Followup: it was

panic: pmap_unwire: invalid (unmapped) va

Only solution was to revert to a 8 July kernel.

Patrick Welche wrote:
> 
> Subject: New panic
> Reply-to: prlw1@cam.ac.uk
> With a new kernel from half an hour ago's source (i386, formerly stable)
> 
> db> ps
>  PID             PPID       PGRP        UID S   FLAGS          COMMAND    WAIT
> >16                15          5          0 2  0x4006         fsck_ffs
> db> t
> Debugger(f7d900cc,8082000,8082000,f7e6fd28,f01ee45f) at Debugger+0x4
> panic(f024ed80,41e,f7d900cc,f7e6fd64,f01dadd2) at panic+0x55
> pmap_unwire(f7d900cc,8082000) at pmap_unwire+0x87
> uvm_fault_unwire_locked(f7d8f21c,8080000,8082000,f7d8b780,f05def80) at uvm_faul
> t_unwire_locked+0xfe
> uvm_fault_unwire(f7d8f21c,8080000,8082000,ffffffff,f7e6fdf0) at uvm_fault_unwir
> e+0x38
> uvm_vsunlock(f7d8b780,8080000,2000,f05def80,2000) at uvm_vsunlock+0x27
> physio(f01f7a90,0,d00,100000,f01f7e2c) at physio+0x28e
> sdread(d00,f7e6fef0,0,0,f7da4304) at sdread+0x1f
> spec_read(f7e6fea4,f7e6feb8,f0177a2d,f7e6fea4,f7e6ff88) at spec_read+0xc3
> ufsspec_read(f7e6fea4,f7e6ff88,2000,f7d9f090,f7e6fea4) at ufsspec_read+0x2d
> vn_read(f7d9f090,f7d9f0ac,f7e6fef0,f05a3880,1) at vn_read+0xad
> dofileread(f7d8b780,4,f7d9f090,8080000,2000) at dofileread+0x93
> sys_read(f7d8b780,f7e6ff88,f7e6ff80,0,0) at sys_read+0x67
> syscall() at syscall+0x149
> --- syscall (number 3) ---
> 0x8070639:
> 
> How does one read the message buffer for the actual panic message from
> within db? This is coming at it on a serial console after the event..
> (Will probably find out from the core dump..)
> 
> Cheers,
> 
> Patrick
> .
>