used.
It turns out that the RX DMA engine does the same last-descriptor-link-
pointer-re-reading trick that the TX DMA engine. That is, the hardware
re-reads the link pointer before it moves onto the next descriptor.
Thus we can't free a descriptor before we move on; it's possible the
hardware will need to re-read the link pointer before we overwrite
it with a new one.
Tested:
* AR5416, STA mode
TODO:
* more thorough AP and STA mode testing!
* test on other pre-AR9380 NICs, just to be sure.
* Break out the RX descriptor grabbing bits from the RX completion
bits, like what is done in the RX EDMA code, so ..
* .. the RX lock can be held during ath_rx_proc(), but not across
packet input.
The hardware can generate its own frames (eg RTS/CTS exchanges, other
kinds of 802.11 management stuff, especially when it comes to 802.11n)
and these also have PWRMGT flags. So if the VAP is asleep but the
NIC is in force-awake for some reason, ensure that the self-generated
frames have PWRMGT set to 1.
Now, this (like basically everything to do with powersave) is still
racy - the only way to guarantee that it's all actually consistent
is to pause transmit and let it finish before transitioning the VAP
to sleep, but this at least gets the basic method of tracking and
updating the state debugged.
Tested:
* AR5416, STA mode
* AR9380, STA mode
fixes and beacon programming / debugging into the ath(4) driver.
The basic power save tracking:
* Add some new code to track the current desired powersave state; and
* Add some reference count tracking so we know when the NIC is awake; then
* Add code in all the points where we're about to touch the hardware and
push it to force-wake.
Then, how things are moved into power save:
* Only move into network-sleep during a RUN->SLEEP transition;
* Force wake the hardware up everywhere that we're about to touch
the hardware.
The net80211 stack takes care of doing RUN<->SLEEP<->(other) state
transitions so we don't have to do it in the driver.
Next, when to wake things up:
* In short - everywhere we touch the hardware.
* The hardware will take care of staying awake if things are queued
in the transmit queue(s); it'll then transit down to sleep if
there's nothing left. This way we don't have to track the
software / hardware transmit queue(s) and keep the hardware
awake for those.
Then, some transmit path fixes that aren't related but useful:
* Force EAPOL frames to go out at the lowest rate. This improves
reliability during the encryption handshake after 802.11
negotiation.
Next, some reset path fixes!
* Fix the overlap between reset and transmit pause so we don't
transmit frames during a reset.
* Some noisy environments will end up taking a lot longer to reset
than normal, so extend the reset period and drop the raise the
reset interval to be more realistic and give the hardware some
time to finish calibration.
* Skip calibration during the reset path. Tsk!
Then, beacon fixes in station mode!
* Add a _lot_ more debugging in the station beacon reset path.
This is all quite fluid right now.
* Modify the STA beacon programming code to try and take
the TU gap between desired TSF and the target TU into
account. (Lifted from QCA.)
Tested:
* AR5210
* AR5211
* AR5212
* AR5413
* AR5416
* AR9280
* AR9285
TODO:
* More AP, IBSS, mesh, TDMA testing
* Thorough AR9380 and later testing!
* AR9160 and AR9287 testing
Obtained from: QCA
The AR9485 chip and AR933x SoC both implement LNA diversity.
There are a few extra things that need to happen before this can be
flipped on for those chips (mostly to do with setting up the different
bias values and LNA1/LNA2 RSSI differences) but the first stage is
putting this code into the driver layer so it can be reused.
This has the added benefit of making it easier to expose configuration
options and diagnostic information via the ioctl API. That's not yet
being done but it sure would be nice to do so.
Tested:
* AR9285, with LNA diversity enabled
* AR9285, with LNA diversity disabled in EEPROM
the RX antenna field.
The AR9285/AR9485 use an LNA mixer to determine how to combine the signals
from the two antennas. This is encoded in the RSSI fields (ctl/ext) for
chain 2. So, let's use that here.
This maps RX antennas 0->3 to the RX mixer configuration used to
receive a frame. There's more that can be done but this is good enough
to diagnose if the hardware is doing "odd" things like trying to
receive frames on LNA2 (ie, antenna 2 or "alt" antenna) when there's
only one antenna connected.
Tested:
* AR9285, STA mode
and if queue mechanism; also fix up (non-11n) TX fragment handling.
This may result in a bit of a performance drop for now but I plan on
debugging and resolving this at a later stage.
Whilst here, fix the transmit path so fragment transmission works.
The TX fragmentation handling is a bit more special. In order to
correctly transmit TX fragments, there's a bunch of corner cases that
need to be handled:
* They must be transmitted back to back, in the same order..
