gvfsd-smb crashed with SIGSEGV in g_simple_async_result_complete()

[Nicolay Doytchev]

Binary package hint: gvfs

This bug exists in Jaunty and the steps to reproduce are the same as for Karmic. Fortunately it is a tad more stable in Karmic but it gvfs-smb crashed eventually today. Reproduce as follows:

0. Mount a remote smb share via Nautilus.
1. Open Rhythmbox / Exaile / (maybe other software that can generate a lot of traffic over smb).
2. Import a folder from the smb share into Rhythmbox / Exaile which contains a lot of music (the one tested here is over 200GB)
3. Wait.
4. Somewhere along the way the smb share will silently unmount and the import process will stop. The only trace of what happened will be visible via dmesg "gvfsd-smb[1873]: segfault at 14 ip 080583f5 sp bfbe2460 error 4 in gvfsd-smb[8048000+20000]" or Apport.

This crash occurs on my setup somewhere between 3GB and 20GB music imported.

I am sure that this bug is not specifically related with music files and importing in players but it is a generic bug in gvfs-smb which is triggered after some amount of data is transferred by it.

ProblemType: Crash
Architecture: i386
Date: Wed Oct 14 19:58:35 2009
DistroRelease: Ubuntu 9.10
ExecutablePath: /usr/lib/gvfs/gvfsd-smb
Package: gvfs-backends 1.4.0-0ubuntu4
ProcCmdline: /usr/lib/gvfs/gvfsd-smb --spawner :1.10 /org/gtk/gvfs/exec_spaw/2
ProcEnviron:
 SHELL=/bin/bash
 LANG=en_US.UTF-8
ProcVersionSignature: Ubuntu 2.6.31-14.46-generic
SegvAnalysis:
 Segfault happened at: 0x80583f5: movzbl 0x14(%eax),%edx
 PC (0x080583f5) ok
 source "0x14(%eax)" (0x00000014) not located in a known VMA region (needed readable region)!
 destination "%edx" ok
SegvReason: reading NULL VMA
Signal: 11
SourcePackage: gvfs
StacktraceTop:
 ?? ()
 ?? ()
 ?? () from /usr/lib/libgio-2.0.so.0
 g_simple_async_result_complete ()
 ?? () from /usr/lib/libgio-2.0.so.0
Title: gvfsd-smb crashed with SIGSEGV in g_simple_async_result_complete()
Uname: Linux 2.6.31-14-generic i686
UserGroups: adm admin cdrom dialout lpadmin plugdev sambashare

[Apport retracing service (apport)]

StacktraceTop:g_vfs_job_emit_finished (job=0x0) at gvfsjob.c:323
send_reply_cb (source_object=0x8edccc8, res=0x8edc4e0,
?? () from /usr/lib/libgio-2.0.so.0
g_simple_async_result_complete ()
?? () from /usr/lib/libgio-2.0.so.0

[Nicolay Doytchev (lightrush)]

I have successfully reproduced it again.

I have installed Valgrind and the debug debs for gvfs and gvfs-backends. I have read how to run Valgrind but gvfs is started in a more complicated way than regular program. If you give specific instructions about doing something that will give more light on this bug I would gladly follow them.

[Sebastien Bacher (seb128)]

Thank you for your bug report. The issue is an upstream one and it would be nice if somebody having it could send the bug the to the people writting the software (https://wiki.ubuntu.com/Bugs/Upstream/GNOME)

[Sebastien Bacher (seb128)]

you can find details on valgrind use there, https://wiki.ubuntu.com/Valgrind
the way to valgrind those usually is to rename the binary and set a wrapper calling valgrind on the real binary instead

[Nicolay Doytchev (lightrush)]

Elegant... Will try it and post the log. I do not program C yet so I have no clue how and if this will be helpful but according to https://wiki.ubuntu.com/Valgrind , this is what I have to do next if I see a segfault. I hope it helps. I could also submit a bugreport upstream. Is it possible to patch this before release if there is enough useful information?

[Nicolay Doytchev (lightrush)]

Creating a wrapper starts the daemon but then the daemon begins to spit DBUS errors and does not function properly. Help?

[Sebastien Bacher (seb128)]

Could you give details on the error?

[Nicolay Doytchev (lightrush)]

Could not open location 'smb://lr-gassus/usb/'

DBus error org.freedesktop.DBus.Error.NoReply: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.

The message appears in a dialog box with a Close button.

[Nicolay Doytchev (lightrush)]

I think this error appears since gvfsd tries to communicate to gvfsd-smb via DBus but since this is no longer the executable of the program the connection cannot be established. If my understanding is right... I am not familiar with the DBus and GTK APIs. Still grinding Java here... C is coming next semester.

[Nicolay Doytchev (lightrush)]

Still here in 1.4.1

[yannack (yannack)]

I experience the same problem (I think). I try copying files from a nautilus SMB share to a nautilus FTP share, and I usually cannot go further than about 500MB (because the SMB share fails, not the FTP)
I try reading musing from the SMB share, sometimes it will stop playing within a few minutes, sometimes longer.
I also believe that suspending the computer will also create the same problem.

Sometimes I am lucky and Nautilus lets me unmount the SMB share and remount it, but most of the time, I have to kill nautilus, because I get the same error message.

Here is what I get through the command line.
gvfs-mount -u smb://admin@192.168.0.125/video/
Error unmounting mount: DBus error org.freedesktop.DBus.Error.NoReply: Did not receive a reply. Possible causes include: the remote application did not send a reply, the message bus security policy blocked the reply, the reply timeout expired, or the network connection was broken.

[Pedro Villavicencio (pedro)]

Thank you for taking the time to report this bug and helping to make Ubuntu better. Is this reproducible with Lucid? Please test and provide us of feedback about it, Thanks in advance..

[Nicolay Doytchev (lightrush)]

I will test it out today and report back.

[Nicolay Doytchev (lightrush)]

The bug does not exist/affect me any more. A full 250GB collection can be imported in Exaile over SMB share without crashing gvfs-smb.

[Nicolay Doytchev (lightrush)]

*Tested on Lucid Beta1 + latest updates.

No EDIT button == double post :|

[Sebastien Bacher (seb128)]

thank you for the update, closing the bug since that works now

[JM (jsmrkhm)]

Apologies if I've submitted this to the wrong place, but I appear to be getting exactly the same error in Maverick Meerkat:

uname -a
Linux -------pc 2.6.35-22-generic #33-Ubuntu SMP Sun Sep 19 20:32:27 UTC 2010 x86_64 GNU/Linux

dmesg|grep error
gvfsd-smb[4331]: segfault at 18 ip 0000000000410304 sp 00007fff8b10d960 error 4 in gvfsd-smb[400000+25000]
gvfsd-smb[4421]: segfault at 18 ip 0000000000410304 sp 00007fff653f8860 error 4 in gvfsd-smb[400000+25000]
gvfsd-smb[4504]: segfault at 18 ip 0000000000410304 sp 00007ffff8b8fe20 error 4 in gvfsd-smb[400000+25000]
...and so on...

(However, this OS is Linux Mint 10, does that mean all responsibility on this site is denied? ;-)

[Taras Perebeynosov (taras-perebeynosov)]

Error reproduced in 12.04

[Lupius (lupius)]

Same bug occured in 12.04 LTS beta fully updated up to 08.04 while trying to play an mp3 located on windows network share via rhytmbox.