Subject: Re: How to disable sending of MAC level ACKs?
Newsgroups: gmane.linux.drivers.madwifi.devel
Date: 2008-07-22 11:01:22 GMT (24 weeks, 3 days, 17 hours and 40 minutes ago)
Hi Dave,
I've been using exactly the approach you describe and it works in ad hoc and infrastructure modes.
I do this and it works nicely. This disables the auto-ACK usually done by the WNIC for *received* frames.
This does the opposite - it causes *sent* frames not to wait for an ACK. You may also need to set the tx count (try0) to "1".
These look pretty good but... you are doing the register write too early (i.e. before the ath_hal_reset). Move it to the end of the function or else you risk the HAL resetting the register back to the default (hw-generated ACK) value. At a guess that is what is happening in your patched version.
Steve
I've been using exactly the approach you describe and it works in ad hoc and infrastructure modes.
In if_ath.c
> First, in the ath_attach(), add the codes "OS_REG_WRITE(ah, 0x8048,
> 0x00000002);", which writes register to disable the hardware auto ACK;
> And in the ath_reset(), do the same thing.
I do this and it works nicely. This disables the auto-ACK usually done by the WNIC for *received* frames.
> Then, in the ath_tx_start(), before the call of
> ath_hal_setuptxdesc(), add the codes "flags |= HAL_TXDESC_NOACK;",
> which sets the flags to stop the hw ACK waiting.
> Done.
This does the opposite - it causes *sent* frames not to wait for an ACK. You may also need to set the tx count (try0) to "1".
* if (ah->ah_abi != HAL_ABI_VERSION) {
* printk(KERN_ERR "%s: HAL ABI mismatch; "
* "driver expects 0x%x, HAL reports 0x%x\n",
* dev->name, HAL_ABI_VERSION, ah->ah_abi);
* error = ENXIO; /* XXX */
* goto bad;
* }
* sc->sc_ah = ah;
* OS_REG_WRITE(ah, 0x8048, 0x00000002);
*
* /*
* * Check if the MAC has multi-rate retry support.
The second reg write (approx line 2180):
*static int
*ath_reset(struct net_device *dev)
*{
* struct ath_softc *sc = dev->priv;
* struct ieee80211com *ic = &sc->sc_ic;
* struct ath_hal *ah = sc->sc_ah;
* struct ieee80211_channel *c;
* HAL_STATUS status;
* OS_REG_WRITE(ah, 0x8048, 0x00000002);
*
* /*
* * Convert to a HAL channel description with the flags
The final flags to prevent nodes waiting for an ack (approx line 7135):
*flags |= HAL_TXDESC_NOACK;
*
* /*
* * Formulate first tx descriptor with tx controls.
* */
* /* XXX check return value? */
* ath_hal_setuptxdesc(ah, ds
* , pktlen /* packet length */
* , hdrlen /* header length */
If anyone who has been successful in disabling L2 ACKs could take a
look at where I may have gone wrong I would be very grateful.
Thanks
Dave
These look pretty good but... you are doing the register write too early (i.e. before the ath_hal_reset). Move it to the end of the function or else you risk the HAL resetting the register back to the default (hw-generated ACK) value. At a guess that is what is happening in your patched version.
Steve
------------------------------------------------------------------------- This SF.Net email is sponsored by the Moblin Your Move Developer's challenge Build the coolest Linux based applications with Moblin SDK & win great prizes Grand prize is a trip for two to an Open Source event anywhere in the world http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________ Madwifi-devel mailing list Madwifi-devel <at> lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/madwifi-devel