Merge tag 'ath-next-20240605' of git://git./linux/kernel/git/kvalo/ath

ath.git patches for v6.11

ath12k

* remove unsupported tx monitor handling

* channel 2 in 6 GHz band support

* Spatial Multiplexing Power Save (SMPS) in 6 GHz band support

* multiple BSSID (MBSSID) and Enhanced Multi-BSSID Advertisements (EMA) support

* dynamic VLAN support

* add panic handler for resetting the firmware state

ath10k

* add qcom,no-msa-ready-indicator Device Tree property

* LED support for various chipsets

Merge tag 'rtw-next-2024-06-04' of https://github.com/pkshih/rtw

rtw-next patches for v6.11

Some fixes and refactors of rtlwifi, rtw88 and rtw89. Only one major change
listed below:

rtlwifi:
 - add new chip support of RTL8192DU

wifi: ath12k: add hw_link_id in ath12k_pdev

Currently, hw_link_id is sent in WMI service ready event but it
is not parsed anywhere.

But, in future, for multi-link operation, this parameter would be
needed by many WMI commands such as WMI beacon template
(WMI_BCN_TMPL_CMDID), WMI vdev start for Multi-link virtual AP
interfaces (WMI_VDEV_START_REQUEST_CMDID), WMI peer assoc command
(WMI_PEER_ASSOC_CMDID) for Multi-link peer and so on.

Hence, add changes to parse and store the hw_link_id received in
WMI service ready event in ath12k_pdev structure.

Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.0.1-00029-QCAHKSWPL_SILICONZ-1

Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com>
Signed-off-by: Harshitha Prem <quic_hprem@quicinc.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240529054955.4105240-1-quic_hprem@quicinc.com

wifi: ath12k: add panic handler

Currently for ath12k PCI devices, firmware could be running in one of
several execution environments, e.g., PBL, SBL and mission mode etc. Among
which PBL is the only stage where PCIe link negotiation could happen. So
normally firmware runs in PBL in order to be enumerated during system reboot.

However it might not work in kernel crash scenario: ath12k target is not
found after warm reboot from kernel crash. This is because when kernel crashes,
ath12k host does nothing to firmware. And during warm reboot, WLAN power
is sustained. So firmware is likely to keep running in mission mode throughout
the bootup process. As a result PCIe link is not established and thus target
not enumerated.

So add a handler in panic notification list for ath12k. When kernel crashes,
this handler gets called and tries to reset target to PBL state. Then PCIe
link negotiation could happen and target gets enumerated.

This change applies to all PCI devices including WCN7850 and QCN9274.

Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-03427-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1.15378.4

Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240529021533.10861-1-quic_bqiang@quicinc.com

wifi: rtw89: chan: Use swap() in rtw89_swap_sub_entity()

Use existing swap() function instead of keeping duplicate source code.

./drivers/net/wireless/realtek/rtw89/chan.c:2336:32-33: WARNING opportunity for swap().

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Closes: https://bugzilla.openanolis.cn/show_bug.cgi?id=9174
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/20240529020244.129027-1-jiapeng.chong@linux.alibaba.com

wifi: brcm80211: remove unused structs

'brcmf_pcie_core_info' was added in
commit 9e37f045d5e7 ("brcmfmac: Adding PCIe bus layer support.")
but never used.

'brcms_c_bit_desc' last use was removed in
commit cdf4352f5c59 ("brcmsmac: Improve tx trace and debug support").

Remove them.

Signed-off-by: Dr. David Alan Gilbert <linux@treblig.org>
Acked-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/20240526234553.286773-1-linux@treblig.org

wifi: brcm80211: use sizeof(*pointer) instead of sizeof(type)

It is preferred to use sizeof(*pointer) instead of sizeof(type)
due to the type of the variable can change and one needs not
change the former (unlike the latter). This patch has no effect
on runtime behavior.

At the same time remove some redundant NULL initializations.

Acked-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Erick Archer <erick.archer@outlook.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://msgid.link/AS8PR02MB7237FF1C2E880D1231684D708BF02@AS8PR02MB7237.eurprd02.prod.outlook.com

wifi: ath12k: do not process consecutive RDDM event

Currently we do reset for each RDDM event from MHI, however there are
cases, see below log, that we get two or more consecutive events, and
it is pointless to do reset for the subsequent ones. What's more, it
makes reset process more likely to fail.

[  103.289864] mhi mhi0: System error detected
[  103.289871] ath12k_pci 0000:03:00.0: mhi notify status reason MHI_CB_EE_RDDM
[  103.293144] mhi mhi0: System error detected
[  103.293150] ath12k_pci 0000:03:00.0: mhi notify status reason MHI_CB_EE_RDDM

Add a check to avoid reset again and again. This is done by tracking previous
MHI status: if we receive a new RDDM event while the previous event is
also the same, we treat it as duplicate and ignore it, because normally
we should at least receive a MHI_CB_EE_MISSION_MODE event between them.

Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0.c5-00481-QCAHMTSWPL_V1.0_V2.0_SILICONZ-3

Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240529015357.9846-1-quic_bqiang@quicinc.com

dt-bindings: net: wireless: ath11k: Drop "qcom,ipq8074-wcss-pil" from example

Convention for examples is to only show what's covered by the binding,
so drop the provider "qcom,ipq8074-wcss-pil". It is also not documented
by a schema which caused a warning.

Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240528134610.4075204-1-robh@kernel.org

wifi: ath12k: fix memory leak in ath12k_dp_rx_peer_frag_setup()

Currently the resource allocated by crypto_alloc_shash() is not
freed in case ath12k_peer_find() fails, resulting in memory leak.

Add crypto_free_shash() to fix it.

This is found during code review, compile tested only.

Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240526124226.24661-1-quic_bqiang@quicinc.com

wifi: rtlwifi: handle return value of usb init TX/RX

Handle error code to cause failed to USB probe result from unexpected
USB EP number, otherwise when USB disconnect skb_dequeue() an uninitialized
skb list and cause warnings below.

usb 2-1: USB disconnect, device number 76
INFO: trying to register non-static key.
The code is fine but needs lockdep annotation, or maybe
you didn't initialize this object before use?
turning off the locking correctness validator.
CPU: 0 PID: 54060 Comm: kworker/0:1 Not tainted 6.9.0-rc7 #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.2-0-gea1b7a073390-prebuilt.qemu.org 04/01/2014
Workqueue: usb_hub_wq hub_event
Call Trace:
 <TASK>
 __dump_stack lib/dump_stack.c:88 [inline]
 dump_stack_lvl+0x116/0x1b0 lib/dump_stack.c:114
 assign_lock_key kernel/locking/lockdep.c:976 [inline]
 register_lock_class+0xc18/0xfa0 kernel/locking/lockdep.c:1289
 __lock_acquire+0x108/0x3bc0 kernel/locking/lockdep.c:5014
 lock_acquire kernel/locking/lockdep.c:5754 [inline]
 lock_acquire+0x1b0/0x550 kernel/locking/lockdep.c:5719
 __raw_spin_lock_irqsave include/linux/spinlock_api_smp.h:110 [inline]
 _raw_spin_lock_irqsave+0x3d/0x60 kernel/locking/spinlock.c:162
 skb_dequeue+0x20/0x180 net/core/skbuff.c:3846
 rtl_usb_cleanup drivers/net/wireless/realtek/rtlwifi/usb.c:706 [inline]
 rtl_usb_deinit drivers/net/wireless/realtek/rtlwifi/usb.c:721 [inline]
 rtl_usb_disconnect+0x4a4/0x850 drivers/net/wireless/realtek/rtlwifi/usb.c:1051
 usb_unbind_interface+0x1e8/0x980 drivers/usb/core/driver.c:461
 device_remove drivers/base/dd.c:568 [inline]
 device_remove+0x122/0x170 drivers/base/dd.c:560
 __device_release_driver drivers/base/dd.c:1270 [inline]
 device_release_driver_internal+0x443/0x620 drivers/base/dd.c:1293
 bus_remove_device+0x22f/0x420 drivers/base/bus.c:574
 device_del+0x395/0x9f0 drivers/base/core.c:3909
 usb_disable_device+0x360/0x7b0 drivers/usb/core/message.c:1418
 usb_disconnect+0x2db/0x930 drivers/usb/core/hub.c:2305
 hub_port_connect drivers/usb/core/hub.c:5362 [inline]
 hub_port_connect_change drivers/usb/core/hub.c:5662 [inline]
 port_event drivers/usb/core/hub.c:5822 [inline]
 hub_event+0x1e39/0x4ce0 drivers/usb/core/hub.c:5904
 process_one_work+0x97b/0x1a90 kernel/workqueue.c:3267
 process_scheduled_works kernel/workqueue.c:3348 [inline]
 worker_thread+0x680/0xf00 kernel/workqueue.c:3429
 kthread+0x2c7/0x3b0 kernel/kthread.c:388
 ret_from_fork+0x45/0x80 arch/x86/kernel/process.c:147
 ret_from_fork_asm+0x1a/0x30 arch/x86/entry/entry_64.S:244
 </TASK>

Reported-by: Shichao Lai <shichaorai@gmail.com>
Closes: https://lore.kernel.org/linux-wireless/CAEk6kZuuezkH1dVRJf3EAVZK-83=OpTz62qCugkpTkswj8JF6w@mail.gmail.com/T/#u
Tested-by: Shichao Lai <shichaorai@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/20240524003248.5952-1-pkshih@realtek.com

wifi: rtlwifi: Enable the new rtl8192du driver

The RTL8192DU is an older Wifi 4 dual band chip. It comes in two
flavours: single MAC single PHY (like most Realtek Wifi 4 USB devices),
and dual MAC dual PHY.

The single MAC single PHY version is 2T2R and can work either in the
2.4 GHz band or the 5 GHz band.

The dual MAC dual PHY version has two USB interfaces and appears to the
system as two separate 1T1R Wifi devices, one working in the 2.4 GHz
band, the other in the 5 GHz band.

This was tested only with a single MAC single PHY device, mostly in
station mode. The speeds in the 2.4 GHz band with 20 MHz channel width
are similar to the out-of-tree driver: 85/51 megabits/second.

Stefan Lippers-Hollmann tested the speed in the 5 GHz band with 40 MHz
channel width: 173/99 megabits/second.

