Re: [Madwifi-devel] rssi not updated correctly

[Pavel R. <pr...@gn...>]

On Wed, 2007-03-14 at 00:09 +0100, Benoit PAPILLAULT wrote:
> Benoit PAPILLAULT wrote:
> > - is my understanding correct?
> > - if so, how to solve this bug?
> 
> According to 802.11-1999, 7.2.2 : "Address 2 always holds the address of 
> the station that is transmitting the frame". So I added a condition in 
> the source code, patch follows.
> 
> Is it correct?

Benoit,

I feel bad as I see your interesting questions not getting replies.
Sometimes interesting questions need more time, and not everybody has it
immediately.  Please create tickets for all outstanding patches if you
haven't done so.

Specifically for this patch:

First of all, please use C comments.  Although C++ comments are allowed
by gcc, they are not common in the kernel code.

Please don't use non-common abbreviations.  What is BPA and what is
comp?  What are unknown nodes?  And how are you avoiding them?

If BPA is your initials, please don't put it into the comments.  This
information doesn't belong to the code, but to the versioning system.
The code would be unreadable if everybody puts the initials for every
change.  Good comments should describe what the code does in general
context.  Further coders would be more interested why this check is
needed, not which problems it avoids (unless the later is non-obvious
from the former).

Then let's consider the implementation.  What are the conditions when
this change would make the difference?  What are the symptoms of the
problem you are trying to fix?

The code disables setting the statistics to some nodes.  But are you
sure that the statistics would be set for the correct node if it exists?

-- 
Regards,
Pavel Roskin