Re: m_copydata panic in vkernel

[Sepherosa Ziehau <sepherosa@xxxxxxxxx>]

On Mon, May 25, 2009 at 1:07 AM, Matthew Dillon
<dillon@apollo.backplane.com> wrote:
>    While running stress2 using trevorjk's vkernel setup with the
>    latest head as of last night (after my second round of pthread
>    fixes for the vkernel) I got a mbuf panic:

Hoho, I tried to reproduce it in real kernel but failed in the last two days :P
I will try vkernel today.

>
>    panic: m_copydata, negative off -1
>    mp_lock = 00000001; cpuid = 1
>    Trace beginning at frame 0x57a9ea04
>    panic(57a9ea28,c,0,579218c0,57a9ea38) at 0x80bb436
>    panic(823d90d,ffffffff,c,0,579218c0) at 0x80bb436
>    m_copydata(0,ffffffff,1,59305dbc,0) at 0x80eacdc
>    tcp_output(579218c0,59305d18,57a9eb44,59305d9c,59305d88) at 0x815e3bc
>    tcp_input(59305d00,14,6,59305d00,584aae20) at 0x815d29a
>    ip_reass(283a0150,14,0,0,0) at 0x8153362
>    ip_input(59305d00,57a9ed60,8136b31,59305d18,1) at 0x8154a73
>    ip_input(59305d18,1,8342c68,57a9ed84,8160775) at 0x8154aa7
>    netmsg_service(59305d18,1,0,1,414003b8) at 0x8136b31
>    tcpmsg_service_loop(0,0,0,0,0) at 0x8160775
>    Debugger("panic")
>
>    The vkernel is bring run on my 2xCPU test box, configured -n 2 -m 128m.
>
>                                        -Matt
>                                        Matthew Dillon
>                                        <dillon@backplane.com>
>



-- 
Live Free or Die