4 person study of orinoco_cs source

[David Gibson]

On Wed, Apr 03, 2002 at 11:20:27AM -0500, Christopher Abiad wrote:
> I'm a part of a group of computer engineering undergrad students that is
> about to start on our 4th year design project.  We have committed to a
> project that involves 802.11b (more details in an email to come soon).
> We've received funding from the school and purchased 802.11b PC cards.
> We've chosen to use Orinoco-based cards because of the GPL driver, which we
> found because of Jean Tourrilhes' page:
> http://www.hpl.hp.com/personal/Jean_Tourrilhes/Linux/Orinoco.html
> 
> Part of our project will require us to change the driver.  We plan to use
> code from David Gibson, which we found at
> http://ozlabs.org/people/dgibson/dldwd/ for our work.
> 
> This weekend we hope to begin to get an understanding of the code's
> workings.  We will report bugs, and suggest any possible optimizations we
> find (?) to this mailing list.  If we get stuck I hope it will be ok to ask
> a question or two as well.  If there are other forums for this kind of
> discussion, please let me know where they are!
> 
> We are by no means experts in the field, but we have taken the time to
> become familiar with the basic structure and workings of network device
> drivers under Linux.  In order to make the best use of our time, we'd
> appreciate any insight that others might share, especially:
> - Recommendations on whether to work from the code in testing/ or in
> orinoco-0.10/.  That is, what level of functionality can I expect from
> testing/?

testing should work at the moment.  In fact it fixes a serious bug or
two in 0.10.  Note that testing does change very frequently and
sometimes bizarrely though - I sync it directly from my working
directory.

Actually you might learn quite a bit from looking at the differences
between the different versions (and correlating it with the changes
listed at the top of the file).

> - Pit-falls and/or especially complex code

The locking is a bit subtle - although there's a bit explanatory
comment in the source.  Actually I think the current locking isn't
really good (not actually broken, but a bit limiting) and I'll
probably be changing it shortly (the new version will probably be
simpler).

> - Recommendations on where to start looking at the code so that it makes the
> most sense

Um... not sure what to recommend there.  I now know the code well
enough that I'm not sure what the best place to start would be.
Probably stay away from ioctl() to start with, though.

-- 
David Gibson			| For every complex problem there is a
david at gibson.dropbear.id.au	| solution which is simple, neat and
				| wrong.  -- H.L. Mencken
http://www.ozlabs.org/people/dgibson