mt76: mt7915: introduce mt7915_mcu_set_txbf()

Use mt7915_mcu_set_txbf() to reduce global functions. This can be
easily extended to support other TxBF commands in further patches.

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: make mt76_update_survey() per phy

Reduce duplicated survey for DBDC.

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: drop the use of repeater entries for station interfaces

There are firmware or hardware issues, which are currently causing tx hangs
when attempting to use these interfaces

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: connac: check band caps in mt76_connac_mcu_set_rate_txpower

Check device band capabilities before configuring single-sku

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: move mt76_get_next_pkt_id in mt76.h

In order to remove duplicated code, move mt76_get_next_pkt_id routine
in mt76.h

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: reduce rx buffer size to 2048

Reduce rx buffer size to 2048 for mt7921/mt7915/mt7615 since we
now support rx amsdu offload

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: connac: fix the maximum interval schedule scan can support

Maximum interval (in seconds) for schedule scan plan supported by
the offload firmware can be U16_MAX.

Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: fix rx fcs error count in testmode

FCS error packets are filtered by default and won't be reported to
driver, so that RX fcs error and PER in testmode always show zero.
Fix this issue by reading fcs error count from hw counter.

We did't fix this issue by disabling fcs error rx filter since it may
let HW suffer some SER errors.

Fixes: 5d8a83f09941 ("mt76: mt7915: implement testmode rx support")
Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: testmode: move chip-specific stats dump before common stats

Move chip-specific stats dumping part before common stats dumping
to provide flexibility for per-chip driver to modify the value of
common stats.

Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: connac: add mt76_connac_mcu_get_nic_capability utility routine

Introduce mt76_connac_mcu_get_nic_capability utility routine to poll
device capabilities returned by mcu fw for CE devices (mt7663/mt7921).
This is a preliminary patch to introduce 6GHz support.

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: make mt7921_set_channel static

Make mt7921_set_channel routine static since it is only used in main.c

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: wake the device before dumping power table

Always wake the device up before dumping the single_sku power table
otherwise the device can hang.

Fixes: ea29acc97c555 ("mt76: mt7921: add dumping Tx power table")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: connac: add mt76_connac_power_save_sched in mt76_connac_pm_unref

Schedule power_save work running mt76_connac_pm_unref in order to reduce
power consumption

Tested-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: connac: fix UC entry is being overwritten

Fix UC entry is being overwritten by BC entry

Tested-by: Deren Wu <deren.wu@mediatek.com>
Co-developed-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: enable VHT BFee capability

Enables VHT beamformee functionality

Signed-off-by: Leon Yen <Leon.Yen@mediatek.com>
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: fix MT_EE_CAL_GROUP_SIZE

Fix wrong offset for pre-calibration data.

Fixes: 495184ac91bb ("mt76: mt7915: add support for applying pre-calibration data")
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7615: update radar parameters

Patch radar parameters to match the SDK to avoid possible false alarms.

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: setup drr group for peers

This is a prerequisite for MU functionality.

Tested-by: Evelyn Tsai <evelyn.tsai@mediatek.com>
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: enable random mac address during sched_scan

Enable src address randomization during scheduled scanning

Co-developed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: remove mt7921_get_wtbl_info routine

Since now the fw reports tx rate events without polling,
mt7921_get_wtbl_info and related structures are no longer used.

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: enable hw offloading for wep keys

Enable wep key hw offloading for sta mode. This patch fixes
WoW support for wep connections.

Tested-by: Deren.Wu <deren.wu@mediatek.com>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: read all eeprom fields from fw in efuse mode

If efuse mode is used, read all values from fw during eeprom init,
which makes it more convinient to check if rf values in efuse are
properly burned.

Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: use mt7915_mcu_get_mib_info() to get survey data

Firmware functions (SCS, MU ...) also require read-clear phy counters,
hence firmware prepares a global task to read shared fields out to a
shared pool to avoid concurrency. Switch to event format accordingly.

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: set MT76_RESET during mac reset

Set MT76_RESET during mt7921_mac_reset in order to avoid packet
transmissions. Move tx scheduling at the end of reset routine.

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: improve error recovery reliability

- Remove no-op code for queue lock/unlock, which is no longer needed
- Set a missing DMA flag
- Wait for full completion of error recovery before restarting tx
- Schedule IRQ tasklet to ensure that IRQ mask gets written

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: allow hw driver code to overwrite wiphy interface_modes

Move wiphy interface_modes configuration in mt76_alloc_device and
mt76_alloc_phy in order to be overwritten by hw specific code
since some drivers do not support all operating modes (mt7921
supports sta only in the current codebase)

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: update statistic in active mode only

wakeup chip every 250ms may cause huge power consumption

try to update statistic counter only if in active status only,
and it would lead fewer power cost

Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: introduce mac tx done handling

Instead of read tx status from mac table, add new mechanisam to hanele
tx done event for data frame, every 250ms

This event indicate the real tx status of this pkt in mac layer and
would help mac80211 correct status more frequently

Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7615: fix potential overflow on large shift