* .. ie, you need to hold the TX lock whilst transmitting this
set of fragments rather than interleaving it with other MSDUs
destined to other nodes;
* The length of the next fragment is required when transmitting, in
order to correctly set the NAV field in the current frame to the
length of the next frame; which requires ..
* .. that we know the transmit duration of the next frame, which ..
* .. requires us to set the rate of all fragments to the same length,
or make the decision up-front, etc.
To facilitate this, I've added a new ath_buf field to describe the
length of the next fragment. This avoids having to keep the mbuf
chain together. This used to work before my 11n TX path work because
the ath_tx_start() routine would be handed a single mbuf with m_nextpkt
pointing to the next frame, and that would be maintained all the way
up to when the duration calculation was done. This doesn't hold
true any longer - the actual queuing may occur at any point in the
future (think ath_node TID software queuing) so this information
needs to be maintained.
Right now this does work for non-11n frames but it doesn't at all
enforce the same rate control decision for all frames in the fragment.
I plan on fixing this in a followup commit.
RTS/CTS has the same issue, I'll look at fixing this in a subsequent
commit.
Finaly, 11n fragment support requires the driver to have fully
decided what the rate scenario setup is - including 20/40MHz,
short/long GI, STBC, LDPC, number of streams, etc. Right now that
decision is (currently) made _after_ the NAV field value is updated.
I'll fix all of this in subsequent commits.
Tested:
* AR5416, STA, transmitting 11abg fragments
* AR5416, STA, 11n fragments work but the NAV field is incorrect for
the reasons above.
TODO:
* It would be nice to be able to queue mbufs per-node and per-TID so
we can only queue ath_buf entries when it's time to assemble frames
to send to the hardware.
But honestly, we should just do that level of software queue management
in net80211 rather than ath(4), so I'm going to leave this alone for now.
* More thorough AP, mesh and adhoc testing.
* Ensure that net80211 doesn't hand us fragmented frames when A-MPDU has
been negotiated, as we can't do software retransmission of fragments.
* .. set CLRDMASK when transmitting fragments, just to ensure.
traffic.
When transmitting non-aggregate traffic, we need to keep the hardware
busy whilst transmitting or small bursts in txdone/tx latency will
kill us.
This restores non-aggregate iperf performance, especially when doing
TDMA.
Tested:
* AR5416<->AR5416, TDMA
* AR5416 STA <-> AR9280 AP
of course.)
There's a few things that needed to happen:
* In case someone decides to set the beacon transmission rate to be
at an MCS rate, use the MCS-aware version of the duration calculation
to figure out how long the received beacon frame was.
* If TxOP enforcing is available on the hardware and we're doing TDMA,
enable it after a reset and set the TDMA guard interval to zero.
This seems to behave fine.
TODO:
* Although I haven't yet seen packet loss, the PHY errors that would be
triggered (specifically Transmit-Override-Receive) aren't enabled
by the 11n HAL. I'll have to do some work to enable these PHY errors
for debugging.
What broke:
* My recent changes to the TX queue handling has resulted in the driver
not keeping the hardware queue properly filled when doing non-aggregate
traffic. I have a patch to commit soon which fixes this situation
(albeit by reminding me about how my ath driver locking isn't working
out, sigh.)
So if you want to test this without updating to the next set of patches
that I commit, just bump the sysctl dev.ath.X.hwq_limit from 2 to 32.
Tested:
* AR5416 <-> AR5416, with ampdu disabled, HT40, 5GHz, MCS12+Short-GI.
I saw 30mbit/sec in both directions using a bidirectional UDP test.
The list-based DMA engine has the following behaviour:
* When the DMA engine is in the init state, you can write the first
descriptor address to the QCU TxDP register and it will work.
* Then when it hits the end of the list (ie, it either hits a NULL
link pointer, OR it hits a descriptor with VEOL set) the QCU
stops, and the TxDP points to the last descriptor that was transmitted.
* Then when you want to transmit a new frame, you can then either:
+ write the head of the new list into TxDP, or
+ you write the head of the new list into the link pointer of the
last completed descriptor (ie, where TxDP points), then kick
TxE to restart transmission on that QCU>
* The hardware then will re-read the descriptor to pick up the link
pointer and then jump to that.
Now, the quirks:
* If you write a TxDP when there's been no previous TxDP (ie, it's 0),
it works.
* If you write a TxDP in any other instance, the TxDP write may actually
fail. Thus, when you start transmission, it will re-read the last
transmitted descriptor to get the link pointer, NOT just start a new
transmission.
