http://bugzilla.novell.com/show_bug.cgi?id=555027 http://bugzilla.novell.com/show_bug.cgi?id=555027#c33 Stanislav Brabec <sbrabec@novell.com> changed: What |Removed |Added ---------------------------------------------------------------------------- Priority|P2 - High |P1 - Urgent Status|REOPENED |ASSIGNED --- Comment #33 from Stanislav Brabec <sbrabec@novell.com> 2010-01-25 12:24:57 CET --- Debugging this issue a bit more, it seems to be another race. Both session manager and metacity gets SIGTERM. If the processing is fast enough, metacity terminates before processing the signal from session manager. But both attempt to close session (smcConn). Bad things will happen, if metacity gets time to process both: It attempts to close already closed session. It causes SIGSEGV and zombie. This is a typical backtrace of this bug: Program received signal SIGSEGV, Segmentation fault. 0xb6f53cc0 in SmcSetProperties (smcConn=0x814a2f0, numProps=1, props=0xbff34ac8) at sm_client.c:373 373 IceGetHeader (iceConn, _SmcOpcode, SM_SetProperties, #0 0xb6f53cc0 in SmcSetProperties (smcConn=0x814a2f0, numProps=1, props=0xbff34ac8) at sm_client.c:373 #1 0x0807d0e1 in meta_session_shutdown () #2 0x080724df in main () Looking at metacity git log, I found a fix for that issue: edeadf6 And an upstream bug: https://bugzilla.gnome.org/show_bug.cgi?id=604867 The patch there seems to fix prompt login issue without breaking autologin issue. Testing package with this patch will appear in home:sbrabec:branches:openSUSE:11.2:Update. It should contain: Mon Jan 25 12:06:32 CET 2010 - sbrabec@suse.cz - Fix a race condition that causes crash during gdm login sometimes and login never completes (bnc#555027, bgo#604867). Please test. It fixed the issue for me, at least during my short test. Well, but one of my backtraces from previous testing look differently: Program received signal SIGSEGV, Segmentation fault. 0xb6ecbc55 in ?? () from /lib/libc.so.6 #0 0xb6ecbc55 in ?? () from /lib/libc.so.6 #1 0xb6ece0ba in malloc () from /lib/libc.so.6 #2 0xb6f450f5 in __vasprintf_chk () from /lib/libc.so.6 #3 0xb71cf656 in g_vasprintf () from /usr/lib/libglib-2.0.so.0 #4 0xb71bb0d7 in g_strdup_vprintf () from /usr/lib/libglib-2.0.so.0 #5 0x0807ee7d in meta_warning () #6 0x0808d54c in meta_workspace_focus_default_window () #7 0x0808bb27 in meta_window_free () #8 0x0806242e in meta_display_unmanage_windows_for_screen () #9 0x08078e70 in meta_screen_free () #10 0x08061eb6 in meta_display_close () #11 0x080724f3 in main () LND: Sending signal 11 to Thread 0xb6a1b750 (LWP 5404) Looking at metacity changelog, there seems to be yet another race: https://bugzilla.gnome.org/show_bug.cgi?id=596200 The patch is a bit large (~9kB), but maybe we should consider it as well: db37deb + 19d85c8 -- Configure bugmail: http://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.