From: Felix Fietkau <nbd@openwrt.org>
To: linux-wireless@vger.kernel.org
Cc: johannes@sipsolutions.net
Subject: [PATCH 1/3] minstrel_ht: improve accuracy of throughput metric at high data rates
Date: Sat, 2 Mar 2013 21:20:12 +0100 [thread overview]
Message-ID: <1362255614-30859-1-git-send-email-nbd@openwrt.org> (raw)
At high data rates the average frame transmission durations are small
enough for rounding errors to matter, sometimes causing minstrel to use
slightly lower transmit rates than necessary.
To fix this, change the unit of the duration value to nanoseconds
instead of microseconds, and reorder the multiplications/divisions when
calculating the throughput metric so that they don't overflow or
truncate prematurely.
At 2-stream HT40 this makes TCP throughput a bit more stable.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
---
net/mac80211/rc80211_minstrel_ht.c | 23 +++++++++++++----------
1 file changed, 13 insertions(+), 10 deletions(-)
diff --git a/net/mac80211/rc80211_minstrel_ht.c b/net/mac80211/rc80211_minstrel_ht.c
index 3af141c..0b5cdd9 100644
--- a/net/mac80211/rc80211_minstrel_ht.c
+++ b/net/mac80211/rc80211_minstrel_ht.c
@@ -26,11 +26,11 @@
/* Number of symbols for a packet with (bps) bits per symbol */
#define MCS_NSYMS(bps) ((MCS_NBITS + (bps) - 1) / (bps))
-/* Transmission time for a packet containing (syms) symbols */
+/* Transmission time (nanoseconds) for a packet containing (syms) symbols */
#define MCS_SYMBOL_TIME(sgi, syms) \
(sgi ? \
- ((syms) * 18 + 4) / 5 : /* syms * 3.6 us */ \
- (syms) << 2 /* syms * 4 us */ \
+ ((syms) * 18000 + 4000) / 5 : /* syms * 3.6 us */ \
+ ((syms) * 1000) << 2 /* syms * 4 us */ \
)
/* Transmit duration for the raw data part of an average sized packet */
@@ -64,9 +64,9 @@
}
#define CCK_DURATION(_bitrate, _short, _len) \
- (10 /* SIFS */ + \
+ (1000 * (10 /* SIFS */ + \
(_short ? 72 + 24 : 144 + 48 ) + \
- (8 * (_len + 4) * 10) / (_bitrate))
+ (8 * (_len + 4) * 10) / (_bitrate)))
#define CCK_ACK_DURATION(_bitrate, _short) \
(CCK_DURATION((_bitrate > 10 ? 20 : 10), false, 60) + \
@@ -211,7 +211,8 @@ static void
minstrel_ht_calc_tp(struct minstrel_ht_sta *mi, int group, int rate)
{
struct minstrel_rate_stats *mr;
- unsigned int usecs = 0;
+ unsigned int nsecs = 0;
+ unsigned int tp;
mr = &mi->groups[group].rates[rate];
@@ -221,10 +222,12 @@ minstrel_ht_calc_tp(struct minstrel_ht_sta *mi, int group, int rate)
}
if (group != MINSTREL_CCK_GROUP)
- usecs = mi->overhead / MINSTREL_TRUNC(mi->avg_ampdu_len);
+ nsecs = 1000 * mi->overhead / MINSTREL_TRUNC(mi->avg_ampdu_len);
- usecs += minstrel_mcs_groups[group].duration[rate];
- mr->cur_tp = MINSTREL_TRUNC((1000000 / usecs) * mr->probability);
+ nsecs += minstrel_mcs_groups[group].duration[rate];
+ tp = 1000000 * ((mr->probability * 1000) / nsecs);
+
+ mr->cur_tp = MINSTREL_TRUNC(tp);
}
/*
@@ -536,7 +539,7 @@ minstrel_calc_retransmit(struct minstrel_priv *mp, struct minstrel_ht_sta *mi,
mr->retry_updated = true;
group = &minstrel_mcs_groups[index / MCS_GROUP_RATES];
- tx_time_data = group->duration[index % MCS_GROUP_RATES] * ampdu_len;
+ tx_time_data = group->duration[index % MCS_GROUP_RATES] * ampdu_len / 1000;
/* Contention time for first 2 tries */
ctime = (t_slot * cw) >> 1;
--
1.8.0.2
next reply other threads:[~2013-03-02 20:20 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-03-02 20:20 Felix Fietkau [this message]
2013-03-02 20:20 ` [PATCH 2/3] mac80211/minstrel_ht: improve max_prob_rate selection Felix Fietkau
2013-03-02 20:20 ` [PATCH 3/3] minstrel_ht: increase sampling frequency Felix Fietkau
2013-03-04 15:06 ` [PATCH 1/3] minstrel_ht: improve accuracy of throughput metric at high data rates Johannes Berg
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=1362255614-30859-1-git-send-email-nbd@openwrt.org \
--to=nbd@openwrt.org \
--cc=johannes@sipsolutions.net \
--cc=linux-wireless@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).