git.openwrt.org Git - openwrt/staging/neocturne.git/blob

   1 From: Peter Seiderer <ps.report@gmx.net>
   2 Date: Mon, 4 Apr 2022 18:54:14 +0200
   3 Subject: [PATCH] mac80211: minstrel_ht: fix where rate stats are stored (fixes
   4  debugfs output)
   5
   6 Using an ath9k card the debugfs output of minstrel_ht looks like the following
   7 (note the zero values for the first four rates sum-of success/attempts):
   8
   9              best    ____________rate__________    ____statistics___    _____last____    ______sum-of________
  10 mode guard #  rate   [name   idx airtime  max_tp]  [avg(tp) avg(prob)]  [retry|suc|att]  [#success | #attempts]
  11 OFDM       1    DP     6.0M  272    1640     5.2       3.1      53.8       3     0 0             0   0
  12 OFDM       1   C       9.0M  273    1104     7.7       4.6      53.8       4     0 0             0   0
  13 OFDM       1  B       12.0M  274     836    10.0       6.0      53.8       4     0 0             0   0
  14 OFDM       1 A    S   18.0M  275     568    14.3       8.5      53.8       5     0 0             0   0
  15 OFDM       1      S   24.0M  276     436    18.1       0.0       0.0       5     0 1            80   1778
  16 OFDM       1          36.0M  277     300    24.9       0.0       0.0       0     0 1             0   107
  17 OFDM       1      S   48.0M  278     236    30.4       0.0       0.0       0     0 0             0   75
  18 OFDM       1          54.0M  279     212    33.0       0.0       0.0       0     0 0             0   72
  19
  20 Total packet count::    ideal 16582      lookaround 885
  21 Average # of aggregated frames per A-MPDU: 1.0
  22
  23 Debugging showed that the rate statistics for the first four rates where
  24 stored in the MINSTREL_CCK_GROUP instead of the MINSTREL_OFDM_GROUP because
  25 in minstrel_ht_get_stats() the supported check was not honoured as done in
  26 various other places, e.g net/mac80211/rc80211_minstrel_ht_debugfs.c:
  27
  28  74                 if (!(mi->supported[i] & BIT(j)))
  29  75                         continue;
  30
  31 With the patch applied the output looks good:
  32
  33               best    ____________rate__________    ____statistics___    _____last____    ______sum-of________
  34 mode guard #  rate   [name   idx airtime  max_tp]  [avg(tp) avg(prob)]  [retry|suc|att]  [#success | #attempts]
  35 OFDM       1    D      6.0M  272    1640     5.2       5.2     100.0       3     0 0             1   1
  36 OFDM       1   C       9.0M  273    1104     7.7       7.7     100.0       4     0 0            38   38
  37 OFDM       1  B       12.0M  274     836    10.0       9.9      89.5       4     2 2           372   395
  38 OFDM       1 A   P    18.0M  275     568    14.3      14.3      97.2       5    52 53         6956   7181
  39 OFDM       1      S   24.0M  276     436    18.1       0.0       0.0       0     0 1             6   163
  40 OFDM       1          36.0M  277     300    24.9       0.0       0.0       0     0 1             0   35
  41 OFDM       1      S   48.0M  278     236    30.4       0.0       0.0       0     0 0             0   38
  42 OFDM       1      S   54.0M  279     212    33.0       0.0       0.0       0     0 0             0   38
  43
  44 Total packet count::    ideal 7097      lookaround 287
  45 Average # of aggregated frames per A-MPDU: 1.0
  46
  47 Signed-off-by: Peter Seiderer <ps.report@gmx.net>
  48 ---
  49
  50 --- a/net/mac80211/rc80211_minstrel_ht.c
  51 +++ b/net/mac80211/rc80211_minstrel_ht.c
  52 @@ -364,6 +364,9 @@ minstrel_ht_get_stats(struct minstrel_pr
  53
  54         group = MINSTREL_CCK_GROUP;
  55         for (idx = 0; idx < ARRAY_SIZE(mp->cck_rates); idx++) {
  56 +               if (!(mi->supported[group] & BIT(idx)))
  57 +                       continue;
  58 +
  59                 if (rate->idx != mp->cck_rates[idx])
  60                         continue;
  61