wireless drivers in staging

John W. Linville linville at tuxdriver.com
Fri Sep 11 19:27:09 UTC 2009


Oops...sorry for the delayed response!

On Fri, Aug 21, 2009 at 09:07:37AM -0700, Greg KH wrote:
> On Fri, Aug 21, 2009 at 10:31:55AM -0400, John W. Linville wrote:
> > Greetings!
> > 
> > Is anyone working specifically on any/all of the wireless drivers
> > in drivers/staging?  I don't mean the odd patch or two, but actually
> > focusing on cleaning-up and/or improving one of the wireless drivers
> > in staging until it becomes ready for mainline.
> 
> Based on the patches I have received over the past few weeks, there are
> a number of developers working on different drivers.
> 
> Here's some that are obviously doing work:
> 	- Bart is working on some of the RealTek drivers (rtl8187se) and
> 	  on merging the different Ralink drivers in the staging tree
> 	  (rt2860, rt2870, rt3070, and rt3090) together into one simpler
> 	  driver.
> 	- Pekka was working on the w35und wireless driver, I don't know
> 	  if he has a device or not.
> 	- Forest and Jim are working on the Via wireless drivers (vt6655
> 	  and vt6656) and are trying to merge them together and then
> 	  clean them up more.
> 	- I personally am working on the RealTek 8192e driver as I have
> 	  a device and a data sheet and have to support it for my
> 	  employer.
> 
> There are others doing lots of general cleanups to the staging wireless
> drivers, but I'm not sure they plan on doing more than that.  If I have
> missed anyone, please feel free to speak up.
> 
> If anyone is looking for stuff to do, I think it's obvious that a number
> of these drivers do need help :)

Thanks for the overview.  I'm glad someone is giving attention to
these drivers.

As you saw recently on LKML, hch has been mentoring a group on
producint a VIA driver.  Also, I understand that someone from Ralink
is now working with the rt2x00 family (which is also growing rt2800usb
and rt2800pci drivers).

Someone had started work on an rtl8187se driver -- I even looked
at it briefly.  Unfortunately that code is even more attrocious
than most vendor code.  Any cleanups there are definitely welcome,
whether they result in a direct port or just a cleaner reference for
some new driver.

> > If so, I would like to hear from you!  Could you reply and let me know
> > what you are doing, what your goals are, if you have a plan, etc?
> > I would just like to make sure everyone is communicating and try to
> > get everyone working together productively.
> 
> I'll speak for myself only on this.  My plan for the rtl8921e driver is
> to clean it up, move it over to the existing wireless stack in the
> kernel, and then submit it for inclusion in your main
> drivers/net/wireless/ tree.
> 
> The code is currently using a very old version of the wireless stack, so
> it isn't going to be that difficult to do.
> 
> I've already done a bit of work on the existing driver, and with the
> help from some people at RealTek, fixed some bugs that were found.
> Hopefully my contacts with them will not disappear as they have in the
> past.

It would be good to get/keep Realtek on-board.  They have provided some
datasheets in the past, but little direct participation.  I presume
you have a datasheet for the rtl8192su?

> > P.S.  If anyone is interested in using a wireless driver in staging
> > as a resource for reverse engineering rather than as a porting target,
> > I would love to hear from you too.
> 
> Personally I like taking an existing, working driver, and slowly moving
> it over to something else, all the while testing that I didn't break
> anything, instead of writing a new one from scratch.
> 
> But that's just me, perhaps because I also don't know the wireless stack
> all that well.  If I did, it might be easier to start from scratch :)
> 
> Might I ask what you need this info for?

Curiousity mostly.  I do get asked about some of these drivers by
users from time to time, so it is good to have an idea which might
eventually make their way to the mainline kernel.

I would like to point-out that not only do drivers need to be ported to
mac80211 (if appropriate for the hardware) but also that non-mac80211
drivers now need to support the cfg80211 API rather than the old
wireless extensions API.  I regret that this may be more work for
those working on ports, but OTOH this is probably a no-op for any
of the drivers currently in staging (since mac80211 takes care of
cfg80211 support for you).

Beyond that, I was hoping to "break the ice" a bit between the
driverdev group and the wireless group.  It would be good if we could
get people working together towards a mutually acceptable set of
drivers rather than having one group toiling on something the other
group considers a "dead end".  Let's hope... :-)

John
-- 
John W. Linville		Someday the world will need a hero, and you
linville at tuxdriver.com			might be all we have.  Be ready.



More information about the devel mailing list