Bug 122662

Summary: nautilus crashes and will not restart when entering gnome
Product: [Fedora] Fedora Reporter: Chuck Mead <csm>
Component: nautilusAssignee: Alexander Larsson <alexl>
Status: CLOSED DUPLICATE QA Contact:
Severity: high Docs Contact:
Priority: high    
Version: rawhideCC: alanh, csm, fedora, jsk_priv, marshall, thomas.walker
Target Milestone: ---   
Target Release: ---   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2006-02-21 19:03:04 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Bug Depends On:    
Bug Blocks: 114961, 123268    
Attachments:
Description Flags
this is an strace of nautilus blowing up with FC2T3 on an x86_64
none
bug-buddy output with some debug info none

Description Chuck Mead 2004-05-06 18:18:31 UTC
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.4) Gecko/20030922

Description of problem:
in run level 3 or 5 when you enter gnome nautilus crashes immediately
and you never get the full gnome desktop. This may be a problem with
nautilus-media on the x86_64 as it appears that package will not build
as it seems to be linking to /usr/lib instead of /usr/lib64.

Version-Release number of selected component (if applicable):
nautilus-2.6.0-4

How reproducible:
Always

Steps to Reproduce:
1. startx or login via gdm
2. gnome starts
3. nautilus dies and will not restart
    

Actual Results:  gnome never really is able to get started as nautilus
kills the desktop when it dies.

Expected Results:  gnome should start and you should have a full gnome
environment.

Additional info:

see the comments in the summary about nautilus-media and take a look
at 122298.

Comment 1 Chuck Mead 2004-05-07 01:37:51 UTC
I ran an strace on this Nautilus problem. Here is the end of it:

mmap(NULL, 10485760, PROT_READ|PROT_WRITE|PROT_EXEC,
MAP_PRIVATE|MAP_ANONYMOUS|0x40, -1, 0) = 0x40000000
mprotect(0x40000000, 4096, PROT_NONE)   = 0
clone(child_stack=0x409ff8b0,
flags=CLONE_VM|CLONE_FS|CLONE_FILES|CLONE_SIGHAND|CLONE_THREAD|CLONE_SYSVSEM|CLONE_SETTLS|CLONE_PARENT_SETTID|CLONE_CHILD_CLEARTID|CLONE_DETACHED,
parent_tidptr=0x409ff9f0, tls=0x409ff960, child_tidptr=0x409ff9f0) = 5008
futex(0x3cc7283c28, FUTEX_WAKE, 1)      = 1
writev(18, [{"GIOP\1\2\1\0\16\3\0\0", 12},
{"h\352\377\277\3\0\0\0\0\0\0\0\34\0\0\0\0\0\0\0\225P\330"..., 782}],
2) = 794
futex(0x5e19e0, FUTEX_WAIT, 0, NULL)    = 0
futex(0x5e19a0, FUTEX_WAKE, 1)          = 0
socket(PF_UNIX, SOCK_STREAM, 0)         = 23
fcntl(23, F_SETFL, O_RDONLY|O_NONBLOCK) = 0
fcntl(23, F_SETFD, FD_CLOEXEC)          = 0
connect(23, {sa_family=AF_UNIX,
path="/tmp/orbit-csm/linc-1392-0-90850df6d638"}, 42) = 0
write(20, "A", 1)                       = 1
futex(0x82d738, FUTEX_WAKE, 1)          = 1
writev(23, [{"GIOP\1\2\1\0\224\1\0\0", 12},
{"8\354\377\277\3\0\0\0\0\0\0\0\34\0\0\0\0\0\0\0q\230H\350"..., 404}],
2) = 416
futex(0x5e19e0, FUTEX_WAIT, 1, NULL)    = 0
futex(0x5e19a0, FUTEX_WAKE, 1)          = 0
writev(23, [{"GIOP\1\2\1\0\234\1\0\0", 12},
{"\230\354\377\277\3\0\0\0\0\0\0\0\34\0\0\0\0\0\0\0q\230"..., 412}],
2) = 424
futex(0x5e19e0, FUTEX_WAIT, 2, NULL)    = 0
futex(0x5e19a0, FUTEX_WAKE, 1)          = 0
writev(23, [{"GIOP\1\2\1\0\214\1\0\0", 12},
{"\370\353\377\277\3\0\0\0\0\0\0\0\34\0\0\0\0\0\0\0q\230"..., 396}],
2) = 408
futex(0x5e19e0, FUTEX_WAIT, 3, NULL)    = 0
futex(0x5e19a0, FUTEX_WAKE, 1)          = 0
--- SIGSEGV (Segmentation fault) @ 0 (0) ---
write(3, "\22\0\20\0\1\0@\2\4\1\0\0\37\0\0\0\10_ID%\0\0\00011c0a"...,
72) = 72
write(3, " \0\2\0\0\0\0\0", 8)          = 8
write(3, "+\0\1\0", 4)                  = 4
read(3, 0x7fbfffe5e0, 32)               = -1 EAGAIN (Resource
temporarily unavailable)
select(4, [3], NULL, NULL, NULL)        = 1 (in [3])
read(3,
"\34\0G\0\1\0@\2\4\1\0\0\365\221\16\0\0\0\0\0\0\0\0\0n\222"..., 32) = 32
read(3, "\1\2J\0\0\0\0\0\25\0\0\1\0\0\0\0\0\0\0\0\0\0\0\0\200\36"...,
32) = 32
clone(child_stack=0,
flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD,
child_tidptr=0x2a9650f1f0) = 5044
wait4(5044, [WIFEXITED(s) && WEXITSTATUS(s) == 0], 0, NULL) = 5044
--- SIGCHLD (Child exited) @ 0 (0) ---
rt_sigreturn(0x5e76e0)                  = 5044
exit_group(1) 


