Fixes gcc '-Wunused-but-set-variable' warning:

drivers/net/wireless/mediatek/mt76/dma.c: In function mt76_dma_rx_fill:
drivers/net/wireless/mediatek/mt76/dma.c:377:6: warning: variable idx set but not used [-Wunused-but-set-variable]

It is not used since commit 17f1de56 ("mt76:
add common code shared between multiple chipsets")
Reported-by: NHulk Robot <hulkci@huawei.com>
Signed-off-by: Nzhengbin <zhengbin13@huawei.com>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

if napi_complete() returns false, it means that polling is still pending.
Interrupts should not fire until the polling is no longer scheduled
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

This reduces the struct size and is useful for adding more flags later
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

mt76 dma layer is supposed to unmap skb data buffers while keep txwi
mapped on hw dma ring. At the moment mt76 wrongly unmap txwi or does
not unmap data fragments in even positions for non-linear skbs. This
issue may result in hw hangs with A-MSDU if the system relies on IOMMU
or SWIOTLB. Fix this behaviour properly unmapping data fragments on
non-linear skbs.

Fixes: 17f1de56 ("mt76: add common code shared between multiple chipsets")
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NKalle Valo <kvalo@codeaurora.org>

Fixes rx of the first frame if a fragmented rx was interrupted by the reset
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

This fixes the following checkpatch warnings:
CHECK: Alignment should match open parenthesis
CHECK: No space is necessary after a cast
Signed-off-by: NRyder Lee <ryder.lee@mediatek.com>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

No functional change intended.

Add SPDX identifiers to all remaining files in /mt76.
Signed-off-by: NRyder Lee <ryder.lee@mediatek.com>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Move netif_napi_del in mt76_dma_cleanup routine since it is done
by all drivers
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

In the tx/rx fastpath, the funciton dma_map_single() rarely fails.
This adds unlikely() optimization to this error check conditional.
Signed-off-by: NRyder Lee <ryder.lee@mediatek.com>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Introduce skb check for dummy address in mt76_dma_tx_cleanup_idx.
This is a preliminary patch to add support for new chipsets (e.g. 7615)
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Introduce skb field in mt76_txwi_cache. Moreover add txwi_flags
to mt76_driver_ops since new chipsets will release mt76_txwi_cache/skbs
at tx completion instead of dma one.
This is a preliminary patch to add mt7615 support
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Pass skb pointer to tx_prepare_skb through mt76_tx_info data structure.
This is a preliminary patch to properly support dma error path for
new chipsets (e.g. 7615)
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Dynamically allocate txwi since new chipsets will use longer txwi
descriptors
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Reduces lock contention from the tx path and improves performance
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

q->tail can be safely updated without locking, because there is no
concurrent access. If called from outside of the tasklet (for flushing),
the tasklet is always disabled.
q->queued can be safely read without locking, as long as the decrement
happens within the locked section.
This patch allows cleaning up tx packets outside of the section that holds
the queue lock for improved performance
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Performance improvement and preparation for adding airtime fairness support
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Switching to readl/writel is faster because it gets rid of an unnecessary
wrapper with extra checks.
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

