Re: NetBSD xen dom0 kernel panic

[yamt%mwd.biglobe.ne.jp@localhost (YAMAMOTO Takashi)]

> A few times my test NetBSD server running NetBSD 5.0_RC2 (amd64) 
> suddenly crashed. (every 2-3 days)
> 
> At the moment of the crash dmesg is:
> 
> panic: lock error
> fatal breakpoint trap in supervisor mode
> trap type 1 code 0 rip ffffffff804f374d cs e030 rflags 246 cr2 
> ffffa00029343000
>   cpl 6 rsp ffffa00028e4e278
> 
> A backtrace:
> 
> b> bt
> breakpoint() at netbsd:breakpoint+0x5
> panic() at netbsd:panic+0x242
> lockdebug_abort() at netbsd:lockdebug_abort+0x42
> mutex_vector_enter() at netbsd:mutex_vector_enter+0x104
> pool_get() at netbsd:pool_get+0x45
> pf_alloc_state_key() at netbsd:pf_alloc_state_key+0x27
> pf_test_rule() at netbsd:pf_test_rule+0xc48
> pf_test() at netbsd:pf_test+0xc3a
> pfil4_wrapper() at netbsd:pfil4_wrapper+0x57
> pfil_run_hooks() at netbsd:pfil_run_hooks+0xa0
> bridge_ipf() at netbsd:bridge_ipf+0x239
> pfil_run_hooks() at netbsd:pfil_run_hooks+0xa0
> bridge_forward() at netbsd:bridge_forward+0xbd
> bridge_input() at netbsd:bridge_input+0x14b
> ether_input() at netbsd:ether_input+0x123
> wm_intr() at netbsd:wm_intr+0x3c0
> evtchn_do_event() at netbsd:evtchn_do_event+0xbe
> call_evtchn_do_event() at netbsd:call_evtchn_do_event+0xd
> hypervisor_callback() at netbsd:hypervisor_callback+0xa3
> pf_purge_expired_states() at netbsd:pf_purge_expired_states+0x6f
> pf_purge_thread() at netbsd:pf_purge_thread+0x5d

inet_pfil_hook and inet6_pfil_hook are not expected to be called
from h/w interrupt context.  ie. BRIDGE_IPF is broken.
it should be fixed to use softint.

YAMAMOTO Takashi

> 
> 
> It doesn't seem completely related to anything else found in the 
> mailinglist and in gnats.