On Wed, 2009-11-25 at 12:35 +0530, Vivek Natarajan wrote:
> For drivers setting IEEE80211_HW_PS_NULLFUNC_STACK, a null data frame
> with PM bit off has to be sent before sending normal data frames.
> If it is done in ps_disable_work, the actual data frame would be
> queued first before this work is executed and hence null data frame
> will be queued later. And also, this null data frame has to be sent
> only after clearing CONF_PS.
> Hence, directly clearing CONF_PS and sending null data frame in
> ieee80211_xmit seems to function properly.
>
> Signed-off-by: Vivek Natarajan <vnatarajan@xxxxxxxxxxx>
> ---
> net/mac80211/tx.c | 8 ++++----
> 1 files changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c
> index 943def2..51f537c 100644
> --- a/net/mac80211/tx.c
> +++ b/net/mac80211/tx.c
> @@ -1433,10 +1433,10 @@ static void ieee80211_xmit(struct ieee80211_sub_if_data *sdata,
>
> if (need_dynamic_ps(local)) {
> if (local->hw.conf.flags & IEEE80211_CONF_PS) {
> - ieee80211_stop_queues_by_reason(&local->hw,
> - IEEE80211_QUEUE_STOP_REASON_PS);
> - ieee80211_queue_work(&local->hw,
> - &local->dynamic_ps_disable_work);
> + local->hw.conf.flags &= ~IEEE80211_CONF_PS;
> + ieee80211_hw_config(local, IEEE80211_CONF_CHANGE_PS);
> + if (local->hw.flags & IEEE80211_HW_PS_NULLFUNC_STACK)
> + ieee80211_send_nullfunc(local, sdata, 0);
Regardless of Kalle's comment I thought I'd point out that you cannot do
this anyway as _hw_config() must be able to sleep.
johannes
Attachment:
signature.asc
Description: This is a digitally signed message part