Fix the following static checker warning:
error: undefined (user controlled) shift '(((1))) << (c->omac_idx)'

Fixes: 402a695b1ae6 ("mt76: mt7615: fix CSA notification for DBDC")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: testmode: remove undefined behaviour in mt76_testmode_alloc_skb

Get rid of an undefined behaviour in mt76_testmode_alloc_skb routine
allocating skb frames

Fixes: 2601dda8faa76 ("mt76: testmode: add support to send larger packet")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: testmode: remove unnecessary function calls in mt76_testmode_free_skb

Get rid of unnecessary function calls in mt76_testmode_free_skb routine
since they are already managed by dev_kfree_skb

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: testmode: fix memory leak in mt76_testmode_alloc_skb

Free all pending frames in case of failure in mt76_testmode_alloc_skb
routine

Fixes: 2601dda8faa76 ("mt76: testmode: add support to send larger packet")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7615: remove useless if condition in mt7615_add_interface()

Get rid of unnecessary if condition in mt7615_add_interface routine

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: use mt7915_mcu_get_txpower_sku() to get per-rate txpower

Get per-rate txpower with mcu command. This is the preparation of
co-driver for the next chipset, which has different tmac power registers
but can share this same command.

Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7615: add .offset_tsf callback

It's much more accurate than .get_tsf + .set_tsf and switch to use
mt76_rmw to operate tsf registers.

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: add .offset_tsf callback

It's much more accurate than .get_tsf + .set_tsf, and switch to use
mt76_rmw to operate tsf registers.

Tested-by: Xing Song <xing.song@mediatek.com>
Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: reset wfsys during hw probe

This patch fixes a mcu hang during device probe on
Marvell ESPRESSObin after a hot reboot.

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: do not schedule hw reset if the device is not running

Do not schedule hw full reset if the device is not fully initialized
(e.g if the channel has not been configured yet). This patch fixes
the kernel crash reported below

[   44.440266] mt7921e 0000:01:00.0: chip reset failed
[   44.527575] Unable to handle kernel paging request at virtual address ffffffc02f3e0000
[   44.535771] Mem abort info:
[   44.538646]   ESR = 0x96000006
[   44.541792]   EC = 0x25: DABT (current EL), IL = 32 bits
[   44.547268]   SET = 0, FnV = 0
[   44.550413]   EA = 0, S1PTW = 0
[   44.553648] Data abort info:
[   44.556613]   ISV = 0, ISS = 0x00000006
[   44.560563]   CM = 0, WnR = 0
[   44.563619] swapper pgtable: 4k pages, 39-bit VAs, pgdp=0000000000955000
[   44.570530] [ffffffc02f3e0000] pgd=100000003ffff003, p4d=100000003ffff003, pud=100000003ffff003, pmd=0000000000000000
[   44.581489] Internal error: Oops: 96000006 [#1] SMP
[   44.606406] CPU: 0 PID: 0 Comm: swapper/0 Tainted: G        W         5.13.0-rc1-espressobin-12875-g6dc7f82ebc26 #33
[   44.617264] Hardware name: Globalscale Marvell ESPRESSOBin Board (DT)
[   44.623905] pstate: 600000c5 (nZCv daIF -PAN -UAO -TCO BTYPE=--)
[   44.630100] pc : __queue_work+0x1f0/0x500
[   44.634249] lr : __queue_work+0x1e8/0x500
[   44.638384] sp : ffffffc010003d70
[   44.641798] x29: ffffffc010003d70 x28: 0000000000000000 x27: ffffff8003989200
[   44.649166] x26: ffffffc010c08510 x25: 0000000000000002 x24: ffffffc010ad90b0
[   44.656533] x23: ffffffc010c08508 x22: 0000000000000012 x21: 0000000000000000
[   44.663899] x20: ffffff8006385238 x19: ffffffc02f3e0000 x18: 00000000000003c9
[   44.671266] x17: 0000000000000000 x16: 0000000000000000 x15: 000009b1a8a3bf90
[   44.678632] x14: 0098968000000000 x13: 0000000000000000 x12: 0000000000000325
[   44.685998] x11: ffffff803fda1928 x10: 0000000000000001 x9 : ffffffc010003e98
[   44.693365] x8 : 0000000000000032 x7 : fff8000000000000 x6 : 0000000000000035
[   44.700732] x5 : 0000000000000000 x4 : 0000000000000000 x3 : ffffffc010adf700
[   44.708098] x2 : ffffff8006385238 x1 : 000000007fffffff x0 : 0000000000000000
[   44.715465] Call trace:
[   44.717982]  __queue_work+0x1f0/0x500
[   44.721760]  delayed_work_timer_fn+0x18/0x20
[   44.726167]  call_timer_fn+0x2c/0x178
[   44.729947]  run_timer_softirq+0x488/0x5c8
[   44.734172]  _stext+0x11c/0x378
[   44.737411]  irq_exit+0x100/0x108
[   44.740830]  __handle_domain_irq+0x60/0xb0
[   44.745059]  gic_handle_irq+0x70/0x2b4
[   44.748929]  el1_irq+0xb8/0x13c
[   44.752167]  arch_cpu_idle+0x14/0x30
[   44.755858]  default_idle_call+0x38/0x168
[   44.759994]  do_idle+0x1fc/0x210
[   44.763325]  cpu_startup_entry+0x20/0x58
[   44.767372]  rest_init+0xb8/0xc8
[   44.770703]  arch_call_rest_init+0xc/0x14
[   44.774841]  start_kernel+0x408/0x424
[   44.778623] Code: aa1403e0 97fff54f aa0003f5 b5fff500 (f9400275)
[   44.784907] ---[ end trace be73c3142d8c36a9 ]---
[   44.789668] Kernel panic - not syncing: Oops: Fatal exception in interrupt

