This flag is needed for the PSE client reset. Fixes watchdog reset issues.
Fixes: c677dda16523 ("wifi: mt76: mt7603: improve watchdog reset reliablity")
Signed-off-by: Felix Fietkau <[email protected]>
Use the correct WMM AC queue instead of the MGMT one to fix potential issues
with aggregation sequence number tracking. Drop non-bufferable packets.
Fixes: fca9615f1a43 ("mt76: mt7603: fix up hardware queue index for PS filtered packets")
Signed-off-by: Felix Fietkau <[email protected]>
When pre-calibration data is missing, do not fail the driver probe.
Instead, just print a warning and fall back to regular calibration.
Signed-off-by: Felix Fietkau <[email protected]>
When queueing packets, only the MT76_RESET flag of the primary PHY
is checked. If the primary PHY is scanning or changing channels, this can
lead to packet loss for tx on the second PHY.
Fix this by passing the phy to the .tx_queue_skb op and using it to check
the correct flag.
Signed-off-by: Felix Fietkau <[email protected]>
Copy cck and ofdm separately in order to avoid __read_overflow2_field
warning.
Fixes: f75e4779d215 ("wifi: mt76: mt7996: add txpower setting support")
Signed-off-by: Felix Fietkau <[email protected]>
Do not clear the interrupt mask register on MT7915, since that prevents
MCU_CMD interrupts from being reported, thus causing timeouts during the
reset sequence. Defer stopping WED until tx/rx processing activity has
stopped.
Signed-off-by: Felix Fietkau <[email protected]>
If WED rx is enabled, rx buffers are added to a buffer pool that can be
filled from multiple page pools. Because buffers freed from rx poll are
not guaranteed to belong to the processed queue's page pool, lockless
caching must not be used in this case.
Cc: [email protected]
Fixes: 2f5c3c77fc9b ("wifi: mt76: switch to page_pool allocator")
Signed-off-by: Felix Fietkau <[email protected]>
On MT76x0, LNA gain should be applied for both external and internal LNA.
On MT76x2, LNA gain should be treated as 0 for external LNA.
Move the LNA type based logic to mt76x2 in order to fix mt76x0.
Fixes: 2daa67588f34 ("mt76x0: unify lna_gain parsing")
Reported-by: Shiji Yang <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
The error handling code was added in order to allow tx enqueue to fail after
calling .tx_prepare_skb. Since this can no longer happen, the error handling
code is unused.
Signed-off-by: Felix Fietkau <[email protected]>
When drv_tx calls race against local tx scheduling, the queue fill status checks
can potentially race, leading to dma queue entries being overwritten.
Fix this by deferring packets from drv_tx calls to the tx worker, in order to
ensure that all regular queue tx comes from the same context.
Signed-off-by: Felix Fietkau <[email protected]>
Enable receiving other-unicast packets
Disable RX header translation in order to proprerly receive data packets
Signed-off-by: Felix Fietkau <[email protected]>
Before preparing the new beacon, check the queue status, flush out all
previous beacons and buffered multicast packets, then (if necessary)
try to recover more gracefully from a stuck beacon condition by making a
less invasive attempt at getting the MAC un-stuck.
Fixes: c8846e101502 ("mt76: add driver for MT7603E and MT7628/7688")
Signed-off-by: Felix Fietkau <[email protected]>
Only trigger PSE reset if PSE was stuck, otherwise it can cause DMA issues.
Trigger the PSE reset while DMA is fully stopped in order to improve
reliabilty.
Fixes: c8846e101502 ("mt76: add driver for MT7603E and MT7628/7688")
Signed-off-by: Felix Fietkau <[email protected]>
It turns out that the code in mt7603_rx_pse_busy() does not detect actual
hardware hangs, it only checks for busy conditions in PSE.
A reset should only be performed if these conditions are true and if there
is no rx activity as well.
Reset the counter whenever a rx interrupt occurs. In order to also deal with
a fully loaded CPU that leaves interrupts disabled with continuous NAPI
polling, also check for pending rx interrupts in the function itself.
Fixes: c8846e101502 ("mt76: add driver for MT7603E and MT7628/7688")
Signed-off-by: Felix Fietkau <[email protected]>
Setting MT_TX_ABORT does not abort any transmission for a wtbl index on its
own. Instead, it modifies the behavior of a queue flush to make it selectively
flush packets for a particular wtbl index.
Adjust powersave filtering to make use of this in order to avoid running into
unnecessary timeouts while flushing
Signed-off-by: Felix Fietkau <[email protected]>
When disabling beacons on a vif, intval is 0. Ensure that dev->mt76.beacon_int
is not overwritten in this case, so that beacons continue to work for other
interfaces.
Signed-off-by: Felix Fietkau <[email protected]>
The IEEE80211_VHT_CAP_EXT_NSS_BW value already indicates support for half-NSS
160 MHz support, so it is wrong to also advertise full 160 MHz support.
Fixes: c2f73eacee3b ("wifi: mt76: mt7915: add back 160MHz channel width support for MT7915")
Signed-off-by: Felix Fietkau <[email protected]>
Capabilities in vif->bss_conf are only initialized in AP mode.
