Bug 13257

Summary: using nouveau on NV17 kills the KDE VNC server
Product: xorg Reporter: Hervé Cauwelier <debian>
Component: Driver/nouveauAssignee: Nouveau Project <nouveau>
Status: CLOSED NOTOURBUG QA Contact: Xorg Project Team <xorg-team>
Severity: normal    
Priority: medium    
Version: unspecified   
Hardware: x86 (IA32)   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
X user log with backtrace of VNC server (krfb) crashing none

Description Hervé Cauwelier 2007-11-15 08:31:06 UTC 
using nouveau on NV17 kills the KDE VNC server

I connect to a distant desktop using VNC. The environment is KDE and I use its VNC server, krfb.

The adapter is as following:

  01:00.0 VGA compatible controller: nVidia Corporation NV17 [GeForce4 420 Go] (rev a3)

Using nv, I used to connect and disconnect several times a day.

Using nouveau, I can only connect once. The second time, vncviewer fails.

First time
----------

VNC Viewer Free Edition 4.1.1 for X - built Sep 10 2007 17:17:04
Copyright (C) 2002-2005 RealVNC Ltd.
See http://www.realvnc.com for information on VNC.

Thu Nov 15 17:23:59 2007
 CConn:       connected to host localhost port 5599

Thu Nov 15 17:24:01 2007
 CConnection: Server supports RFB protocol version 3.3
 CConnection: Using RFB protocol version 3.3
 TXImage:     Using default colormap and visual, TrueColor, depth 24.
 CConn:       Using pixel format depth 6 (8bpp) rgb222
 CConn:       Using ZRLE encoding


Second time
-----------

VNC Viewer Free Edition 4.1.1 for X - built Sep 10 2007 17:17:04
Copyright (C) 2002-2005 RealVNC Ltd.
See http://www.realvnc.com for information on VNC.

Thu Nov 15 17:25:12 2007
 CConn:       connected to host localhost port 5599
 main:        End of stream


I only close the vncviewer window between the tries.

I tried "-log *:stdout:100" to get maximum information but nothing more is printed.

Nothing is printed in syslog nor Xorg.0.log at the moment I try.

Last ouput of syslog is:

Nov 15 17:03:49 ourson kernel: [54925.848000] [drm] nouveau_fifo_free: freeing fifo 1
Nov 15 17:03:51 ourson kernel: [54926.684000] [drm:nouveau_fifo_free] *ERROR* Failed to idle channel 1 before destroy.Prepare for strangeness..
Nov 15 17:03:51 ourson kernel: [54926.684000] [drm] cleanning a channel with graph in current context
Nov 15 17:03:51 ourson kernel: [54926.684000] [drm] nouveau_fifo_free: freeing fifo 0
Nov 15 17:03:51 ourson kernel: [54927.524000] [drm:nouveau_fifo_free] *ERROR* Failed to idle channel 0 before destroy.Prepare for strangeness..
Nov 15 17:05:09 ourson NetworkManager: <info>  Updating allowed wireless network lists.
Nov 15 17:05:10 ourson NetworkManager: <WARN>  nm_dbus_get_networks_cb(): error received: org.freedesktop.NetworkManagerInfo.NoNetworks - org.freedesktop.NetworkManagerInfo.NoNetworks.
Nov 15 17:17:02 ourson /USR/SBIN/CRON[14947]: (root) CMD (   cd / && run-parts --report /etc/cron.hourly)
Nov 15 17:22:10 ourson kernel: [56028.228000] agpgart: Found an AGP 2.0 compliant device at 0000:00:00.0.
Nov 15 17:22:10 ourson kernel: [56028.228000] agpgart: Putting AGP V2 device at 0000:00:00.0 into 4x mode
Nov 15 17:22:10 ourson kernel: [56028.228000] agpgart: Putting AGP V2 device at 0000:01:00.0 into 4x mode
Nov 15 17:22:10 ourson kernel: [56028.308000] [drm] Allocating FIFO number 0
Nov 15 17:22:10 ourson kernel: [56028.308000] [drm] nouveau_fifo_alloc: initialised FIFO 0
Nov 15 17:22:10 ourson kernel: [56028.308000] [drm] Allocating FIFO number 1
Nov 15 17:22:10 ourson kernel: [56028.308000] [drm] nouveau_fifo_alloc: initialised FIFO 1
Nov 15 17:22:34 ourson NetworkManager: <info>  Updating allowed wireless network lists.
Nov 15 17:22:34 ourson NetworkManager: <WARN>  nm_dbus_get_networks_cb(): error received: org.freedesktop.NetworkManagerInfo.NoNetworks - org.freedesktop.NetworkManagerInfo.NoNetworks.

