Slow rates that have >95% success probability do not need to be
monitored continuously. When the channel conditions change rapidly, the
slow sampling results are useless anyway. When conditions change slowly,
they will be monitored by gradual downgrading of the actively used
rates. This patch slightly improves throughput under good conditions.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Under load, otherwise stable rates can easily fluctuate because of
collisions. In my tests on a clean channel, the success probability of
the max throughput rate often stays somewhere between 90% and 100% under
load. This can cause some unnecessary switching to lower rates.
This patch improves stability by treating success probability values
between 90% and 100% the same.
In my tests on a 3x3 HT20 link with lots of TCP traffic, it improves the
average throughput by a few mbit/s.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
While minstrel bootstraps and fills the success probabilities of each
rate the lowest rate has typically a very high success probability
(often 100% in our tests).
Its statistics are never updated but considered to setup the mrr chain.
In our tests we see that especially the 3rd mrr stage (which is that
rate providing highest success probability) is filled with the lowest rate
because its initial high sucess probability is never updated. By design
the 4th mrr stage is filled with the lowest rate so often 3rd and 4th
mrr stage are equal.
This patch follows minstrels general approach of assuming as little
as possible about rate dependencies. Consequently we include the
lowest rate into the random sampling table to get balanced up-to-date
statistics of all rates and therefore balanced decisions.
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Thomas Huehn <thomas@net.t-labs.tu-berlin.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Both rate control algorithms (minstrel and minstrel_ht) calculate
averages based on EWMA. Shift function minstrel_ewma() into
rc80211_minstrel.h and make use of it in both minstrel version.
Also shift the default EWMA level (75%) definition to the header file
and clean up variable usage.
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Thomas Huehn <thomas@net.t-labs.tu-berlin.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
The last minstrel_ht changes increased the sampling frequency for
potentially useful rates to decrease the response time to rate
fluctuations. This caused an increase in sampling frequency that can
slightly reduce throughput, so this patch limits the sampling attempts
to one per rate instead of two.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
A sample attempt should only count in mi->sample_tries if the sample
attempt wasn't skipped based on slower rate criteria.
This patch increases the sampling frequency for potentially desirable
rates and thus enables faster recovery from interference or collisions.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
If a rate is below the max_tp_rate, sample it frequently if:
- it is above max_tp_rate2, or
- it is above max_prob_rate and is a candidate for max_prob_rate
(has fewer streams than max_tp_rate).
This helps the retry chain recover more quickly from bad statistics
caused by collisions or interference, and slightly reduces throughput
fluctuations with higher rates.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Try to sample all available rates, as sample attempts do not cost much
airtime and are appropriately spaced based on the average A-MPDU length.
This helps with faster recovery on rate fluctuations.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
max_prob_rate should be selected to be very reliable, however limiting
it to single-stream on 3-stream devices is a bit much.
Allow max_prob_rate to use one stream less than the max_tp_rate.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
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>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Instead of modifying the HT SMPS capability field
for stations, track the SMPS mode explicitly in a
new field in the station struct and use it in the
drivers that care about it. This simplifies the
code using it.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
For VHT, many more bandwidth changes are possible. As a first
step, stop toggling the IEEE80211_HT_CAP_SUP_WIDTH_20_40 flag
in the HT capabilities and instead introduce a bandwidth field
indicating the currently usable bandwidth to transmit to the
station. Of course, make all drivers use it.
To achieve this, make ieee80211_ht_cap_ie_to_sta_ht_cap() get
the station as an argument, rather than the new capabilities,
so it can set up the new bandwidth field.
If the station is a VHT station and VHT bandwidth is in use,
also set the bandwidth accordingly.
Doing this allows us to get rid of the supports_40mhz flag as
the HT capabilities now reflect the true capability instead of
the current setting.
While at it, also fix ieee80211_ht_cap_ie_to_sta_ht_cap() to not
ignore HT cap overrides when MCS TX isn't supported (not that it
really happens...)
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
When MCS rates start to get bad in 2.4 GHz because of long range or
strong interference, CCK rates can be a lot more robust.