For other modes, they should be enabled by default, in order to avoid a
mismatch.
Fixes: 885f7af7e544 ("wifi: mt76: mt7915: remove mt7915_mcu_beacon_check_caps()")
Signed-off-by: Felix Fietkau <[email protected]>
According to the documentation, ieee80211_rx_list must not run concurrently
with ieee80211_tx_status (or its variants).
Cc: [email protected]
Fixes: 88046b2c9f6d ("mt76: add support for reporting tx status with skb")
Reported-by: Brian Coverstone <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
This helps avoid cleartext leakage of already queued or powersave buffered
packets, when a reassoc triggers the key deletion.
Cc: [email protected]
Signed-off-by: Felix Fietkau <[email protected]>
A number of users reported that this support was working fine before
it got removed. Add it back, but leave out the unsupported 80+80 mode.
Fixes: ac922bd60ace ("wifi: mt76: mt7915: remove BW160 and BW80+80 support")
Signed-off-by: Felix Fietkau <[email protected]>
When adding WED support, mt76_dma_add_buf was accidentally changed to set
the skip_buf0 flag for tx buffers on the wrong queue descriptor entry.
Additionally, there is a rxwi leak when rx buffer allocation fails.
Fix this and make the code more readable by adding a separate function for
adding rx buffers.
Reported-by: Mikhail Gavrilov <[email protected]>
Tested-by: Mikhail Gavrilov <[email protected]>
Fixes: cd372b8c99c5 ("wifi: mt76: add WED RX support to mt76_dma_{add,get}_buf")
Signed-off-by: Felix Fietkau <[email protected]>
When mt7921_abort_roc is called with dev->mutex held, it can deadlock while
calling cancel_work_sync(&phy->roc_work), because the work function could
be waiting to acquire the mutex.
Fix this by flushing the work before taking the mutex
Reported-by: Mikhail Gavrilov <[email protected]>
Fixes: 034ae28b56f1 ("wifi: mt76: mt7921: introduce remain_on_channel support")
Fixes: 41ac53c899bd ("wifi: mt76: mt7921: introduce chanctx support")
Signed-off-by: Felix Fietkau <[email protected]>
mt7663 efuse has 0x600 bytes instead of 0x400. Increase the size in order
to fix issues with incomplete calibration data
Signed-off-by: Felix Fietkau <[email protected]>
Checking the relevant rxd bits for the checksum information only indicates
if the checksum verification was performed by the hardware and doesn't show
actual checksum errors. Checksum errors are indicated in the info field of
the DMA descriptor. Fix packets erroneously marked as CHECKSUM_UNNECESSARY
by checking the extra bits as well.
Those bits are only passed to the driver for MMIO devices at the moment, so
limit checksum offload to those.
Fixes: 2122dfbfd0bd ("mt76: mt7615: add rx checksum offload support")
Fixes: 94244d2ea503 ("mt76: mt7915: add rx checksum offload support")
Fixes: 0e75732764e8 ("mt76: mt7921: enable rx csum offload")
Signed-off-by: Felix Fietkau <[email protected]>
When 802.3 decap offload is enabled, the hardware indicates header translation
failure, whenever either the LLC-SNAP header was not found, or a VLAN header
with an unregcognized tag is present.
In that case, the hardware inserts a 2-byte length fields after the MAC
addresses. For VLAN packets, this tag needs to be removed. However,
for 802.3 LLC packets, the length bytes should be preserved, since there
is no separate ethertype field in the data.
This fixes an issue where the length field was omitted for LLC frames, causing
them to be malformed after hardware decap.
Fixes: 1eeff0b4c1a6 ("mt76: mt7915: fix decap offload corner case with 4-addr VLAN frames")
Reported-by: Chad Monroe <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
- remove stbc, because mt76x0 doesn't support it and mt76x2 uses the same
values as HT/VHT
- reduce vht array to 2 elements, because VHT MCS0-7 are the same as HT
Signed-off-by: Felix Fietkau <[email protected]>
Some users have reported being unable to connect to MT76x0 APs running mt76
after a commit enabling the VHT extneded NSS BW feature.
Fix this regression by ensuring that this feature only gets enabled on drivers
that support it
Fixes: d9fcfc1424aa ("mt76: enable the VHT extended NSS BW feature")
Signed-off-by: Felix Fietkau <[email protected]>
The code was accidentally shifting register values down by tid % 32 instead of
(tid * field_size) % 32.
Signed-off-by: Felix Fietkau <[email protected]>
Previously the code used skb_set_queue_mapping for management or non-bufferable
powersave frames that need to be sent to a different hardware queue.
This can confuse AQL, which expects the value to remain the same until the tx
status report.