Comment 2 Chuck Mead 2004-05-07 01:47:08 UTC
Created attachment 100056 [details]
this is an strace of nautilus blowing up with FC2T3 on an x86_64

Comment 3 Chuck Mead 2004-05-08 05:11:11 UTC
Created attachment 100104 [details]
bug-buddy output with some debug info

this is the output from bug-buddy including some debug info. I created a brand
new user account to get this so it's default profile stuff.

Comment 4 Joerg Skottke 2004-05-09 14:10:23 UTC
CC me.

Comment 5 Joerg Skottke 2004-05-09 14:10:55 UTC
Argh.

Comment 6 Justin M. Forbes 2004-05-10 04:56:16 UTC
I have not been able to reproduce this on any of my x86_64 systems
(the rebuild issue would not impact beehive, so that is not the
problem).  Have you tried deleting your preferences, or replicating on
a newly created user with no existing environment?

Comment 7 Joerg Skottke 2004-05-10 06:09:28 UTC
In my case the system is newly installed, only root and one normal
user exist. 

Does the gnome configuration get written when a user is created or
when he logs in for the first time?

The problem occurs for both root and my local user at first login.

However, the problem is 100% reproducible, i even installed the OS
twice to ensure nothing went wrong during installation. 

Comment 8 Bastien Nocera 2004-05-10 13:57:58 UTC
Could you get a backtrace of the crash?
If nautilus-media is installed, could you try removing it and see if
it makes any difference?

Comment 9 Chuck Mead 2004-05-10 14:04:57 UTC
an strace and a bug-buddy attachment are already here. Also... in my
case I have done 4 installations. the first was an upgrade, the 2nd
and 3rd were fresh installs, and on the 4th I repartitioned the box
too. It does not matter what I do... nautilus crashes on root or a
regular user.

Comment 10 Chuck Mead 2004-05-10 14:05:41 UTC
oh... I tried uninstalling nautilus media and it made no difference.

Comment 11 Thomas Walker 2004-05-15 12:54:31 UTC
Fresh install of FC2 Test3 on an Athlon64.  Think I got all of the
symbols there (deep bt w/ quite a ew libs in there).  Looks like
somthing in gnome-vfs is not 64 bit clean.  Note that this is likely
related to another x86_64 nautilus bug (I'll find it in a minute)
because it starts to tumble after read_drives_from_daemon()

