Re: pf-related network problem

[Jan Lentfer <Jan.Lentfer@xxxxxx>]

Sepherosa Ziehau schrieb:
> On Tue, Jan 25, 2011 at 4:58 PM, Peter Avalos <peter@theshell.com> wrote:
> > While attempting to reload a huge table (about 300k entries), my network
> > locks up and I get thousand of these messages on console:
> >
> > pfr_unroute_kentry: delete failed.
>
> I believe it has been fixed on master:
> b4628cf91a5f54965e6a4221f243e315aa7a6860

I can confirm that the pfctl -ef /etc/pf.conf; pfctl -F Tables cycle now 
succeeds without any errors on my SMP VM.

BUT, if I kldunload pf now I get

1 tables deleted.
df386devel# kldunload pf
kthread 0xd2ab5f58 pfpurge has exited


Fatal trap 12: page fault while in kernel mode
cpuid = 1; lapic.id = 01000000
fault virtual address	= 0x500ffd7
fault code		= supervisor read, page not present
instruction pointer	= 0x8:0x500ffd7
stack pointer	        = 0x10:0xd04bf7c4
frame pointer	        = 0x10:0xd04bf7d8
code segment		= base 0x0, limit 0xfffff, type 0x1b
			= DPL 0, pres 1, def32 1, gran 1
processor eflags	= interrupt enabled, resume, IOPL = 0
current process		= 927 (kldunload)
current thread          = pri 10
 <- SMP: XXX
trap number		= 12
panic: page fault
cpuid = 1
Trace beginning at frame 0xd04bf6d0
panic(ffffffff) at panic+0x164
panic(c05edca4,c061e127,0,0,fffff) at panic+0x164
trap_fatal(500ffd7,0,d2852620,d294ac98,0) at trap_fatal+0x307
trap_pfault(cfc03260,d04bf754,0,0,0) at trap_pfault+0x124
trap(d04bf77c) at trap+0x4df
calltrap() at calltrap+0xd
--- trap 0, eip = 0, esp = 0x10246, ebp = 0 ---
boot() called on cpu#1
Uptime: 1m23s
Physical memory: 1014 MB
Dumping 140 MB:panic: Bad link elm 0xd04bf44c next->prev != elm
cpuid = 1
boot() called on cpu#1
Uptime: 1m23s
spin_lock: 0xcd86c2f4, indefinite wait!
panic: Bad tailq NEXT(0xcd86c2f8->tqh_last) != NULL
cpuid = 1
boot() called on cpu#1
Uptime: 1m23s
spin_lock: 0xcd86c2f4, indefinite wait!
panic: Bad tailq NEXT(0xcd86c2f8->tqh_last) != NULL
cpuid = 1
boot() called on cpu#1
[..]

That does not happen if I do exactly the same but without table <p2p> 
persist file "/etc/p2p_small.pf" in pf.conf.


Jan