This patch adds a pseudo MCS group containing CCK rates (long preamble
in the lower 4 slots, short preamble in the upper slots).
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
[make minstrel_ht_get_stats static]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
It's more of an unnecessary micro-optimization and it prevents switching
from long-GI to short-GI in HT20/single-stream for the lowest rate
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
When sample_idx is set to a value other than -1 it activates
the IEEE80211_TX_CTL_RATE_CTRL_PROBE flag which disables
frame aggregation. To allow frame aggregation during fixed
rate it is necessary to set max_tp_rate, max_tp_rate2 and
max_prob_rate instead of sample_idx.
Signed-off-by: Sylvain Roger Rieunier <sylvain.roger.rieunier@gmail.com>
[reword commit message a bit]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
msp has type struct minstrel_ht_sta_priv not struct minstrel_ht_sta.
(This incorporates the fixup originally posted as "mac80211: fix kzalloc
memory corruption introduced in minstrel_ht". -- JWL)
Reported-by: Fengguang Wu <wfg@linux.intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Thomas Huehn <thomas@net.t-labs.tu-berlin.de>
Acked-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Removes hw.conf.channel usage from the following functions:
* ieee80211_mandatory_rates
* ieee80211_sta_get_rates
* ieee80211_frame_duration
* ieee80211_rts_duration
* ieee80211_ctstoself_duration
This is in preparation for multi-channel operation.
Signed-off-by: Michal Kazior <michal.kazior@tieto.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The channel type argument to the rate_update()
callback isn't really the correct way to give
the rate control algorithm about the desired
RX bandwidth of the peer.
Remove this argument, and instead update the
STA capabilities with 20/40 appropriately. The
SMPS update done by this callback works in the
same way, so this makes the callback cleaner.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The current max throughput rate is known to be good as otherwise it
wouldn't be the max throughput rate. Since rate sampling can introduce
some overhead (by adding RTS for example or due to not aggregating the
frame) don't sample the max throughput rate.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Disable multi stream rates (MCS > 7) when a STA is in static SMPS mode
since it has only one active rx chain. Hence, it doesn't even make
sense to sample multi stream rates.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Remove unused function parameters in the following functions:
minstrel_calc_rate_ewma()
minstrel_ht_calc_tp()
minstrel_aggr_check()
minstrel_ht_set_rate()
Signed-off-by: Patrick Kelle <patrick.kelle81@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Finding the group index for a specific rate is done by looping through
all groups and returning if the correct one is found. This code is
called for each tx'ed frame and thus it makes sense to reduce its
runtime.
Do this by calculating the group index by this formula based on the SGI
and HT40 flags as well as the stream number:
idx = (HT40 * 2 * MINSTREL_MAX_STREAMS) +
(SGI * MINSTREL_MAX_STREAMS) +
(streams - 1)
Hence, the groups are ordered by th HT40 flag first, then by the SGI
flag and afterwards by the number of used streams.
This should reduce the runtime of minstrel_ht_get_group_idx
considerable.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The drivers are not required to fill in rate->count if rate->idx is set
to -1. Hence, we should first check rate->idx before accessing
rate->count.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
I noticed a possible issue in the max_tp_rate2 management of
minstrel_ht. In particular, if we look up just among max_tp_rate2 of
each group it will be possible that the selected rate will not be the
mcs with second maximum throughput. I wrote this simple patch.
Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi83@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Minstrel HT tries very hard to establish a BA session with
each peer once there's some data on the way. However the
stack does not inform minstrel if an aggregation session
is already in place, so it keeps trying and wastes good
cycles in the tx status path.
[ 8149.946393] Open BA session requested for $AP tid 0
[ 8150.048765] Open BA session requested for $AP tid 0
[ 8150.174509] Open BA session requested for $AP tid 0
[ 8150.274376] Open BA session requested for $AP tid 0
...
Signed-off-by: Christian Lamparter <chunkeey@googlemail.com>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Found several threads about fixed rate mode in minstrel_ht for test
environments, but no patches for it.
This patch provides such a mode through debugfs.
