Re: Status of 2.6.22-rc3 ?

On Sunday 27 May 2007, you wrote:
> On Sun, May 27, 2007 at 10:17:33PM +0200, Helge Deller wrote:
> > On Sunday 27 May 2007, Kyle McMartin wrote:
> > > On Sun, May 27, 2007 at 02:38:57PM +0200, Helge Deller wrote:
> > > > Synthesizing the initial hotplug events...done.
> > > > Waiting for /dev to be fully populated...<2>Backtrace:
> > > >  [<1011326c>] unwind_frame_regs+0x48/0x220
> > > >  [<10113578>] unwind_once+0x10/0x48
> > > >  [<1011284c>] get_wchan+0x2c/0x68
> > > >  [<101b9308>] do_task_stat+0x24c/0x500
> > > >  [<101b4f34>] proc_info_read+0x88/0x108
> > > >  [<10179db0>] vfs_read+0x8c/0x138
> > > >  [<1017a10c>] sys_read+0x5c/0xa4
> > > >  [<1010c080>] syscall_exit+0x0/0x28
> > > > BUG: spinlock recursion on CPU#0, udevd/560
> > > >  lock: 1056e08c, .magic: dead4ead, .owner: udevd/560, .owner_cpu: 0
> > > > Backtrace:
> > > >  [<1026c5d0>] _raw_spin_lock+0x50/0xa4
> > > >  [<101076cc>] parisc_terminate+0x58/0xe8
> > > >  [<10107cbc>] handle_interruption+0x560/0x58c
> > > >  [<1010b080>] intr_check_sig+0x0/0x34
> > > > BUG: spinlock lockup on CPU#0, udevd/560, 1056e08c
> > > 
> > > What compiler version did you use to build?
> > 
> > hppa-linux-gcc (GCC) 3.3.4 (Debian 1:3.3.4-11)
> > (cross-compiling from i386 -> parisc 32bit)
> > 
> 
> Ok. I'll try to get a machine plugged in this week and take a look.
> 
> Hmm, is spinlock debugging enabled? Does it make a difference if you
> turn that off?

Yes, they were enabled, but turning them off gave the same problem:

Waiting for /dev to be fully populated...
Backtrace:
 [<101124fc>] unwind_frame_regs+0x48/0x220
 [<10112808>] unwind_once+0x10/0x48
 [<10111b34>] get_wchan+0x2c/0x68
 [<101aed2c>] do_task_stat+0x244/0x4f8
 [<101aacb8>] proc_info_read+0x88/0x108
 [<10173d70>] vfs_read+0x8c/0x138
 [<101740cc>] sys_read+0x5c/0xa4
 [<1010c080>] syscall_exit+0x0/0x28
Backtrace:
 [<101124fc>] unwind_frame_regs+0x48/0x220
 [<10112808>] unwind_once+0x10/0x48
 [<10106ea8>] do_show_stack+0x28/0xac
 [<101074f8>] parisc_terminate+0x84/0xc8
 [<10107a9c>] handle_interruption+0x560/0x58c
 [<1010b080>] intr_check_sig+0x0/0x34
(...the last backtrace continues forever...)

Helge