[PATCH] FreeBSD backend

[David Zeuthen]

On Sun, 2006-11-05 at 17:01 -0500, Joe Marcus Clarke wrote:
> Not to steal Artem's thunder, but I have completed initial work on
> merging our HAL patches with HAL git.  In the same vein, I have included
> a diff with all of the "global" code changes.  A link to the full patch
> is also attached.
> 
> Last time Jean-Yves Lefort submitted a "global" code patch, he was told
> to separate it into individual patch files.  I am hoping that simply
> describing what each patch file in this one consolidate patch does will
> be sufficient.

Thanks for working on this. Some comments inline after briefly looking
at the patch... I'll look more at it later..

> hald-runner/main.c : Add PATH elements needed for FreeBSD (patch by
> Jean-Yves Lefort <jylefort at FreeBSD.org>)

hald-runner/main.c: 
-	tmpstr = g_strdup_printf("PATH=/sbin:/usr/sbin:/bin:/usr/bin:%s", getenv("PATH"));
+	tmpstr = g_strdup_printf("PATH=/sbin:/usr/sbin:/bin:/usr/bin:/usr/local/sbin:/usr/X11R6/sbin:/bin:/usr/bin:/usr/local/bin:/usr/X11R6/bin:%s", getenv("PATH"));

Any reason in particular you need all these paths? I think /usr/local
might make sense but I'm not sure why /usr/X11R6/* would be ever needed.
It's not like you'll have an X connection or helpers (callouts, method
calls etc.) should ever need to communicate with X.

> fdi/policy/10osvendor/20-storage-methods.fdi : Add support for FreeBSD
> mount options (is this the best place to add these?)

I think so, yes. Though the actual program doing the mounting, such as
gnome-mount, needs to be updated to use the right mount options.

Though.. right now gnome-mount just allows the user to enter the mount
options in a text box... at some point we'll add nicer UI that will be
constructed based on the contents of valid_mount_options... the we need
some code to properly interpret each mount option. But that shouldn't be
hard.

> hald/hald.c : Add a new function, osspec_privileged_init, that is called
> before hald drops privileges.  This is needed for FreeBSD to be able to
> fully probe hardware (patch by Jean-Yves Lefort <jylefort at FreeBSD.org>)

Oh, ok, this is for things like opening device files for receiving
events from the kernel? I guess that makes sense, yeah, it's just not
something we've been using on Linux...

Thanks again!

    David