Date:   Fri, 26 Nov 1999 05:59:05 -0500 (EST)
From:   Alexander Viro <viro@math.psu.edu>
To:     linux-fsdevel@vger.rutgers.edu
Subject: QNXFS must die.

	Folks, this is no joke. Driver is broken and unless documentation
on the filesystem layout will show up it will remain broken. By some
accident it works on 2.2 if files are small and not too fragmented.
Porting it to 2.3 might make sense if authors would care to do _something_
about the thing. It didn't happen.
	Driver went into the main tree about 17 month ago. Since then
there was _NO_ patches from the authors. And it's not like driver was in
the working order. Moreover, the first time when authors cared to reply to
email was in October, when they were asked whether they want it to stay in
the tree. Reply went along the lines "QSSL had cut all contacts long ago
and we can't reach them. But when we'll manage to get something out of
them we'll do something." Well, it's the ebd of November. Result: no
documentation, no patches, no nothing. No sound from the authors, BTW.
	The bottom line: fs/qnx4 is completely unmaintained. If documentation
will not appear RSN I _really_ think that rm -rf fs/qnx4 include/linux/qnx*
is the best possible way for 2.3. From bits that could be found on search
engines it looks like proper implementation will not be easy - it's an
extent-based filesystem with inode table and block bitmap sitting in
files, moreover, inode may be embedded into directory entry and can
migrate into the inode table. Worse yet, the layout reconstructable from
the qnx*.h is flexible enough to leave a lot of ambiguity in support of
fragmented files. As much as it may hurt feature-list lovers, unless there
will be documentation on filesystem layout fs/qnx4 must die. Keeping this
code around makes no sense - it's broken, unmaintained and apprently going
to remain that way.


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/