This adjusts FW TX block allocation for connected stations in PS.
Firmware congestion is measured in allocated packets instead of blocks.
Allow a link in PS to queue up to 2 packets to the FW.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
When selecting packets for transmission, prefer the ACs that are least
occupied in the FW. When packets for multiple ACs are present in the FW,
it decides which to transmit according to WMM QoS parameters.
With these changes, lower priority ACs should not be starved when higher
priority traffic is present.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Track the number of freed packets in each AC when receiving an interrupt
from the FW. This paves the way for tracking allocated packets per AC.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Start IBSS role when the interface type is IBSS.
As with sta role, use the dev role until the role
is started.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Using hlid=0 in AP mode is a bug. Dynamically allocate HLIDs.
Set the "first sta hlid" as 3. This will have to be changed
when multiple vifs will be supported.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
The ROC command asks the fw stay on the channel of the given
hlid. it currently has 2 primary functions:
1. Allow tx/rx from the device role.
In order to tx/rx packets while the stations is not associated
(e.g. auth req/resp), the device role has to be used, along
with ROC on its link.
Keep the logic similiar to the one used in dummy_join. However,
since we can't scan while we ROC, we add CROC before starting
a scan, and ROC again (if needed) on scan complete.
2. Keeping the antenna for a specific link.
We ROC until the connection was completed (after EAPOLs exchange)
in order to prevent BT coex operations from taking the antenna
and failing the connection (after this stage, psm can be used).
During association, we ROC on the station role, and then CROC
the device role, thus assuring being ROC during all the connection
process.
Delete the WL1271_FLAG_JOINED flag, and use a roc bitmap
to indicate what roles are currently ROCed.
Add wl12xx_roc/croc functions in order to wrap the roc/croc
commands while taking care of the roc bitmap.
The current ROC/CROC state-machine is a bit complicated. In
the future we'll probably want to use wpa_supplicant to control
the ROC during connection.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Allow handling of DUMMY_PACKET_EVENT_ID also in ap mode.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
system_hlid is a const hlid (always 0), used by the fw and driver
for packets which are not bound to specific role (e.g. dynamic
memory packets).
indicate it as always allocated.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Update the rx/tx descriptors according to the new fw api
(fw >= 6/7.3.0.0.75)
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Change the commands and events according to the new fw api
(fw >= 6/7.3.0.0.75).
The main change is the replacement of JOIN/DISCONNECT commands,
with ROLE_START/ROLE_STOP commands.
The use of these commands should be preceded by the ROLE_ENABLE
command (allocating role resources), and followed by the
ROLE_DISABLE command (freeing role resources).
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
The new fw doesn't support rx_filtering configuration (as a
stand-alone command. the rx filtering is done automatically
according to the active role).
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
In order to keep to driver compiling during the patchset,
while avoiding one-huge-patch, temporarily disable some
advanced ap functions.
These changes will be reverted later in the patchset, as
part of the patches for advanced ap functions support.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
This does not make sense in fw >= 6/7.3.0.0.75 (wl127x/wl128x) -
we don't use Tx blocks to measure FW occupancy anymore.
This reverts commit 9e374a37b6.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Split tx_queue_count to count per-AC skb's queued, instead of relying on
the skb-queue len. The skb queues used were only valid in STA-mode, as
AP-mode uses per-link queues.
This fixes a major regression in AP-mode, caused by the patch
"wl12xx: implement Tx watermarks per AC". With that patch applied, we
effectively had no regulation of Tx queues in AP-mode. Therefore a
sustained high rate of Tx could cause exhaustion of the skb memory pool.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Each AC is stopped when its queue is filled up to the high watermark,
and restarted when its queue it lower than the low watermark. This
ensures congested ACs are not able to starve other ACs.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
When selecting packets for transmission, prefer the ACs that are least
occupied in the FW. When packets for multiple ACs are present in the FW,
it decides which to transmit according to WMM QoS parameters.
With these changes, lower priority ACs should not be starved when higher
priority traffic is present.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
When operating with TKIP encryption, the function wl1271_tx_fill_hdr()
relocates the 802.11 header to the start of the frame, and leaves room
for the security header.