So the correct thing to do here is:
* ALWAYS use the holding descriptor (ie, the last transmitted descriptor
that we've kept safe) and use the link pointer in _THAT_ to transmit
the next frame.
* NEVER write to the TxDP after you've done the initial write.
* .. also, don't do this whilst you're also resetting the NIC.
With this in mind, the following patch does basically the above.
* Since this encapsulates Sam's issues with the QCU behaviour w/ TDMA,
kill the TDMA special case and replace it with the above.
* Add a new TXQ flag - PUTRUNNING - which indicates that we've started
DMA.
* Clear that flag when DMA has been shutdown.
* Ensure that we're not restarting DMA with PUTRUNNING enabled.
* Fix the link pointer logic during TXQ drain - we should always ensure
the link pointer does point to something if there's a list of frames.
Having it be NULL as an indication that DMA has finished or during
a reset causes trouble.
Now, given all of this, i want to nuke axq_link from orbit. There's now HAL
methods to get and set the link pointer of a descriptor, so what we
should do instead is to update the right link pointer.
* If there's a holding descriptor and an empty TXQ list, set the
link pointer of said holding descriptor to the new frame.
* If there's a non-empty TXQ list, set the link pointer of the
last descriptor in the list to the new frame.
* Nuke axq_link from orbit.
Note:
* The AR9380 doesn't need this. FIFO TX writes are atomic. As long as
we don't append to a list of frames that we've already passed to the
hardware, all of the above doesn't apply. The holding descriptor stuff
is still needed to ensure the hardware can re-read a completed
descriptor to move onto the next one, but we restart DMA by pushing in
a new FIFO entry into the TX QCU. That doesn't require any real
gymnastics.
Tested:
* AR5210, AR5211, AR5212, AR5416, AR9380 - STA mode.
PS-POLL support.
This implements PS-POLL awareness i nthe
* Implement frame "leaking", which allows for a software queue
to be scheduled even though it's asleep
* Track whether a frame has been leaked or not
* Leak out a single non-AMPDU frame when transmitting aggregates
* Queue BAR frames if the node is asleep
* Direct-dispatch the rest of control and management frames.
This allows for things like re-association to occur (which involves
sending probe req/resp as well as assoc request/response) when
the node is asleep and then tries reassociating.
* Limit how many frames can set in the software node queue whilst
the node is asleep. net80211 is already buffering frames for us
so this is mostly just paranoia.
* Add a PS-POLL method which leaks out a frame if there's something
in the software queue, else it calls net80211's ps-poll routine.
Since the ath PS-POLL routine marks the node as having a single frame
to leak, either a software queued frame would leak, OR the next queued
frame would leak. The next queued frame could be something from the
net80211 power save queue, OR it could be a NULL frame from net80211.
TODO:
* Don't transmit further BAR frames (eg via a timeout) if the node is
currently asleep. Otherwise we may end up exhausting management frames
due to the lots of queued BAR frames.
I may just undo this bit later on and direct-dispatch BAR frames
even if the node is asleep.
* It would be nice to burst out a single A-MPDU frame if both ends
support this. I may end adding a FreeBSD IE soon to negotiate
this power save behaviour.
* I should make STAs timeout of power save mode if they've been in power
save for more than a handful of seconds. This way cards that get
"stuck" in power save mode don't stay there for the "inactivity" timeout
in net80211.
* Move the queue depth check into the driver layer (ath_start / ath_transmit)
rather than doing it in the TX path.
* There could be some naughty corner cases with ps-poll leaking.
Specifically, if net80211 generates a NULL data frame whilst another
transmitter sends a normal data frame out net80211 output / transmit,
we need to ensure that the NULL data frame goes out first.
This is one of those things that should occur inside the VAP/ic TX lock.
Grr, more investigations to do..
Tested:
* STA: AR5416, AR9280
* AP: AR5416, AR9280, AR9160
of "right".)
Flip back on the "always continue TX DMA using the holding descriptor"
code - by always setting ATH_BUF_BUSY and never setting axq_link to NULL.
Since the holding descriptor is accessed via txq->axq_link and _that_
is done behind the TXQ lock rather than the TX path lock, the holding
descriptor stuff itself needs to be behind the TXQ lock.
So, do the mental gymnastics needed to do this.
I've not seen any of the hardware failures that I was seeing when
I last tried to do this.
Tested:
* AR5416, STA mode
but partly to just tidy up things.
The problem here - there are too many TX buffers in the queue! By the
time one needs to transmit an EAPOL frame (for this PR, it's the response
to the group rekey notification from the AP) there are no ath_buf entries
free and the EAPOL frame doesn't go out.
