From Bugzilla Helper: User-Agent: Mozilla/5.0 (X11; U; Linux 2.2.17 i686; en-US; 0.7) Gecko/20010123 BuildID: My build of 0.7 with --disable-debug --enable-optimize when moving mouse over mozilla you get high cpu usage Reproducible: Always Steps to Reproduce: 1. go to page mentioned above. 2. start top to watch cpu user usage. 3. move mouse up and down over the left fenced side section (with a lot of links). 4. watch cpu usage while moving mose Actual Results: Cpu usage on my machine (AMD 800 MHz 256MB memory) goes up to nearly 50 %. Expected Results: On netscape 4.75 when doing exactly the same thing it goes to 4.5% only. Ok I recompiled with the options mentioned above. The effect got somewhat less (on ordinary page with less links usage goes now upto 25% while without options it went above 40%. But with above page you can still see huge difference with netscape 4.75 (48% vs 4.5%), over ten times more cpu usage. So there IS something seriously wrong in your code in system level. And yes there is the stupid looping poll still: > ps ax ... 21066 ? S 0:04 netscape 21067 ? S 0:00 (dns helper) 21070 ? S 0:00 sh /usr/local/mozilla/dist/bin/run-mozilla.sh /usr/lo 21075 ? S 0:55 /usr/local/mozilla/dist/bin/mozilla-bin 21076 ? S 0:00 /usr/local/mozilla/dist/bin/mozilla-bin 21077 ? S 0:00 /usr/local/mozilla/dist/bin/mozilla-bin 21078 ? S 0:00 /usr/local/mozilla/dist/bin/mozilla-bin ... > strace -tt -p 21075 ... 01:38:35.029435 gettimeofday({980293115, 29626}, NULL) = 0 01:38:35.029968 gettimeofday({980293115, 30159}, NULL) = 0 01:38:35.030505 ioctl(8, FIONREAD, [0]) = 0 01:38:35.031050 poll([{fd=8, events=POLLIN}, {fd=15, events=POLLIN}, {fd=5, events=POLLIN}], 3, 7) = 0 01:38:35.047181 gettimeofday({980293115, 47372}, NULL) = 0 01:38:35.047720 gettimeofday({980293115, 47911}, NULL) = 0 01:38:35.048260 ioctl(8, FIONREAD, [0]) = 0 01:38:35.048801 poll([{fd=8, events=POLLIN}, {fd=15, events=POLLIN}, {fd=5, events=POLLIN}], 3, 0) = 0 01:38:35.049405 gettimeofday({980293115, 49598}, NULL) = 0 01:38:35.049943 gettimeofday({980293115, 50133}, NULL) = 0 01:38:35.050480 ioctl(8, FIONREAD, [0]) = 0 01:38:35.051023 poll([{fd=8, events=POLLIN}, {fd=15, events=POLLIN}, {fd=5, events=POLLIN}], 3, 7) = 0 01:38:35.067174 gettimeofday({980293115, 67368}, NULL) = 0 01:38:35.067719 gettimeofday({980293115, 67906}, NULL) = 0 01:38:35.068257 ioctl(8, FIONREAD, [0]) = 0 01:38:35.068802 poll([{fd=8, events=POLLIN}, {fd=15, events=POLLIN}, {fd=5, events=POLLIN}], 3, 0) = 0 ... And this goes on indefenetly while I'm doing nothing, just watching. In comparision same with netscape: > strace -tt -p 21066 ... 01:43:00.867954 oldselect(18, [5 6 8 9 17], NULL, [5], {137, 890000} ... And it stays sleeping (like it should) till I move mouse over the netscape. So something isn't right in mozilla in this matter either! Pekka Sarnila

Confirmed Platform: PC OS: Linux 2.2.16 Mozilla Build: 200101223 Marking NEW.

any guess for the component?

over to event handling for investigation

Well as far as comparisons with Netscape4.x, Mozilla simply does a lot more with pseudo-styles relating to mouse positioning over content than 4.x ever could do so I'd expect more CPU. I'll admit 50% is a problem but from non-crashing, non-feature breaking nature I'm going to mark this future for now. This bug has been marked "future" because the original netscape engineer working on this is over-burdened. If you feel this is an error, that you or another known resource will be working on this bug,or if it blocks your work in some way -- please attach your concern to the bug for reconsideration.

For us, marking it as future is indeed a big problem. We rely on performance for our application and such a resources leak is quite disastrous to us. Please, provide a fix as soon as possible. GD

QA contact updated

On a P3/500 i see top show CPU usage between 10% and 13% when mousing over the links as described. Never saw it go higher than 13%. Tested w/ day old CVS build, build params: ac_add_options --disable-accessibility ac_add_options --disable-bidi ac_add_options --enable-strip-libs ac_add_options --disable-ldap ac_add_options --disable-dtd-debug ac_add_options --disable-debug ac_add_options --disable-tests ac_add_options --disable-logging ac_add_options --enable-crypto ac_add_options --enable-optimize=-O2 ac_add_options --disable-jsd ac_add_options --disable-venkman mk_add_options --mcpu=i686 mk_add_options --march=i686

I have similar experience under Windows 2000. For me is even worse since the CPU usage stays really high (70-100% depending on number of windows/tabs open) even after I minimize Mozilla.

.

Is this bugfile still active? It seems that http://www.mozilla.org/quality/help/bugzilla-helper.html no longer exists. If this bug is still active, I wish someone could re-edit the summary accordingly to make it more useful when searching for duplicates. Thanks.

No url, no testcase, no replies, no specifics in this bug. Since nothing seems to be updated and/or nothing seems to be happening in this bugfile... Resolving as WORKSFORME