[574] in NetBSD-Development
Re: bitbucket NFS mounting was acting strange.. could this be the cause?
daemon@ATHENA.MIT.EDU (Charles M. Hannum)
Sat Feb 18 12:26:41 1995
Date: Sat, 18 Feb 1995 12:26:07 -0500
From: "Charles M. Hannum" <mycroft@ai.mit.edu>
To: yoav@MIT.EDU
Cc: eichin@MIT.EDU, netbsd-dev@MIT.EDU
I noticed a few days ago that doing a readdir() on the root directory
of `bitbucket' on other systems appears to return an entry with no
name. The NetBSD NFS client is going to balk at this because: a) it
violates the protocol, and b) it makes no sense. (BTW, if you do a
`ls -F' of the directory under Linux, you get `ls: : No such file or
directory'. On a BSDish system, `' is equivalent to `.', so when
ls(1) stat(2)s the file, it will get the info for `.'; hence the
duplicate inode number.)
I thought jhawk was supposed to tcpdump(8) the returned packets and
see whether the NFS server was exercising some loophole in the
protocol (or at least send me the data so I could do so), but he
hasn't that I know of.
