]> asedeno.scripts.mit.edu Git - linux.git/commitdiff
mac80211: minstrel_ht: reduce unnecessary rate probing attempts
authorFelix Fietkau <nbd@nbd.name>
Tue, 20 Aug 2019 09:54:47 +0000 (11:54 +0200)
committerJohannes Berg <johannes.berg@intel.com>
Wed, 21 Aug 2019 09:09:20 +0000 (11:09 +0200)
On hardware with static fallback tables (e.g. mt76x2), rate probing attempts
can be very expensive.
On such devices, avoid sampling rates slower than the per-group max throughput
rate, based on the assumption that the fallback table will take care of probing
lower rates within that group if the higher rates fail.

To further reduce unnecessary probing attempts, skip duplicate attempts on
rates slower than the max throughput rate.

Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20190820095449.45255-2-nbd@nbd.name
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
net/mac80211/rc80211_minstrel_ht.c

index ba230b037257160fa0d7663d13be37404965e679..ad5da9a71da0a7440c56b9200a8adf5c6942baa6 100644 (file)
@@ -1059,6 +1059,21 @@ minstrel_get_sample_rate(struct minstrel_priv *mp, struct minstrel_ht_sta *mi)
            minstrel_get_duration(mi->max_prob_rate) * 3 < sample_dur)
                return -1;
 
+
+       /*
+        * For devices with no configurable multi-rate retry, skip sampling
+        * below the per-group max throughput rate, and only use one sampling
+        * attempt per rate
+        */
+       if (mp->hw->max_rates == 1 &&
+           (minstrel_get_duration(mg->max_group_tp_rate[0]) < sample_dur ||
+            mrs->attempts))
+               return -1;
+
+       /* Skip already sampled slow rates */
+       if (sample_dur >= minstrel_get_duration(tp_rate1) && mrs->attempts)
+               return -1;
+
        /*
         * Make sure that lower rates get sampled only occasionally,
         * if the link is working perfectly.