The only place that currently uses the altered skb queue mapping is the txwi
write function. Change the code to pass the hardware queue id as a function
parameter instead.
Signed-off-by: Felix Fietkau <[email protected]>
For some reason, mt7615 reacts badly to repeatedly enabling/disabling the radar
detector without also switching the channel.
This results in very bad throughput on DFS channels, because
hw->conf.radar_enabled can get toggled a few times after CAC ends.
Fix this by always leaving the DFS detector enabled on DFS channels and instead
suppress unwanted detection events.
Fixes: 2c86f6752046 ("mt76: mt7615: fix/rewrite the dfs state handling logic")
Signed-off-by: Felix Fietkau <[email protected]>
Increment the counter only when writing beacons to the hardware in order
to avoid triggering restarts if beacons are disabled.
Additionally, avoid resetting the MAC if stopping it failed
Signed-off-by: Felix Fietkau <[email protected]>
Set EXTRA_INFO_NEW for the first time only to prevent adding the same
starec entry, otherwise the entry might be removed in fw.
Signed-off-by: Felix Fietkau <[email protected]>
After initially establishing a connection, it can produce multi-second latency
spikes and tx hangs when pushing traffic.
It should work better for MT7916 and MT7986, so leave it enabled there
Signed-off-by: Felix Fietkau <[email protected]>
napi_schedule() can call __raise_softirq_irqoff(), which can perform softirq
handling, so it must not be called in a pure process context with BH enabled.
Signed-off-by: Felix Fietkau <[email protected]>
napi_schedule() can call __raise_softirq_irqoff(), which can perform softirq
handling, so it must not be called in a pure process context with BH enabled.
Signed-off-by: Felix Fietkau <[email protected]>
There is a small race window where ongoing tx activity can lead to a skb
getting added to the status tracking idr after that idr has already been
cleaned up, which will keep the wcid linked in the status poll list.
Fix this by only adding status skbs if the wcid pointer is still assigned
in dev->wcid, which gets cleared early by mt76_sta_pre_rcu_remove
Fixes: bd1e3e7b693c ("mt76: introduce packet_id idr")
Tested-by: Ben Greear <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
The driver needs to check if the format is 802.2 vs 802.3 in order to set
a tx descriptor flag. skb->protocol can't be used, since it may not be properly
initialized for packets coming in from a packet socket.
Fix misdetection by checking the ethertype from the skb data instead
Reported-by: Thibaut VARÈNE <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Fixes potential latency / packet drop issues in cases where a BA session has
not (yet) been established.
Fixes: e195dad14115 ("mt76: add support for 802.3 rx frames")
Signed-off-by: Felix Fietkau <[email protected]>
When the QoS ack policy was set to non explicit / psmp ack, frames are treated
as not being part of a BA session, which causes extra latency on reordering.
Fix this by only bypassing reordering for packets with no-ack policy
Signed-off-by: Felix Fietkau <[email protected]>
WED support requires using non-coherent DMA, whereas the PCI device might
be configured for coherent DMA.
The WED driver will take care of changing the PCI HIF coherent IO setting
on attach.
Signed-off-by: Felix Fietkau <[email protected]>
Clean up and fix error paths in mt7915_register_device
Initialize second DBDC tx queue in mt7915_dma_init
Signed-off-by: Felix Fietkau <[email protected]>
Instead of taking q->lock for every single cleaned up frame, batch up to
16 cleaned up frames before taking it.
Signed-off-by: Felix Fietkau <[email protected]>
dma unmap is already needed for cleanup anyway, so we don't need the extra
tracking and can save a bit of memory here
Signed-off-by: Felix Fietkau <[email protected]>
When operating on a mix of DFS and non-DFS channels, the driver only checks
the CAC status of the control channel. This causes beacons/tx to fail if the
control channel is on a non-DFS channel.
Fix this by calling cfg80211_reg_can_beacon to determine the DFS status of
all affected channels
Signed-off-by: Felix Fietkau <[email protected]>
Instead of just taking the maximum per-chain signal strength values,
add an approximation for the sum of the combined signal.
This should more accurately reflect the real signal strength, especially
if the per-chain signal strength values are close to each other
Signed-off-by: Felix Fietkau <[email protected]>
Client mode on DFS channels was broken, because the old code was activating
the DFS detector on radar channels while leaving it in CAC state.
This was caused by making the decision based on the channel radar flag,
instead of hw->conf.radar_enabled.
In order to properly deal with the various corner cases, rip out the state
handling code and replace it with something that's much easier to reason
about.
Signed-off-by: Felix Fietkau <[email protected]>
Instead of checking if any queue has not made progress since the last run,
only trigger hang detection if one of the queues has not made any progress
in 10 subsequent runs. This should reduce false positive firmware restarts
Signed-off-by: Felix Fietkau <[email protected]>
On AP mode devices, check beacon tx counters to detect MAC errors.
When an error is detected, stop the MAC before resetting it
Signed-off-by: Felix Fietkau <[email protected]>
Felix Fietkau