Some functions in the Tx path rely on the location of the header,
namely, for purposes of roaming in STA mode and connecting new stations
in AP mode. Call these functions only after the header is relocated.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Do not reset the security sequence number when issuing a join command or
interface is removed. Instead, reset the counter only during the unjoin
command.
Added the notion of counter wrap-around to the LSB number in
wl1271_tx_complete_packet.
Added post recovery padding to adjust for potential security number
progress during the recovery process by the firmware and avoid
potential interop issues in encrypted networks.
Signed-off-by: Oz Krakowski <ozk@ti.com>
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
When resuming (after wowlan), we want the rx packets (which is
usually the wake-up packet itself) to be passed to mac80211 only
after the resume notifier was completed, and mac80211 is up and
running (otherwise, the packets will be dropped).
By enqueueing the netstack_work to a freezable workqueue, we can
guarantee the rx processing to occur only after mac80211 was resumed.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
wl1271_flush_deferred_work(), which calls ieee80211_rx() and
ieee80211_tx_status(), is called from a process context.
hence, use ieee80211_tx_status_ni() instead of ieee80211_tx_status().
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
When rx_streaming.interval is non-zero, use automatic rx streaming.
Enable rx streaming on the each rx/tx packet, and disable it
rx_streaming.duration msecs later.
When rx_streaming.always=0 (default), rx streaming is enabled only
when there is a coex operation.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
When operating as AP, the TX queues are not stopped when we start
recovery. mac80211 is notified only after the fact. When there is
pending TX, it will be queued even after the FW is down. This leads to
situations where the TX queues are stopped (because of the TX-watermark
mechanism), and are never woken up when we return from recovery.
Fix this by explicitly stopping the TX queues when before initiating
recovery.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
The driver stops sending TX packets when either there aren't enough
memory blocks, or it runs out of TX descriptors. The driver continues to
send packets to the FW only when more memory blocks are available.
The FW might free TX descriptors without freeing the corresponding
memory blocks, especially when dynamic memory is enabled. In cases where
memory blocks are not freed at all, the driver will keep waiting for
more memory blocks indefinitely.
Fix this by clearing the WL1271_FLAG_FW_TX_BUSY flag when there are
available TX descriptors.
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
commit d05c806 ("wl12xx: rearrange some ELP wake_up/sleep calls")
introduced a bug in which wl1271_ps_elp_wakeup() was called instead
of wl1271_ps_elp_sleep() after completing the tx work.
Reported-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
The total number of TX memory blocks may change when the dynamic memory
option is enabled. The current implementation only tracks the available
memory blocks, which over-complicates TX blocks accounting.
By tracking the number of allocated blocks, calculation of the number of
available blocks becomes simpler and cleaner. It simply equals the total
number of TX memory blocks minus the allocated ones.
Also, remove some unnecessary castings and use union member accesses
instead.
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
The current implementation allocates a skb each time one is requested by
the firmware. Since dummy packets are handled differently than regular
packets, the skb needs to be marked. Currently, this is done by
setting the pkt_type member to 5. This might not be safe, as we cannot
be sure that there won't be any other packets with this pkt_type value.
Since the packet does not change from one request to another, we can
simply allocate a dummy packet template and always send it. All changes
to the skb done during packet preparation must be reverted, so the same
skb can be reused.
The dummy packets are not transmitted, therefore there's no need to set
the BSSID or our own MAC address.
In addition, the header portion of the packet was zeroed by mistake, so
fix that as well.
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Simplify and clean up the block size alignment code:
1. Set the block size according to the padding field type, as it cannot
exceed the maximum value this field can hold.
2. Move the alignment code into a function instead of duplicating it in
multiple places.
3. In the current implementation, the block_size member can be
misleading because a zero value actually means that there's no need to
align. Declare a block size alignment quirk instead.
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
The wl12xx device normally drops all frames coming from BSSID
it is not joined with.
This behavior is configured today by the wl12xx driver in response
to a handful of ieee80211_bss_change and ieee80211_conf_changed
notification flags, such as BSS_CHANGED_ASSOC, BSS_CHANGED_BSSID,
IEEE80211_CONF_CHANGE_IDLE, etc..
This breaks when we roam to a new BSSID, where authentication frames
are sent before any BSS_CHANGED/CONF_CHANGED flags are received.
