Bug#405179: gnome: Hangs horribly on attempt to switch user

[Vincent Lönngren]

mån 2007-01-01 klockan 20:57 +0100 skrev Sven Arvidsson:
> On Mon, 2007-01-01 at 16:07 +0100, Vincent Lönngren wrote:
> > No, I haven't found any keyboard input that works. I've tried
> > control-alt-F1 to control-alt-F8 and control-backspace. I don't remember
> > now if I tried control-alt-delete before giving up and pulling the plug.
> 
> Is it a hard lock up, or can you still log in from another computer
> using SSH?
> 

I haven't been able to try that because I don't have a computer to log
in from. I did some experimentation, though, and here is what I found.

Everything seems locked for a couple of minutes - no disc activity, no
reaction to any keyboard input and so on - and then everything proceeds
like it should! The GDM login screen is shown. There is some output on
virtual console number 8, which I attach. The next time, and on any
subsequent attempts, it works right away, without hanging. It doesn't
hang again until after a reboot, and it always hangs the first time
after reboot.

The console - NVIDIA framebuffer - gets kind of messed up. Everything
looks ok when you switch to a virtual terminal, but the cursor is a
random pattern of garble, different each time it blinks. Any text
entered is garbled. When I used the gpm pointer to copy the text of the
error message, all the text got garbled on the screen. Switching to
another virtual terminal and back made the text look ok again.

I don't know what to make of all this. Some kind of conflict between the
nv driver and the framebuffer? Enabling UseFBDev in /etc/X11/xorg.conf
makes X crash on this system, if that's any help.
-- 
Vincent Lönngren <bice77 at malmo2.net>
-------------- next part --------------
BUG: soft lockup detected on CPU#0!
 [<c011cd6a>] update_process_times+0x2a/0x70
 [<c0105b9b>] timer_interrupt+0x2b/0x60
 [<c012fe43>] handle_IRQ_event+0x33/0x60
 [<c012feeb>] __do_IRQ+0x7b/0xe0
 [<c0104994>] do_IRQ+0x54/0xa0
 =======================
 [<c0102cca>] common_interrupt+0x1a/0x20
 [<c01de3cb>] NVDmaWait+0x3b/0x190
 [<c01de88b>] nvidiafb_imageblit+0x36b/0x3a0
 [<c01d59af>] bit_putcs+0x36f/0x610
 [<c0110c80>] default_wake_function+0x0/0x10
 [<c01d4657>] fbcon_switch+0x3e7/0x5a0
 [<c01d8839>] nvidia_write_clut+0x59/0x80
 [<c01cfd1b>] fbcon_putcs+0x19b/0x2d0
 [<c01d5640>] bit_putcs+0x0/0x610
 [<c02128a2>] do_update_region+0xf2/0x150
 [<c02151e7>] redraw_screen+0x137/0x1e0
 [<c01641d3>] dput+0xb3/0x120
 [<c020d5c8>] complete_change_console+0x28/0xe0
 [<c020ecdb>] vt_ioctl+0x150b/0x1850
 [<c015c7ef>] link_path_walk+0x6f/0xe0
 [<c020d7d0>] vt_ioctl+0x0/0x1850
 [<c0208d25>] tty_ioctl+0x105/0xd30
 [<c015ca58>] do_path_lookup+0x98/0x210
 [<c015b656>] may_open+0x66/0x220
 [<c01c60f0>] proc_bus_pci_open+0x0/0x40
 [<c01c6102>] proc_bus_pci_open+0x12/0x40
 [<c014c50d>] __dentry_open+0x16d/0x1e0
 [<c02b59ea>] pci_write+0x2a/0x30
 [<c01c663c>] proc_bus_pci_write+0xfc/0x170
 [<c0208c20>] tty_ioctl+0x0/0xd30
 [<c015ea49>] do_ioctl+0x69/0x70
 [<c015eaac>] vfs_ioctl+0x5c/0x270
 [<c01c6540>] proc_bus_pci_write+0x0/0x170
 [<c015ecfd>] sys_ioctl+0x3d/0x70
 [<c0102ad9>] sysenter_past_esp+0x56/0x79
DMA Wait Lockup