As already done for mt76_dma_tx_queue_skb_raw, add static qualifier to
mt76_dma_tx_queue_skb and introduce mt76_tx_queue_skb in order to run
mt76_dma_tx_queue_skb in driver code
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Add mt76_tx_info as auxiliary data structure to pass values
to tx_prepare_skb pointer. This is a preliminary patch to add
support for new chipsets (e.g. mt7615)
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Move skb dma mapping before configuring txwi since new chipsets (mt7615)
will need skb dma addresses in order to properly configure txwi
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Introduce tx_aligned4_skbs in mt76_driver_ops and move
mt76_insert_hdr_pad in mt76_dma_tx_queue_skb. This is a preliminary
patch in order to unify tx dma mapping for mt76x02 and new chipsets
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Add mt76_txq_id field to mt76_queue_entry in order to properly
track outstanding frames for mt7615 that relies on a single hw queue
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Introduce mt76_sw_queue data structure in order to support new
chipsets (e.g. mt7615) that have a shared hardware queue for all traffic
identifiers. mt76_sw_queue will be used to track outstanding packets
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Remove mt76_queue dependency from tx_complete_skb function pointer and
rely on mt76_tx_qid instead. Remove flush from tx_complete_skb
signature. This is a preliminary patch to introduce mt76_sw_queue
support
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Remove mt76_queue dependency from tx_prepare_skb function pointer and
rely on mt76_tx_qid instead. This is a preliminary patch to introduce
mt76_sw_queue support
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Remove mt76_queue dependency from tx_queue_skb function pointer and
rely on mt76_tx_qid instead. This is a preliminary patch to introduce
mt76_sw_queue support
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Move mt76x02_init_tx_queue and mt76x02_init_rx_queue in mt76
module in order to be reused adding support for mt7603 driver
and remove duplicated code.
Squash mt76x02_init_tx_queue and mt76x02_init_rx_queue in
mt76_dma_alloc_queue
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Useful in case the hardware reset clobbers these values
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Introduce mt76_queue stopped parameter in order to run
ieee80211_wake_queue only when mac80211 queues have been
previously stopped and avoid to disable interrupts when
it is not necessary
Signed-off-by: NLorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

There is no need to retun 0 in mt76_dma_attach(), so switch it to void.
Signed-off-by: NRyder Lee <ryder.lee@mediatek.com>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Remove add_buf function pointer in mt76_queue_ops data structure since
it is no longer used
Signed-off-by: NLorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Move mt76_dma_tx_queue_skb_raw routine in dma.c and add the
corresponding entry in mt76_queue_ops data structure.
mt76_dma_tx_queue_skb_raw  will be reused adding support for
mt7603 driver
Signed-off-by: NLorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Call mt76_dma_add_buf routine directly in mt76_dma_tx_queue_skb
and avoid indirect call if not necessary
Signed-off-by: NLorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Precompute data length in order to avoid to allocate the related
skb data structure if reported length does not fit in queue buf_size
Signed-off-by: NLorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Remove napi from mt76_dma_rx_fill routine signature since
it is a leftover of a previous implementation and it is no
longer used
Signed-off-by: NLorenzo Bianconi <lorenzo.bianconi@redhat.com>
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

So far the code only validates the buffer size of the first skb.
Extend this check to cover additional fragments as well, in case the size
is corrupted during a DMA reset.
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

While the queue is being cleaned up, the stack must not attempt to add
any extra packets
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

There is some code in the mac80211 tx status processing code that could
potentially call back into the tx codepath.
To avoid deadlocks, make sure that no tx related spinlocks are taken
during the ieee80211_tx_status call.
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

MT76x2/MT76x0 has somewhat unreliable tx status reporting, and for that
reason the driver currently does not report per-skb tx ack status at all.
This breaks things like client idle polling, which relies on the tx ack
status of a transmitted nullfunc frame.

This patch adds code to report skb-attached tx status if requested by
mac80211 or the rate control module. Since tx status is polled from a
simple FIFO register, the code needs to account for the possibility of
tx status events getting lost.

The code keeps a list of skbs for which tx status is required and passes
them to mac80211 once tx status has been filled in and the DMA queue is
done with it.
If a tx status event is not received after one second, the status rates
are cleared, and a succesful ACK is indicated to avoid spurious disassoc
during assoc or client polling.
Signed-off-by: NFelix Fietkau <nbd@nbd.name>

Using the NAPI or netdev frag cache along with other drivers can lead to
32 KiB pages being held for a long time, despite only being used for
very few page fragments.

This can happen if the driver grabs one or two fragments for rx ring
refill, while other drivers use (and free up) the remaining fragments.
The 32 KiB higher-order page can only be freed once all users have freed
their fragments.

Depending on the traffic patterns, this can waste a lot of memory and
look a lot like a memory leak.
Signed-off-by: NFelix Fietkau <nbd@nbd.name>