Re: netfs part of a console server with server-client model

[Marcus Brinkmann]

On Sun, Jun 02, 2002 at 05:13:57AM +0200, Marcus Brinkmann wrote:
> What I attach to this mail is a file that incorporates some the virtual
> console abstraction in console/console.c in the Hurd source tree, and new
> parts which provide a netfs based filesystem that looks like this:

Ok, here is a new version.  Below is a diff from the old version to the new
one.  It fixes a couple of bugs (for one, a potential deadlock in
netfs_node_norefs vs. netfs_attempt_lookup, which might be interesting for
netfs Hurd hackers to know about, furthermore a glitch in the vcons list
processing, and some mode bit settings, which could be done even better,
though, but it is only cosmetic).

I also added the input node that was suggested by Nils.  An example:

$ ls con/2
total 0
crw-r-----    1 marcus   marcus     0,   0 Jun  3 03:17 console
-rw-r-----    1 marcus   marcus       2000 Jun  3 03:17 display
prw-r-----    1 marcus   marcus          0 Jun  3 03:17 input

You can see that the input node is shown as a fifo, which I think is
appropriate.  All of this is entirely fake, of course.  I did some more
stress testing, and this version seems to be quite robust (eg, listing 100
virtual consoles in parallel works fine).

I will soon merge in the real meat for the input and console device.  The
display device will be more difficult, as 1. the code isn't written yet, and
2. the interface is not designed ;)  But I certainly will add some cheesy
polling interface for testing rather soon, and a simple display client that
piggybacks on the existing Hurd console to go with.

Thanks,
Marcus

-- 
`Rhubarb is no Egyptian god.' Debian http://www.debian.org brinkmd@debian.org
Marcus Brinkmann              GNU    http://www.gnu.org    marcus@gnu.org
Marcus.Brinkmann@ruhr-uni-bochum.de
http://www.marcus-brinkmann.de