Subject: Re: something strange with mbuf length...
To: Konstantin KABASSANOV <>
From: Manuel Bouyer <>
List: tech-net
Date: 04/25/2006 21:34:42
On Mon, Apr 24, 2006 at 12:34:59PM +0200, Konstantin KABASSANOV wrote:
> Here is another panic from this morning. 
> It is however strange that the panic happens within psql, since this command
> uses IPv6... In this case, it seems the panic is not related to the current
> process... 
>  login: panic: tcp_output
> Stopped in pid 930.1 (psql) at  netbsd:cpu_Debugger+0x4:        popl    %ebp
> db> bt
> cpu_Debugger(c0516c5c,34,1,c6a93dbc,c012cd24) at netbsd:cpu_Debugger+0x4
> panic(c02cb23a,c05a2000,c05a0038,0,0) at netbsd:panic+0x10f
> tcp_output(c063147c,0,832c,24fe1b83,60000) at netbsd:tcp_output+0xd6c
> tcp_input(c0516c00,14,6,c055281c,0) at netbsd:tcp_input+0x2a1f
> ip_input(c0516c00,0,c0100c40,4,c054d800) at netbsd:ip_input+0x752
> ipintr(c6a90010,c0220030,c6a40010,10,c6a93f78) at netbsd:ipintr+0x106
> DDB lost frame for netbsd:Xsoftnet+0x41, trying 0xc6a93f0c
> Xsoftnet() at netbsd:Xsoftnet+0x41

Sure; at this point the kernel is in interrupt context, not process context.

Manuel Bouyer, LIP6, Universite Paris VI. 
     NetBSD: 26 ans d'experience feront toujours la difference