crashed firefox:

http://crash-stats.mozilla.com/report/index/9bf804cf-8aa0-11dd-bb9b-0013211cbf8a?p=1
WFM (including clicking Reading Pane part) using Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9.1b1pre) Gecko/20080925033548 Minefield/3.1b1pre
(In reply to comment #25)
> Not sure if this crash is related, but with the latest 3.1 nightly build,
> WITHOUT the JIT enabled, clicking through my ext.js based app crashed firefox:
> 
> http://crash-stats.mozilla.com/report/index/9bf804cf-8aa0-11dd-bb9b-0013211cbf8a?p=1

That's a different stack trace and you mentioned that JIT is DISABLED. You might want to file a separate bug.
I also think this is fixed with the latest tm sources.  Mark, feel free to re-open this bug if you disagree.
Status: REOPENED → RESOLVED
Closed: 16 years ago16 years ago
Resolution: --- → FIXED
Using the tinderbox/tracemonkey build ftp://ftp.mozilla.org/pub/firefox/tinderbox-builds/tracemonkey-linux/1222377731/firefox-3.1b1pre.en-US.linux-i686.tar.bz2 I saw the feed reader sample page load and display, which is a lot more than I initially saw (or saw with previous builds), however, Firefox still crashed momentarily after the page loaded.

The crash log is:

http://crash-stats.mozilla.com/report/index/86206491-8b4d-11dd-8a96-001cc45a2ce4?p=1

I suspect this may actually be a new problem?  If so, I'll open a new ticket for it.
Status: RESOLVED → REOPENED
Resolution: FIXED → ---
Ok, can you please file a new bug and make sure that "TM:" is in the subject line and mark it as critical. I will pick it up. Please link it here in the comments. I will close this bug here (doesn't crash for me, just verified again, need to figure out why your config still crashes, might be a plugin interaction).
Status: REOPENED → RESOLVED
Closed: 16 years ago16 years ago
Resolution: --- → WORKSFORME
Just entered the additional ticket as https://bugzilla.mozilla.org/show_bug.cgi?id=457107
Crash Signature: [@ nanojit::LIns::isop]
You need to log in before you can comment on or make changes to this bug.