Fixes: 0c1ce9884607 ("mt76: mt7921: add wifi reset support")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: return proper error value in mt7921_mac_init

Return possible error values in mt7921_mac_init routine

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: connac: add bss color support for sta mode

Add bss color support for sta mode

Signed-off-by: Jayden.Kuo <jayden.kuo@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>

mt76: mt7921: enable runtime pm by default

mt7921 is mainly used in CE/IoT market so enable runtime-pm by default

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: fix OMAC idx usage

OMAC idx have to be same with BSS idx according to firmware usage.

Fixes: e0f9fdda81bd ("mt76: mt7921: add ieee80211_ops")
Reviewed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Deren Wu <deren.wu@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>

mt76: mt7921: fix invalid register access in wake_work

Make sure mt7921_pm_wake_work wouldn't be scheduled after the driver is
in suspend mode to fix the following the kernel crash.

[ 3515.390012] mt7921e 0000:01:00.0: calling pci_pm_suspend+0x0/0x22c @ 2869, parent: 0000:00:00.0
[ 3515.390015] mt7921e 0000:01:00.0: mt7921_pci_suspend +
[ 3515.396395] anx7625 3-0058: anx7625_suspend+0x0/0x6c returned 0 after 0 usecs
[ 3515.405965] mt7921e 0000:01:00.0: mt7921_pci_suspend -
[ 3515.411336] usb 1-1.4: usb_dev_suspend+0x0/0x2c returned 0 after 1 usecs
[ 3515.411513] SError Interrupt on CPU7, code 0xbe000011 -- SError
[ 3515.411515] CPU: 7 PID: 2849 Comm: kworker/u16:27 Not tainted 5.4.114 #44
[ 3515.411516] Hardware name: MediaTek Asurada rev1 board (DT)
[ 3515.411517] Workqueue: mt76 mt7921_pm_wake_work [mt7921e]
[ 3515.411518] pstate: 80c00009 (Nzcv daif +PAN +UAO)
[ 3515.411519] pc : mt76_mmio_rr+0x30/0xf0 [mt76]
[ 3515.411520] lr : mt7921_rr+0x38/0x44 [mt7921e]
[ 3515.411520] sp : ffffffc015813c50
[ 3515.411521] x29: ffffffc015813c50 x28: 0000000000000402
[ 3515.411522] x27: ffffffe5a2012138 x26: ffffffe5a1eea018
[ 3515.411524] x25: 00000000328be505 x24: 00000000000a0002
[ 3515.411525] x23: 0000000000000006 x22: ffffffbd29b7a300
[ 3515.411527] x21: ffffffbd29b7a300 x20: 00000000000e0010
[ 3515.411528] x19: 00000000eac08f43 x18: 0000000000000000
[ 3515.411529] x17: 0000000000000000 x16: ffffffe5a16b2914
[ 3515.411531] x15: 0000000000000010 x14: 0000000000000010
[ 3515.411532] x13: 00000000003dd3a2 x12: 0000000000010000
[ 3515.411533] x11: ffffffe597abec14 x10: 0000000000000010
[ 3515.411535] x9 : ffffffe597abeba8 x8 : ffffffc013ce0010
[ 3515.411536] x7 : 000000b2b5593519 x6 : 0000000000300000
[ 3515.411537] x5 : 0000000000000000 x4 : 0000000000000032
[ 3515.411539] x3 : 0000000000000000 x2 : 0000000000000004
[ 3515.411540] x1 : 00000000000e0010 x0 : ffffffbd29b7a300
[ 3515.411542] Kernel panic - not syncing: Asynchronous SError Interrupt
[ 3515.411543] CPU: 7 PID: 2849 Comm: kworker/u16:27 Not tainted 5.4.114 #44
[ 3515.411544] Hardware name: MediaTek Asurada rev1 board (DT)
[ 3515.411544] Workqueue: mt76 mt7921_pm_wake_work [mt7921e]
[ 3515.411545] Call trace:
[ 3515.411546]  dump_backtrace+0x0/0x14c
[ 3515.411546]  show_stack+0x20/0x2c
[ 3515.411547]  dump_stack+0xa0/0xfc
[ 3515.411548]  panic+0x154/0x350
[ 3515.411548]  panic+0x0/0x350
[ 3515.411549]  arm64_serror_panic+0x78/0x84
[ 3515.411550]  do_serror+0x0/0x118
[ 3515.411550]  do_serror+0xa4/0x118
[ 3515.411551]  el1_error+0x84/0xf8
[ 3515.411552]  mt76_mmio_rr+0x30/0xf0 [mt76]
[ 3515.411552]  mt7921_rr+0x38/0x44 [mt7921e]
[ 3515.411553]  __mt76_poll_msec+0x5c/0x9c [mt76]
[ 3515.411554]  __mt7921_mcu_drv_pmctrl+0x50/0x94 [mt7921e]
[ 3515.411555]  mt7921_mcu_drv_pmctrl+0x38/0xb0 [mt7921e]
[ 3515.411555]  mt7921_pm_wake_work+0x34/0xd4 [mt7921e]
[ 3515.411556]  process_one_work+0x208/0x3c8
[ 3515.411557]  worker_thread+0x23c/0x3e8
[ 3515.411557]  kthread+0x144/0x178
[ 3515.411558]  ret_from_fork+0x10/0x18
[ 3515.418831] SMP: stopping secondary CPUs
[ 3515.418832] Kernel Offset: 0x2590c00000 from 0xffffffc010000000
[ 3515.418832] PHYS_OFFSET: 0xffffffc400000000
[ 3515.418833] CPU features: 0x080026,2a80aa18
[ 3515.418834] Memory Limit: none
[DL] 00000000 00000000 010701

