Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: pool corruption panic in pf



At 2:45 Uhr +0100 10.9.2010, Mindaugas Rasiukevicius wrote:
>Hauke Fath <hauke%Espresso.Rhein-Neckar.DE@localhost> wrote:
>> after J??rg fixed the sbrk() issue, the ss20 gets some way into booting
>> multi user, then panics over a corrupted memory pool:
>>
>> [...]
>> Configuring network interfaces: hme0 hme1 pppoe0.
>> Adding interface aliases:
>> pppoe0: connected to DARX43-erx
>> Enabling pf firewall.
>> panic: pool 'pfrktable' is IPL_NONE, but called from interrupt context
>
>It is not a pool corruption.  The pool should be created with IPL_NET.

The following patch

Index: pf_table.c
===================================================================
RCS file: /cvsroot/src/sys/dist/pf/net/pf_table.c,v
retrieving revision 1.16
diff -u -u -r1.16 pf_table.c
--- pf_table.c  12 Apr 2010 13:57:38 -0000      1.16
+++ pf_table.c  12 Sep 2010 18:44:40 -0000
@@ -200,9 +200,9 @@
 {
 #ifdef __NetBSD__
        pool_init(&pfr_ktable_pl, sizeof(struct pfr_ktable), 0, 0, 0,
-           "pfrktable", &pool_allocator_nointr, IPL_NONE);
+           "pfrktable", &pool_allocator_nointr, IPL_NET);
        pool_init(&pfr_kentry_pl, sizeof(struct pfr_kentry), 0, 0, 0,
-           "pfrkentry", &pool_allocator_nointr, IPL_NONE);
+           "pfrkentry", &pool_allocator_nointr, IPL_NET);
        pool_init(&pfr_kentry_pl2, sizeof(struct pfr_kentry), 0, 0, 0,
            "pfrkentry2", NULL, IPL_SOFTNET);
 #else

fixes the panic.

        hauke


--
"It's never straight up and down"     (DEVO)




Home | Main Index | Thread Index | Old Index