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.)
* a flags field that lets me know what's going on;
* the hardware ratecode, unmolested by conversion to a bitrate;
* the HAL rs_flags field, useful for debugging;
* specifically mark aggregate sub-frames.
This stuff sorely needs tidying up - it's missing some important
stuff (eg numdelims) and it would be nice to put the flags at the
beginning rather than at the end.
Tested:
* AR9380, STA mode, 2x2 HT40, monitoring RSSI and EVM values
I can 100% reliably trigger this on TID 1 traffic by using iperf -S 32
<client fields> to create traffic that maps to TID 1.
The reference driver doesn't do this check.
I stumbled across this whilst trying to debug another weird hang reported
on the freebsd-wireless list.
Whilst here, add in the STBC check to ath_rateseries_setup().
Whilst here, fix the short preamble flag to be set only for legacy rates.
Whilst here, comment that we should be using the full set of decisions
made by ath_rateseries_setup() rather than recalculating them!
routine.
There were still corner cases where the EWMA update stats are being
called on a rix which didn't have an intermediary stats update; thus
no packets were counted against it. Sigh.
This should fix the crashes I've been seeing on recent -HEAD.
* If both ends have negotiated (at least) one stream;
* Only if it's a single stream rate (MCS0-7);
* Only if there's more than one TX chain enabled.
Tested:
* AR9280 STA mode -> Atheros AP; tested both MCS2 (STBC) and MCS12 (no STBC.)
Verified using athalq to inspect the TX descriptors.
TODO:
* Test AR5416 - no STBC should be enabled;
* Test AR9280 with one TX chain enabled - no STBC should be enabled.
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.
rate.
This fixes two things:
* The intermediary rates now also have their EWMA values changed;
* The existing code was using the wrong value for longtries - so the
EWMA stats were only adjusted for the first rate and not subsequent
rates in a MRR setup.
TODO:
* Merge the EWMA updates into update_stats() now..
* Remove ar5416UpdateChainmasks();
* Remove the TX chainmask override code from the ar5416 TX descriptor
setup routines;
* Write a driver method to calculate the current chainmask based on the
operating mode and update the driver state;
* Call the HAL chainmask method before calling ath_hal_reset();
* Use the currently configured chainmask in the TX descriptors rather than
the hardware TX chainmasks.
Tested:
* AR5416, STA/AP mode - legacy and 11n modes
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.
an incorrectly calculated RTS duration value when transmitting aggregates.
These earlier 802.11n NICs incorrectly used the ACK duration time when
calculating what to put in the RTS of an aggregate frame. Instead it
should have used the block-ack time. The result is that other stations
may not reserve enough time and start transmitting _over_ the top of
the in-progress blockack field. Tsk.
This workaround is to popuate the burst duration field with the delta
between the ACK duration the hardware is using and the required duration
for the block-ack. The result is that the RTS field should now contain
the correct duration for the subsequent block-ack.
This doesn't apply for AR9280 and later NICs.
Obtained from: Qualcomm Atheros
Specifically - never jack the TX FIFO threshold up to the absolute
maximum; always leave enough space for two DMA transactions to
appear.
This is a paranoia from the Linux ath9k driver. It can't hurt.
Obtained from: Linux ath9k
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.
This has reduced the number of TX delimiter and data underruns when
doing large UDP transfers (>100mbit).
This stops any HAL_INT_TXURN interrupts from occuring, which is a good
sign!
Obtained from: Qualcomm Atheros
* Delete this debugging print - I used it when debugging the initial
TX descriptor chaining code. It now works, so let's toss it.
It just confuses people if they enable TX descriptor debugging as they
get two slightly different versions of the same descriptor.
* Indenting.
part of ts_status. Thus:
* make sure we decode them from ts_flags, rather than ts_status;
* make sure we decode them regardless of whether there's an error or not.
This correctly exposes descriptor configuration errors, TX delimiter
underruns and TX data underruns.
actually do have to reinitialise the RX side of things after an RX
descriptor EOL error.
* Revert a change of mine from quite a while ago - don't shortcut the
RX initialisation path. There's a RX FIFO bug in the earlier chips
(I'm not sure when it was fixed in this series, but it's fixed
with the AR9380 and later) which causes the same RX descriptor to
be written to over and over. This causes the descriptor to be
marked as "done", and this ends up causing the whole RX path to
go very strange. This should fixed the "kickpcu; handled X packets"
message spam where "X" is consistently small.
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.
when they're being called from the TX completion handler.
Going (back) through the taskqueue is just adding extra locking and
latency to packet operations. This improves performance a little bit
on most NICs.
It still hasn't restored the original performance of the AR5416 NIC
but the AR9160, AR9280 and later NICs behave very well with this.
Tested:
* AR5416 STA (still tops out at ~ 70mbit TCP, rather than 150mbit TCP..)
* AR9160 hostap (good for both TX and RX)
* AR9280 hostap (good for both TX and RX)
crappy 802.11n performance, sigh.)
With the AR5416, aggregates need to be limited to 8KiB if RTS/CTS is
enabled. However, larger aggregates were going out with RTSCTS enabled.
