2012-03-19 13:37:23

by Ronald Wahl

Subject: [PATCH] mac80211: when receiving DTIM disable power-save mode only if it was enabled

When receiving DTIM we currently disable power save mode in the
hardware unconditionally, i.e. also when the hardware was not sleeping.
This causes trouble with at least one wireless chipset (Ralink RT3572).
When the hardware is not sleeping and we send a wakeup command (e.g.
this happens after a scan) then a significant decrease of the link
quality or a disconnect may occur.
Disabling power save mode only when it was enabled prevents this issue.

Signed-off-by: Ronald Wahl <[email protected]>
---
Index: linux/net/mac80211/mlme.c
===================================================================
--- linux/net/mac80211/mlme.c (revision 213137)
+++ linux/net/mac80211/mlme.c (working copy)
@@ -1922,9 +1922,11 @@
if (local->hw.flags & IEEE80211_HW_PS_NULLFUNC_STACK) {
if (directed_tim) {
if (local->hw.conf.dynamic_ps_timeout > 0) {
- local->hw.conf.flags &= ~IEEE80211_CONF_PS;
- ieee80211_hw_config(local,
- IEEE80211_CONF_CHANGE_PS);
+ if (local->hw.conf.flags & IEEE80211_CONF_PS) {
+ local->hw.conf.flags &= ~IEEE80211_CONF_PS;
+ ieee80211_hw_config(local,
+ IEEE80211_CONF_CHANGE_PS);
+ }
ieee80211_send_nullfunc(local, sdata, 0);
} else {
local->pspolling = true;

2012-03-19 20:20:05

by Gertjan van Wingerde

Subject: Re: [PATCH] mac80211: when receiving DTIM disable power-save mode only if it was enabled

On 19 mrt. 2012, at 14:37, Ronald Wahl <[email protected]> wrote:

> When receiving DTIM we currently disable power save mode in the
> hardware unconditionally, i.e. also when the hardware was not sleeping.
> This causes trouble with at least one wireless chipset (Ralink RT3572).
> When the hardware is not sleeping and we send a wakeup command (e.g.
> this happens after a scan) then a significant decrease of the link
> quality or a disconnect may occur.
> Disabling power save mode only when it was enabled prevents this issue.
>
> Signed-off-by: Ronald Wahl <[email protected]>

Reviewed-by: Gertjan van Wingerde <[email protected]>

Obviously rt2x00 is the only one that seems to need this, but simply replicating maintaining the PS state inside rt2x00 while mac80211 already does this doesn't seem right to me.

> ---
> Index: linux/net/mac80211/mlme.c
> ===================================================================
> --- linux/net/mac80211/mlme.c (revision 213137)
> +++ linux/net/mac80211/mlme.c (working copy)
> @@ -1922,9 +1922,11 @@
> if (local->hw.flags & IEEE80211_HW_PS_NULLFUNC_STACK) {
> if (directed_tim) {
> if (local->hw.conf.dynamic_ps_timeout > 0) {
> - local->hw.conf.flags &= ~IEEE80211_CONF_PS;
> - ieee80211_hw_config(local,
> - IEEE80211_CONF_CHANGE_PS);
> + if (local->hw.conf.flags & IEEE80211_CONF_PS) {
> + local->hw.conf.flags &= ~IEEE80211_CONF_PS;
> + ieee80211_hw_config(local,
> + IEEE80211_CONF_CHANGE_PS);
> + }
> ieee80211_send_nullfunc(local, sdata, 0);
> } else {
> local->pspolling = true;
> --
> To unsubscribe from this list: send the line "unsubscribe linux-wireless" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html