Hi Everyone,

  I recently squashed some big bugs in the Solaris port. I can now execute files in /coda (mmap now works.) I haven't (yet) had a kernel panic with the current solaris-coda module. But I do know there are things that don't work. For one, the simple perl program:

open(FH, ">xxx");
print FH "abcdef\n";
close(FH)

works just fine on a local file system, but on coda, it creates the file xxx, but doesn't write any thing to it.

Has anyone ever run into a problem like this under coda?

Also, the file size total at the top of an "ls -l" is completely wrong. I haven't yet found which call is not correct.

In any case, the solaris code is now working for me on both a sparc 5 and an ultra 10 both running solaris 2.6. I hope to get Solaris 7 up and running on a machine in the near future.

These changes for solaris (for lwp, coda, and solaris-coda) are big enough that as soon as I test out a few more things, I think we should make a new release of each.

By the way, I'm currently having iomgr.c problems on both my Solaris systems. venus stops with

../../src/iomgr.c:293: failed assertion `0'

Shouldn't there be something better to do there than panic? Like go back to the select and wait again? Or is it really the right thing to do there?

-- 
Phil Nelson                    NetBSD: http://www.netbsd.org
e-mail: phil@cs.wwu.edu        Coda: http://www.coda.cs.cmu.edu
http://www.cs.wwu.edu/~phil