Help needed with DR11-W driver

[Chris Torek]

>... After about 30 seconds of exerciser transfers, someone
>(presumably the driver) corrupts something and we get messages
>like (usually inode 2):
>
>bad dir entry ino 2 offset 0: mangled entry

The most common (?) cause of those symptoms is something tromping
on the buffer cache.  (I managed to do this once in the course of 
writing the mass driver, by using geteblk(10240); nothing in the
kernel checked the size of the request.  I fixed *that* too....)
-- 
In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 4251)
UUCP:	seismo!umcp-cs!chris
CSNet:	chris at umcp-cs		ARPA:	chris at maryland