Now, the problem!
* Enforcing the TX buffer limitation _before_ we dequeue the frame?
Bad idea. Because..
* .. it means I can't check whether the mbuf has M_EAPOL set.
The solution(s):
* De-queue the frame first
* Don't bother doing the TX buffer minimum free check until after
we know whether it's an EAPOL frame or not.
* If it's an EAPOL frame, allocate the buffer from the mgmt pool
rather than the default pool.
Whilst I'm here:
* Add a tweak to limit how many buffers a single node can acquire.
* Don't enforce that for EAPOL frames.
* .. set that to default to 1/4 of the available buffers, or 32,
whichever is more sane.
This doesn't fix issues due to a sleeping node or a very poor performing
node; but this doesn't make it worse.
Tested:
* AR5416 STA, TX'ing 100+ mbit UDP to an AP, but only 50mbit being received
(thus the TX queue fills up.)
* .. with CCMP / WPA2 encryption configured
* .. and the group rekey time set to 10 seconds, just to elicit the
behaviour very quickly.
PR: kern/138379
Each set of frames pushed into a FIFO is represented by a list of
ath_bufs - the first ath_buf in the FIFO list is marked with
ATH_BUF_FIFOPTR; the last ath_buf in the FIFO list is marked with
ATH_BUF_FIFOEND.
Multiple lists of frames are just glued together in the TAILQ as per
normal - except that at the end of a FIFO list, the descriptor link
pointer will be NULL and it'll be tagged with ATH_BUF_FIFOEND.
For non-EDMA chipsets this is a no-op - the ath_txq frame list (axq_q)
stays the same and is treated the same.
For EDMA chipsets the frames are pushed into axq_q and then when
the FIFO is to be (re) filled, frames will be moved onto the FIFO
queue and then pushed into the FIFO.
So:
* Add a new queue in each hardware TXQ (ath_txq) for staging FIFO frame
lists. It's a TAILQ (like the normal hardware frame queue) rather than
the ath9k list-of-lists to represent FIFO entries.
* Add new ath_buf flags - ATH_TX_FIFOPTR and ATH_TX_FIFOEND.
* When allocating ath_buf entries, clear out the flag value before
returning it or it'll end up having stale flags.
* When cloning ath_buf entries, only clone ATH_BUF_MGMT. Don't clone
the FIFO related flags.
* Extend ath_tx_draintxq() to first drain the FIFO staging queue, _then_
drain the normal hardware queue.
Tested:
* AR9280, hostap
* AR9280, STA
* AR9380/AR9580 - hostap
TODO:
* Test on other chipsets, just to be thorough.
related issues.
Moving the TX locking under one lock made things easier to progress on
but it had one important side-effect - it increased the latency when
handling CABQ setup when sending beacons.
This commit introduces a bunch of new changes and a few unrelated changs
that are just easier to lump in here.
The aim is to have the CABQ locking separate from other locking.
The CABQ transmit path in the beacon process thus doesn't have to grab
the general TX lock, reducing lock contention/latency and making it
more likely that we'll make the beacon TX timing.
The second half of this commit is the CABQ related setup changes needed
for sane looking EDMA CABQ support. Right now the EDMA TX code naively
assumes that only one frame (MPDU or A-MPDU) is being pushed into each
FIFO slot. For the CABQ this isn't true - a whole list of frames is
being pushed in - and thus CABQ handling breaks very quickly.
The aim here is to setup the CABQ list and then push _that list_ to
the hardware for transmission. I can then extend the EDMA TX code
to stamp that list as being "one" FIFO entry (likely by tagging the
last buffer in that list as "FIFO END") so the EDMA TX completion code
correctly tracks things.
Major:
* Migrate the per-TXQ add/removal locking back to per-TXQ, rather than
a single lock.
* Leave the software queue side of things under the ATH_TX_LOCK lock,
(continuing) to serialise things as they are.
* Add a new function which is called whenever there's a beacon miss,
to print out some debugging. This is primarily designed to help
me figure out if the beacon miss events are due to a noisy environment,
issues with the PHY/MAC, or other.
* Move the CABQ setup/enable to occur _after_ all the VAPs have been
looked at. This means that for multiple VAPS in bursted mode, the
CABQ gets primed once all VAPs are checked, rather than being primed
on the first VAP and then having frames appended after this.
Minor:
* Add a (disabled) twiddle to let me enable/disable cabq traffic.
It's primarily there to let me easily debug what's going on with beacon
and CABQ setup/traffic; there's some DMA engine hangs which I'm finally
trying to trace down.