#0  __libc_free (mem=0x500000000) at malloc.c:3356
#1  0x0000003faf52a42f in g_free (mem=0x500000000) at gmem.c:186
#2  0x0000003fb432d2ba in ORBit_free_T (mem=0x500000000) at
allocators.c:187
#3  0x0000003fb432d15a in ORBit_freekids_via_TypeCode_T (mem=0x8d38b0,
    tc=0x5a9e00) at allocators.c:88
#4  0x0000003fb432d126 in ORBit_freekids_via_TypeCode_T (mem=0x8d38b0,
    tc=0x35a97661c0) at allocators.c:65
#5  0x0000003fb432d297 in ORBit_free_T (mem=0x500000000) at
allocators.c:197
#6  0x0000003fb432d0b5 in ORBit_freekids_via_TypeCode_T (mem=0x8d3810,
    tc=0x35a9766240) at allocators.c:96
#7  0x0000003fb432d297 in ORBit_free_T (mem=0x500000000) at
allocators.c:197
#8  0x0000003fb432d2fd in ORBit_free (mem=0x8d3810) at allocators.c:213
#9  0x0000003fb432d1f9 in CORBA_free (mem=0x500000000) at allocators.c:138
#10 0x00000035a9646234 in read_drives_from_daemon (
    volume_monitor_client=0x8d3e60) at
gnome-vfs-volume-monitor-client.c:135
#11 0x00000035a9646399 in gnome_vfs_volume_monitor_client_init (
    volume_monitor_client=0x8cf430) at
gnome-vfs-volume-monitor-client.c:182
#12 0x0000003fafb21dc4 in g_type_create_instance (type=0) at gtype.c:1595
#13 0x0000003fafb0ea89 in g_object_constructor (type=21474836480,
    n_construct_properties=0, construct_params=0x0) at gobject.c:1044
#14 0x0000003fafb0e242 in g_object_newv (object_type=7059216,
n_parameters=0,
    parameters=0x0) at gobject.c:941
#15 0x0000003fafb0ea63 in g_object_new_valist (object_type=7059216,
    first_property_name=0x0, var_args=0x7fbffff080) at gobject.c:984
#16 0x0000003fafb0dfb7 in g_object_new (object_type=7059216,
    first_property_name=0x0) at gobject.c:822
#17 0x00000035a9648330 in _gnome_vfs_get_volume_monitor_internal
(create=1)
    at gnome-vfs-volume-monitor.c:224
#18 0x00000035a964836e in gnome_vfs_get_volume_monitor ()
    at gnome-vfs-volume-monitor.c:238
#19 0x0000000000429d92 in nautilus_application_instance_init (
    application=0x8cb270) at nautilus-application.c:183
#20 0x0000003fafb21dc4 in g_type_create_instance (type=0) at gtype.c:1595
#21 0x0000003fafb0ea89 in g_object_constructor (type=21474836480,
    n_construct_properties=1, construct_params=0x8d2c70) at gobject.c:1044
#22 0x0000003fb3d31bdf in bonobo_object_query_local_interface ()
   from /usr/lib64/libbonobo-2.so.0
#23 0x0000003fafb0e242 in g_object_newv (object_type=9244096,
    n_parameters=7279424, parameters=0x0) at gobject.c:941
#24 0x0000003fafb0ea63 in g_object_new_valist (object_type=9244096,
    first_property_name=0x0, var_args=0x7fbffff470) at gobject.c:984
#25 0x0000003fafb0dfb7 in g_object_new (object_type=9244096,
    first_property_name=0x0) at gobject.c:822
#26 0x0000000000429e52 in nautilus_application_new ()
    at nautilus-application.c:202
#27 0x000000000043549e in main (argc=1, argv=0x7fbffff8b8)
    at nautilus-main.c:319



Comment 12 Thomas Walker 2004-05-15 12:56:03 UTC
#122087 looks to be the same.

Comment 13 Chuck Mead 2004-05-15 16:22:50 UTC
Since it works if I remove some things from my /etc/fstab file I
figured I should post that file here:

LABEL=/                 /                       ext3    defaults     
  1 1
LABEL=/boot             /boot                   ext3    defaults     
  1 2
none                    /dev/pts                devpts  gid=5,mode=620
 0 0
none                    /dev/shm                tmpfs   defaults     
  0 0
LABEL=/home             /home                   ext3    defaults     
  1 2
LABEL=/opt              /opt                    ext3    defaults     
  1 2
none                    /proc                   proc    defaults     
  0 0
none                    /sys                    sysfs   defaults     
  0 0
LABEL=/usr              /usr                    ext3    defaults     
  1 2
LABEL=/usr/local        /usr/local              ext3    defaults     
  1 2
LABEL=/var              /var                    ext3    defaults     
  1 2
/dev/hdc3               swap                    swap    defaults     
  0 0
/dev/hda5               swap                    swap    defaults     
  0 0
/dev/cdrom              /mnt/cdrom              udf,iso9660
noauto,owner,kudzu,ro 0 0
/dev/cdrom1             /mnt/cdrom1             udf,iso9660
noauto,owner,kudzu,ro 0 0
/dev/fd0                /mnt/floppy             auto   
noauto,owner,kudzu 0 0

If I remove the two cdrom's and the floppy it works.


Comment 14 Marshall Lewis 2004-05-15 17:52:48 UTC
I have basically the same behavior, I only have to remove my 2nd cdrom
though (which is a usb dvd writer).

My install and upgrades worked perfectly, but today I attached a usb
keyboard, and upgraded the bios on my workstation, that's when the
problems started appearing (I had experienced the problem with an
earlier test release, but it when away on a fresh install).