Fixes: 1d8efc741df80 ("mt76: mt7921: introduce Runtime PM support")
Co-developed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: avoid unnecessary consecutive WiFi resets

Avoid unnecessary consecutive WiFi resets by dropping reset
request when reset work is working.

Co-developed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: connac: update BA win size in Rx direction

Update BA size used data transimission in the Rx direction to improve Rx
throughput.

Reviewed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Leon Yen <leon.yen@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>

mt76: mt7615: add thermal sensor device support

Similar to mt7915, switching to use standard hwmon sysfs.
For reading temperature, cat /sys/class/ieee80211/phy*/hwmon*/temp1_input

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: add thermal cooling device support

Thermal cooling device support is added to control the temperature by
throttling the data transmission for the given duration. Throttling is
done by adjusting Tx period by given percentage of time. The thermal
device allows user to configure duty cycle.

Throttling can be disabled by setting the duty cycle to 0. The cooling
device can be found under /sys/class/thermal/cooling_deviceX/.
Corresponding soft link to this device can be found under phy folder

To set duty cycle as 80%,
echo 80 > /sys/class/ieee80211/phy*/cooling_device/cur_state

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: add thermal sensor device support

This provides userspace with a unified interface, hwmon sysfs, to monitor
temperature in the hardware and can be adapted to system monitoring tools.

For reading temperature, cat /sys/class/ieee80211/phy*/hwmon*/temp1_input

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: add back connection monitor support

Hw beacon cmd to the mt7921 firmware doesn't only filter out the beacon,
but also performs its own connection monitoring, including periodic
keep-alives to the AP and probing the AP on beacon loss. Will indicate
the host with the event when the firmware detects the connection is lost.

Fixes: 1d8efc741df8 ("mt76: mt7921: introduce Runtime PM support")
Reviewed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Deren Wu <deren.wu@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>

mt76: mt7921: consider the invalid value for to_rssi

It is possible the RCPI from the certain antenna is an invalid value,
especially packets are receiving while the system is frequently entering
deep sleep mode, so consider calculating RSSI with the reasonable upper
bound to avoid report the wrong value to the mac80211 layer.

Fixes: 163f4d22c118 ("mt76: mt7921: add MAC support")
Reviewed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: connac: fix WoW with disconnetion and bitmap pattern

Update MCU command usage to fix WoW configuration with disconnection
and bitmap pattern and to avoid magic number.

Fixes: ffa1bf97425b ("mt76: mt7921: introduce PM support")
Reviewed-by: Lorenzo Bianconi <lorenzo@kernel.org>
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>

mt76: mt7921: fix reset under the deep sleep is enabled

To fix possibly the race to access register between the WiFi reset
and the other context that is caused by explicitly cancelling ps_work
and wake_work to break PM_STATE consistency.

Deep sleep would cause the hardware into the inactive state,
so we forcely put device drv_own state before we start to reset.

