[openwrt/openwrt] mac80211: fix rounding error in minstrel_ht throughput calculation

LEDE Commits lede-commits at lists.infradead.org
Fri Jan 15 07:05:51 EST 2021


nbd pushed a commit to openwrt/openwrt.git, branch master:
https://git.openwrt.org/b0ad07e9a0fcdf59430547773e4ccc2416a2185a

commit b0ad07e9a0fcdf59430547773e4ccc2416a2185a
Author: Felix Fietkau <nbd at nbd.name>
AuthorDate: Fri Jan 15 13:05:25 2021 +0100

    mac80211: fix rounding error in minstrel_ht throughput calculation
    
    Fixes rate selection with lower data rates
    
    Signed-off-by: Felix Fietkau <nbd at nbd.name>
---
 ...nstrel_ht-fix-rounding-error-in-throughpu.patch | 34 ++++++++++++++++++++++
 1 file changed, 34 insertions(+)

diff --git a/package/kernel/mac80211/patches/subsys/345-mac80211-minstrel_ht-fix-rounding-error-in-throughpu.patch b/package/kernel/mac80211/patches/subsys/345-mac80211-minstrel_ht-fix-rounding-error-in-throughpu.patch
new file mode 100644
index 0000000000..1df5dec039
--- /dev/null
+++ b/package/kernel/mac80211/patches/subsys/345-mac80211-minstrel_ht-fix-rounding-error-in-throughpu.patch
@@ -0,0 +1,34 @@
+From: Felix Fietkau <nbd at nbd.name>
+Date: Fri, 15 Jan 2021 12:15:06 +0100
+Subject: [PATCH] mac80211: minstrel_ht: fix rounding error in throughput
+ calculation
+
+On lower data rates, the throughput calculation has a significant rounding
+error, causing rates like 48M and 54M OFDM to share the same throughput
+value with >= 90% success probablity.
+
+This is because the result of the division (prob_avg * 1000) / nsecs
+is really small (8 in this example).
+
+Improve accuracy by moving over some zeroes, making better use of the full
+range of u32 before the division.
+
+Signed-off-by: Felix Fietkau <nbd at nbd.name>
+---
+
+--- a/net/mac80211/rc80211_minstrel_ht.c
++++ b/net/mac80211/rc80211_minstrel_ht.c
+@@ -445,10 +445,9 @@ minstrel_ht_get_tp_avg(struct minstrel_h
+ 	 * (prob is scaled - see MINSTREL_FRAC above)
+ 	 */
+ 	if (prob_avg > MINSTREL_FRAC(90, 100))
+-		return MINSTREL_TRUNC(100000 * ((MINSTREL_FRAC(90, 100) * 1000)
+-								      / nsecs));
+-	else
+-		return MINSTREL_TRUNC(100000 * ((prob_avg * 1000) / nsecs));
++		prob_avg = MINSTREL_FRAC(90, 100);
++
++	return MINSTREL_TRUNC(100 * ((prob_avg * 1000000) / nsecs));
+ }
+ 
+ /*



More information about the lede-commits mailing list