Comment 15 Thomas Walker 2004-05-16 15:39:57 UTC
Same here... if I comment out hdc (atapi cdrw) nautilus works.  Its
worth noting that I get seek errors from this device when I boot (duh,
it has no media in it) but it otherwise seems to work fine.
I'll take a look at this a bit more closely under gdb when I get a
chance (its just too nice outside to be staring at things in debuggers
today!)

Comment 16 Chuck Mead 2004-05-16 17:04:57 UTC
Well how about posting your fstab files guys... it might help with the
bugfix to see what's in common on all of this!

Comment 17 Marshall Lewis 2004-05-17 01:57:21 UTC
Here's my fstab, but I don't think that it show's anything overly
interesting..   sda is a 18gig drive on an adaptec scsi controller,
hde is a sata drive on a via controller (builtin), cdrom is a regular
ide dvd drive, cdrom1 is a usb dvd writer (commented out at the moment
so nautilus would work).

/dev/sda2     /            ext3        defaults              1 1
/dev/sda1     /boot        ext3        defaults              1 2
/dev/hde3     /home        ext3        defaults              1 2
none          /dev/pts     devpts      gid=5,mode=620        0 0
none          /dev/shm     tmpfs       defaults              0 0
none          /proc        proc        defaults              0 0
none          /sys         sysfs       defaults              0 0
/dev/sda3     swap         swap        defaults              0 0
/dev/cdrom    /mnt/cdrom   udf,iso9660 noauto,owner,kudzu,ro 0 0
/dev/fd0      /mnt/floppy  auto        noauto,owner,kudzu    0 0
#/dev/cdrom1  /mnt/cdrom1  udf,iso9660 noauto,owner,kudzu,ro 0 0


This fstab with the bottom line uncommented had been working fine for
me until I did a bios upgrade and attached a usb keyboard (I did both
things at the same time, so I'm not sure which caused the problem).  I
did try putting the keyboard back on the ps/2 port and down grading
the bios, but nautilus still wouldn't start.  Oh, and I did create a
fresh test user to verify that it wasn't something in my user's config.


Comment 18 Thomas Walker 2004-05-17 11:40:35 UTC
Offending line is:
/dev/cdrom1             /mnt/cdrom1             udf,iso9660
noauto,owner,kudzu,ro 0 0
(The default that kudzu sets up).  If I remove 'owner' nautilus is
able to start up.  Replace it with 'user' (what I really want because
local user secuirty isn't of a concern here) and it crashes again.
The backtrace I posted earlier seems to indicate stack corruption (the
mem value to be freed keeps flipping between a "valid" address of
0x8d3810 and 0x500000000).  The last one, between frames 2 and 3 is
strange, you can print the value of the pointer you're passing into
ORBit_free_T(mem) before the function call and it looks good but once
inside the "free", the address is corrupt... from there on out you
wind up trying to 'free' in invalid address and get a SEGV as expected.
Someone with a little more gonme-vfs/ORBit/CORBA experience might be
able to spot what is going on here...

Comment 19 Pete Bradbury 2004-05-20 19:56:23 UTC
Ditto above removing owner from fstab entry brought back nautilus as
well as this gedit could open file tree to locate the fstab file.

Comment 20 Chuck Mead 2004-05-20 20:02:47 UTC
I have the exact same thing. If I remove the "owner" feature from the
options line then nautilus runs perfectly!

Comment 21 Alan Hamilton 2004-05-27 23:08:10 UTC
*** Bug 123154 has been marked as a duplicate of this bug. ***

Comment 22 Thorsten Leemhuis 2004-06-12 06:43:40 UTC
I think
bug 123807 
bug 123066 
bug 122087 
are duplicates / related of this bug?!

Comment 23 Alan Hamilton 2004-06-17 07:32:58 UTC
A patch has been posted over at http://bugzilla.gnome.org/show_bug.
cgi?id=138986">Gnome.org.  The bug is in ORBit2's 
ORBit_freekids_via_TypeCode_T().  It doesn't adjust for the alignment 
of the structures in a 64 bit environment.  Thanks to Chris Landrieu 
for finding this.

I tried the fix on my Core 2 system and it works.

Comment 24 Joerg Skottke 2004-06-17 08:21:27 UTC
Good to hear. I will try it within the week.

Comment 25 Alexander Larsson 2004-08-26 08:48:46 UTC
The new ORBit2 is in rawhide, so this should be ok now.

Comment 26 Alexander Larsson 2004-09-06 12:48:39 UTC

*** This bug has been marked as a duplicate of 123807 ***

Comment 27 Red Hat Bugzilla 2006-02-21 19:03:04 UTC
Changed to 'CLOSED' state since 'RESOLVED' has been deprecated.