When this happens the hardware silently drops the authentication
responses, and the roaming fails.
Ideally this aggressive filtering behavior of the device should be disabled
upon a notification from mac80211. Such notification will take place
after multi-channel support will be added: mac80211 will likely send a
remain-on-channel notification to drivers when entering sensitive
states (like authentication), otherwise the firmware might jump to
different channels (to serve a different role).
Until those notifications materialize, disable the hw BSSID filter
when authentication requests are sent, so roaming would work.
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
Reviewed-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
ELP (Extremely/Enhanced Low Power, or something like that ;)) refers to
the powerstate of the 12xx chip, in which very low power is consumed,
and no commands (from the host) can be issued until the chip is woken up.
Wakeup/sleep commands must be protected by a wl->mutex, so it's generally
a good idea to call wakeup/sleep along with the mutex lock/unlock (where
needed). However, in some places the wl12xx driver calls wakeup/sleep in
some "inner" functions. This result in some "nested" wakeup/sleep calls
which might end up letting the chip go to sleep prematurely (e.g. during
event handling).
Fix it by rearranging the elp calls to come along with mutex_lock/unlock.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
The firmware requires dummy packets to be sent using TID 7
(WL1271_TID_MGMT). Instead of hardcoding it in the tx_fill_hdr()
function, set it when creating the packet itself.
This requires Eliad's fix to set the actual TID in the TX descriptor.
Cc: Ido Yariv <ido@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
When passing a tx frame, the driver incorrectly set desc->tid
with the ac instead of the actual tid.
It has some serious implications when using 802.11n + QoS,
as the fw starts a BlockAck with the wrong tid (which finally
cause beacon loss and disconnection / some fw crash)
Fix it by using the actual tid stored in skb->priority.
Reported-by: Shahar Levi <shahar_levi@ti.com>
Signed-off-by: Eliad Peller <eliad@wizery.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
All the new firmware versions (>=6.1.3.50.58 for STA and >=6.2.0.0.47
for AP) use 1 spare TX block. We still want to support older
firmwares that require 2 spare blocks, so added a quirk to handle the
difference.
Also implemented a generic way of setting quirks that depend on the
firmware revision.
Signed-off-by: Luciano Coelho <coelho@ti.com>
Support sending dummy packet to wl128x FW as results of
dummy packet event. That is part of dynamic TX mem blocks mechanism.
Only send dummy packet when not in AP mode.
[Even though the DUMMY_PACKET_EVENT_ID and the
STA_REMOVE_COMPLETE_EVENT_ID events are defined to the same value, we
need to treat them separately in the code. Keep the check and enable
STA_REMOVE_COMPLETE_EVENT_ID for AP mode and DUMMY_PACKET_EVENT_ID for
STA. Moved one warning to a cleaner place. Use WL1271_TID_MGMT for
dummy packets -- Luca]
Signed-off-by: Shahar Levi <shahar_levi@ti.com>
Reviewed-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Reduced bus transactions in the Tx & Rx path.
[Removed unnecessary check wl->chip.id != CHIP_ID_1283_PG20 when
checking the quirk -- Luca]
Signed-off-by: Shahar Levi <shahar_levi@ti.com>
Reviewed-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
New acx command that sets: Rx fifo enable reduced bus transactions
in RX path. Tx bus transactions padding to SDIO block size that
improve preference in Tx and essential for working with SDIO HS (48Mhz).
The max SDIO block size is 256 when working with Tx bus transactions
padding to SDIO block.
Add new ops to SDIO & SPI that handles the win size change in case of
transactions padding (relevant only for SDIO).
[Fix endianess issues; simplify sdio-specific block_size handling;
minor changes in comments; use "aligned_len" in one calculation
instead of "pad" to avoid confusion -- Luca]
Signed-off-by: Shahar Levi <shahar_levi@ti.com>
Reviewed-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
To achieve maximal throughput, it is very important to react to
interrupts as soon as possible. Currently the interrupt handler wakes up
a worker for handling interrupts in process context. A cleaner and more
efficient design would be to request a threaded interrupt handler. This
handler's priority is very high, and can do blocking operations such as
SDIO/SPI transactions.
