Re: 2.6.18-rc1-mm1 inconsistent lock state in netpoll_send_skb

eth0: RealTek RTL-8029 found at 0xa400, IRQ 10, 00:40:95:46:6E:2D.
netconsole: device eth0 not up yet, forcing it
netconsole: carrier detect appears untrustworthy, waiting 4 seconds

=================================
[ INFO: inconsistent lock state ]
---------------------------------
inconsistent {in-softirq-W} -> {softirq-on-W} usage.
swapper/1 [HC0[0]:SC0[0]:HE1:SE1] takes:
 (&dev->_xmit_lock){-+..}, at: [<c0252bcb>] netpoll_send_skb+0x6b/0xdb
{in-softirq-W} state was registered at:
  [<c012cfb3>] lock_acquire+0x60/0x80
  [<c02923a0>] _spin_lock+0x19/0x28
  [<c02547dd>] dev_watchdog+0x11/0xaf
  [<c011df77>] run_timer_softirq+0xed/0x145
  [<c011b0c9>] __do_softirq+0x46/0x9c
  [<c0104fc1>] do_softirq+0x4d/0xab
irq event stamp: 616842
hardirqs last  enabled at (616841): [<c0292668>] _spin_unlock_irqrestore+0x36/0x3c
hardirqs last disabled at (616842): [<c02926b9>] _spin_lock_irqsave+0xf/0x32
softirqs last  enabled at (616802): [<c024afd5>] dev_mc_upload+0x36/0x3a
softirqs last disabled at (616798): [<c02923ba>] _spin_lock_bh+0xb/0x2d

other info that might help us debug this:
1 lock held by swapper/1:
 #0:  (&dev->_xmit_lock){-+..}, at: [<c0252bcb>] netpoll_send_skb+0x6b/0xdb

stack backtrace:
 [<c0104d2c>] show_trace+0xd/0x10
 [<c0104d48>] dump_stack+0x19/0x1d
 [<c012b5cd>] print_usage_bug+0x1cc/0x1d9
 [<c012c15c>] mark_lock+0x22d/0x349
 [<c012c2bf>] mark_held_locks+0x47/0x65
 [<c012c468>] trace_hardirqs_on+0xef/0x119
 [<c022395a>] ei_start_xmit+0x242/0x260
 [<c0252be9>] netpoll_send_skb+0x89/0xdb
 [<c02538a6>] netpoll_send_udp+0x1fd/0x207
 [<c0223b18>] write_msg+0x42/0x6a
 [<c01167f3>] __call_console_drivers+0x3b/0x48
 [<c0116a0c>] release_console_sem+0x118/0x1ed
 [<c0116d77>] register_console+0x190/0x197
 [<c0223ac2>] init_netconsole+0x4e/0x62
 [<c0100378>] init+0x88/0x1e1
 [<c0101005>] kernel_thread_helper+0x5/0xb
netconsole: network logging started

full dmesg and .config attached.
~~
laurent