[PATCH] libertas: rename escape_essid() to _escape_essid() in assoc.c

Dan Williams dcbw at redhat.com
Fri May 18 11:27:46 EDT 2007


On Thu, 2007-05-17 at 00:30 -0400, Andres Salomon wrote:
> Jeff Garzik wrote:
> > Andres Salomon wrote:
> >> Hi,
> >>
> >> I just committed this to olpc-2.6; it should probably be in the libertas
> >> tree as well.  Pavel noticed that including both libertas and the 80211
> >> stack made things blow up.
> >>
> >>
> >>
> >> ------------------------------------------------------------------------
> >>
> >> >From 3ecd1083f2e095172f0b69f0082ed41ff58a3a84 Mon Sep 17 00:00:00 2001
> >> From: Andres Salomon <dilinger at debian.org>
> >> Date: Tue, 8 May 2007 15:28:55 -0400
> >> Subject: [PATCH] libertas: rename escape_essid() to _escape_essid() in assoc.c
> >>
> >> We took escape_essid from the 80211 stack so that we wouldn't have to include
> >> it (until there's a generic 80211 lib), but there are name conflicts if both
> >> modules are included.  Thanks to Pavel Machek for pointing that out.
> >>
> >> Aside from renaming it, I've also made it static (since it's not used outside
> >> of assoc.c) and updated the comment describing its origin.
> >>
> >> Signed-off-by: Andres Salomon <dilinger at debian.org>
> > 
> > seems like the correct fix is to depend on the code in the upstream 
> > kernel, not avoid it.
> > 
> 
> Of course; the rename fix was designed to be temporary.  Dan had a
> reason for libertas having its own copy of the function (I forget what
> it was, though).

That's still the only function in the ieee80211 that libertas would
depend on, and I don't feel that's critical enough mass to load in
ieee80211.  I guess it's only ~30k or so as reported by lsmod, which
isn't horrible.  Andres, that does mean we have to include ieee80211 in
OLPCs config as well.

The plan is to eventually depend on lib80211, which will include a bunch
of stuff that's moved from _both_ mac80211 and ieee80211 that many
drivers likely need.

Dan





More information about the libertas-dev mailing list