Introduce mt76s_txrx_wq workqueue and move tx processing from kthread to
a dedicated work. This is preliminary patch to improve mt7663s throughput
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Free the mcu skb in case of error in mt76s_tx_queue_skb_raw routine
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Reschedule runtime-pm after receiving a tx interrupt.
Update runtime-pm last activity before injecting packets
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Do not tx packets in mt76_txq_send_burst() or mt76_txq_schedule_list()
if the device is in runtime-pm
Signed-off-by: Lorenzo Bianconi <[email protected]>
Reported-by: kernel test robot <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
wq queue is always updated holding mt76 spinlock. Grab mt76 lock in
mt7615_queue_key_update() before putting a new element at the end of the
queue.
Fixes: eb99cc95c3b65 ("mt76: mt7615: introduce mt7663u support")
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Initialize set_drv_ctrl and set_fw_ctrl function pointers in
mt7663s_mcu_init. This is a preliminary patch to enable runtime-pm for
mt7663s chipset.
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Introduce set_drv_ctrl and set_fw_ctrl function pointers in
mt7615_mcu_ops data structure. This is a preliminary patch to enable
runtime-pm for non-pci chipsets
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
MT_EE_WIFI_CONF field can be used to detect if the chipset is MT7615D.
Thus, add support to automatically register ext_phy if DBDC is detected.
Signed-off-by: Shayne Chen <[email protected]>
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]>
In the implementation of __mt76x02u_mcu_send_msg() the skb is consumed
all execution paths except one. Release skb before returning if
test_bit() fails.
Signed-off-by: Navid Emamdoost <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Free the second mcu skb if __mt76_mcu_skb_send_msg() fails to transmit
the first one in mt7615_mcu_wtbl_sta_add().
Fixes: 99c457d902cf9 ("mt76: mt7615: move mt7615_mcu_set_bmc to mt7615_mcu_ops")
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
In order to avoid possible overflows, move tx queue accounting from
mt7663s_tx_run_queue() to mt76s_tx_queue_skb_raw()/mt76s_tx_queue_skb()
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Introduce support for mt7663s 802.11ac 2x2:2 chipset to mt7615 driver.
Co-developed-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Sean Wang <[email protected]>
Introduce mt76_sdio module as common layer to add mt7663s support
Co-developed-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Sean Wang <[email protected]>
As usb, sdio relies on mt76 workqueue to configure tx rate or upload
keys to the hw. This is a preliminary patch to add SDIO support to
mt76 driver
Co-developed-by: Sean Wang <[email protected]>
Signed-off-by: Sean Wang <[email protected]>
Signed-off-by: Lorenzo Bianconi <[email protected]>
Introduce mt76_skb_adjust_pad to reuse the code adding sdio support to
mt7615 driver and remove code duplication. Move 4B header configuration
for usb devices out of mt76_skb_adjust_pad
Signed-off-by: Lorenzo Bianconi <[email protected]>
usb and sdio relies on SF architecture. This is a preliminary patch to
add SDIO support to mt76 driver
Co-developed-by: Sean Wang <[email protected]>
Signed-off-by: Sean Wang <[email protected]>
Signed-off-by: Lorenzo Bianconi <[email protected]>
Smatch complains that "wcidx" value comes from the network and thus
cannot be trusted. In this case, it actually seems to come from the
firmware. If your wireless firmware is malicious then probably no
amount of carefulness can protect you.
On the other hand, these days we still try to check the firmware as much
as possible. Verifying that the index is within bounds will silence a
static checker warning. And it's harmless and a good exercise in kernel
hardening. So I suggest that we do add a bounds check.
Fixes: e57b7901469f ("mt76: add mac80211 driver for MT7915 PCIe-based chipsets")
Signed-off-by: Dan Carpenter <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Fix potential memory leak in mcu message handler on error condition.
Fixes: 0e6a29e477f3 ("mt76: mt7615: add support to read temperature from mcu")
Acked-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Sean Wang <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Fix memory leak in set key.
Fixes: eb99cc95c3b6 ("mt76: mt7615: introduce mt7663u support")
Signed-off-by: Sean Wang <[email protected]>
Acked-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Reschedule runtime-pm delayed work if there is a new activity when ps
delayed work is already scheduled
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Do not schedule ps_work during hw scanning or hw scheduled frequency
scan
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
On usb device rate configuration for sampling is performed relying on a
workqueue since it is not possible to access the device in the interrupt
context. Move the configuration of the probe_rate flag in the workqueue
in order to keep probe sampling in sync with actual rate configuration
Fixes: eb99cc95c3b6 ("mt76: mt7615: introduce mt7663u support")
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
mt7615_driver_own can fail if it runs too close to mt7615_fw_own. In
order to improve mt7615_driver_own reliability, retry to get runtime-pm
ownership if mt7615_driver_own fails
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 idle-timeout knob in mt7615 debugfs in order to configure the
idle time to switch to low-power state
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
In order to reduce number of received interrupts and power consumption,
enable hw beacon filter if runtime-pm is enabled
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Introduce runtime-pm knob in mt7615 debugfs in order to enable/disable
runtime pm available in offload firmware
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Check MT76_STATE_PM in mt7615_driver_own/mt7615_firmware_own
in order to not requested power ownership if it is already granted
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Double-check if the device is in low-power state before accessing
registermap in mt7615_sta_rate_tbl_update() and in
mt7615_led_set_config()
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Run mt7615_pm_wake()/mt7615_pm_power_save_sched() directly in order to
wake the device from low power state in mt7615_mac_sta_{add,remove}
since they run holding mt76 mutex in common mt76 code
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Queue frames pushed by mac80211 running mt7615_tx if the device is
low-power state. Run wake workqueue in order to swicth to full-power
before transmitting pending frames
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Make sure the device is in full-power before pulling frames from
mac80211 queues
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Make sure to run mt7615_mcu_set_roc() holding mt76 mutex in order to
wake the device from low power state and avoid races
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Make sure to run mt7615_mcu_set_wmm() holding mt76 mutex in order to
wake the device from low power state and avoid races
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Make sure to run mt7615_dfs_init_radar_detector in mt7615_regd_notifier
holding mt76 mutex in order to avoid races and set the device in full
power
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Set device in full power before performing hw scan or hw scheduled scan
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Make sure to wake the device in mt7615_pci_suspend in order to properly
configure device registers before suspend
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Introduce mt7615_update_survey utility routine in order to compute
survey stats without waking up the device since it runs holding mt76 lock.
Run mt7615_pm_wake directly in mt7615_update_channel since it can run
with mt76.mutex held if called by mac80211
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
Introduce runtime-pm power_save delayed work used to enable
low-power after an inactivity period
Signed-off-by: Lorenzo Bianconi <[email protected]>
Signed-off-by: Felix Fietkau <[email protected]>