The patch also ignore the reset request when the procedure is in
progress to avoid the consecutive WiFi resets.

localhost ~ # [ 2932.073966] SError Interrupt on CPU7, code 0xbe000011
[ 2932.073967] CPU: 7 PID: 8761 Comm: kworker/u16:2 Not tainted 5.4.112 #30
[ 2932.073968] Hardware name: MediaTek Asurada rev1 board (DT)
[ 2932.073968] Workqueue: phy0 ieee80211_reconfig_filter [mac80211]
[ 2932.073969] pstate: 80400089 (Nzcv daIf +PAN -UAO)
[ 2932.073969] pc : el1_irq+0x78/0x180
[ 2932.073970] lr : mt76_mmio_rmw+0x30/0x5c [mt76]
[ 2932.073970] sp : ffffffc01142bad0
[ 2932.073970] x29: ffffffc01142bc00 x28: ffffff8f96fb1e00
[ 2932.073971] x27: ffffffd2cdc12138 x26: ffffffd2cdaeb018
[ 2932.073972] x25: 0000000000000000 x24: ffffff8fa8e14c08
[ 2932.073973] x23: 0000000080c00009 x22: ffffffd2a5603918
[ 2932.073974] x21: ffffffc01142bc10 x20: 0000007fffffffff
[ 2932.073975] x19: 0000000000000000 x18: 0000000000000400
[ 2932.073975] x17: 0000000000000400 x16: ffffffd2cd2b87dc
[ 2932.073976] x15: 0000000000000000 x14: 0000000000000000
[ 2932.073977] x13: 0000000000000001 x12: 0000000000000001
[ 2932.073978] x11: 0000000000000001 x10: 000000000010e000
[ 2932.073978] x9 : 0000000000000000 x8 : ffffffc013921404
[ 2932.073979] x7 : 000000b2b5593519 x6 : 0000000000300000
[ 2932.073980] x5 : 0000000000000000 x4 : ffffffc01142bbc8
[ 2932.073980] x3 : 00000000000001f0 x2 : 0000000000000000
[ 2932.073981] x1 : 0000000000021404 x0 : ffffff8fa8e12300
[ 2932.073982] Kernel panic - not syncing: Asynchronous SError Interrupt
[ 2932.073983] CPU: 7 PID: 8761 Comm: kworker/u16:2 Not tainted 5.4.112 #30
[ 2932.073983] Hardware name: MediaTek Asurada rev1 board (DT)
[ 2932.073984] Workqueue: phy0 ieee80211_reconfig_filter [mac80211]
[ 2932.073984] Call trace:
[ 2932.073985]  dump_backtrace+0x0/0x14c
[ 2932.073985]  show_stack+0x20/0x2c
[ 2932.073985]  dump_stack+0xa0/0xf8
[ 2932.073986]  panic+0x154/0x360
[ 2932.073986]  test_taint+0x0/0x44
[ 2932.073986]  arm64_serror_panic+0x78/0x84
[ 2932.073987]  do_serror+0x0/0x118
[ 2932.073987]  do_serror+0xa4/0x118
[ 2932.073987]  el1_error+0x84/0xf8
[ 2932.073988]  el1_irq+0x78/0x180
[ 2932.073988]  mt76_mmio_rr+0x30/0xf0 [mt76]
[ 2932.073988]  mt76_mmio_rmw+0x30/0x5c [mt76]
[ 2932.073989]  mt7921_rmw+0x4c/0x5c [mt7921e]
[ 2932.073989]  mt7921_configure_filter+0x138/0x160 [mt7921e]
[ 2932.073990]  ieee80211_configure_filter+0x2f0/0x3e0 [mac80211]
[ 2932.073990]  ieee80211_reconfig_filter+0x1c/0x28 [mac80211]
[ 2932.073990]  process_one_work+0x208/0x3c8
[ 2932.073991]  worker_thread+0x23c/0x3e8
[ 2932.073991]  kthread+0x140/0x17c
[ 2932.073992]  ret_from_fork+0x10/0x18
[ 2932.074071] SMP: stopping secondary CPUs
[ 2932.074071] Kernel Offset: 0x12bc800000 from 0xffffffc010000000
[ 2932.074072] PHYS_OFFSET: 0xfffffff180000000
[ 2932.074072] CPU features: 0x080026,2a80aa18
[ 2932.074072] Memory Limit: none

Co-developed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: connac: fw_own rely on all packet memory all being free

If the device is MMIO-based, we must ensure all TxD/TxP on the host
memory all being consumed by the device prior to safely switching to
fw_own state.

Fixes: ec7bd7b4a9c0 ("mt76: connac: check wake refcount in mcu_fw_pmctrl")
Reviewed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: Don't alter Rx path classifier

Keep Rx path classifier the mt7921 firmware prefers to allow frames pass
through MCU.

Fixes: 5c14a5f944b9 ("mt76: mt7921: introduce mt7921e support")
Reviewed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: fix mt7921_wfsys_reset sequence

WiFi subsytem reset should control MT_WFSYS_SW_RST_B and then poll the
same register until the bit WFSYS_SW_INIT_DONE bit is set.

Fixes: 0c1ce9884607 ("mt76: mt7921: add wifi reset support")
Reviewed-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Sean Wang <sean.wang@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7663: enable hw rx header translation

As already done for mt7615 and mt7915, enable rx header translation
offload for mt7663 in order to reduce cpu load in the rx path.

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7615: free irq if mt7615_mmio_probe fails

As already done for mt7915 and mt7921, free registered irq line if
mt7615_mmio_probe routine fails

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt76x0: use dev_debug instead of dev_err for hw_rf_ctrl

BIT(0) in MT_EE_NIC_CONF_1 is use to notify the driver if the radio RF
switch is controlled through a gpio. Use dev_debug instead of dev_err
to log this info.

