On Thu, 2005-09-01 at 10:59 -0700, Jean Tourrilhes wrote:
> Jiri Benc wrote :
> > On Thu, 01 Sep 2005 11:09:16 +0100, Pedro Ramalhais wrote:
> > > Right, that would need a new interface where all parameters are passed
> > > at once,
> >
> > Then you will lose the possibility of having default parameters.
>
> Just for your information, it's actually trivial to cache
> parameters in the driver and to apply them in one go using the WE
> commit mechanism. It's actually even strongly advised, as it make the
> startup performance much better (fewer reset of the hardware).
> Many drivers do implement this, such as orinoco.c, atmel.c,
> airo.c (partial) and ray_cs.c. The performance benefit of implementing
> it in orinoco.c was actually noticeable.
> The current API is flexible and allow you to have it both
> way. When I designed it, I actually thought about passing a big struct
> with all the parameters to the driver and rejected it because too
> inflexible. It's even documented in iw_handler.h, and I list other
> drawbacks of the approach.
>
> Pedro, if you want more detail of the commit stuff, please
> yell...
> Have fun...
>
> Jean
>
Oops, my brain had censored that part of the iwconfig manual.
"commit Some cards may not apply changes done through Wireless
Extensions immediately (they may wait to aggregate the changes or
apply it only when the card is brought up via ifconfig).
This command (when available) forces the card to apply all pending
changes.
This is normally not needed, because the card will
eventually apply the changes, but can be useful for debugging."
Reading that, and since i imagine that few people use it, it comes to
mind that we could abuse this, and redefine it to be THE association
command. ex:
iwconfig eth0 commit would cause association and no other command would
cause it.
I also fail to see what other uses it could have besides setting the
configuration on the card all at once. Which wireless settings could be
left to be applied on the card "later"? hmmm, the only thing coming to
mind is waiting for a scan to finish, does that count? That would be
better exposed as a BLOCKING vs NONBLOCKING setting.
Thanks!
--
Pedro Ramalhais <[EMAIL PROTECTED]>
-
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to [EMAIL PROTECTED]
More majordomo info at http://vger.kernel.org/majordomo-info.html