The following was going on:
* The first buffer in the list would have RTS/CTS enabled in
bf->bf_state.txflags;
* The aggregate would be formed;
* The "copy over the txflags from the first buffer" logic that I added
blanked the RTS/CTS TX flags fields, and then copied the bf_first
RTS/CTS flags over;
* .. but that'd cause bf_first to be blanked out! And thus the flag
was cleared;
* So the rest of the aggregate formation would run with those flags
cleared, and thus > 8KiB aggregates were formed.
The driver is now (again) correctly limiting aggregate formation for
the AR5416 but there are still other pending issues to resolve.
Tested:
* AR5416, STA mode
Right now, ic_curchan seems to be updated rather quickly (ie, during
the ioctl) and before the driver gets notified of what's going on.
So what I was seeing was:
* NIC was in channel X;
* It generates PHY errors for channel X;
* an ioctl comes along from userland and changes things to channel Y;
* .. this updates ic_curchan, but hasn't yet reset the hardware;
* in parallel, RX is occuring and it looks at ic_curchan;
* .. which is channel Y, so events get stamped with that now.
Sigh.
the separate ath0 TX taskq.
Whilst here, make sure that the TX software scheduler is also
running out of the TX task, rather than the ath0 taskqueue.
Make sure that the tx taskqueue is blocked/unblocked as necessary.
This allows for a little more parallelism on multi-core machines,
as well as (eventually) supporting a higher task priority for TX
tasks, allowing said TX task to preempt an already running RX or
TX completion task.
Tested:
* AR5416, AR9280 hostap and STA modes
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.
chip hangs.
* Always do a reset in ath_bmiss_proc(), regardless of whether the
hardware is "hung" or not. Specifically, for spectral scan, there's
likely a whole bunch of potential hangs that we don't (yet) recognise
in the HAL. So to avoid staying RX deaf persisting until the station
disassociates, just do a no-loss reset.
* Set sc_beacons=1 in STA mode. During a reset, the beacon programming
isn't done. (It's likely I need to set sc_syncbeacons during a hang
reset, but I digress.) Thus after a reset, there's no beacon timer
programming to send a BMISS interrupt if beacons aren't heard ..
thus if the AP disappears, you won't get notified and you'll have to
reset your interface.
This hasn't yet fixed all of the hangs that I've seen when debugging
spectral scan, but it's certainly reduced the hang frequency and it
should improve general STA stability in very noisy environments.
Tested:
* AR9280, STA mode, spectral scan off/on
PR: kern/175227
when an interface is going down.
Right now it's quite possible (but very unlikely!) that ath_reset()
or similar is called, leading to a beacon config call, in parallel with
the last VAP being destroyed.
This likely should be fixed by making sure the bmiss/bstuck/watchdog
taskqueues are canceled whenever the last VAP is destroyed.
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
This is intended to support reporting FFT results during active channel
scans, for users who would like to fiddle around with writing applications
that do both FFT visualisation _and_ AP scanning.
* add a new ioctl to enable/trigger spectral scan at channel change/reset;
* set do_spectral consistently if it's enabled, so a channel set/reset
will carry forth the correct PHY error configuration so frames
are actually received;
* for NICs that don't do spectral scan, don't bother checking the
spectral scan state on channel change/reset.
Tested:
* AR9280 - STA and scanning;
* AR5416 - STA, ensured that the SS code doesn't panic
This includes the HAL routines to setup, enable/activate/disable spectral
scan and configure the relevant registers.
This still requires driver interaction to enable spectral scan reporting.
Specifically:
* call ah_spectralConfigure() to configure and enable spectral scan;
* .. there's currently no way to disable spectral scan... that will have
to follow.
* call ah_spectralStart() to force start a spectral report;
* call ah_spectralStop() to force stop an active spectral report.
The spectral scan results appear as PHY errors (type 0x5 on the AR9280,
same as radar) but with the spectral scan bit set (0x10 in the last byte
of the frame) identifying it as a spectral report rather than a radar
FFT report.
Caveats:
* It's likely quite difficult to run spectral _and_ radar at the same
time. Enabling spectral scan disables the radar thresholds but
leaves radar enabled. Thus, the driver (for now) needs to ensure
that only one or the other is enabled.
* .. it needs testing on HT40 mode.
Tested:
* AR9280 in STA mode, HT/20 only
TODO:
* Test on AR9285, AR9287;
* Test in both HT20 and HT40 modes;
* .. all the driver glue.
Obtained from: Qualcomm Atheros
* Finish adding the HAL capability to announce whether a NIC supports
spectral scan or not;
* Add spectral scan methods to the HAL structure;
* Add HAL_SPECTRAL_PARAM for configuration of the spectral scan logic.
The capability ID and HAL_SPECTRAL_PARAM struct are from Qualcomm
Atheros.
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.
enforcing the TXOP and TBTT limits:
* Frames which will overlap with TBTT will not TX;
* Frames which will exceed TXOP will be filtered.
This is not enabled by default; it's intended to be enabled by the
TDMA code on 802.11n capable chipsets.
now this works for non-debug and debug builds.
