Re: [PATCH] bcm5974-0.58: name changes, open/close and suspend/resume serialized

[Dmitry Torokhov <dmitry.torokhov@gmail.com>]

On Fri, Jul 25, 2008 at 08:25:29PM +0200, Henrik Rydberg wrote:
> Hello Dmitry,
> > 
> > Thank you for the changes.  You don't have to track whether device is
> > manually suspended or not - usb_submit_urb will fail and that is it.
> > 
> 
> Perfect.
> 
> > Attached are 2 patches. First cleans suspend state tracking as it is not
> > really needed and does some formatting and other minor changes and 2nd
> > implements runtime pm for the device. Could you please try them and if
> > everything still works I will apply the driver.
> 
> Patch 1:
> 
>   Looking good, just a few comments:
> 
>   * I notice that clamp_val is new since 2.6.24.3

Probably, I am not sure...

>   * Regarding the open-fails comment: is EACCES correct?

We don't return EACCESS anymore, I don't think. We will return what
bcm5974_start_traffic returns which in turn returns result of
usb_submit_urb and the likes.

I don't think returning -EACCESS was proper - it usually indicates
permisison issues, not the fact that the hardware is unavailable.

> 
> Patch 2:
> 
>   Perfect.
> 
> Thank you so much, Dmitry - it's a GO!
> 

Great, thank you!

-- 
Dmitry