It was also tested briefly in AP mode. It's emitting beacons and my
phone can connect to it.

Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/5f2da7ee-876a-42fc-8fec-ec5386fa8c26@gmail.com

wifi: rtlwifi: Add rtl8192du/sw.c

This contains the new module's entry point.

Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/1fdc83ac-4a6f-4645-8837-6e38683b6c84@gmail.com

wifi: rtlwifi: Constify rtl_hal_cfg.{ops,usb_interface_cfg} and rtl_priv.cfg

This allows the drivers to declare the structs rtl_hal_cfg, rtl_hal_ops,
and rtl_hal_usbint_cfg as const.

Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/1b29994f-3d07-4297-875d-57c3a87a1ec6@gmail.com

wifi: rtlwifi: Add rtl8192du/dm.{c,h}

These contain functions related to the dynamic mechanism, which runs
every two seconds to adjust to changes in the environment.

Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/a010ed6d-efb0-456d-bd4a-dfc6da23e7aa@gmail.com

wifi: rtlwifi: Add rtl8192du/fw.{c,h} and rtl8192du/led.{c,h}

fw.c contains a function for loading the firmware.
led.c contains a function for controlling the LED.

Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/a52b8907-267c-4811-90ca-5eb852a2aa09@gmail.com

wifi: rtlwifi: Add rtl8192du/rf.{c,h}

These contain one RF configuration function and some functions related
to dual MAC operation.

Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/36258d72-11fd-4f70-84c6-782972f706ed@gmail.com

wifi: rtlwifi: Add rtl8192du/trx.{c,h}

These contain routines related to sending frames to the chip.

Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/bd2c24a0-a1fe-493a-8079-d83fda81a0d5@gmail.com

wifi: rtlwifi: Add rtl8192du/phy.{c,h}

These contain mostly the calibration and channel switching routines
for RTL8192DU.

Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/299a5ab7-46bf-410b-bbeb-e252eeeeb34c@gmail.com

wifi: rtlwifi: Add rtl8192du/hw.{c,h}

These contain mostly hardware init/deinit routines for RTL8192DU.

Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/282afe18-f37d-45f7-9275-68c39d1fde09@gmail.com

wifi: rtlwifi: Add new members to struct rtl_priv for RTL8192DU

These are needed for the dual MAC version of RTL8192DU.

The two mutexes are used to avoid concurrent access to the hardware
from the two USB interfaces.

The two arrays are filled by one interface during LC calibration and
accessed by the other interface during channel switching.

Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/d1bf581b-de41-4d09-bf56-768c2fc29dd3@gmail.com

wifi: rtlwifi: Add rtl8192du/table.{c,h}

These contain the MAC, BB, RF, and AGC initialisation tables for
RTL8192DU.

Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/5ea548b9-5ec6-4358-aa5d-94071f2341cb@gmail.com

wifi: rtlwifi: rtl8192d: Use "rtl92d" prefix

Some functions moved from rtl8192de still use the "rtl92de" prefix.
Rename them.

Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/abe3624d-f986-4a24-966c-67defa962e3f@gmail.com

wifi: nl80211: clean up coalescing rule handling

There's no need to allocate a tiny struct and then
an array again, just allocate the two together and
use __counted_by(). Also unify the freeing.

Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240523120213.48a40cfb96f9.Ia02bf8f8fefbf533c64c5fa26175848d4a3a7899@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: handle HW restart during ROC

If we have a HW restart in the middle of a ROC period,
then there are two cases:
 - if it's a software ROC, we really don't need to do
   anything, since the ROC work will still be queued
   and will run later, albeit with the interruption
   due to the restart;
 - if it's a hardware ROC, then it may have begun or
   not, if it did begin already we can only remove it
   and tell userspace about that.

In both cases, this fixes the warning that would appear
in ieee80211_start_next_roc() in this case.

In the case of some drivers such as iwlwifi, the part of
restarting is never going to happen since the driver will
cancel the ROC, but flushing the work to ensure nothing
is pending here will also result in no longer being able
to trigger the warning in this case.

Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240523120352.f1924b5411ea.Ifc02a45a5ce23868dc7e428bad8d0e6996dd10f4@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: check ieee80211_bss_info_change_notify() against MLD

It's not valid to call ieee80211_bss_info_change_notify() with
an sdata that's an MLD, remove the FIXME comment (it's not true)
and add a warning.

Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240523121140.97a589b13d24.I61988788d81fb3cf97a490dfd3167f67a141d1fd@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: move Bz and Gl iwl_dev_info entries

The iwl_dev_info entries for these devices were incorrectly positioned
within the array, out of chronological order. Move them arround.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Link: https://msgid.link/20240527190228.a75e9a1eb86c.I003980f8ca60a96019657e396e848a07df6b92a4@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: move amsdu_size parsing to iwlwifi

The code that is parsing the amsdu_size module parameter and mapping it
to the corresponding Rx buffer size is common to all opmodes.
Move it into a function in iwlwifi, as preparation to a new op mode we
are working on.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Link: https://msgid.link/20240527190228.a3430d32923d.Iab3c22ef0df685f72f22dafc47021f0dc7bd6fa5@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: remove redundant prints

