From: Toke Høiland-Jørgensen <[email protected]>
The previous commit added the ability to throttle stations when they queue
too much airtime in the hardware. This commit enables the functionality by
calculating the expected airtime usage of each packet that is dequeued from
the TXQs in mac80211, and accounting that as pending airtime.
The estimated airtime for each skb is stored in the tx_info, so we can
subtract the same amount from the running total when the skb is freed or
recycled. The throttling mechanism relies on this accounting to be
accurate (i.e., that we are not freeing skbs without subtracting any
airtime they were accounted for), so we put the subtraction into
ieee80211_report_used_skb(). As an optimisation, we also subtract the
airtime on regular TX completion, zeroing out the value stored in the
packet afterwards, to avoid having to do an expensive lookup of the station
from the packet data on every packet.
This patch does *not* include any mechanism to wake a throttled TXQ again,
on the assumption that this will happen anyway as a side effect of whatever
freed the skb (most commonly a TX completion).
Signed-off-by: Toke Høiland-Jørgensen <[email protected]>
---
net/mac80211/status.c | 33 +++++++++++++++++++++++++++++++++
net/mac80211/tx.c | 21 +++++++++++++++++++++
2 files changed, 54 insertions(+)
diff --git a/net/mac80211/status.c b/net/mac80211/status.c
index ab8ba5835ca0..ae15c8fd2421 100644
--- a/net/mac80211/status.c
+++ b/net/mac80211/status.c
@@ -676,6 +676,28 @@ static void ieee80211_report_used_skb(struct ieee80211_local *local,
if (dropped)
acked = false;
+ if (info->tx_time_est) {
+ struct sta_info *sta = NULL, *s;
+ struct rhlist_head *tmp;
+
+ rcu_read_lock();
+
+ for_each_sta_info(local, hdr->addr1, s, tmp) {
+ /* skip wrong virtual interface */
+ if (!ether_addr_equal(hdr->addr2, s->sdata->vif.addr))
+ continue;
+
+ sta = s;
+ break;
+ }
+
+ ieee80211_sta_update_pending_airtime(local, sta,
+ skb_get_queue_mapping(skb),
+ info->tx_time_est << 2,
+ true);
+ rcu_read_unlock();
+ }
+
if (info->flags & IEEE80211_TX_INTFL_MLME_CONN_TX) {
struct ieee80211_sub_if_data *sdata;
@@ -986,6 +1008,17 @@ static void __ieee80211_tx_status(struct ieee80211_hw *hw,
ieee80211_sta_register_airtime(&sta->sta, tid,
info->status.tx_time, 0);
+ if (info->tx_time_est) {
+ /* Do this here to avoid the expensive lookup of the sta
+ * in ieee80211_report_used_skb().
+ */
+ ieee80211_sta_update_pending_airtime(local, sta,
+ skb_get_queue_mapping(skb),
+ info->tx_time_est << 2,
+ true);
+ info->tx_time_est = 0;
+ }
+
if (ieee80211_hw_check(&local->hw, REPORTS_TX_ACK_STATUS)) {
if (info->flags & IEEE80211_TX_STAT_ACK) {
if (sta->status_stats.lost_packets)
diff --git a/net/mac80211/tx.c b/net/mac80211/tx.c
index 12653d873b8c..1405304d8994 100644
--- a/net/mac80211/tx.c
+++ b/net/mac80211/tx.c
@@ -3542,6 +3542,9 @@ struct sk_buff *ieee80211_tx_dequeue(struct ieee80211_hw *hw,
WARN_ON_ONCE(softirq_count() == 0);
+ if (!ieee80211_txq_airtime_check(hw, txq))
+ return NULL;
+
begin:
spin_lock_bh(&fq->lock);
@@ -3652,6 +3655,24 @@ struct sk_buff *ieee80211_tx_dequeue(struct ieee80211_hw *hw,
}
IEEE80211_SKB_CB(skb)->control.vif = vif;
+
+ if (local->airtime_flags & AIRTIME_USE_AQL) {
+ u32 airtime;
+
+ airtime = ieee80211_calc_expected_tx_airtime(hw, vif, txq->sta,
+ skb->len);
+ if (airtime) {
+ /* We only have 10 bits in tx_time_est, so store airtime
+ * in increments of 4us and clamp the maximum to 2**12-1
+ */
+ airtime = min_t(u32, airtime, 4095) & ~3U;
+ info->tx_time_est = airtime >> 2;
+ ieee80211_sta_update_pending_airtime(local, tx.sta,
+ txq->ac, airtime,
+ false);
+ }
+ }
+
return skb;
out:
On Wed, 2019-10-23 at 11:59 +0200, Toke Høiland-Jørgensen wrote:
>
> + if (info->tx_time_est) {
> + struct sta_info *sta = NULL, *s;
> + struct rhlist_head *tmp;
> +
> + rcu_read_lock();
> +
> + for_each_sta_info(local, hdr->addr1, s, tmp) {
> + /* skip wrong virtual interface */
> + if (!ether_addr_equal(hdr->addr2, s->sdata->vif.addr))
> + continue;
> +
> + sta = s;
> + break;
> + }
I guess that is better than looking up the sdata and then using
sta_info_get(), but I think I'd like to see this wrapped into a function
(even if it's an inline) in sta_info.{c,h}.