* Clear bf_next when flushing frames; it should quieten some warnings
that show up when a node goes away.
Tested:
* AR9280, STA/hostap, up to 4 vaps (staggered)
* AR5416, STA/hostap, up to 4 vaps (staggered)
TODO:
* (Lots) more AR9380 and later testing, as I may have missed something here.
* Leverage this to fix CABQ hanling for AR9380 and later chips.
* Force bursted beaconing on the chips that default to staggered beacons and
ensure the CABQ stuff is all sane (eg, the MORE bits that aren't being
correctly set when chaining descriptors.)
"complete RX frames."
The 128 entry RX FIFO is really easy to fill up and miss refilling
when it's done in the ath taskq - as that gets blocked up doing
RX completion, TX completion and other random things.
So the 128 entry RX FIFO now gets emptied and refilled in the ath_intr()
task (and it grabs / releases locks, so now ath_intr() can't just be
a FAST handler yet!) but the locks aren't held for very long. The
completion part is done in the ath taskqueue context.
Details:
* Create a new completed frame list - sc->sc_rx_rxlist;
* Split the EDMA RX process queue into two halves - one that
processes the RX FIFO and refills it with new frames; another
that completes the completed frame list;
* When tearing down the driver, flush whatever is in the deferred
queue as well as what's in the FIFO;
* Create two new RX methods - one that processes all RX queues,
one that processes the given RX queue. When MSI is implemented,
we get told which RX queue the interrupt came in on so we can
specifically schedule that. (And I can do that with the non-MSI
path too; I'll figure that out later.)
* Convert the legacy code over to use these new RX methods;
* Replace all the instances of the RX taskqueue enqueue with a call
to a relevant RX method to enqueue one or all RX queues.
Tested:
* AR9380, STA
* AR9580, STA
* AR5413, STA
Since this is being done during buffer free, it's a crap shoot whether
the TX path lock is held or not. I tried putting the ath_freebuf() code
inside the TX lock and I got all kinds of locking issues - it turns out
that the buffer free path sometimes is called with the lock held and
sometimes isn't. So I'll go and fix that soon.
Hence for now the holdingbf buffers are protected by the TXBUF lock.
When working on TDMA, Sam Leffler found that the MAC DMA hardware
would re-read the last TX descriptor when getting ready to transmit
the next one. Thus the whole ATH_BUF_BUSY came into existance -
the descriptor must be left alone (very specifically the link pointer
must be maintained) until the hardware has moved onto the next frame.
He saw this in TDMA because the MAC would be frequently stopping during
active transmit (ie, when it wasn't its turn to transmit.)
Fast-forward to today. It turns out that this is a problem not with
a single MAC DMA instance, but with each QCU (from 0->9). They each
maintain separate descriptor pointers and will re-read the last
descriptor when starting to transmit the next.
So when your AP is busy transmitting from multiple TX queues, you'll
(more) frequently see one QCU stopped, waiting for a higher-priority QCU
to finsh transmitting, before it'll go ahead and continue. If you mess
up the descriptor (ie by freeing it) then you're short of luck.
Thanks to rpaulo for sticking with me whilst I diagnosed this issue
that he was quite reliably triggering in his environment.
This is a reimplementation; it doesn't have anything in common with
the ath9k or the Qualcomm Atheros reference driver.
Now - it in theory doesn't apply on the EDMA chips, as long as you
push one complete frame into the FIFO at a time. But the MAC can DMA
from a list of frames pushed into the hardware queue (ie, you concat
'n' frames together with link pointers, and then push the head pointer
into the TXQ FIFO.) Since that's likely how I'm going to implement
CABQ handling in hostap mode, it's likely that I will end up teaching
the EDMA TX completion code about busy buffers, just to be "sure"
this doesn't creep up.
Tested - iperf ap->sta and sta->ap (with both sides running this code):
* AR5416 STA
* AR9160/AR9220 hostap
To validate that it doesn't break the EDMA (FIFO) chips:
* AR9380, AR9485, AR9462 STA
Using iperf with the -S <tos byte decimal value> to set the TCP client
side DSCP bits, mapping to different TIDs and thus different TX queues.
TODO:
* Make this work on the EDMA chips, if we end up pushing lists of frames
to the hardware (eg how we eventually will handle cabq in hostap/ibss
mode.)
The HAL already included the STBC fields; it just needed to be exposed
to the driver and net80211 stack.
This should allow single-stream STBC TX and RX to be negotiated; however
the driver and rate control code currently don't do anything with it.