Some work can be deferred, mostly calls to mac80211 APIs
(ieee80211_rx_ni and ieee80211_tx_status). By deferring such work to a
different worker, we can keep the irq handler thread more I/O
responsive. In addition, on multi-core systems the two threads can be
scheduled on different cores, which will improve overall performance.
The use of WL1271_FLAG_IRQ_PENDING & WL1271_FLAG_IRQ_RUNNING was
changed. For simplicity, always query the FW for more pending
interrupts. Since there are relatively long bursts of interrupts, the
extra FW status read overhead is negligible. In addition, this enables
registering the IRQ handler with the ONESHOT option.
Signed-off-by: Ido Yariv <ido@wizery.com>
Reviewed-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
On newer hardware revisions, there is no need to write the host's
counter at the end of a RX transaction. The same applies to writing the
number of packets at the end of a TX transaction.
It is generally a good idea to avoid unnecessary SDIO/SPI transfers.
Throughput and CPU usage are improved when avoiding these.
Send the host's RX counter and the TX packet count only if needed, based
on the hardware revision.
[Changed WL12XX_QUIRK_END_OF_TRANSACTION to use BIT(0) -- Luca]
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Ohad Ben-Cohen <ohad@wizery.com>
Reviewed-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Frames in the tx_frames array include extra private headers, which must
be removed before passing the skbs to ieee80211_tx_status.
Fix this by removing any private headers in wl1271_tx_reset, similar to
how this is done in wl1271_tx_complete_packet.
Signed-off-by: Ido Yariv <ido@wizery.com>
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Update the PS mode of each link according to a bitmap polled from
fw_status. Manually notify mac80211 about PS mode changes in connected
stations.
mac80211 will only be notified about PS start when the station is in PS
and there is a small number of TX blocks from this link ready in HW.
This is required for waking up the remote station since the TIM is
updated entirely by FW.
When a station enters mac80211-PS-mode, we drop all the skbs in the
low-level TX queues belonging to this sta with STAT_TX_FILTERED
to keep our queues clean.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Count the number of FW TX blocks allocated per link. We add blocks to a
link counter when allocated for a TX descriptor. We remove blocks
according to counters in fw_status indicating the number of freed blocks
in FW. These counters are polled after each IRQ.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Report a TX rate idx of -1 and count 0 when returning untransmitted
skbs to mac80211 using ieee80211_tx_status(). Otherwise mac80211
tries to use the returned (essentially garbage) status.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
When operating in AP-mode we require a per link tx-queue.
This allows us to implement HW assisted PS mode for links,
as well as regulate per-link FW TX blocks consumption.
Split each link into ACs to support future QoS for AP-mode.
AC queues are emptied in priority and per-link queues are
scheduled in a simple round-robin fashion.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
If a sta starts transmitting immediately after authentication, sometimes
the FW deauthenticates it. Fix this by marking the sta "in-connection"
in FW before sending the autentication response.
The "in-connection" entry is automatically removed when connection
succeeds or after a timeout.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Use the standard ALIGN() macro instead of redefining similar macros.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Instead of looking for supported_rates change on every tx packet,
just extract the supported_rates after association completes (station only).
Remove wl1271.sta_rate_set and WL1271_FLAG_STA_RATES_CHANGED which are
not used anymore.
Signed-off-by: Eliad Peller <eliad@wizery.com>
Reviewed-by: Juuso Oikarinen <juuso.oikarinen@nokia.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Encryption key configuration is different for AP/STA modes.
AP encryption keys are recorded when the BSS is not started. On BSS
start they are propagated to the AP (in wl1271_ap_init_hwenc).
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Reviewed-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
When in AP mode set appropriate HLID and rate policy for each skb.
Respond to supported-rates related changes in op_tx only when acting
as STA.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Reviewed-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Split HW init sequence into AP/STA specific parts
The AP specific init sequence includes configuration of templates, rate
classes, power mode, etc. Also unmask AP specific events in the event mbox.
Separate the differences between AP and STA init into mode
specific functions called from wl1271_hw_init. The first is called after
radio configuration and the second after memory configuration.
Signed-off-by: Arik Nemtsov <arik@wizery.com>
Reviewed-by: Luciano Coelho <coelho@ti.com>
Signed-off-by: Luciano Coelho <coelho@ti.com>
Arik Nemtsov