Age | Commit message (Collapse) | Author |
|
Run dma cleanup immediately if the queue is almost full, instead of waiting
for the tx interrupt
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
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 <nbd@nbd.name>
|
|
Introduce mt7915 dbdc support. If dbdc is available, mt7915 primary phy
will work on 2.4GHz band, while secondary one on 5GHz band.
Co-developed-by: Shayne Chen <shayne.chen@mediatek.com>
Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Co-developed-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
This is a preliminary patch to properly support mt7915 dbdc
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
This is a preliminary patch to introduce dbdc support to mt7915 devices
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
This is a preliminary patch to move properly support mt7915 dbdc
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
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 <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
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 <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
This is a preliminary patch to move data queues in mt76_phy and properly
support dbdc
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Introduce mt76_init_mcu_queue utility routine in order to allocate
dedicate q_mcu pointers for mcu hw queues. This is a preliminary patch
to move data queues in mt76_phy and add dbdc support to mt7915 driver
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
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 <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
This is a preliminary patch to move data queues in mt76_phy and properly
support dbdc
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
This is a preliminary patch to move data queues in mt76_phy and properly
support dbdc for mt7915
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Move mt76_init_tx_queue in mac80211.c since it is shared by all
drivers.
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Prepare for setting different LTF (1x, 2x, and 4x) and GI (0.8us,
1.6us, 3.2us) combinations in HE rate mode.
Reviewed-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Add the parameter to control STBC function in testmode.
Reviewed-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Split tx and status path in mt76-usb module relying on mt76 workers APIs
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
In order to improve parallelism, convert rx path in mt76-usb module to
mt76 workers APIs
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Move mt76_mcu_send_firmware routine in common code since it is actually
shared between mt7615, mt7663, mt7603 and mt7915 devices
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
sched quota is now updated in mt7663s_txrx_worker so there is no more
need of sched.lock mutex
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Can be used by the caller to get the response data directly instead of using the
hack of storing the result in internal data structures from .mcu_parse_response
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Remove it from mt7615 and mt7915 and implement it in core code instead
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
This removes some code duplication and prepares the code for making
the MCU API more flexible
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Do not free skb inside that function
Preparation for further cleanup
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Preparation for further cleanup
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Preparation for further cleanup
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Enable MWDS mode in firmware as well and fix txp->rept_wds_wcid for wcid >= 255
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
In order to improve driver throughput, convert status_work and net_work
to mt76 worker APIs.
Remove txrx_wq sdio workqueue since it is no longer used
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
In order to improve driver throughput, convert txrx_work to mt76 worker
threads APIs.
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Move mt7663 tx and rx processing in the same workqueue in order to
reduce jitter that can hit tcp performances. This is a preliminary patch
to switch to mt76 workers APIs
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Introduce sdio tx aggregation to reduce bus transaction ands improve tx
throughput. For the moment the skb are copied in a dedicated buffer
since mmc APIs do not support sg table for zero-copy.
Since skb data are already copied in xmit_buff[], avoid linearization in
ma80211 layer. Relying on tx aggregation, we improve tx tpt of ~65%.
Tested-by: Sean Wang <sean.wang@mediatek.com>
Co-developed-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Move frame pad computation out of mt76_skb_adjust_pad routine.
This is a preliminary patch to introduce sdio tx aggregation.
Tested-by: Sean Wang <sean.wang@mediatek.com>
Co-developed-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Since the switch to using AQL by default, mtxq->retry_q is never filled anymore
Remove it to get rid of a few more unnecessary cycles in the tx path
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
This way we can make some functions static
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Accessing them from uncached memory can be expensive, so it's cheaper to
cache them
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
This improves performance by allowing the scheduler to move the tx scheduling
work to idle CPUs. Since tx scheduling work is very latency sensitive and
kept short via AQL, sched_set_fifo_low is used to keep worker priority above
normal tasks
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Increase data queue size to improve performance.
Reduce PS/mgmt queue size
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
All members except for the struct mt76_queue pointer have been removed
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Now that AQL works well on all mt76 drivers, completely replace the arbitrary
burst sizing and number of bursts logic for tx scheduling.
For the short period of time in which AQL does not work well yet, limit each
stations to 16 in-flight packets that have no estimated tx time.
This should avoid filling the queue if a station connects and queues up a
large number of packets before rate control information is available, especially
with hardware rate control
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Since txq selection was moved to mac80211, it is no longer used
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
It is not needed
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Cleanup and preparation for changing tx scheduling behavior
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Use buffer allocated with kmalloc instead of with stack to fix kernel
crash due to Unable to handle kernel paging request at virtual address
ffffffc0095cbce8.
