heads-up! ATH9K_RATE_CONTROL config change

[Andy Whitcroft]

On Fri, Jun 14, 2013 at 11:56:26AM -0700, Kamal Mostafa wrote:
> Consumers of the linux-3.8.y and linux-3.5.y extended stable branches
> should take note of this patch, which is queued up for inclusion in
> 3.8.13.3 and 3.5.7.15:
> 
>     ath9k: Use minstrel rate control by default
> 
>     commit 5efac94999ff218e0101f67a059e44abb4b0b523 upstream.
> 
> Its purpose is to make ath9k to use a different rate control algorithm,
> as the current default ath9k algorithm has been deemed too problematic.
> 
> This change is implemented by the removal of the config option
>   ATH9K_RATE_CONTROL  (default=y)
> 
> and the introduction of a new config option
>   ATH9K_LEGACY_RATE_CONTROL  (default=n)
> 
> 
> Take note!  Do not set ATH9K_LEGACY_RATE_CONTROL=y just because your old
> config had ATH9K_RATE_CONTROL=y.  Most consumers of 3.5.y and 3.8.y
> should probably accept the default ATH9K_LEGACY_RATE_CONTROL=n so as to
> keep consistent with mainline and the stable lines.
> 
> More discussion about this:
> http://thread.gmane.org/gmane.linux.kernel.stable/55302
> http://thread.gmane.org/gmane.linux.kernel.stable/55315

I would note that the new option description implies that even if you
enable this option it will not be used by default.  Indeed the change
itself seems to drop part of the change, which is to switch to this new
rate control by default.

I beleive that means enabling this legacy option should be safe in the
sense we will still default to the standard rate control algorithm, but
the flawed algorithm can still be selected should anyone have a use
case.

-apw