612885 - Around 50% CPU utilized, interface unresponsive, high memory usage

Status: RESOLVED WORKSFORME

(Core :: JavaScript Engine, defect)

Description

[relinquish]

User-Agent:       Mozilla/5.0 (Windows NT 5.1; rv:2.0b7) Gecko/20100101 Firefox/4.0b7
Build Identifier: Mozilla/5.0 (Windows NT 5.1; rv:2.0b7) Gecko/20100101 Firefox/4.0b7

Interface becomes completely unresponsive when visiting this site (and a few others, which I can't recall at the moment). CPU utilization ranges between 49% and 51%. Memory shoots up to near a gigabyte. Can anyone reproduce this?

Reproducible: Always

Steps to Reproduce:
1. Open link in FF4b7
2. Crash

Comment 1

[Maniac Vlad Florin (:vladmaniac)]

Seen this also. Plus, I get the non-responsive script alert message. 

Unfortunately, we don't have yet a concrete reproducing algorithm.  
Uploaded a screenshot of the alert message.

Comment 2

[Maniac Vlad Florin (:vladmaniac)]

Attachment: busy script alert

Comment 3

[(mostly gone) XtC4UaLL [:xtc4uall]]

I can repro this with Beta 7.
It is WFM using a MC Build built from http://hg.mozilla.org/mozilla-central/rev/ad227939db82 and a recent TM Hourly built from http://hg.mozilla.org/tracemonkey/rev/7b8898c9b54c

If you want you can try a Nightly Build via http://nightly.mozilla.org/ or wait for the next Beta.

Comment 4

[juan becerra [:juanb]]

Changing to New since it's been confirmed. Reporter, could you try the more recent nightly and see if this is still a problem in that build?

Comment 5

[relinquish]

Attachment: Testcase: HTML and de-obfuscated js

I de-obfuscated the js from the site that caused the crash.

Comment 6

[relinquish]

Confirmed crash on 17-Nov-2010 build of FFb8pre (Minefield). Firefox 3 doesn't seem to be affected.

I've determined that the issue tied Dean Edwards' js packer code. Note line 24 (of my de-obfuscated version):
p = p.replace(new RegExp('\\b' + e(c) + '\\b', 'g'), k[c]);

When this line is removed from the source, the stalling issue no longer happens.

Comment 7

[relinquish]

I should actually clarify: it is more of a stalling than crash. It stalled my browser for at least one minute if I let it sit on minefield, and then became responsive after that time.

Comment 8

[Shriram (irc: Mavericks)]

(In reply to relinquish from comment #7)
> I should actually clarify: it is more of a stalling than crash. It stalled
> my browser for at least one minute if I let it sit on minefield, and then
> became responsive after that time.

Do you still face this issue ?

Comment 9

[Wayne Mery (:wsmwk)]

WFM on modest vista laptop with current trunk build.
Some high CPU, but not excessive memory