I've had my own filesystem deadlock on me before with every kernel I've =
compiled, but I haven't been able to duplicate the conditions.  It seems =
most likely to happen when two processes in different shells try to =
access the same directory (or directories related within one generation) =
at the same time, i.e., running a compile in one xterm and checking the =
status of the files in another.  My system is an i386 with a pair of =
SCSI drives.

----------
From: 	Robert Black[SMTP:r.black@ic.ac.uk]
Sent: 	Monday, June 10, 1996 12:45 PM
To: 	current-users@NetBSD.ORG
Subject: 	Processes Hanging in a Disk Wait

Hello,

Some time ago I know there was a problem with several ports with =
processes
occassionally hanging in a disk-wait (in vflushbuf). Does anyone know =
whether
this problem was ever resolved?

If the problem hasn't been resolved I can now reproduce the problem =
reliably so
if anyone has any bright ideas of things I might like to try to track =
this down
I will give them a whirl. (unfortunately I'm not really up to speed on =
how the
vnodes work and my 4.4 book doesn't arrive until next week)

Cheers

Rob Black

PS I am using the arm32 port. I don't think this is relevant though =
because I'm
sure this problem has been discussed with other ports (okay, so it could =
be an
spl bug...). If anyone can point me in the direction of any other ports =
known
to be affected I would be grateful.

--=20