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]>
Add the capability to define a custom rx napi callback for each driver.
This is a preliminary patch to properly support runtime-pm on rx side
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
With buf uninitialized in mt76_dma_tx_queue_skb_raw, its field skip_unmap
could potentially inherit a non-zero value from stack garbage.
If this happens, it will cause DMA mappings for MCU command frames to not be
unmapped after completion
Fixes: 27d5c528a7ca ("mt76: fix double DMA unmap of the first buffer on 7615/7915")
Cc: [email protected]
Signed-off-by: Felix Fietkau <[email protected]>
Export mt76_dma_rx_cleanup routine in mt76_queue_ops data structure.
This is a preliminary patch to introduce mt7921 chip reset support.
Co-developed-by: Sean Wang <[email protected]>
Signed-off-by: Sean Wang <[email protected]>
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Introduce mt76_dma_queue_reset utility routine to reset a given hw
queue. This is a preliminary patch to introduce mt7921 chip reset
support.
Co-developed-by: Sean Wang <[email protected]>
Signed-off-by: Sean Wang <[email protected]>
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
When running out of room in the tx queue after calling drv->tx_prepare_skb,
the buffer list will already have been modified on MT7615 and newer drivers.
This can leak a DMA mapping and will show up as swiotlb allocation failures
on x86.
Fix this by moving the queue length check further up. This is less accurate,
since it can overestimate the needed room in the queue on MT7615 and newer,
but the difference is small enough to not matter in practice.
Signed-off-by: Felix Fietkau <[email protected]>
If the fragment is discarded in mt76_add_fragment() since shared_info
frag array is full, discard truncated frames and do not forward them to
mac80211.
Signed-off-by: Lorenzo Bianconi <[email protected]>
Fix a memory leak in mt76_add_fragment routine returning the buffer
to the page_frag_cache when we receive a new fragment and the
skb_shared_info frag array is full.
Fixes: b102f0c522cf6 ("mt76: fix array overflow on receiving too many fragments for a packet")
Signed-off-by: Lorenzo Bianconi <[email protected]>
Acked-by: Felix Fietkau <[email protected]>
Add testmode support for DBDC NICs (both MT7615D and MT7915D work).
Testmode data and parameters are moved from per-dev to per-phy
for maintaining the value of each band.
Signed-off-by: Shayne Chen <[email protected]>
Acked-by: Ryder Lee <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Instead of stopping and waking only a single queue, handle all phy tx queues
mapped ot the same hardware queue.
Also allow the driver to block tx queues
Signed-off-by: Felix Fietkau <[email protected]>
Move hw data queues in mt76_phy from mt76_dev since mt7915 supports per
phy hw queues in dbdc mode
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Move mcu queue to a dedicated array q_mcu in mt76_dev structure.
This is a preliminary patch to move data queues in mt76_phy and properly
support dbdc
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
This is a preliminary patch to move data queues in mt76_phy and properly
support dbdc
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Rely on mt76_queue instead of qid in tx_queue_skb signature.
This is a preliminary patch to move data queues in mt76_phy and add
dbdc support to mt7915 driver
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
This is a preliminary patch to move data queues in mt76_phy and properly
support dbdc
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
On MT7615 and MT7915, packets received from DMA can include TX_FREE events,
which free up space for tx. In order to process those packets faster, keep
keep a longer internal queue of packets and pass as many packets to the
stack as the NAPI budget allows
Signed-off-by: Felix Fietkau <[email protected]>
With threaded NAPI support it becomes more likely that tx cleanup and tx
enqueuing will run in parallel. In order to avoid queue starvation issues
under load, update q->queued immediately.
Signed-off-by: Felix Fietkau <[email protected]>
With threaded NAPI, the rx handler function is no longer bound to the CPU that
fired the interrupt, which significantly helps to spread the workload over
multiple CPUs, especially when multiple devices are using threaded NAPI at the
same time.
Signed-off-by: Felix Fietkau <[email protected]>
A small part of the first skb buffer is passed to the firmware for parsing
via DMA, while the full buffer is passed as part of the TXP.
Avoid calling DMA unmap on the first part (with a different length than map)
Signed-off-by: Felix Fietkau <[email protected]>
If the hardware receives an oversized packet with too many rx fragments,
skb_shinfo(skb)->frags can overflow and corrupt memory of adjacent pages.
This becomes especially visible if it corrupts the freelist pointer of
a slab page.
Cc: [email protected]
Signed-off-by: Felix Fietkau <[email protected]>
Fixes a theoretical issue where it could potentially overwrite an existing
descriptor entry (and leaking its skb)
Signed-off-by: Felix Fietkau <[email protected]>
Avoids a theoretical corner case where the hardware could try to process
a stale descriptor after a watchdog reset
Signed-off-by: Felix Fietkau <[email protected]>
To avoid having the hardware potentially write to memory behind stale
descriptors, set the dma-done flag on all of them before cleaning up
allocated rx buffers
Signed-off-by: Felix Fietkau <[email protected]>
This is preparation for supporting multiple wiphys per device to support the
concurrent dual-band feature of MT7615D
Signed-off-by: Felix Fietkau <[email protected]>
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 17f1de56df05 ("mt76:
add common code shared between multiple chipsets")
Reported-by: Hulk Robot <[email protected]>
Signed-off-by: zhengbin <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Felix Fietkau