> just in case this is different from the 'locking against myself' fixes
> recently committed ...
> 
> i386 NetBSD 1.3K, slightly trimmed kernel based on GENERIC off the CDROM
> image recently posted (thanks). I haven't layered any filesystems.
> WARNING hand-copied; low S/N; no ECC
> 
> _panic(f0181d25,20002,f4b72d68,30002,1048) at _panic+0x57
> _lockmgr(f4ba,30002,f4b72df4,f4bd1668,f01aa6f3) at _lockmgr+0x264
> _ufs_lock(f4bd165c) at _ufs_lock+0x22
> _vn_lock(f4b72d68,20002) at _vn_lock+0x3f
> _uvn_io((f4b72d68,f4bd1c04,1,2,0) at _uvn_io+0x11d
> _uvn_get(f4b72d68,0,f4bd1c80,f4bd1c80,f4bd1c84,0) at _uvn_get+0x196
> _uvm_fault(f4a15630,400a,0,1,0) at _uvm_fault+0x6f
> _trap() at _trap+0x40f
> --- trap (number 6) ---
> _copyin(2a28000,2000,f4bd1eec,1,f4b72df8) at _copyin+0x33
> _ffs_write(f4bd1ea0,f4bd1f88,189612,f46c2004,f4bd1ea0) at _ffs_write+0x355
> _vn_write(f4a2357c,f4a23594,f4bd1eec,f05d9f80,1) at _vn_write+0xd3
> _dofilewrite(f4bc2004,3,f4a2357c,400a3000,189612) at _dofilewrite+0x81
> _sys_write(f4bc2004,f4bd1f88,f4bd1f80,0,0) at _sys_write+0x43
> _syscall() at _syscall+0x20e
> --- syscall (number 4) ---

This looks exactly like the case fixed in the "locking against myself"
fix I committed yesterday.  In this case, it's presumably a write to a
file, out of an mmap'ed region backed by the file.

Do you understand how you triggered this?  The reason I ask is that
the "fix" I committed merely prevents a panic; the read() or write()
call will still fail, so whatever application was doing this will
still be unhappy after the fix, but won't take the whole system down
with it.

					- Bill