Upon start, the different opmodes are currently printing:
1. HW rev, which is already print by iwlwifi
2. The HW name (e.g. "Intel(R) Wi-Fi 7 BE201 320MHz")

cleanup things such as the hw rev won't be printed again, the HW name will
be printed by iwlwifi instead of each opmode.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Link: https://msgid.link/20240527190228.5c589c7abed0.I3590c4ee0ee99d1b207852c32d25d326afb327dd@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: fix a wrong comment

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Link: https://msgid.link/20240527190228.2105ffcf747d.Ic838959b812b6cb4cbb856e8c0bcaad2f46ac71b@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: declare band variable in the scope

band is not used outside of the for loop, declare it inside it.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Link: https://msgid.link/20240527190228.b634d184980a.I9d9e71125f9a77ba4f9a33a7a2ff1e8e78e17767@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: show full firmware ID in debugfs

The firmware prefix is unused today, but it might still be
useful to have some information. Since the prefix will get
the intel/ directory in some cases, replace the prefix by
the full FW id which has all the information.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.527a1c72996d.If7588b854149d51605031fc9a70a650534351ef4@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: add debug data for MPDU counting

It is hard to debug issues of EMLSR entry/exit due to low throughput.
Add debug data.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Link: https://msgid.link/20240527190228.77ef3c2654dc.I1796a3995da2a49dd5102d33766af1ad416dd60b@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: use only beacon BSS load for active links

For active links, don't take information that may have been
received in probe responses, as those are not protected. For
inactive links, there may not be a choice.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.6947dc462fa9.I2076961211d6785c8a15b4308e0e87a413148222@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: add support for version 10 of the responder config command

This version adds the band to the responder config command.

Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.af09129ebd74.I9356e2504a4c19961d4856494416ae49b36bfe62@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: add support for version 14 of the range request command

This version adds a testing option to send an incorrect SAC in the
first NDPA.

Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.8f36ccb1cc74.I9b76461634f308e75dc3cd016d0b0bce812c6e9d@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: modify iwl_mvm_ftm_set_secured_ranging() parameters

Modify iwl_mvm_ftm_set_secured_ranging() parameters to support
multiple versions of the target struct.
This is done as preparation for moving to the new range request
version.

Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.56d6aee320b3.I5a52fa93cd791d0229b392a20f076b7cebb110cd@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: initiator: move setting the sta_id into a function

Move setting the target's sta_id (and related flags) into a function
to support different versions of the target struct.
This is done as preparation for moving to the new range request
version.

Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.18e3a6a6f1cb.I85e3ee607b3947448532bc16730f8898a11c92b8@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: Remove debug related code

This code is intended for internal testing. Remove it.

Fixes: 30ce039094b5 ("wifi: iwlwifi: mvm: Don't allow EMLSR when the RSSI is low")
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.c3b89828efb2.I7ded6348d2fb0cb7e103c4eee033451924c3461d@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: initiator: move setting target flags into a function

Move setting the target flags into a dedicated function to support
different versions of the target struct.
This is done as preparation for moving to the new range request
version.

Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.9d22b61ce589.I7dbe596b4f677638d9a48c3f39b0826a9e35bea4@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: handle TPE advertised by AP

6 GHz BSS SP client shall respect TX power limits advertised
by the AP in TPE elements, send the data to the firmware using
the AP_TX_POWER_CONSTRAINTS_CMD command, so do that.

Co-developed-by: Mukesh Sisodiya <mukesh.sisodiya@intel.com>
Signed-off-by: Mukesh Sisodiya <mukesh.sisodiya@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.32f1e2e1447c.I58ac91c38585362aa42bb4a8a59c7d88e67bc40b@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: move a constant to constants.h

Move IWL_MVM_MIN_BEACON_INTERVAL_TU to constants.h

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.395f5b24ff82.Id78a5a404c3ae15cbbc950b1f1d7e500a8b92a43@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: remove IWL_MVM_USE_NSSN_SYNC

This is not used and not needed anymore.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.c5edfb967e35.I34a3d54a1dc90d9b2937b0bcbbc508fe3cd41773@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: Enable p2p low latency

Enable p2p low latency by default.

Signed-off-by: Avri Altman <avri.altman@intel.com>
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.21b431bdfc26.I018e330d1e6a76e967b25e9542c2260f0f6ed2a8@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: report 64-bit radiotap timestamp

Use the new RX_FLAG_MACTIME_IS_RTAP_TS64 in mac80211 to report
a 64-bit timestamp in the mactime for radiotap only, in case we
report the synchronized PTP clock timestamp (otherwise we really
only have 32 bits anyway).

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Reviewed-by: Gregory Greenman <gregory.greenman@intel.com>
Reviewed-by: Benjamin Berg <benjamin.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.4fc44356e7df.Ic8642d35f1090a415e09299a1bd409f51dfb6351@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: add an option to use ptp clock for rx timestamp

Add a debugfs option to use the ptp clock time for rx device
timestamps. This can be useful to e.g. synchronize multiple NICs or
reporting the timestamp in the system clock instead of gp2.

Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Reviewed-by: Gregory Greenman <gregory.greenman@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240527190228.b197b4f3f9ed.I578ee916d5f517a2e3a89e890a8ea24065946427@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: ath12k: fix wrong definition of CE ring's base address

Base address of CE ring is defined as u32, currently this works
because DMA mask configured as 32 bit:

#define ATH12K_PCI_DMA_MASK 32

However this mask could be changed once firmware bugs are fixed
to fully support 36 bit DMA addressing. So to protect against any
future changes to the DMA mask, change the type of the fields that
are dependent upon it.

This is found during code review. Compile tested only.

Fixes: d889913205cf ("wifi: ath12k: driver for Qualcomm Wi-Fi 7 devices")
Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240524024021.37711-1-quic_bqiang@quicinc.com

wifi: ath12k: fix race due to setting ATH12K_FLAG_EXT_IRQ_ENABLED too early

Commit 5082b3e3027e ("wifi: ath11k: fix race due to setting
ATH11K_FLAG_EXT_IRQ_ENABLED too early") fixes a race in ath11k
driver. Since ath12k shares the same logic as ath11k, currently
the race also exists in ath12k: in ath12k_pci_ext_irq_enable(),
ATH12K_FLAG_EXT_IRQ_ENABLED is set before NAPI is enabled.
In cases where only one MSI vector is allocated, this results
in a race condition: after ATH12K_FLAG_EXT_IRQ_ENABLED is set
but before NAPI enabled, CE interrupt breaks in. Since IRQ is
shared by CE and data path, ath12k_pci_ext_interrupt_handler()
is also called where we call disable_irq_nosync() to disable
IRQ. Then napi_schedule() is called but it does nothing because
NAPI is not enabled at that time, meaning that
ath12k_pci_ext_grp_napi_poll() will never run, so we have
no chance to call enable_irq() to enable IRQ back. Since IRQ
is shared, all interrupts are disabled and we would finally
get no response from target.

So port ath11k fix here, this is done by setting
ATH12K_FLAG_EXT_IRQ_ENABLED after all NAPI and IRQ work are
done. With the fix, we are sure that by the time
ATH12K_FLAG_EXT_IRQ_ENABLED is set, NAPI is enabled.

Note that the fix above also introduce some side effects:
if ath12k_pci_ext_interrupt_handler() breaks in after NAPI
enabled but before ATH12K_FLAG_EXT_IRQ_ENABLED set, nothing
will be done by the handler this time, the work will be
postponed till the next time the IRQ fires.

This is found during code review.

Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0.c5-00481-QCAHMTSWPL_V1.0_V2.0_SILICONZ-3

Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240524023642.37030-1-quic_bqiang@quicinc.com

wifi: ath11k: fix wrong definition of CE ring's base address

Base address of CE ring is defined as u32, currently this works
because coherent DMA mask configured as 32 bit:

#define ATH11K_PCI_COHERENT_DMA_MASK 32

However this mask could be changed once firmware bugs are fixed
to fully support 36 bit DMA addressing. So to protect against any
future changes to the DMA mask, change the type of the fields that
are dependent upon it.

This is found during code review. Compile tested only.

Fixes: d5c65159f289 ("ath11k: driver for Qualcomm IEEE 802.11ax devices")
Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240524021558.34452-1-quic_bqiang@quicinc.com

wifi: ath12k: Fix devmem address prefix when logging

Currently when ath12k QMI logging is enabled, messages such as the
following can be logged:

ath12k_pci 0000:03:00.0: devmem [0] start ox113000 size 20480

Replace ox% with 0x% to get a proper hex address prefix:

ath12k_pci 0000:03:00.0: devmem [0] start 0x113000 size 20480

Signed-off-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240523-ox-v1-1-112ae7350059@quicinc.com

wifi: rtw89: Fix P2P behavior for WiFi 7 chips

Previously we used CCK rate when scanning for P2P on WiFi 7 chips.
Fix this by explicitly setting the rate to OFDM 6Mbps.

Signed-off-by: Po-Hao Huang <phhuang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/20240522024531.11401-1-pkshih@realtek.com

wifi: rtw89: add LDPC and STBC to rx_status and radiotap known fields for monitor mode

Add LDPC and STBC to rx_status by hardware PPDU status, and set them as
known fields of radiotap, so we can check these capabilities in monitor
mode.

Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/20240522023830.9798-2-pkshih@realtek.com

wifi: rtw89: fill STBC and LDPC capabilities to TX descriptor

Fill STBC and LDPC fields to get better performance when transmitting under
noisy environment. Old chips like 8852AE use firmware command to enable
these two capabilities, but newer chips must use TX descriptor.

Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/20240522023830.9798-1-pkshih@realtek.com

wifi: rtw89: 8852c: correct logic and restore PCI PHY EQ after device resume

PCI PHY EQ value is missing after card off/on, so update the value after
device resume. The original commit only updates once at probe stage, which
could lead problem after suspend/resume.

The logic should be read a value from one register and write to another
register with a mask to avoid affecting unrelated bits.

