Hide Forgot
abrt version: 1.1.14 architecture: i686 Attached file: backtrace cmdline: /usr/lib/firefox-3.6/firefox component: firefox crash_function: PR_JoinThread executable: /usr/lib/firefox-3.6/firefox kernel: 2.6.35.10-74.fc14.i686.PAE package: firefox-3.6.13-1.fc14 rating: 4 reason: Process /usr/lib/firefox-3.6/firefox was killed by signal 11 (SIGSEGV) release: Fedora release 14 (Laughlin) Attached file: reproduce time: 1296650806 uid: 502
Created attachment 476559 [details] File: backtrace
Created attachment 476560 [details] File: reproduce
Removing nspluginwrapper makes one of the SE messages go away, but setsched still fails.
dwalsh said on bug 673224: """ sandbox -X firefox will blow up because the standard SELinux type does not allow connections to http ports, and I guess firefox does not handle this well. Using sandbox_web_t for the type should allow the access. """ The topic of this issue is that firefox should handle that it is blocked without crashing.
I think it is failing because we are not allowing setsched call. Maybe we should allow this in sandbox_x_t and then you could run firefox without network access.
We're using mozilla crash reporter now, ABRT is no more used for Firefox/Thunderbird. If you can reliably reproduce the crash (you have a testcase, reproduction steps, etc.) please reopen the bug and attach the reproduction info and assign it directly to me (stransky). Thanks!
I canøt reproduce the crash on f16 - perhaps because the policy has changed.