[rt2x00-users] [PATCH] rt2x00: Fix beacon de-synchronization while update beacon (rt73usb)

[Igor Perminov]

On Sat, 01/08/2009 at 15:21 +0200, Ivo van Doorn wrote:
> Hi,
> 
> > When beacon is being updated to refresh TIM (AP mode), beacon frames are
> > de-synchronizing (i.e. two neighbor beacon frames - before and after
> > update - are being transmitted with a wrong time interval).
> > That is because rt73usb_write_beacon should clear only
> > TXRX_CSR9_BEACON_GEN flag, but not TXRX_CSR9_TSF_TICKING and
> > TXRX_CSR9_TBTT_ENABLE.
> 
> Sorry for the late reply, after discovering that your mails are ending up
> in my spambox I have now gone through the entire folder to see what
> other mails were incorrectly marked as spam. :S
> 
> > Signed-off-by: Igor Perminov <igor.perminov at inbox.ru>
> > ---
> > diff -urN a/drivers/net/wireless/rt2x00/rt73usb.c b/drivers/net/wireless/rt2x00/rt73usb.c
> > --- a/drivers/net/wireless/rt2x00/rt73usb.c	2009-07-16 23:12:43.000000000 +0400
> > +++ b/drivers/net/wireless/rt2x00/rt73usb.c	2009-07-23 00:56:07.000000000 +0400
> > @@ -1545,8 +1545,6 @@
> >  	 * otherwise we might be sending out invalid data.
> >  	 */
> >  	rt2x00usb_register_read(rt2x00dev, TXRX_CSR9, &reg);
> > -	rt2x00_set_field32(&reg, TXRX_CSR9_TSF_TICKING, 0);
> > -	rt2x00_set_field32(&reg, TXRX_CSR9_TBTT_ENABLE, 0);
> >  	rt2x00_set_field32(&reg, TXRX_CSR9_BEACON_GEN, 0);
> >  	rt2x00usb_register_write(rt2x00dev, TXRX_CSR9, reg);
> 
> Could you update this patch for the other drivers as well?
> 
> Thanks,
> 
> Ivo

Hi,

As for rt2400pci, rt2500pci, rt61pci and rt2800usb, beacon on/off code
is almost the same and I can update the patch for these drivers easily.

And as for rt2500usb, this driver has an extra code in
rt2500usb_kick_tx_queue (lines 6 - 14 in the code snippet below):

 1:	rt2500usb_register_read(rt2x00dev, TXRX_CSR19, &reg);
 2:	if (!rt2x00_get_field16(reg, TXRX_CSR19_BEACON_GEN)) {
 3:		rt2x00_set_field16(&reg, TXRX_CSR19_TSF_COUNT, 1);
 4:		rt2x00_set_field16(&reg, TXRX_CSR19_TBCN, 1);
 5:		rt2x00_set_field16(&reg, TXRX_CSR19_BEACON_GEN, 1);
 6:		/*
 7:		 * Beacon generation will fail initially.
 8:		 * To prevent this we need to register the TXRX_CSR19
 9:		 * register several times.
10:		 */
11:		rt2500usb_register_write(rt2x00dev, TXRX_CSR19, reg);
12:		rt2500usb_register_write(rt2x00dev, TXRX_CSR19, 0);
13:		rt2500usb_register_write(rt2x00dev, TXRX_CSR19, reg);
14:		rt2500usb_register_write(rt2x00dev, TXRX_CSR19, 0);
15:		rt2500usb_register_write(rt2x00dev, TXRX_CSR19, reg);
16:	}

This extra code will also de-synchronize beacon frames.

It is surprisingly, but my rt2500usb stick generates beacon frames fine
even without this part of code, I've tested it.
May be this was a workaround for some early version of a buggy rt2500usb
hardware?

So, we have three alternatives:
A) Leave rt2500usb code as is, not fixing the beacon synchronization
problem for it.
B) Exclude the code that switches TXRX_CSR19 to zero in
rt2500usb_kick_tx_queue with a risk of regression.
C) Same as B, but allow to turn on that code by a module parameter.

What do you think of it?

Igor