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7615: fix NULL pointer dereference in tx_prepare_skb()

Fix theoretical NULL pointer dereference in mt7615_tx_prepare_skb and
mt7663_usb_sdio_tx_prepare_skb routines. This issue has been identified
by code analysis.

Fixes: 6aa4ed7927f11 ("mt76: mt7615: implement DMA support for MT7622")
Fixes: 4bb586bc33b98 ("mt76: mt7663u: sync probe sampling with rate configuration")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: fix possible NULL pointer dereference in mt76_tx

Even if this is not a real issue since mt76_tx is never run with wcid set
to NULL, fix a theoretical NULL pointer dereference in mt76_tx routine

Fixes: db9f11d3433f7 ("mt76: store wcid tx rate info in one u32 reduce locking")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: add .set_bitrate_mask() callback

Add runtime configuration for bitrate mask. This update firmware rate
control to add a boundary on top of table to limit the rate selection
for each peer, so when user set bitrates vht-mcs-5 1:9, which actually
means nss = 1 mcs = 0~9. This only applies to data frames as for other
mgmt, mcast, bcast still use legacy rates as it is.

Note that driver does not support GI configuration.

Example:
iw dev wlan0 set bitrates vht-mcs-5 1:9 he-mcs-5 2:7
iw dev wlan0 set bitrates legacy-5 6 he-mcs-5 2:0-11

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: cleanup mt7915_mcu_sta_rate_ctrl_tlv()

Remove obsoleted codes. This is the preparation for .set_bitrate_mask().

Signed-off-by: Ryder Lee <ryder.lee@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: enable rx csum offload

As already done for mt7615 and mt7915, enable hw rx checksum offload.

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: enable rx header traslation offload

As already done for mt7615 and mt7915, enable rx header translation
offload for mt7921 in order to reduce cpu load in the rx path.

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: connac: add missing configuration in mt76_connac_mcu_wtbl_hdr_trans_tlv

Add missing configuration parameters in mt76_connac_mcu_wtbl_hdr_trans_tlv
routine

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7921: enable rx hw de-amsdu

Enable hw rx-amsdu de-aggregation support available in 7921 devices.
This is a preliminary patch to enable rx checksum offload.

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: fix a signedness bug in mt7915_mcu_apply_tx_dpd()

"idx" needs to be signed for the error handling to work.

Fixes: 495184ac91bb ("mt76: mt7915: add support for applying pre-calibration data")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: fix tssi indication field of DBDC NICs

Correct the bitfield which indicates TSSI on/off for MT7915D NIC.

Signed-off-by: Evelyn Tsai <evelyn.tsai@mediatek.com>
Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: move mt76_rates in mt76 module

Move mt76_rates array in mt76 module and remove duplicated code since it
is shared by all drivers

Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: add support for tx status reporting

For now, this only reports ACK status

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: rework tx rate reporting

Instead of attaching the last reported rate to tx packets, use
ieee80211_tx_status_ext to immediately pass the rate to mac80211 after
receiving it from the firmware.
Preparation for implementing full tx status reporting

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: dma: use ieee80211_tx_status_ext to free packets when tx fails

Fixes AQL issues on full queues, especially with 802.3 encap offload

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: improve tx status codepath

Use ieee80211_tx_status_ext instead of ieee80211_free_skb and
ieee80211_tx_status. This makes it compatible with 802.3 encap offload
and improves performance by removing a redundant sta lookup

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: intialize tx queue entry wcid to 0xffff by default

Avoid accidentally mapping them to WCID 0 on completion

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7603: avoid use of ieee80211_tx_info_clear_status

It overwrites mt76_tx_cb data in the skb

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7615: avoid use of ieee80211_tx_info_clear_status

It overwrites mt76_tx_cb data in the skb

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7615: fix fixed-rate tx status reporting

Rely on the txs fixed-rate bit instead of info->control.rates

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: move mt7915_queue_rx_skb to mac.c

It is not really DMA specific, and moving it makes it possible to make some
functions in mac.c static

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: disable ASPM

The same is done on the other chips already, so it is very likely needed
on MT7915 as well

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mt76: mt7915: add MSI support

Move IRQ processing to a tasklet, similar to MT7615/MT7663

Signed-off-by: Felix Fietkau <nbd@nbd.name>

Merge tag 'wireless-drivers-next-2021-06-16' of git://git./linux/kernel/git/kvalo/wireless-drivers-next

Kalle Valo says:

====================
wireless-drivers-next patches for v5.14

First set of patches for v5.14. Major new features are here support
WCN6855 PCI in ath11k and WoWLAN support for wcn36xx. Also smaller
fixes and cleanups all over.

ath9k

* provide STBC info in the received frames

brcmfmac

* fix setting of station info chains bitmask

* correctly report average RSSI in station info

rsi

* support for changing beacon interval in AP mode

ath11k

* support for WCN6855 PCI hardware

wcn36xx

* WoWLAN support with magic packets and GTK rekeying
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'marvell-prestera-flower-match-all'

Vadym Kochan says:

====================
Marvell Prestera add flower and match all support

Add ACL infrastructure for Prestera Switch ASICs family devices to
offload cls_flower rules to be processed in the HW.

