Hi folks,

My stock netbsd-1.5.2 kernel running on a Quadra 950 paniced in the file
system last night shortly after I backed-up a gigabyte or so from my
powerbook via a Finder-copy over netatalk. The panic was along the lines of
"invalid directory inode entry XXXXXXXX offset 0", along with some message
about being in find, and in directory /home (which, BTW, is on a different
file system and disk than the shared AFP directories).

There were quite a number of console messages about FIFO overruns during the
backup; I'm presuming these are harmless with the exception that my TCP
throughput is degraded by the lost packets.

I've a couple of questions to bother you all with:

1) is this a well-known bug? If not, then

2) the kernel debugger didn't give a stack trace; what do I type to get one?
The netbsd problem-report web-page claims to only want the PC, but I suspect
a stack backtrace would be helpful.

3) I've been lax in backing-up non-irreplacable data, expecting NetBSD to be
quite reliable. What has your experience been with NetBSD and
netatalk-asun-<mumble> on 68K Macs? Relatively bug-free, or problems? Am I
suffering from the dread NCR driver file system rot?

Thanks.

Regards,
-Steve

Steve Byan
<smb@world.std.com>