Right now the only way to set the chainmask is to set the hardware
configured chainmask through capabilities. This is fine for forcing
the chainmask to be something other than what the hardware is capable
of (eg to reduce TX/RX to one connected antenna) but it does change what
the HAL hardware chainmask configuration is.
For operational mode changes, it (may?) make sense to separately control
the TX/RX chainmask.
Right now it's done as part of ar5416_reset.c - ar5416UpdateChainMasks()
calculates which TX/RX chainmasks to enable based on the operating mode.
(1 for legacy and whatever is supported for 11n operation.) But doing
this in the HAL is suboptimal - the driver needs to know the currently
configured chainmask in order to correctly enable things for each
TX descriptor. This is currently done by overriding the chainmask
config in the ar5416 TX routines but this has to disappear - the AR9300
HAL support requires the driver to dynamically set the TX chainmask based
on the TX power and TX rate in order to meet mini-PCIe slot power
requirements.
So:
* Introduce a new HAL method to set the operational chainmask variables;
* Introduce null methods for the previous generation chipsets;
* Add new driver state to record the current chainmask separate from
the hardware configured chainmask.
Part #2 of this will involve disabling ar5416UpdateChainMasks() and moving
it into the driver; as well as properly programming the TX chainmask
based on the currently configured HAL chainmask.
Tested:
* AR5416, STA mode - both legacy (11a/11bg) and 11n rates - verified
that AR_SELFGEN_MASK (the chainmask used for self-generated frames like
ACKs and RTSes) is correct, as well as the TX descriptor contents is
correct.
The default is to limit them to what the hardware is capable of.
Add sysctl twiddles for both the non-RTS and RTS protected aggregate
generation.
Whilst here, add some comments about stuff that I've discovered during
my exploration of the TX aggregate / delimiter setup path from the
reference driver.
My changed had some rather significant behavioural changes to throughput.
The two issues I noticed:
* With if_start and the ifnet mbuf queue, any temporary latency
would get eaten up by some mbufs being queued. With ath_transmit()
queuing things to ath_buf's, I'd only get 512 TX buffers before I
couldn't queue any further frames.
* There's also some non-zero latency involved with TX being pushed
into a taskqueue via direct dispatch. Any time the scheduler didn't
immediately schedule the ath TX task would cause extra latency.
Various 1ge/10ge drivers implement both direct dispatch (if the TX
lock can be acquired) and deferred task transmission (if the TX lock
can't be acquired), with frames being pushed into a drbd queue.
I'll have to do this at some point, but until I figure out how to
deal with 802.11 fragments, I'll have to wait a while longer.
So what I saw:
* lots of extra latency, specially under load - if the taskqueue
wasn't immediately scheduled, things went pear shaped;
* any extra latency would result in TX ath_buf's taking their sweet time
being replenished, so any further calls to ath_transmit() would drop
mbufs.
* .. yes, there's no explicit backpressure here - things are just dropped.
Eek.
With this, the general performance has gone up, but those subtle if_start()
related race conditions are back. For some reason, this is doubly-obvious
with the AR5416 NIC and I don't quite understand why yet.
There's an unrelated issue with AR5416 performance in STA mode (it's
fine in AP mode when bridging frames, weirdly..) that requires a little
further investigation. Specifically - it works fine on a Lenovo T40
(single core CPU) running a March 2012 9-STABLE kernel, but a Lenovo T60
(dual core) running an early November 2012 kernel behaves very poorly.
The same hardware with an AR9160 or AR9280 behaves perfectly.
This is easily possible now that the TX is protected by a single
lock, rather than a per-TXQ (and thus per-TID) lock.
Only set CLRDMASK if none of the destinations are filtered.
This likely will need some tuning when it comes time to do UASPD/PS-POLL
TX, however at that point it should be manually set anyway.
Tested:
* AR9280, STA mode
TODO:
* More thorough testing in AP mode
* test other chipsets, just to be safe/sure.
if_start().
This removes the overlapping data path TX from occuring, which
solves quite a number of the potential TX queue races in ath(4).
It doesn't fix the net80211 layer TX queue races and it doesn't
fix the raw TX path yet, but it's an important step towards this.
This hasn't dropped the TX performance in my testing; primarily
because now the TX path can quickly queue frames and continue
along processing.
This involves a few rather deep changes:
* Use the ath_buf as a queue placeholder for now, as we need to be
able to support queuing a list of mbufs (ie, when transmitting
fragments) and m_nextpkt can't be used here (because it's what is
joining the fragments together)
* if_transmit() now simply allocates the ath_buf and queues it to
a driver TX staging queue.
