[PATCH 0/6] ath5k: various misc fixes

[PATCH 0/6] ath5k: various misc fixes

[Bob Copeland]

Hi, here are some various updates for ath5k, some minor fixes, it's 
probably all 2.6.30 material.

- patch 1 fixes a locking bug, but for consistency, not to fix any reported
  issue
- patches 2 and 4 are cleanups
- patch 3 makes reset() more closely match the HAL, it may help with reset
  errors but really is unlikely to fix any problem
- patches 5 and 6 add missing support for some rate flags that have been
  ignored by ath5k since forever

Re: [ath5k-devel] [PATCH 0/6] ath5k: various misc fixes

[Maxim Levitsky]

On Thu, 2009-01-22 at 08:44 -0500, Bob Copeland wrote:
> Hi, here are some various updates for ath5k, some minor fixes, it's 
> probably all 2.6.30 material.
> 
> - patch 1 fixes a locking bug, but for consistency, not to fix any reported
>   issue
> - patches 2 and 4 are cleanups
> - patch 3 makes reset() more closely match the HAL, it may help with reset
>   errors but really is unlikely to fix any problem
> - patches 5 and 6 add missing support for some rate flags that have been
>   ignored by ath5k since forever

I tried using those 4 patches that were your attempt to fix hangs
the first one, probably helps a bit, but with them I got those 'failed
to wakeup PHY' errors that persist even after suspend/resume cycle.

Now I only use the 'continue reset sequence' and while chip locks up
from time to time, I don't see the messages about failed PHY wakeup.

Best regards,
	Maxim Levitsky

Re: [ath5k-devel] [PATCH 0/6] ath5k: various misc fixes

[Bob Copeland]

On Thu, 22 Jan 2009 17:11:58 +0200, Maxim Levitsky wrote
> On Thu, 2009-01-22 at 08:44 -0500, Bob Copeland wrote:
> > Hi, here are some various updates for ath5k, some minor fixes, it's 
> > probably all 2.6.30 material.
> > 
> > - patch 1 fixes a locking bug, but for consistency, not to fix any reported
> >   issue
> > - patches 2 and 4 are cleanups
> > - patch 3 makes reset() more closely match the HAL, it may help with reset
> >   errors but really is unlikely to fix any problem
> > - patches 5 and 6 add missing support for some rate flags that have been
> >   ignored by ath5k since forever
> 
> I tried using those 4 patches that were your attempt to fix hangs
> the first one, probably helps a bit, but with them I got those 'failed
> to wakeup PHY' errors that persist even after suspend/resume cycle.
> 
> Now I only use the 'continue reset sequence' and while chip locks up
> from time to time, I don't see the messages about failed PHY wakeup.

Cool, thanks for the feedback.  For the record that is patch 3 in this
series.  I sent everything except the one that puts a spinlock around
reset; that still hangs for me so I'd rather not send it until I know
why.  Consequently, the other 2 patches (1 and 2) don't have any major
effect.

I suspect patch 3 helps but doesn't cure anything.

-- 
Bob Copeland %% www.bobcopeland.com