Re: [PATCH] b43legacy: Change the hardware radio enable logic and cleanup code

[Michael Buesch <mb@bu3sch.de>]

On Friday 21 September 2007 03:10:07 Larry Finger wrote:
> This change cleans up the radio-related messages in several ways.
> 
> (1) The state of the rfkill switch is assumed to be on, rather than
>     tested. Now, any user without such a switch will not see any
>     messages. For devices with such a switch, a message will be
>     logged only if the initial state is off, or if the switch is toggled.
> (2) The routine for testing the switch state is no longer inline.
> (3) The LED handling routine is simplified.
> (4) The "Radio turned off" message that has confused some users has been
>     changed to "Radio initialized".
> 
> This patch is patterned after a similar change to b43 by Michael Buesch.
> 
> Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>

> Index: wireless-dev/drivers/net/wireless/b43legacy/leds.c
> ===================================================================
> --- wireless-dev.orig/drivers/net/wireless/b43legacy/leds.c
> +++ wireless-dev/drivers/net/wireless/b43legacy/leds.c
> @@ -182,6 +182,7 @@ void b43legacy_leds_update(struct b43leg
>  	unsigned long interval = 0;
>  	u16 ledctl;
>  	unsigned long flags;
> +	bool radio_enabled = (phy->radio_on && dev->radio_hw_enable);

Just that you are aware of it:
This is a racy access to the radio_hw_enable variable, as we don't take
the mutex here (we can't and we don't want). But it's OK, since nobody
cares if the LED is racing for a second and displays the wrong state
for a second.
Same goes for the phy->radio_on access.

The patch is ACKed by me.

-- 
Greetings Michael.