* TX is now moved into a taskqueue function.
* The TX taskqueue function now dequeues and transmits frames.
* Fragments are handled correctly here - as the current API passes
the fragment list as one mbuf list (joined with m_nextpkt) through
to the driver if_transmit().
* For the couple of places where ath_start() may be called (mostly
from net80211 when starting the VAP up again), just reimplement
it using the new enqueue and taskqueue methods.
What I don't like (about this work and the TX code in general):
* I'm using the same lock for the staging TX queue management and the
actual TX. This isn't required; I'm just being slack.
* I haven't yet moved TX to a separate taskqueue (but the taskqueue is
created); it's easy enough to do this later if necessary. I just need
to make sure it's a higher priority queue, so TX has the same
behaviour as it used to (where it would preempt existing RX..)
* I need to re-review the TX path a little more and make sure that
ieee80211_node_*() functions aren't called within the TX lock.
When queueing, I should just push failed frames into a queue and
when I'm wrapping up the TX code, unlock the TX lock and
call ieee80211_node_free() on each.
* It would be nice if I could hold the TX lock for the entire
TX and TX completion, rather than this release/re-acquire behaviour.
But that requires that I shuffle around the TX completion code
to handle actual ath_buf free and net80211 callback/free outside
of the TX lock. That's one of my next projects.
* the ic_raw_xmit() path doesn't use this yet - so it still has
sequencing problems with parallel, overlapping calls to the
data path. I'll fix this later.
Tested:
* Hostap - AR9280, AR9220
* STA - AR5212, AR9280, AR5416
I couldn't think of a way to maintain the hardware TXQ locks _and_ layer
on top of that per-TXQ software queuing and any other kind of fine-grained
locks (eg per-TID, or per-node locks.)
So for now, to facilitate some further code refactoring and development
as part of the final push to get software queue ps-poll and u-apsd handling
into this driver, just do away with them entirely.
I may eventually bring them back at some point, when it looks slightly more
architectually cleaner to do so. But as it stands at the present, it's
not really buying us much:
* in order to properly serialise things and not get bitten by scheduling
and locking interactions with things higher up in the stack, we need to
wrap the whole TX path in a long held lock. Otherwise we can end up
being pre-empted during frame handling, resulting in some out of order
frame handling between sequence number allocation and encryption handling
(ie, the seqno and the CCMP IV get out of sequence);
* .. so whilst that's the case, holding the lock for that long means that
we're acquiring and releasing the TXQ lock _inside_ that context;
* And we also acquire it per-frame during frame completion, but we currently
can't hold the lock for the duration of the TX completion as we need
to call net80211 layer things with the locks _unheld_ to avoid LOR.
* .. the other places were grab that lock are reset/flush, which don't happen
often.
My eventual aim is to change the TX path so all rejected frame transmissions
and all frame completions result in any ieee80211_free_node() calls to occur
outside of the TX lock; then I can cut back on the amount of locking that
goes on here.
There may be some LORs that occur when ieee80211_free_node() is called when
the TX queue path fails; I'll begin to address these in follow-up commits.
events.
This is primarily for the TX EDMA and TX EDMA completion. I haven't yet
tied it into the EDMA RX path or the legacy TX/RX path.
Things that I don't quite like:
* Make the pointer type 'void' in ath_softc and have if_ath_alq*()
return a malloc'ed buffer. That would remove the need to include
if_ath_alq.h in if_athvar.h.
* The sysctl setup needs to be cleaned up.
I've tried serialising TX using queues and such but unfortunately
due to how this interacts with the locking going on elsewhere in the
networking stack, the TX task gets delayed, resulting in quite a
noticable throughput loss:
* baseline TCP for 2x2 11n HT40 is ~ 170mbit/sec;
* TCP for TX task in the ath taskq, with the RX also going on - 80mbit/sec;
* TCP for TX task in a separate, second taskq - 100mbit/sec.
So for now I'm going with the Linux wireless stack approach - lock tx
early. The linux code does in the wireless stack, before the 802.11
state stuff happens and before it's punted to the driver.
But TX locking needs to also occur at the driver layer as the TX
completion code _also_ begins to drain the ifnet TX queue.
Whilst I'm here, add some KTR traces for the TX path.
Note:
* This really should be done at the net80211 layer (as well, at least.)
But that'll have to wait for a little more thought to happen.
the power save queue.
* introduce some new ATH_NODE lock protected fields, tracking the
net80211 psq and TIM state;
* when doing buffer transitions - ie, when sending and completing
buffers - check the state of the SWQ and update the TIM appropriately.
