Since the driver split, there's no need for no_agg_framecnt_info since
all devices have this set to false. Secondly, the compressed block ack
handling code was broken. Fix this.
(1) A shift less than zero simply implies that the buffer wrapped, this
is expected. Remove the incorrect comment.
(2) The (agg->frame_count > (64-sh)) condition can happen if the last
frame is dropped. E.g., if I send 7 frames and the 6th is received but
the 7th is lost, the other side may only shift the window 6, not 7
frames since the last bit is a 0. This is perfectly fine behavior and
doesn't invalidate the feedback.
(3) Store the feedback from a Compressed BA in the first newly received
frame, rather than the start of the window. This way it will get
processed by the rate selection code. Feedback stored in a non-received
frame is likely to get overwritten by the retransmission.
This is based on the approach taken by minstrel_ht.
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
When filling out its rate scale table, iwlwifi repeats the first HT rate
IWL_HT_NUMBER_TRY times. The hardware scheduler will stop using
aggregation for any frame that fails LINK_QUAL_AGG_DISABLE_START_DEF
times. Currently, both these constants equal 3.
If iwlwifi probes a faster rate than the link supports, all frames in a
(potentially tens of frames large) batch will fail IWL_HT_NUMBER_TRY
times. Because this happens to be as large as
LINK_QUAL_AGG_DISABLE_START_DEF, all frames will then be sent
individually. This leads to a short, but performance-degrading window
where the legacy stop-and-wait MAC takes over.
Bounding the initial rate by (LINK_QUAL_AGG_DISABLE_START_DEF-1)
attempts makes the third try use a lower rate and hence more be likely
to succeed. This somewhat mitigates the above described behavior.
Signed-off-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
iwlagn's hardware scheduler needs to be set up
with the right aggregation frame limit and
buffer sizes. To achieve this, we need to move
the hardware queue setup to when the session
becomes operational.
Tested-by: Daniel Halperin <dhalperi@cs.washington.edu>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Since the driver split, there's no need for
function pointers any more for aggregation
queue setup and teardown as all devices now
share the same code. Simplify this.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Some clients seems to keep track of their reorder window even after an
aggregation session has been disabled. This causes issues if there are
still retried but not completed frames pending for the TID.
To ensure that rx does not stall in such situations, set sendbar to 1
for any frame purged from the TID queue on teardown.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
ath9k calls ath9k_hw_stoptxdma every time it sends a beacon, however there
is not much point in doing that if the previous beacon and mcast traffic
went out properly. On AR9380, calling that function too often can result
in an increase of stuck beacons due to differences in the handling of the
queue enable/disable functionality.
With this patch, the queue will only be explicitly stopped if the previous
data frames were not sent successfully. With the beacon code being the
only remaining user of ath9k_hw_stoptxdma, this function can be simplified
in order to remove the now pointless attempts at waiting for transmission
completion, which would never happen at this point due to the different
method of tx scheduling of the beacon queue.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This patch simplifies the flush op and reuses ath_drain_all_txq for
flushing out pending frames if necessary. It also uses a global timeout
of 200ms instead of the per-queue 60ms timeout.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
In some situations, stopping Tx DMA frequently fails, leading to messages
like this:
ath: Failed to stop TX DMA in 100 msec after killing last frame
ath: Failed to stop TX DMA!
This patch uses a few MAC features to abort DMA globally instead of iterating
over all hardware queues and attempting to stop them individually.
Not only is that faster and works with a shorter timeout, it also makes the
process much more reliable.
With this change, I can no longer trigger these messages on AR9380,
and on AR9280 they become much more rare.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Add support to iwlagn for off-channel TX. The
microcode API for this is a bit strange in that
it uses a hacked-up scan command, so the scan
code needs to change quite a bit to accomodate
that and be able to send it out.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
defs.meshie.val.mesh_id is 32 chars long. It's not supposed to be NUL
terminated. This code puts a terminator on the end to make it easier to
print to sysfs. The problem is that if the mesh_id fills the entire
buffer the original code puts the terminator one spot past the end.
The way the original code was written, there was a check to make sure
that maxlen was less than PAGE_SIZE. Since we know that maxlen is at
most 34 chars, I just removed the check.
