Re: LORs with ipfw

On Thu, Jul 01, 2004 at 08:38:06PM +0000, Bjoern A. Zeeb wrote:
> On Thu, 1 Jul 2004, Wiktor Niesiobedzki wrote:
> 
> Hi,
> can you please give a date from when you current is or if it's a
> release please give the release number ?
This is CURRENT as of about 2004.07.01 17:00:00 UTC.

> I will add these two to my page tomorrow morning. I think I remember
> at least one of those but that must have been fixed in late 2003 so
> I am quite unsure.
Kernel is rather standard, IPFW + DUMMYNET, no POLLING. PF and IPF compiled in.

> > Those two are missing from your list:
> > lock order reversal
> >  1st 0xc14d5c84 inp (tcpinp) _at_ /usr/src/sys/netinet/tcp_syncache.c:1142
> >  2nd 0xc06d5a08 IPFW static rules (IPFW static rules) _at_ /usr/src/sys/netinet/ip_fw2.c:1828
> > Stack backtrace:
> > backtrace(0,ffffffff,c05ec8e0,c05eb940,c05c1a9c) at backtrace+0x12
> > witness_checkorder(c06d5a08,9,c06d3c8c,724) at witness_checkorder+0x53b
> > _mtx_lock_flags(c06d5a08,0,c06d3c8c,724,cb9689a8) at _mtx_lock_flags+0x57
> > ipfw_chk(cb968a84,cb968a64,cb968ac8,0,0) at ipfw_chk+0x2d8
> > ip_output(c1388600,0,0,0,0) at ip_output+0x633
> > syncache_respond(c1707000,c1388600) at syncache_respond+0x2d7
> > syncache_add(cb968bd8,cb968c38,c14aa02a,cb968bd4,c147d000,c14d7700,cb968c38,c14aa03e,8,1,c14aa02a) at syncache_add+0x4af
> > tcp_input(c147d000,14,0,14,100a8c0) at tcp_input+0x768
> > ip_input(c147d000) at ip_input+0x832
> > netisr_processqueue(c060ff98,c10d6440,c10dd400,cb968d1c,c048f724) at netisr_processqueue+0x6e
> > swi_net(0) at swi_net+0x85
> > ithread_loop(c10dd400,cb968d48,c10dd400,c048f5f0,0) at ithread_loop+0x134
> > fork_exit(c048f5f0,c10dd400,cb968d48) at fork_exit+0x98
> > fork_trampoline() at fork_trampoline+0x8
> > --- trap 0x1, eip = 0, esp = 0xcb968d7c, ebp = 0 ---
> >
> >
> > And:
> > lock order reversal
> >  1st 0xc0610d8c tcp (tcp) _at_ /usr/src/sys/netinet/tcp_input.c:611
> >  2nd 0xc06d5a08 IPFW static rules (IPFW static rules) _at_ /usr/src/sys/netinet/ip_fw2.c:1828
> > Stack backtrace:
> > backtrace(0,ffffffff,c05ec908,c05eb940,c05c1a9c) at backtrace+0x12
> > witness_checkorder(c06d5a08,9,c06d3c8c,724) at witness_checkorder+0x53b
> > _mtx_lock_flags(c06d5a08,0,c06d3c8c,724,c05ead20) at _mtx_lock_flags+0x57
> > ipfw_chk(cb968acc,0,cb968b10,0,0) at ipfw_chk+0x2d8
> > ip_output(c1388e00,0,0,0,0) at ip_output+0x633
> > tcp_respond(0,c1388e40,c1388e54,c1388e00,f1576cfc,0,14) at tcp_respond+0x2ac
> > tcp_input(c1388e00,14,0,14,100007f) at tcp_input+0x27b4
> > ip_input(c1388e00) at ip_input+0x832
> > netisr_processqueue(c060ff98,c10d6440,c10dd400,cb968d1c,c048f724) at netisr_processqueue+0x6e
> > swi_net(0) at swi_net+0x85
> > ithread_loop(c10dd400,cb968d48,c10dd400,c048f5f0,0) at ithread_loop+0x134
> > fork_exit(c048f5f0,c10dd400,cb968d48) at fork_exit+0x98
> > fork_trampoline() at fork_trampoline+0x8
> > --- trap 0x1, eip = 0, esp = 0xcb968d7c, ebp = 0 ---
> >
> >
> > I tried to look out in some archives/googles for this ones, but didn't find
> > anything. OTOH i'm more less sure, that they are quite old and (possibly)
> > harmful.

Cheers,

Wiktor Niesiobedzki