After an evaluation, GNOME has moved from Bugzilla to GitLab. Learn more about GitLab.
No new issues can be reported in GNOME Bugzilla anymore.
To report an issue in a GNOME project, go to GNOME GitLab.
Do not go to GNOME Gitlab for: Bluefish, Doxygen, GnuCash, GStreamer, java-gnome, LDTP, NetworkManager, Tomboy.
Bug 747310 - Crash on session startup
Crash on session startup
Status: RESOLVED FIXED
Product: gdm
Classification: Core
Component: general
3.16.x
Other Linux
: Normal critical
: ---
Assigned To: GDM maintainers
GDM maintainers
Depends on:
Blocks:
 
 
Reported: 2015-04-03 17:48 UTC by Vadim Rutkovsky
Modified: 2015-04-16 13:55 UTC
See Also:
GNOME target: ---
GNOME version: ---


Attachments
Stacktrace (6.59 KB, text/x-log)
2015-04-03 17:48 UTC, Vadim Rutkovsky
  Details
manager: Don't double-free x11_display_name (1.35 KB, patch)
2015-04-15 11:51 UTC, Jan Alexander Steffens (heftig)
committed Details | Review

Description Vadim Rutkovsky 2015-04-03 17:48:29 UTC
Created attachment 300903 [details]
Stacktrace

Crashes on Continuous at https://git.gnome.org/browse/gdm/commit/?id=1edb0e4ae897cb1c53af3f3b2b913d5baf731580
Comment 1 Vadim Rutkovsky 2015-04-03 17:50:13 UTC
Please revert https://git.gnome.org/browse/gnome-continuous/commit/?id=a1c4953699612fa4e22033306e34cae5241360eb after the issue is fixed
Comment 2 Jasper St. Pierre (not reading bugmail) 2015-04-03 17:59:06 UTC
segfaulting in malloc? sounds like memory corruption. can you get a valgrind trace?
Comment 3 Jan Alexander Steffens (heftig) 2015-04-15 10:30:58 UTC


  • #0 _int_malloc
    from /usr/lib/libc.so.6
  • #1 malloc
    from /usr/lib/libc.so.6
  • #2 g_malloc
    at gmem.c line 97
  • #3 g_memdup
    at gstrfuncs.c line 384
  • #4 g_variant_type_new_tuple
    at gvarianttype.c line 1112
  • #5 g_variant_make_tuple_type
    at gvariant.c line 804
  • #6 g_variant_builder_end
    at gvariant.c line 3603
  • #7 g_variant_valist_new
    at gvariant.c line 5083
  • #8 g_variant_new_va
    at gvariant.c line 5256
  • #9 g_variant_new
  • #10 invoke_get_property_in_idle_cb
    at gdbusconnection.c line 4180
  • #11 g_main_dispatch
    at gmain.c line 3122
  • #12 g_main_context_dispatch
    at gmain.c line 3737
  • #13 g_main_context_iterate
    at gmain.c line 3808
  • #14 g_main_loop_run
    at gmain.c line 4002

Comment 4 Jan Alexander Steffens (heftig) 2015-04-15 11:29:18 UTC
Bisection points me to either e5a0e92f (manager: find session at registration time) or 507aefdf (manager: set display name on session object at registration time).
Comment 5 Jan Alexander Steffens (heftig) 2015-04-15 11:51:58 UTC
Created attachment 301619 [details] [review]
manager: Don't double-free x11_display_name

Found a double-free in the manager code.
Comment 6 Ray Strode [halfline] 2015-04-16 13:53:10 UTC
Review of attachment 301619 [details] [review]:

seems right to me.  I had a somewhat different fix for this in tree at one point before I pushed, I must have accidentally dropped it instead of squashing it or something.
Comment 7 Ray Strode [halfline] 2015-04-16 13:54:57 UTC
Attachment 301619 [details] pushed as fae288e - manager: Don't double-free x11_display_name