* Add a comment reminding me (or someone) to audit all of the relevant
math to ensure there's no weird wrapping issues still lurking about.
But yes, this does seem to be mostly working.
Pointy-hat-to: adrian, yet again
* add some further debugging prints, which are quite nice to have
* add in ALQ hooks (optional!) to allow for the TDMA information to be
logged in-line with the TX and RX descriptor information.
The existing logic wrapped programming nexttbtt at 65535 TU.
This is not good enough for the 11n chips, whose nexttbtt register
(GENERIC_TIMER_0) has an initial value from 0..2^31-1 TSF.
So converting the TU to TSF had the counter wrap at (65535 << 10) TSF.
Once this wrap occured, the nexttbtt value was very very low, much
lower than the current TSF value. At this point, the nexttbtt timer
would constantly fire, leading to the TX queue being constantly gated
open.. and when this occured, the sender was not correctly transmitting
in its slot but just able to continuously transmit. The master would
then delay transmitting its beacon until after the air became free
(which I guess would be after the burst interval, before the next burst
interval would quickly follow) and that big delta in master beacon TX
would start causing big swings in the slot timing adjustment.
With this change, the nexttbtt value is allowed to go all the way up
to the maximum value permissable by the 32 bit representation.
I haven't yet tested it to that point; I really should. The AR5212
HAL now filters out values above 65535 TU for the beacon configuration
(and the relevant legal values for SWBA, DBA and NEXTATIM) and the
AR5416 HAL just dutifully programs in what it should.
With this, TDMA is now useful on the 802.11n chips.
Tested:
* AR5416, AR9280 TDMA slave
* AR5413 TDMA slave
what the maximum legal values are.
The current beacon timer configuration from TDMA wraps things at
HAL_BEACON_PERIOD-1 TU. For the 11a chips this is fine, but for
the 11n chips it's not enough resolution. Since the 11a chips have a
limit on what's "valid", just enforce this so when I do write larger
values in, they get suitably wrapped before programming.
Tested:
* AR5413, TDMA slave
Todo:
* Run it for a (lot) longer on a clear channel, ensure that no strange
slippages occur.
* Re-validate this on STA configurations, just to be sure.
After chatting with the MAC team, the TSF writes (at least on the 11n
MACs, I don't know about pre-11n MACs) are done as 64 bit writes that
can take some time. So, doing a 32 bit TSF write is definitely not
supported. Leave a comment here which explains that.
Whilst here, add a comment which outlines that after a reset or TSF
write, the TSF write may take a while (up to 50uS) to update.
A write or reset shouldn't be done whilst the previous one is in
flight. Also (and this isn't currently done) a read shouldn't
occur until the SLEEP32_TSF_WRITE_STAT is clear. Right now we're
not doing that, mostly because we haven't been doing lots of TSF
resets/writes until recently.
TSF write.
The TSF_L32 update is fine for the AR5413 (and later, I guess) 11abg NICs
however on the 11n NICs this didn't work. The TSF writes were causing
a much larger time to be skipped, leading to the timing to never
converge.
I've tested this 64 bit TSF read, adjust and write on both the
11n NICs and the AR5413 NIC I've been using for testing. It works
fine on each.
This patch allows the AR5416/AR9280 to be used as a TDMA member.
I don't yet know why the AR9280 is ~7uS accurate rather than ~3uS;
I'll look into it soon.
Tested:
* AR5413, TDMA slave (~ 3us accuracy)
* AR5416, TDMA slave (~ 3us accuracy)
* AR9280, TDMA slave (~ 7us accuracy)
on the 802.11n NICs.
The 802.11n NICs return a TBTT value that continues far past the 16 bit
HAL_BEACON_PERIOD time (in TU.) The code would constrain nextslot to
HAL_BEACON_PERIOD, but it wasn't constraining nexttbtt - the pre-11n
NICs would only return TU values from 0 -> HAL_BEACON_PERIOD. Thus,
when nexttbtt exceeded 64 milliseconds, it would not wrap (but nextslot
did) which lead to a huge tsfdelta.
So until the slot calculation is converted to work in TSF rather than
a mix of TSF and TU, "make" the nexttbtt values match the TU assumptions
for pre-11n NICs.
This fixes the crazy deltatsf calculations but it doesn't fix the
non-convergent tsfdelta issue. That'll be fixed in a subsequent commit.
encryption types.
The AR5210 only has four WEP key slots, in contrast to what the
later MACs have (ie, the keycache.) So there's no way to store a "clear"
key.
Even if the driver is taught to not allocate CLR key entries for
the AR5210, the hardware will actually attempt to decode the encrypted
frames with the (likely all 0!) WEP keys.
So for now, disable the hardware encryption entirely and just so it
all in software. That allows both WEP -and- WPA to actually work.
If someone wishes to try and make hardware WEP _but_ software WPA work,
they'll have to create a HAL capability to enable/disable hardware
encryption based on the current STA/Hostap mode. However, making
multi-vap work with one WEP and one WPA VAP will require hardware
encryption to be disabled anyway.