I have a nice .xsession-errors that I'll attach afterwards.
Comment 1 Hervé Cauwelier 2007-11-15 08:33:04 UTC 
Created attachment 12574 [details]
X user log with backtrace of VNC server (krfb) crashing
Comment 2 Maarten Maathuis 2008-02-01 11:45:47 UTC 
At least a better backtrace or some more information is needed, if you want any hope of this getting solved.
Comment 3 Hervé Cauwelier 2008-02-01 13:27:01 UTC 
No, I wouldn't be available to track this down. I dropped all KDE and don't intend to try it again for long.
Comment 4 Adam Jackson 2008-02-24 18:22:40 UTC 
Mass reopen.  The "LATER" resolution is lame, I'm deleting it.  Consider LATER to have arrived.
Comment 5 Maarten Maathuis 2008-03-08 05:40:41 UTC 
Any other victim of this bug?
Comment 6 Stuart Bennett 2009-01-26 19:18:45 UTC 
Reproduced using vncviewer 4.1.3, ddx cf65b875, drm e6a062c2, and krfb from KDE 3.5.7 (openSUSE 10.3) on nv34m.

Backtrace looks like:

#0  0xffffe410 in __kernel_vsyscall ()
#1  0xb6bd38f5 in *__GI_raise (sig=6)
    at ../nptl/sysdeps/unix/sysv/linux/raise.c:64
#2  0xb6bd51e1 in *__GI_abort () at abort.c:88
#3  0xb6c0a80c in __libc_message (do_abort=2,
    fmt=0xb6cc0f28 "*** glibc detected *** %s: %s: 0x%s ***\n")
    at ../sysdeps/unix/sysv/linux/libc_fatal.c:170
#4  0xb6c104b6 in malloc_printerr (action=2,
    str=0xb6cc1020 "double free or corruption (!prev)",
    ptr=<value optimized out>) at malloc.c:5887
#5  0xb6c12179 in *__GI___libc_free (mem=0x8275fb0) at malloc.c:3622
#6  0xb6e040f9 in _XDestroyImage (ximage=0x8274058) at ImUtil.c:438
#7  0x08056a0c in ~XUpdateScanner (this=0x8272f48)
    at ./krfb/krfb/xupdatescanner.cc:203
#8  0x0805781d in RFBController::stopServer (this=0xbf8f86d0, xtestUngrab=true)
    at ./krfb/krfb/rfbcontroller.cc:481
#9  0x0805a6b7 in ~RFBController (this=0xbf8f86d0)
    at ./krfb/krfb/rfbcontroller.cc:381
#10 0x0805f180 in main (argc=Cannot access memory at address 0x52f4
) at ./krfb/krfb/main.cpp:188
#11 0xb6bbffe0 in __libc_start_main (main=0x805e920 <main>, argc=3,
    ubp_av=0xbf8f8c64, init=0x8081c10 <__libc_csu_init>,
    fini=0x8081c00 <__libc_csu_fini>, rtld_fini=0xb7f20f60 <_dl_fini>,
    stack_end=0xbf8f8c5c) at libc-start.c:222
#12 0x08055601 in _start ()

I'm content to call this a krfb bug; it seems ridiculously easy to make the tested KDE3 version fail in a number of ways.  The KDE4 (4.1.3) version of krfb doesn't show it (though at depths != 16 it suffers http://bugs.kde.org/show_bug.cgi?id=162493 ).
Comment 7 Hervé Cauwelier 2009-01-27 01:01:21 UTC 
Thanks for your report.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.