29 May
2007
29 May
'07
5:49 p.m.
On Di, 29.05.07 21:37 David Baron <d_baron(a)012.net.il> wrote:
Could you please
a) add a Re: to the subject before answering yourself, so this gets a
thread, instead of single mails.
and b) maybe CC linux-kernel(a)vger.kernel.org as I do not think Ingo
Molnar is subscribed to this list here, but it would be good if he also
gets your reports in order to fix it.
Thanks.
Ran for nearly
two hours with none. Nada.
Then they started up again. What triggered them?
The only thing in the logcheck at the same time as
the first outputs
was a pptp connect after the original connection fell. Relevant?
The offending functions are icmp_.... functions, most likely the more
used ipv4 flavor of icmp.c.
Example:
BUG: using smp_processor_id() in preemptible [00000000] code:
IRQ-9/1110 caller is icmp_xmit_unlock+0xb/0x30
[<c01faab9>] debug_smp_processor_id+0xa9/0xb0
[<c02e531b>] icmp_xmit_unlock+0xb/0x30
[<c02e598e>] icmp_send+0x11e/0x3f0
....