Signed-off-by: Dan Carpenter <error27@gmail.com>
Acked-by: Dan Williams <dcbw@redhat.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Increasing the wait count makes the nf load pass in
most of the cases.
Signed-off-by: Vivek Natarajan <vnatarajan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Increase the delay to make sure the initialization of pll
passes.
Signed-off-by: Vivek Natarajan <vnatarajan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Set some GPIO pins to Pull-down mode to save power.
Signed-off-by: Vivek Natarajan <vnatarajan@atheros.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The hardware rx filter flag triggered by FIF_PROMISC_IN_BSS is overly broad
and covers even frames with PHY errors. When this flag is enabled, this message
shows up frequently during scanning or hardware resets:
ath: Could not stop RX, we could be confusing the DMA engine when we start RX up
Since promiscuous mode is usually not particularly useful, yet enabled by
default by bridging (either used normally in 4-addr mode, or with hacks
for various virtualization software), we should sacrifice it for better
reliability during normal operation.
This patch leaves it enabled if there are active monitor mode interfaces, since
it's very useful for debugging.
Signed-off-by: Felix Fietkau <nbd@openwrt.org>
Cc: stable@kernel.org
Signed-off-by: John W. Linville <linville@tuxdriver.com>
While working on PS I've noticed elp_work is kicking rather often, and
sometimes the chip is put to sleep before 5ms delay expires. This
seems to happen because by the time wl1251_ps_elp_wakeup is called
elp_work might still be pending. After wakeup is done, the processing
may take some time, during which 5ms might expire and elp_work might
get scheduled. In this case, ss soon as 1st thread finishes work and
releases the mutex, elp_work will then put the device to sleep without
5ms delay. In addition 1st thread will queue additional elp_work
needlessly.
Fix this by cancelling work in wl1251_ps_elp_wakeup instead.
Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
Acked-by: Kalle Valo <kvalo@adurom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
wl1251_ps_set_elp() only does acx_sleep_auth call and takes the chip
from/to ELP, however all callers of wl1251_ps_set_mode() have already
taken the chip out of ELP and puts it back to ELP when they finish.
This makes ELP calls (and register writes they result in) superfluous.
So remove wl1251_ps_set_elp function and call acx_sleep_auth directly.
Signed-off-by: Grazvydas Ignotas <notasas@gmail.com>
Acked-by: Kalle Valo <kvalo@adurom.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Use kcalloc or kzalloc rather than the combination of kmalloc and memset.
Thanks coccicheck for detecting this.
(http://coccinelle.lip6.fr/)
Signed-off-by: Shan Wei <shanwei@cn.fujitsu.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This updates the p54spi Kconfig help text.
The driver works well on n8x0, so remove the words "experimental" and "untested".
Signed-off-by: Michael Buesch <mb@bu3sch.de>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
It seems ath5k has issues receiving broadcast packets (ARPs) when
using multiple STA interfaces associated with multiple APs.
This patch ensures the NIC is always in PROMISC mode if there
are more than 1 stations associated.
Signed-off-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Usually H/W generate statistics notify once per about 100ms, but
sometimes we can receive notify in shorter time, even 2 ms.
This can be problem for plcp health and ack health checking.
I.e. with 2 plcp errors happens randomly in 2 ms duration, we
exceed plcp delta threshold equal to 100 (2*100/2).
Also checking ack's in short time, can results not necessary false
positive and firmware reset, for example when channel is noised and
we do not receive ACKs frames or when remote device does not send
ACKs at the moment.
Patch change code to do statistic check and possible recovery only
if 99ms elapsed from last check. Forced delay should assure we have
good statistic data to estimate hardware state.
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Make iwl_good_plcp_health code easiest to read.
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Put generic rx_handlers (except iwlagn_rx_reply_compressed_ba) to
iwl-rx.c . Make functions static and change prefix from iwlagn_ to
iwl_ . Beautify iwl_setup_rx_handlers and do some other minor coding
style changes.
Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com>
Acked-by: Wey-Yi Guy <wey-yi.w.guy@intel.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
No use printing addresses of pointers, just print the
pointers themselves.
Signed-off-by: Ben Greear <greearb@candelatech.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Now that all accesses to the data_queue structures is done via the specialized
rt2x00queue_get_tx_queue function or via direct accesses, there is no
need for the rt2x00queue_get_queue function anymore, so remove it.
Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
In the spirit of optimizing the code to get the queue structure of TX queues,
also optimize the code to get beacon queues. We can simply use the bcn queue
field of the rt2x00_dev structure instead of using the rt2x00queue_get_queue
function.
Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The ATIM queue is considered to be a TX queue by the drivers that support
the queue. Therefore include support for the ATIM queue to the
rt2x00queue_get_tx_queue function so that the drivers that support the ATIM
queue can also use that function.
Add the support in such a way that drivers that do not support the ATIM
queue are not penalized in their efficiency.
Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Current code for the atim queue is strange, as it is considered in the
rt2x00_dev structure as a second beacon queue.
Normalize this by letting the atim queue have its own struct data_queue
pointer in the rt2x00_dev structure.
Signed-off-by: Gertjan van Wingerde <gwingerde@gmail.com>
Acked-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
We don't use interrupt threads anymore. Fix the comment.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
The PCI device irqmask is locked by a spin_lock. Currently
spin_lock_irqsave is used everywhere. To reduce the locking overhead
replace spin_lock_irqsave in hard irq context with spin_lock and in
soft irq context with spin_lock_irq.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
When setting up multiple BSSIDs in AP mode on an rt2800pci device we
previously used the STAs AID to select an appropriate key slot. But
since the AID is per VIF we can end up with two STAs having the same AID
and thus using the same key index. This resulted in one STA overwriting
the key information of another STA.
Fix this by simply searching for the next unused entry in the pairwise
key table.
Also bring the key table init in sync with deleting keys by initializing
the key table entries to 0 instead of 1.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This makes the code less error-prone.
Acked-by: Gertjan van Wingerde <gwingerde@gmail.com>
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
ieee80211_get_tx_rate is not valid for HT rates. Hence, restructure the
TX desciptor creation to be aware of MCS rates. The generic TX desciptor
creation now cares about the rate_mode (CCK, OFDM, MCS, GF).
As a result, ieee80211_get_tx_rate gets only called for legacy rates.
Acked-by: Gertjan van Wingerde <gwingerde@gmail.com>
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
"ifs" is only used by no-HT devices. Move it into the plcp substruct and
fill in the value only for no-HT devices.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Acked-by: Gertjan van Wingerde <gwingerde@gmail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Some fields only need to be u8 and for ifs and txop we can use the
already available enums.
Acked-by: Gertjan van Wingerde <gwingerde@gmail.com>
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
HT and no-HT rt2x00 devices use a partly different TX descriptor.
Optimize the tx desciptor memory layout by putting the PLCP and HT
substructs into a union and introduce a new driver flag to decide which
TX desciptor format is used by the device.
This saves us the expensive PLCP calculation fOr HT devices and the HT
descriptor setup on no-HT devices.
Acked-by: Gertjan van Wingerde <gwingerde@gmail.com>
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Newer devices like rt2800* own a hardware sequence counter and thus
don't need to use a software sequence counter at all. Add a new driver
flag to shortcut the software sequence number generation on devices that
don't need it.
rt61pci, rt73usb and rt2800* seem to make use of a hw sequence counter
while rt2400pci and rt2500* need to do it in software.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Acked-by: Gertjan van Wingerde <gwingerde@gmail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
rt2x00queue_write_tx_frame is unlikely to fail. Tell the compiler.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
This special case shouldn't happen very often. Only if a frame that
is not intended to be aggregated ends up in an AMPDU _and_ was intended
to be sent at a different MCS rate as the aggregate. Hence, using
unlikely is justified.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Since tx_info->control.vif was already accessed before it cant't be NULL
here.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
These conditions are unlikely to happen, tell the compiler.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
In some cases (tx path for example) we don't need to check for non-tx
queues in rt2x00queue_get_queue. Hence, introduce a new method
rt2x00queue_get_tx_queue that is only valid for tx queues and use it in
places where only tx queues are valid.
Furthermore, this new method is quite short and as such can be inlined
to avoid the function call overhead.
This only converts the txdone functions of drivers that don't use an ATIM
queue and the generic tx path.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com>
Signed-off-by: Ivo van Doorn <IvDoorn@gmail.com>
Signed-off-by: John W. Linville <linville@tuxdriver.com>
Daniel Halperin