pool corruption panic in pf

[Hauke Fath <hauke%Espresso.Rhein-Neckar.DE@localhost>]

Hi,

after Jörg 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

Stupped in fid 222.1 (xf) 06dfc     =0tb3d530u_Dsfsg=0x3a6<PERR=0x0,LV  0x3
AT=0x5,FT=0x1, AV0O %g
db{0}> t
cpu_Debugger(0xf030fea0, 0xf02ffa30, 0x17, 0xf033e000, 0xf038dc00, 0x104)
at netbsd:pool_get+0xa8
pool_get(0xf037ee98, 0x2, 0x0, 0x8, 0x20, 0xf0dd3800) at
netbsd:pfr_create_ktable+0x1c
pfr_create_ktable(0x0, 0x0, 0x0, 0xf0dc9800, 0x2, 0x1) at
netbsd:pfr_set_addrs+0x88
pfr_set_addrs(0x1, 0xf0dd3800, 0x1, 0xf396b8d4, 0x0, 0x4c86bf59) at
netbsd:pfi_table_update+0x50
pfi_table_update(0xf44759b0, 0xf0dd4900, 0x80, 0x8, 0xf0db8085, 0x4) at
netbsd:pfi_dynaddr_update+0x4c
pfi_dynaddr_update(0xf44759b0, 0xf4474f70, 0xf0db82a8, 0xf0e0ac78,
0xf0dd488c, 0xf0da7500) at netbsd:pfi_kif_update+0x14
pfi_kif_update(0xf0dd4900, 0xf0e0ac78, 0xf0db8280, 0x0, 0x0, 0xf0e0acf8) at
netbsd:pfi_kifaddr_update+0x3c
pfi_kifaddr_update(0xf0dd4900, 0x8040691a, 0x0, 0xf0e0ac78, 0xf0385c00,
0xf0dd4800) at netbsd:pfil_ifaddr_wrapper+0x44
pfil_ifaddr_wrapper(0x0, 0x8040691a, 0xf0dc9800, 0x8, 0xf396bb60,
0xf0dd488c) at netbsd:pfil_run_hooks+0xe4
pfil_run_hooks(0xf0db3a60, 0x8040691a, 0xf0dc9800, 0x8, 0xf0346a88, 0x0) at
netbsd:sppp_ipcp_tlu+0x138
sppp_ipcp_tlu(0xf0dc9800, 0xf02ee800, 0x5, 0xf0dc9800, 0x2, 0x1) at
netbsd:sppp_cp_input+0x78c
sppp_cp_input(0xf02bf138, 0xf0dc9800, 0xf0d81d00, 0x2c, 0x0, 0x320) at
netbsd:sppp_input+0x6f0
sppp_input(0xf0dc9800, 0xf0d81d00, 0xf00f6cb0, 0x0, 0xf0379000, 0x0) at
netbsd:pppoe_softintr_handler+0x390
pppoe_softintr_handler(0x0, 0xf033ec00, 0xf033ec00, 0xf02f7c00, 0xf0d321c0,
0xf0378c00) at netbsd:softint_overlay+0x564
softint_overlay(0xf02e1000, 0x20000020, 0x40, 0xf02f7800, 0xf02e1000,
0xf363b340) at netbsd:lwp_userret+0x1d8
lwp_userret(0xf363b340, 0x8, 0x4033b000, 0x3, 0xf038f158, 0xf0318400) at
netbsd:mem_access_fault4m+0x130
mem_access_fault4m(0x9, 0x38a, 0x403371c0, 0xf396bfb0, 0x0, 0x5) at 0xf000625c
db{0}>

(Note the peculiar speling in "Stupped in fid" above.)

This happens with both a UP and a MP kernel. Userland is -current, but the
content of /etc is still netbsd-4, if it matters.

Any ideas?

        hauke


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