ACL implementation is based on tc filter api. The flower classifier
is supported to configure ACL rules/matches/action.

Supported actions:

    - drop
    - trap
    - pass

Supported dissector keys:

    - indev
    - src_mac
    - dst_mac
    - src_ip
    - dst_ip
    - ip_proto
    - src_port
    - dst_port
    - vlan_id
    - vlan_ethtype
    - icmp type/code

- Introduce matchall filter support
- Add SPAN API to configure port mirroring.
- Add tc mirror action.

At this moment, only mirror (egress) action is supported.

Example:
    tc filter ... action mirred egress mirror dev DEV

v2:
    Fixed "newline at EOF warnings" from "git am" by
        re-applying with --whitespace=fix

    patch #1:
        1) Set TC HW Offload always enabled without disable it     [suggested by Vladimir Oltean]
           by user. It reduced the logic by removing feature
           handling and acl block disable counting.

    patch #2:
        1) Removed extra not needed diff with prestera_port and    [suggested by Vladimir Oltean]
           prestera_switch  lines exchanging in prestera_acl.h

        2) Fix local variables ordering to reverse chrostmas tree  [suggested by Vladimir Oltean]

        3) Use tc_cls_can_offload_and_chain0() in                  [suggested by Vladimir Oltean]
           prestera_span_replace()

        4) Removed TODO about prio check                           [suggested by Vladimir Oltean]

        5) Rephrase error message if prestera_netdev_check()       [suggested by Vladimir Oltean]
           fails in prestera_span_replace()
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

net: marvell: prestera: Add matchall support

- Introduce matchall filter support
- Add SPAN API to configure port mirroring.
- Add tc mirror action.

At this moment, only mirror (egress) action is supported.

Example:
    tc filter ... action mirred egress mirror dev DEV

Co-developed-by: Volodymyr Mytnyk <vmytnyk@marvell.com>
Signed-off-by: Volodymyr Mytnyk <vmytnyk@marvell.com>
Signed-off-by: Serhiy Boiko <serhiy.boiko@plvision.eu>
Signed-off-by: Vadym Kochan <vkochan@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net: marvell: Implement TC flower offload

Add ACL infrastructure for Prestera Switch ASICs family devices to
offload cls_flower rules to be processed in the HW.

ACL implementation is based on tc filter api. The flower classifier
is supported to configure ACL rules/matches/action.

Supported actions:

    - drop
    - trap
    - pass

Supported dissector keys:

    - indev
    - src_mac
    - dst_mac
    - src_ip
    - dst_ip
    - ip_proto
    - src_port
    - dst_port
    - vlan_id
    - vlan_ethtype
    - icmp type/code

Co-developed-by: Volodymyr Mytnyk <vmytnyk@marvell.com>
Signed-off-by: Volodymyr Mytnyk <vmytnyk@marvell.com>
Signed-off-by: Serhiy Boiko <serhiy.boiko@plvision.eu>
Signed-off-by: Vadym Kochan <vkochan@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'net-smc-stats'

Karsten Graul says:

====================
net/smc: Add SMC statistic support

Please apply the following patch series for smc to netdev's net-next tree.

This v2 is a resend of the code contained in v1 but with an updated
cover letter to describe why we have chosen to use the generic netlink
mechanism to access the smc protocol's statistic data.

The patchset adds statistic support to the SMC protocol. Per-cpu
variables are used to collect the statistic information for better
performance and for reducing concurrency pitfalls. The code that is
collecting statistic data is implemented in macros to increase code
reuse and readability.
The generic netlink mechanism in SMC is extended to provide the
collected statistics to userspace.
Network namespace awareness is also part of the statistics
implementation.

SMC is a protocol interacting with PCI devices (like RoCE Cards) and
runs on top of the TCP protocol. As SMC is a network protocol and not
an ethernet device driver, we decided to use the generic netlink
interface. This should be comparable to what other protocols in the
net subsystem like tipc, ncsi, ieee802154 or tcp, et al, do.
There is already an established internal generic netlink interface
mechanism in SMC which is used to collect SMC Protocol internal
information. This patchset extends that existing mechanism.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

net/smc: Make SMC statistics network namespace aware

Make the gathered SMC statistics network namespace aware, for each
namespace collect an own set of statistic information.

Signed-off-by: Guvenc Gulce <guvenc@linux.ibm.com>
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/smc: Add netlink support for SMC fallback statistics

Add support to collect more detailed SMC fallback reason statistics and
provide these statistics to user space on the netlink interface.

Signed-off-by: Guvenc Gulce <guvenc@linux.ibm.com>
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/smc: Add netlink support for SMC statistics

Add the netlink function which collects the statistics information and
delivers it to the userspace.

