On Tue, 2009-05-26 at 22:05 -0700, Jeff Hansen wrote:
> Some of the acwmin and acwmax values were backwards in the original patch,
> which cut TX performance (on my device at least) down from ~25-30Mbps to
> 2Mbps. Performance is back up after this commit.
Thanks for reporting these and verifying that this fixes the issue.
However, not all of these are typos.. In addition,
linux-wireless@xxxxxxxxxxxxxxx should be cc'ed.
> diff --git a/net/mac80211/util.c b/net/mac80211/util.c
> @@ -657,20 +657,20 @@ void ieee80211_set_wmm_default(struct ieee80211_sub_if_data *sdata)
>
> switch (queue) {
> case 3: /* AC_BK */
> - qparam.cw_max = aCWmin;
> - qparam.cw_min = aCWmax;
> + qparam.cw_max = aCWmax;
> + qparam.cw_min = aCWmin;
> qparam.txop = 0;
> qparam.aifs = 7;
> break;
> default: /* never happens but let's not leave undefined */
> case 2: /* AC_BE */
> - qparam.cw_max = aCWmin;
> - qparam.cw_min = aCWmax;
> + qparam.cw_max = aCWmax;
> + qparam.cw_min = aCWmin;
> qparam.txop = 0;
> qparam.aifs = 3;
> break;
These are correct fixes.
> case 1: /* AC_VI */
> - qparam.cw_max = aCWmin;
> + qparam.cw_max = aCWmax;
> qparam.cw_min = (aCWmin + 1) / 2 - 1;
> if (use_11b)
> qparam.txop = 6016/32;
> @@ -679,7 +679,7 @@ void ieee80211_set_wmm_default(struct ieee80211_sub_if_data *sdata)
> qparam.aifs = 2;
> break;
> case 0: /* AC_VO */
> - qparam.cw_max = (aCWmin + 1) / 2 - 1;
> + qparam.cw_max = (aCWmax + 1) / 2 - 1;
> qparam.cw_min = (aCWmin + 1) / 4 - 1;
> if (use_11b)
> qparam.txop = 3264/32;
These were actually already correct (yes, it may look like a typo, but
it is not, both of VI and VO are indeed using aCWmin in calculation for
both min and max values) and should not be changed.
- Jouni
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