* when clearing the TIM bit, if the SWQ is not empty then delay clearing
it.
This is racy, but it's no less racy than the current net80211 power
save queue management code. Specifically, with multiple TX threads,
it's quite plausible that parallel state updates will race and the
TIM will be left in an inconsistent state. I'll address that in
a follow-up commit.
stashed away in ath_node.
As much as I tried to stuff that behind the ATH_NODE lock, unfortunately
the locking is just too plain hairy (for me! And I wrote it!) to do
cleanly. Hence using atomics here instead of a lock. The ATH_NODE lock
just isn't currently used anywhere besides the rate control updates.
If in the future everything gets migrated back to using a single ATH_NODE
lock or a single global ATH_TX lock (ie, a single TX lock for all TX and
TX completion) then fine, I'll remove the atomics.
it run out of multiple concurrent contexts.
Right now the ath(4) TX processing is a bit hairy. Specifically:
* It was running out of ath_start(), which could occur from multiple
concurrent sending processes (as if_start() can be started from multiple
sending threads nowdays.. sigh)
* during RX if fast frames are enabled (so not really at the moment, not
until I fix this particular feature again..)
* during ath_reset() - so anything which calls that
* during ath_tx_proc*() in the ath taskqueue - ie, TX is attempted again
after TX completion, as there's now hopefully some ath_bufs available.
* Then, the ic_raw_xmit() method can queue raw frames for transmission
at any time, from any net80211 TX context. Ew.
This has caused packet ordering issues in the past - specifically,
there's absolutely no guarantee that preemption won't occuring _during_
ath_start() by the TX completion processing, which will call ath_start()
again. It's a mess - 802.11 really, really wants things to be in
sequence or things go all kinds of loopy.
So:
* create a new task struct for TX'ing;
* make the if_start method simply queue the task on the ath taskqueue;
* make ath_start() just be called by the new TX task;
* make ath_tx_kick() just schedule the ath TX task, rather than directly
calling ath_start().
Now yes, this means that I've taken a step backwards in terms of
concurrency - TX -and- RX now occur in the same single-task taskqueue.
But there's nothing stopping me from separating out the TX / TX completion
code into a separate taskqueue which runs in parallel with the RX path,
if that ends up being appropriate for some platforms.
This fixes the CCMP/seqno concurrency issues that creep up when you
transmit large amounts of uni-directional UDP traffic (>200MBit) on a
FreeBSD STA -> AP, as now there's only one TX context no matter what's
going on (TX completion->retry/software queue,
userland->net80211->ath_start(), TX completion -> ath_start());
but it won't fix any concurrency issues between raw transmitted frames
and non-raw transmitted frames (eg EAPOL frames on TID 16 and any other
TID 16 multicast traffic that gets put on the CABQ.) That is going to
require a bunch more re-architecture before it's feasible to fix.
In any case, this is a big step towards making the majority of the TX
path locking irrelevant, as now almost all TX activity occurs in the
taskqueue.
Phew.
the ATH_TXQ_* macros.
* Introduce the new macros;
* rename the TID queue and TID filtered frame queue so the compiler
tells me I'm using the wrong macro.
These should correspond 1:1 to the existing code.
net80211 node power save state.
* Add an ATH_NODE_UNLOCK_ASSERT() check
* Add a new node field - an_is_powersave
* Pause/unpause the queue based on the node state
* Attempt to handle net80211 concurrency issues so the queue
doesn't get paused/unpaused more than once at a time from
the net80211 power save code.
Whilst here (and breaking my usual rule), set CLRDMASK when a queue
is unpaused, regardless of whether the queue has some pending traffic.
This means the first frame from that TID (now or later) will hvae
CLRDMASK set.
Also whilst here, bump the swretrymax counters whenever the
filtered frames code expires a frame. Again, breaking my rule, but
this is just a statistics thing rather than a functional change.
This doesn't fix ps-poll (but it doesn't break it too much worse
than it is at the present) or correcting the TID updates.
That's next on the list.
Tested:
* AR9220 AP (Atheros AP96 reference design)
* Macbook Pro and LG Optimus 1 Android phone, both setting
and clearing power save state (but not using PS-POLL.)
This should eventually be unified with ATH_DEBUG() so I can get both
from one macro; that may take some time.
Add some new probes for TX and TX completion.
These are intended for software TX filtering support, where the NIC
decides there has been too many successive failues to a destination
and will filter it.
Although the filtering is done per-destination (via the keycache),
the state and queue is kept per-TID for now. It simplifies the overall
architecture design and locking.
Whilst here, add ATH_TID_UNLOCK_ASSERT().
