Bug 732525 – crash after useradd

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 732525 - crash after useradd


Summary:	crash after useradd


Status:	RESOLVED FIXED

Product:	gnome-initial-setup
Classification:	Applications
Component:	general
Version:	unspecified
Hardware:	Other Linux

Importance:	Normal critical
Target Milestone:	---
Assigned To:	GNOME Initial Setup maintainer(s)
QA Contact:	GNOME Initial Setup maintainer(s)

URL:
Whiteboard:

Depends on:
Blocks:

Reported:	2014-06-30 22:40 UTC by Colin Walters
Modified:	2014-08-01 09:16 UTC

See Also:
GNOME target:	---
GNOME version:	---

Description Colin Walters 2014-06-30 22:40:07 UTC

Hitting this in latest Continuous:

+ Trace 233748

Thread 1 (Thread 0x7f2931937980 (LWP 425))

#0 __strcmp_sse2
at ../sysdeps/x86_64/strcmp.S line 212
#1 validate
at ../../../../gnome-initial-setup/pages/password/gis-password-page.c line 131
#2 g_closure_invoke
at ../../gobject/gclosure.c line 768
#3 signal_emit_unlocked_R
at ../../gobject/gsignal.c line 3553
#4 g_signal_emit_valist
at ../../gobject/gsignal.c line 3309
#5 g_signal_emit
at ../../gobject/gsignal.c line 3365
#6 g_object_dispatch_properties_changed
at ../../gobject/gobject.c line 1056
#7 g_object_notify_by_spec_internal
at ../../gobject/gobject.c line 1150
#8 g_object_notify
at ../../gobject/gobject.c line 1197
#9 gis_driver_set_username
at ../../gnome-initial-setup/gis-driver.c line 148
#10 gis_account_page_local_apply
at ../../../../gnome-initial-setup/pages/account/gis-account-page-local.c line 552
#11 gis_page_apply_begin
at ../../gnome-initial-setup/gis-page.c line 303
#12 _g_closure_invoke_va
at ../../gobject/gclosure.c line 831
#13 g_signal_emit_valist
at ../../gobject/gsignal.c line 3218
#14 g_signal_emit
at ../../gobject/gsignal.c line 3365
#15 button_activate_timeout
at ../../gtk/gtkbutton.c line 2025
#16 gdk_threads_dispatch
at ../../gdk/gdk.c line 634
#17 g_timeout_dispatch
at ../../glib/gmain.c line 4472
#18 g_main_dispatch
at ../../glib/gmain.c line 3064
#19 g_main_context_dispatch
at ../../glib/gmain.c line 3663
#20 g_main_context_iterate
at ../../glib/gmain.c line 3734
#21 g_main_context_iteration
at ../../glib/gmain.c line 3795
#22 g_application_run
at ../../gio/gapplication.c line 2115
#23 main
at ../../gnome-initial-setup/gnome-initial-setup.c line 226

Comment 1 Jasper St. Pierre (not reading bugmail) 2014-06-30 22:51:32 UTC

The only way this can happen is if gtk_entry_get_text returns NULL or an invalid pointer. Smells like memory corruption to me. I can't reproduce here on my machine. I might test on a continuous VM in a bit.

Comment 2 Adam Williamson 2014-07-16 17:01:00 UTC

I've been hitting a crash which *sounds* similar but for which the backtrace does not look the same, in Fedora Rawhide / 21:

https://bugzilla.redhat.com/show_bug.cgi?id=1116478

it's accepted as an F21 Alpha blocker currently. The 'user experience' is that the first time you hit the user creation step after a fresh install, g-i-s crashes, leaving you in an unusable position. If you then reboot the system, g-i-s runs again, and user creation succeeds the second time. This seems to be 100% reproducible for me.

Comment 3 Adam Williamson 2014-07-25 00:44:16 UTC

the fix mclasen applied and built in Fedora as gnome-initial-setup-3.12.1-3.fc21 appears to fix my case, I forced it into an install via chroot before booting it for the first time, and g-i-s completed straight through without crashing after the username screen, first time I've seen it do that since the bug started happening. thanks.

Comment 4 Vadim Rutkovsky 2014-08-01 09:16:12 UTC

Doesn't crash anymore on latest Continuous