[ 156.977349] Unable to handle kernel paging request at virtual address ffffffc0095cbce8
[ 156.985270] Mem abort info:
[ 156.988059] ESR = 0x96000045
[ 156.991104] Exception class = DABT (current EL), IL = 32 bits
[ 156.997013] SET = 0, FnV = 0
[ 157.000057] EA = 0, S1PTW = 0
[ 157.003190] Data abort info:
[ 157.006061] ISV = 0, ISS = 0x00000045
[ 157.009887] CM = 0, WnR = 1
[ 157.012850] swapper pgtable: 4k pages, 39-bit VAs, pgdp = 0000000042adcba2
[ 157.019715] [ffffffc0095cbce8] pgd=0000000000000000, pud=0000000000000000
[ 157.026499] Internal error: Oops: 96000045 [#1] PREEMPT SMP
[ 157.032065] Modules linked in: mt7663s mt7663_usb_sdio_common mt7615_common
...
[ 157.073007] Process CompositorTileW (pid: 1625, stack limit = 0x000000003f2389fc)
[ 157.080484] CPU: 0 PID: 1625 Comm: CompositorTileW Not tainted 4.19.137 #36
[ 157.092219] pstate: 80000085 (Nzcv daIf -PAN -UAO)
[ 157.097012] pc : __memcpy+0xc0/0x180
[ 157.100585] lr : swiotlb_tbl_unmap_single+0x84/0x14c
[ 157.105540] sp : ffffff8008003cb0
[ 157.108845] x29: ffffff8008003cb0 x28: ffffff9c1a211f60
[ 157.114149] x27: ffffff9c19ecc018 x26: 0000000000001000
[ 157.119452] x25: ffffff9c1a378000 x24: 0000000000000001
[ 157.124755] x23: ffffff9c1a378000 x22: 00000000000001ff
[ 157.130058] x21: 0000000000000000 x20: 00000000fbefe800
[ 157.135360] x19: 0000000000000070 x18: 0000000000000000
[ 157.140663] x17: 0000000000000000 x16: 0000000000000000
[ 157.145965] x15: 0000000000000000 x14: 0000000000000000
[ 157.151267] x13: 0000000000000000 x12: 000000000000000d
[ 157.156569] x11: 000000000000000c x10: 0000000a7befe800
[ 157.161873] x9 : fffffff680000000 x8 : 0000000000000000
[ 157.167175] x7 : 0000000100000003 x6 : ffffffc0095cbce8
[ 157.172479] x5 : 0000000000000000 x4 : 0000000000000000
[ 157.177781] x3 : 0000000000000002 x2 : fffffffffffffff0
[ 157.183085] x1 : ffffffca7befe810 x0 : ffffffc0095cbce8
[ 157.188389] Call trace:
[ 157.190832] __memcpy+0xc0/0x180
[ 157.194053] swiotlb_unmap_sg_attrs+0xa8/0xb0
[ 157.198406] __swiotlb_unmap_sg_attrs+0x8c/0xa4
[ 157.202931] msdc_unprepare_data+0x6c/0x84
[ 157.207019] msdc_request_done+0x58/0x98
[ 157.210934] msdc_data_xfer_done+0x1a8/0x1d0
[ 157.215195] msdc_irq+0x12c/0x17c
[ 157.218505] __handle_irq_event_percpu+0xd8/0x298
[ 157.223202] handle_irq_event+0x60/0xdc
[ 157.227031] handle_fasteoi_irq+0xa4/0x1d4
[ 157.231120] __handle_domain_irq+0x84/0xc4
[ 157.235210] gic_handle_irq+0x124/0x1a4
[ 157.239038] el0_irq_naked+0x4c/0x54
[ 157.242608] Code: 14000028 f1020042 5400024a a8c12027 (a88120c7)
[ 157.248693] ---[ end trace 28b8090135b0a2e1 ]---
[ 157.265589] Kernel panic - not syncing: Fatal exception in interrupt
[ 157.271944] SMP: stopping secondary CPUs
[ 157.275865] Kernel Offset: 0x1c10e00000 from 0xffffff8008000000
[ 157.281779] CPU features: 0x0,2188200c
[ 157.285519] Memory Limit: none
Fixes: a66cbdd6573d ("mt76: mt7615: introduce mt7663s support")
Co-developed-by: YN Chen <YN.Chen@mediatek.com>
Signed-off-by: YN Chen <YN.Chen@mediatek.com>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
This disables the software A-MSDU aggregation in mac80211 and enables hardware
offloading
Suggested-by: Yiwei Chung <yiwei.chung@mediatek.com>
Suggested-by: YF Luo <yf.luo@mediatek.com>
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Co-developed-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
mt7615 and newer drivers do not need this, since they use sequence number offload
Moving this code also reduces the number of callsites to make it easier to review
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
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 <nbd@nbd.name>
|
|
Remove unused state variable in mt76_sdio structure
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
In order to maximize parallelism, split status work in tx status work
and rx net work
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
As it has been done for tx and rx processing, move tx/rx status
processing into mt76s_txrx_wq workqueue
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|
|
Move rx processing to mt76s_txrx_wq in order to minimize the interval when
the sdio bus is locked during rx
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
|