[PATCH] linking wpa supplicant sm with hostapd

[Leonardo Maccari]

On Tue, Sep 20, 2005 at 08:49:14PM -0700, Jouni Malinen wrote:
> On Tue, Sep 20, 2005 at 12:23:02PM +0200, Leonardo Maccari wrote:
> 
> It needs to be callable from other files, so in that sense, yes, it
> cannot be static. I don't link chaning just pmksa_cache_add() while
> leaving other functions using their current names. Looks like this could
> be easier to resolve on the hostapd side either by renaming the
> pmksa_cache_() functions or maybe just adding another function as a
> wrapper for pmksa_cache_add() and making this conflicting function
> static.

that's an easy way, but wouldn't it be better to have the names of
functions and files respecting also which side they belong to
(supplicant/authenticator)?

> 
> > #define WPA_REPLAY_COUNTER_LEN 8
> 
> I would assume you mean to wpa_common.h, not .c.. Yes, that would be

yes.

> possible. I was just cleaning up wpa_supplicant by moving wpa.c-specific
> defines away from wpa.h, but if there is a good reason to share these
> with hostapd, using a common header file would be fine.

If I want to link wpa.c from wpa_supplicant into hostapd I cannot just
add the object into the makefile becouse there are double definitions
with wpa.c in hostapd code. 

for the same reason I had to redefine struct wpa_eapol_key since they
differ from supplicant to hostapd (why they do?).

> One thing to note here is that I'm trying to start stabilizing 0.4.x
> branch soon, so this kind of moving code from one place to other is
> going to become less desired before a separate branch for 0.4.x is
> created. In addition, I would like to delay making the new branch as
> long as possible in order to avoid having to maintain two branches that
> are still under active development.

this sounds like "do it now or wait a lot" :)
ok.

ciao,
leonardo.

-- 
   Key fingerprint = 3129 C583 F03B 2E73 0115  C040 3489 0185 B592 19FE
 Obviously -thisaintpartofmyaddress- is not part of my real email address