> + airtime = ieee80211_calc_expected_tx_airtime(hw, vif, txq->sta,
> + skb->len);
> + if (airtime) {
> + /* We only have 10 bits in tx_time_est, so store airtime
> + * in increments of 4us and clamp the maximum to 2**12-1
> + */
> + airtime = min_t(u32, airtime, 4095) & ~3U;
> + info->tx_time_est = airtime >> 2;
> + ieee80211_sta_update_pending_airtime(local, tx.sta,
> + txq->ac, airtime,
> + false);
I wonder if it'd be better to pass the shifted value to
ieee80211_sta_update_pending_airtime() to avoid all the shifting in all
places?
You could even store the shifted value in "aql_tx_pending" and
"aql_total_pending_airtime" etc., it's completely equivalent, and only
shift it out for people looking at debugfs.
johannes
Johannes Berg <[email protected]> writes:
> On Wed, 2019-10-23 at 11:59 +0200, Toke Høiland-Jørgensen wrote:
>>
>> + if (info->tx_time_est) {
>> + struct sta_info *sta = NULL, *s;
>> + struct rhlist_head *tmp;
>> +
>> + rcu_read_lock();
>> +
>> + for_each_sta_info(local, hdr->addr1, s, tmp) {
>> + /* skip wrong virtual interface */
>> + if (!ether_addr_equal(hdr->addr2, s->sdata->vif.addr))
>> + continue;
>> +
>> + sta = s;
>> + break;
>> + }
>
> I guess that is better than looking up the sdata and then using
> sta_info_get(), but I think I'd like to see this wrapped into a function
> (even if it's an inline) in sta_info.{c,h}.
OK, can do.
>> + airtime = ieee80211_calc_expected_tx_airtime(hw, vif, txq->sta,
>> + skb->len);
>> + if (airtime) {
>> + /* We only have 10 bits in tx_time_est, so store airtime
>> + * in increments of 4us and clamp the maximum to 2**12-1
>> + */
>> + airtime = min_t(u32, airtime, 4095) & ~3U;
>> + info->tx_time_est = airtime >> 2;
>> + ieee80211_sta_update_pending_airtime(local, tx.sta,
>> + txq->ac, airtime,
>> + false);
>
> I wonder if it'd be better to pass the shifted value to
> ieee80211_sta_update_pending_airtime() to avoid all the shifting in all
> places?
>
> You could even store the shifted value in "aql_tx_pending" and
> "aql_total_pending_airtime" etc., it's completely equivalent, and only
> shift it out for people looking at debugfs.
My reasoning for doing it this way was that we have another set of APIs
dealing with airtime which doesn't do any shifting; so keeping the APIs
in the same unit (straight airtime) seemed less confusing.
We could add (inline) setter and getter functions for the tx_time_est
field instead to avoid sprinkling shifts all over the place? :)
-Toke
On Fri, 2019-11-08 at 12:01 +0100, Toke Høiland-Jørgensen wrote:
>
> My reasoning for doing it this way was that we have another set of APIs
> dealing with airtime which doesn't do any shifting; so keeping the APIs
> in the same unit (straight airtime) seemed less confusing.
Fair enough.
> We could add (inline) setter and getter functions for the tx_time_est
> field instead to avoid sprinkling shifts all over the place? :)
It doesn't really bother me that much, but yeah, perhaps that's easier.
johannes