Signed-off-by: Guvenc Gulce <guvenc@linux.ibm.com>
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/smc: Add SMC statistics support

Add the ability to collect SMC statistics information. Per-cpu
variables are used to collect the statistic information for better
performance and for reducing concurrency pitfalls. The code that is
collecting statistic data is implemented in macros to increase code
reuse and readability.

Signed-off-by: Guvenc Gulce <guvenc@linux.ibm.com>
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

mlxsw: spectrum_router: remove redundant continue statement

The continue statement at the end of a for-loop has no effect,
remove it.

Addresses-Coverity: ("Continue has no effect")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

Merge branch 'nfp-ct-part-two'

Simon Horman says:

====================
Next set of conntrack patches for the nfp driver

Louis Peens says:

This follows on from the previous series of a similar nature.
Looking at the diagram as explained in the previous series
this implements changes up to the point where the merged
nft entries are saved. There are still bits of stubbed
out code where offloading of the flows will be implemented.

+-------------+                      +----------+
| pre_ct flow +--------+             | nft flow |
+-------------+        v             +------+---+
                  +----------+              |
                  | tc_merge +--------+     |
                  +----------+        v     v
+--------------+       ^           +-------------+
| post_ct flow +-------+       +---+nft_tc merge |
+--------------+               |   +-------------+
                               |
                               |
                               |
                               v
                        Offload to nfp
====================

Signed-off-by: David S. Miller <davem@davemloft.net>

nfp: flower-ct: implement action_merge check

Fill in code stub to check that the flow actions are valid for
merge. The actions of the flow X should not conflict with the
matches of flow X+1. For now this check is quite strict and
set_actions are very limited, will need to update this when
NAT support is added.

Signed-off-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: Yinjun Zhang <yinjun.zhang@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

nfp: flower-ct: fill ct metadata check function

Fill in check_meta stub to check that ct_metadata action fields in
the nft flow matches the ct_match data of the post_ct flow.

Signed-off-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: Yinjun Zhang <yinjun.zhang@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

nfp: flower-ct: fill in ct merge check function

Replace merge check stub code with the actual implementation. This
checks that the match parts of two tc flows does not conflict.
Only overlapping keys needs to be checked, and only the narrowest
masked parts needs to be checked, so each key is masked with the
AND'd result of both masks before comparing.

Signed-off-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: Yinjun Zhang <yinjun.zhang@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

nfp: flower-ct: implement code to save merge of tc and nft flows

Add in the code to merge the tc_merge objects with the flows
received from nft. At the moment flows are just merged blindly
as the validity check functions are stubbed out, this will
be populated in follow-up patches.

Signed-off-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: Yinjun Zhang <yinjun.zhang@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

nfp: flower-ct: add nft_merge table

Add table and struct to save the result of the three-way merge
between pre_ct,post_ct, and nft flows. Merging code is to be
added in follow-up patches.

Signed-off-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: Yinjun Zhang <yinjun.zhang@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

nfp: flower-ct: make a full copy of the rule when it is a NFT flow

The nft flow will be destroyed after offload cb returns. This means
we need save a full copy of it since it can be referenced through
other paths other than just the offload cb, for example when a new
pre_ct or post_ct entry is added, and it needs to be merged with
an existing nft entry.

Signed-off-by: Yinjun Zhang <yinjun.zhang@corigine.com>
Signed-off-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

nfp: flower-ct: add nft flows to nft list

Implement code to add and remove nft flows to the relevant list.
Registering and deregistering the callback function for the nft
table is quite complicated. The safest is to delete the callback
on the removal of the last pre_ct flow. This is because if this
is also the latest pre_ct flow in software it means that this
specific nft table will be freed, so there will not be a later
opportunity to do this. Another place where it looks possible
to delete the callback is when the last nft_flow is deleted,
but this happens under the flow_table lock, which is also taken
when deregistering the callback, leading to a deadlock situation.

This means the final solution here is to delete the callback
when removing the last pre_ct flow, and then clean up any
remaining nft_flow entries which may still be present, since
there will never be a callback now to do this, leaving them
orphaned if not cleaned up here as well.

Signed-off-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: Yinjun Zhang <yinjun.zhang@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

nfp: flower-ct: add nft callback stubs

Add register/unregister of the nft callback. For now just add
stub code to accept the flows, but don't do anything with it.
Decided to accept the flows since netfilter will keep on trying
to offload a flow if it was rejected, which is quite noisy.
Follow-up patches will start implementing the functions to add
nft flows to the relevant tables.

Signed-off-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: Yinjun Zhang <yinjun.zhang@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

nfp: flower-ct: add delete flow handling for ct

Add functions to handle delete flow callbacks for ct flows. Also
accept the flows for offloading by returning 0 instead of -EOPNOTSUPP.
Flows will still not actually be offloaded to hw, but at this point
it's difficult to not accept the flows and also exercise the cleanup
paths properly. Traffic will still be handled safely through the
fallback path.

Signed-off-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: Yinjun Zhang <yinjun.zhang@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>