Signed-off-by: John W. Linville <linville@tuxdriver.com>
When transmitting a frame, the transmitter waits a random number of
slots between 0 and cw. Thus, the contention time is (cw / 2) * t_slot
which we can represent instead as (cw * t_slot) >> 1. Also fix a few
other accounting bugs around contention time, and add comments.
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
When a client connects in HT mode but does not provide any valid MCS
rates, the function that finds the next sample rate gets stuck in an
infinite loop.
Fix this by falling back to legacy rates if no usable MCS rates are found.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Devices without multi rate retry support won't be able to use all rates
as specified by mintrel_ht. Hence, we can simply skip setting up further
rates as the devices will only use the first one.
Also add a special case for devices with only two possible tx rates. We
use sample_rate -> max_prob_rate for sampling and max_tp_rate ->
max_prob_rate by default.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Waiting until the status is received can cause the same rate to be
probed multiple times consecutively.
Cc: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
mac80211 does the same afterwards anyway. Hence, just drop
this redundant code.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Allow drivers or rate control algorithms to specify BlockAck session
timeout when initiating an ADDBA transaction. This is useful in cases
where maintaining persistent BA sessions does not incur any overhead.
The current timeout value of 5000 TUs is retained for all non ath9k/ath9k_htc
drivers.
Signed-off-by: Sujith Manoharan <Sujith.Manoharan@atheros.com>
Reviewed-by: Johannes Berg <johannes@sipsolutions.net>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This should help with latency issues which can happen when
using aggregation.
Cc: Felix Fietkau <nbd@openwrt.org>
Cc: Matt Smith <matt.smith@atheros.com>
Cc: Senthil Balasubramanian <senthilkumar@atheros.com>
Signed-off-by: Luis R. Rodriguez <lrodriguez@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
- reduce the number of retransmission attempts for sample rates
- sample lower rates less often
- do not use RTS/CTS for sampling frames
- increase the time between sampling attempts
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This patch fixes two problems with the minstrel_ht rate control
algorithms handling of A-MPDU frames:
1. The ampdu_len field of the tx status is not always initialized for
non-HT frames (and it would probably be unreasonable to require all
drivers to do so). This could cause rate control statistics to be
corrupted. We now trust the ampdu_len and ampdu_ack_len fields only when
the frame is marked with the IEEE80211_TX_STAT_AMPDU flag.
2. Successful transmission attempts where only recognized when the A-MPDU
subframe carrying the rate control status information was marked with the
IEEE80211_TX_STAT_ACK flag. If this information happed to be carried on a
frame that failed to be ACKed then the other subframes (which may have
succeeded) where not correctly registered. We now update rate control
statistics regardless of whether the subframe carrying the information was
ACKed or not.
Cc: <stable@kernel.org>
Signed-off-by: Björn Smedman <bjorn.smedman@venatech.se>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The intent was to free "msp->ratelist" here. "msp->sample_table" is
always NULL at this point.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
If sta is NULL, we will have problems long before we get here...
Reported-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Cc: Felix Fietkau <nbd@openwrt.org>
The check should be against current top2 rate, instead of
current top rate.
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The throughput should be considered when updating rate
with best probability.
Signed-off-by: Ming Lei <tom.leiming@gmail.com>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
net/mac80211/rc80211_minstrel_ht.c:440:46: warning: incorrect type in argument 2 (different signedness)
net/mac80211/rc80211_minstrel_ht.c:440:46: expected int *idx
net/mac80211/rc80211_minstrel_ht.c:440:46: got unsigned int *<noident>
net/mac80211/rc80211_minstrel_ht.c:446:46: warning: incorrect type in argument 2 (different signedness)
net/mac80211/rc80211_minstrel_ht.c:446:46: expected int *idx
net/mac80211/rc80211_minstrel_ht.c:446:46: got unsigned int *<noident>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Acked-by: Felix Fietkau <nbd@openwrt.org>
Currently we allocate some memory for each TX
aggregation session and additionally keep a
state bitmap indicating the state it is in.
By using RCU to protect the pointer, moving
the state into the structure and some locking
trickery we can avoid locking when the TX agg
session is fully operational.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Felix Fietkau