User Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:29.0) Gecko/20100101 Firefox/29.0 Build ID: 20151104000845 Steps to reproduce: Just open 4 new tabs via Ctrl+Click on links from Google search. I have near 1500 tabs and 100 of them are active. about:memory shown 5 GiB memory allocated. I run Firefox for a long time (several days with "sudo pm-suspend") and browser rich pages (coursera.org, etc). Following report measured during Firefox hangs: https://cleopatra.io/#report=c869295afa8d5f001ca2642e8dbf85a695b62ba4 Following report measured some time later, when Firefox become responsive again: https://cleopatra.io/#report=6b7fb1a0c492f616f3eb08dceb14c3f82e88fc8f Both report show that near 29% / 32% time spent in Timer::Fire => (JS:RunScript + JS::UnmarkGrayGCThingRecursively) and profiler shown site icons / URLs. I assume that activity just regular JS activity on browser tabs. I see difference in another leaf - "NM_observeActivity" call, here it is: https://dxr.mozilla.org/mozilla-central/source/devtools/shared/webconsole/network-monitor.js#664 When Firefox hangs it spent 46% of time in that function while second report spent near 29% in __poll call of glibc.so. Actual results: Firefox hangs for several minutes with long or no response.
Sorry, I fake **user agent** as some sites don't work with Iceweasel string. It is no longer in Debian experimental (at 42.0-1 for AMD64), but here is old signed announce for build: https://packages.qa.debian.org/i/iceweasel/news/20151104T070026Z.html
You need to log in before you can comment on or make changes to this bug.