Fixes: a78d33a1286c ("wifi: rtw89: 8852c: disable PCI PHY EQ to improve compatibility")
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://msgid.link/20240521040139.20311-1-pkshih@realtek.com

wifi: ath12k: fix firmware crash during reo reinject

When handling fragmented packets, the ath12k driver reassembles each
fragment into a normal packet and then reinjects it into the HW ring.
However, a firmware crash occurs during this reinjection process.
The issue arises because the driver populates peer metadata in
reo_ent_ring->queue_addr_lo, while the firmware expects the physical
address obtained from the corresponding peer’s queue descriptor. Fix it
by filling peer's queue descriptor's physical address in queue_addr_lo.

Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.1.1-00209-QCAHKSWPL_SILICONZ-1

Fixes: d889913205cf ("wifi: ath12k: driver for Qualcomm Wi-Fi 7 devices")
Signed-off-by: P Praneesh <quic_ppranees@quicinc.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240520070045.631029-4-quic_ppranees@quicinc.com

wifi: ath12k: fix invalid memory access while processing fragmented packets

The monitor ring and the reo reinject ring share the same ring mask index.
When the driver receives an interrupt for the reo reinject ring, the
monitor ring is also processed, leading to invalid memory access. Since
monitor support is not yet enabled in ath12k, the ring mask for the monitor
ring should be removed.

Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.1.1-00209-QCAHKSWPL_SILICONZ-1

Fixes: d889913205cf ("wifi: ath12k: driver for Qualcomm Wi-Fi 7 devices")
Signed-off-by: P Praneesh <quic_ppranees@quicinc.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240520070045.631029-3-quic_ppranees@quicinc.com

wifi: ath12k: change DMA direction while mapping reinjected packets

For fragmented packets, ath12k reassembles each fragment as a normal
packet and then reinjects it into HW ring. In this case, the DMA
direction should be DMA_TO_DEVICE, not DMA_FROM_DEVICE. Otherwise,
an invalid payload may be reinjected into the HW and
subsequently delivered to the host.

Given that arbitrary memory can be allocated to the skb buffer,
knowledge about the data contained in the reinjected buffer is lacking.
Consequently, there’s a risk of private information being leaked.

Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.1.1-00209-QCAHKSWPL_SILICONZ-1

Fixes: d889913205cf ("wifi: ath12k: driver for Qualcomm Wi-Fi 7 devices")
Co-developed-by: Baochen Qiang <quic_bqiang@quicinc.com>
Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
Signed-off-by: P Praneesh <quic_ppranees@quicinc.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240520070045.631029-2-quic_ppranees@quicinc.com

wifi: ath11k: restore country code during resume

We got report that regdomain is not correct after
return from hibernation:

Before hibernation:
% iw reg get
[...]
phy#0 (self-managed)
country CH: DFS-ETSI
        (2402 - 2482 @ 40), (N/A, 20), (N/A)
        (5170 - 5250 @ 80), (N/A, 23), (N/A), NO-OUTDOOR, AUTO-BW
        (5250 - 5330 @ 80), (N/A, 23), (0 ms), NO-OUTDOOR, DFS, AUTO-BW
        (5490 - 5590 @ 80), (N/A, 30), (0 ms), DFS, AUTO-BW
        (5590 - 5650 @ 40), (N/A, 30), (600000 ms), DFS, AUTO-BW
        (5650 - 5730 @ 80), (N/A, 30), (0 ms), DFS, AUTO-BW
        (5735 - 5875 @ 80), (N/A, 14), (N/A), AUTO-BW

After hibernation:
% iw reg get
[...]
phy#0 (self-managed)
country na: DFS-UNSET
        (2402 - 2472 @ 40), (N/A, 20), (N/A)
        (2457 - 2482 @ 20), (N/A, 20), (N/A), PASSIVE-SCAN
        (5170 - 5330 @ 160), (N/A, 20), (N/A), AUTO-BW, PASSIVE-SCAN
        (5490 - 5730 @ 160), (N/A, 20), (N/A), AUTO-BW, PASSIVE-SCAN
        (5735 - 5895 @ 160), (N/A, 20), (N/A), AUTO-BW, PASSIVE-SCAN
        (5945 - 7125 @ 160), (N/A, 30), (N/A), AUTO-BW, PASSIVE-SCAN

The reason is, during resume, firmware is reinitialized but host does
not send current country code to firmware. So default reg rules with
country code set to 'na' is uploaded to host, as shown above.

Fix it by restoring country code to firmware during resume.

Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.30

Fixes: 166a490f59ac ("wifi: ath11k: support hibernation")
Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240520024148.5472-3-quic_bqiang@quicinc.com

wifi: ath11k: refactor setting country code logic

ath11k_wmi_send_set_current_country_cmd() is called in several places
and all of them are just simply repeating the same logic.

Refactor to make code clean.

Compile tested only.

Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240520024148.5472-2-quic_bqiang@quicinc.com

wifi: ath12k: improve the rx descriptor error information

The ath12k_dp_get_rx_desc() failure log currently contains the same
information across multiple Rx data paths and lacks sufficient detail for
debugging purposes. To address this, change the ath12k_dp_get_rx_desc()
failure log to include cookie information along with Rx path details.
This will provide more specific data for debugging purposes.

Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.0.1-00029-QCAHKSWPL_SILICONZ-1
Tested-on: WCN7850 HW2.0 PCI WLAN.HMT.1.0.c5-00481-QCAHMTSWPL_V1.0_V2.0_SILICONZ-3

Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240516000807.1704913-3-quic_periyasa@quicinc.com

wifi: ath12k: refactor rx descriptor CMEM configuration

Currently, data path Rx descriptor placed in the CMEM memory respective
to the Rx offset. This Rx descriptor CMEM placement is not meeting the
requirements to support inter device MLO. In inter device MLO, multiple
devices participate in a group. The device specific Rx descriptor buffers
transmit to multiple device REO rings. To distinguish between device
specific Rx descriptor buffers, the CMEM setup configuration need to
choose a different placement based on the unique identifier (device index).

Below are the CMEM configuration placement representation:

Intra-Device MLO scenario:

        Device 0                                Device 1

     +-------------+                         +-------------+
     |             |                         |             |
     |     Tx      |                         |     Tx      |
     |             |                         |             |
     +-------------+                         +-------------+
     |             |                         |             |
     |     Rx      |                         |     Rx      |
     |             |                         |             |
     +-------------+                         +-------------+

Inter-Device MLO scenario:

        Device 0                                Device 1

     +-------------+                         +-------------+
     |             |                         |             |
     |     Tx      |                         |     Tx      |
     |             |                         |             |
     +-------------+                         +-------------+
     |             |                         |             |
     |     Rx      |                         |   Reserved  |
     |  Device 0   |                         |             |
     |             |                         |             |
     +-------------+                         +-------------+
     |             |                         |     Rx      |
     |   Reserved  |                         |   Device 1  |
     |             |                         |             |
     +-------------+                         +-------------+

Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.0.1-00029-QCAHKSWPL_SILICONZ-1
Tested-on: WCN7850 HW2.0 PCI WLAN.HMT.1.0.c5-00481-QCAHMTSWPL_V1.0_V2.0_SILICONZ-3

Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com>
Link: https://msgid.link/20240516000807.1704913-2-quic_periyasa@quicinc.com

wifi: mac80211: send DelBA with correct BSSID

In MLO, the deflink BSSID is clearly invalid. Since we fill
the addresses as MLD addresses and translate later, use the
AP address here instead.

This fixes an issue that happens with HW restart, where the
DelBA frame is transmitted, but not processed correctly due
to the wrong BSSID (or even just discarded entirely).  As a
result, the BA sessions are kept alive; however, as other
state is reset during HW restart, this then fails (reorder,
etc.) and data doesn't go through until new BA sessions are
established.

Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240510112601.f4e1effdea29.I98e81f22166b68d4b6211191bcaaf8531b324a77@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: reset negotiated TTLM on disconnect

The negotiated TTLM data must be reset on disconnect, otherwise
it may end up getting reused on another connection. Fix that.

Fixes: 8f500fbc6c65 ("wifi: mac80211: process and save negotiated TID to Link mapping request")
Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240506211858.04142e8fe01c.Ia144457e086ebd8ddcfa31bdf5ff210b4b351c22@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: don't stop TTLM works again

There's no need to stop works that have already been
stopped during disconnect, so don't.

Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240506211034.f8434be19f56.I021afadc538508da3bc8f95c89f424ca62b94bef@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: cancel TTLM teardown work earlier

It shouldn't be possible to run this after disconnecting, so
cancel the work earlier.

Fixes: a17a58ad2ff2 ("wifi: mac80211: add support for tearing down negotiated TTLM")
Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240506211034.096a10ccebec.I5584a21c27eb9b3e87b9e26380b627114b32ccba@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: cancel multi-link reconf work on disconnect

This work shouldn't run after we're disconnecting. Cancel it earlier
(and then don't cancel it in stop later.)

Fixes: 8eb8dd2ffbbb ("wifi: mac80211: Support link removal using Reconfiguration ML element")
Reviewed-by: Ilan Peer <ilan.peer@intel.com>
Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240506211034.ac754794279f.Ib9fbb1dab50c6b67f6de9be09a6c452ce89bbd50@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: fix TTLM teardown work

The worker calculates the wrong sdata pointer, so if it ever
runs, it'll crash. Fix that.

Fixes: a17a58ad2ff2 ("wifi: mac80211: add support for tearing down negotiated TTLM")
Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240506211853.e6471800c76d.I8b7c2d6984c89a11cd33d1a610e9645fa965f6e1@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: cfg80211: use __counted_by where appropriate

Annotate 'sub_specs' of 'struct cfg80211_sar_specs', 'channels'
of 'struct cfg80211_sched_scan_request', 'channels' of 'struct
cfg80211_wowlan_nd_match', and 'matches' of 'struct
cfg80211_wowlan_nd_info' with '__counted_by' attribute. Briefly
tested with clang 18.1.1 and CONFIG_UBSAN_BOUNDS running iwlwifi.

Signed-off-by: Dmitry Antipov <dmantipov@yandex.ru>
Link: https://msgid.link/20240517153332.18271-1-dmantipov@yandex.ru
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: Add EHT UL MU-MIMO flag in ieee80211_bss_conf

Add flag for Full Bandwidth UL MU-MIMO for EHT. This is utilized
to pass EHT MU-MIMO configurations from user space to driver
in AP mode.

Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.0.1-00029-QCAHKSWPL_SILICONZ-1

Signed-off-by: Pradeep Kumar Chitrapu <quic_pradeepc@quicinc.com>
Link: https://msgid.link/20240515181327.12855-2-quic_pradeepc@quicinc.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: refactor chanreq.ap setting

There are now three places setting up chanreq.ap which always
depends on the mode (EHT being used or not) and override flag.
Refactor that code into a common function with a comment, to
make that clearer.

Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240506215543.5cd6a209e58a.I3be318959d9e2df5dccd2d0938c3d2fcc6688030@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: handle TPE element during CSA

Handle the transmit power envelope (TPE) element during
channel switch, applying it when the channel switch is
done.

Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240506215543.486c33157d18.Idf971ad801b6961c177bdf42cc323fd1a4ca8165@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: handle wider bandwidth OFDMA during CSA

During channel switch, track the AP configuration in the
chanreq, so that wider bandwidth OFDMA is taken into
account correctly, since multiple channel contexts may
be needed due to sharing not being possible due to
wider bandwidth OFDMA.

Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240506215543.b2c5a72dac1b.I69f65cb2e75d4a49a174b1aede68bf8ff0a3cab3@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: collect some CSA data into sub-structs

Collect the CSA data in ieee80211_link_data_managed and
ieee80211_link_data into a csa sub-struct to clean up a
bit and make adding new things more obvious.

Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240506215543.29f954b1f576.I9a683a9647c33d4dd3011aade6677982428c1082@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: mlme: handle cross-link CSA

If we see a channel switch announcement on one link for
another, handle that case and start the CSA. The driver
can react to this in whatever way it needs. The stack
will have the ability to track it via the RNR/MLE in the
reporting link's beacon if it sees it for inactive links
and adjust everything accordingly.

Note that currently the timings for the CSA aren't set,
the values are only used by the Intel drivers, and they
don't need this for newer devices that support MLO, so
I've left it out for now.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240415112355.4d34b6a31be7.Ie8453979f5805873a8411c99346bcc3810cd6476@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: cfg80211: restrict operation during radar detection

Just like it's not currently possible to start radar
detection while already operating, it shouldn't be
possible to start operating while radar detection is
running. Fix that.

Also, improve the check whether operating (carrier
might not be up if e.g. attempting to join IBSS).

Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240506211158.ae8dca3d0d6c.I7c70a66a5fbdbc63a78fee8a34f31d1995491bc3@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: mac80211: move radar detect work to sdata

At some point we thought perhaps this could be per link, but
really that didn't happen, and it's confusing. Radar detection
still uses the deflink to allocate the channel, but the work
need not be there. Move it back.

Reviewed-by: Miriam Rachel Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240506211311.43bd82c6da04.Ib39bec3aa198d137385f081e7e1910dcbde3aa1b@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: Add support for LARI_CONFIG_CHANGE_CMD v11

Add support for enable/disable 11be for china and South Korea by
reading DSM function 12 from UEFI or ACPI and sending it to the FW.

Signed-off-by: Anjaneyulu <pagadala.yesu.anjaneyulu@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512152312.037613d03c87.I4e5ee63c0dbdd4bd6c48daee7b0c88462b702423@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: move code from iwl-eeprom-parse to dvm

Move code that is DVM only to dvm.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512152312.9a1b6ef116e0.I217a513f544d5288a7760d265f51419e81abfd9d@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: kill iwl-eeprom-read

This is used by dvm only, move to dvm.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512152312.72f4bf256c8d.I7433bda9b0fc9eece5210db2cb90c2f03973f5ec@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: don't skip link selection

If we exit EMLSR due to a IWL_MVM_ESR_EXIT*, a MLO scan followed by a
link selection is scheduled with a delay of 30 seconds.
If during that 30 seconds EMLSR was blocked and unblocked
(IWL_MVM_ESR_BLOCKED*), we would still want to get the needed data from
the MLO scan and select link accordingly, and not return immediately to
EMLSR.

Fixes: 2f33561ea8f9 ("wifi: iwlwifi: mvm: trigger link selection after exiting EMLSR")
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Link: https://msgid.link/20240512152312.caab27a8dd8f.I63f67e213d5e05416f71513a8d914917d59aa44f@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: add mvm-specific guard

Add and use an mvm-specific guard for the mvm mutex.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512152312.7391425b3af2.I1b109f56753dd3f35602dd2c5cf6e1e35aee313d@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: simplify TX tracing

There's no need to calculate again whether data should
be included or be handled externally, just check if any
space for it was already reserved.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512152312.a3d3b4796460.I85bd3029baee24ebf0be04db7d6bf01834090869@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: tracing: fix condition to allocate buf1

There's no need to allocate buf1 if we're not going to
write anything to it, and the condition for writing is
(correctly) "hdr_len > 0 && !iwl_trace_data()". Also
allocate the event space only under the same condition.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512152312.a11e6490d053.Ie090beea70d2f271bb22c9b0e287dbd342fb2a96@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: leave a print in the logs when we call fw_nmi()

When we crash the firmware, we need to know why we decided to do so.
Almost all the callsites of iwl_force_nmi() print something in the logs
that explain why the driver decided to crash the firmware.
Debugfs doesn't print anything and it is then hard to understand why the
firmware has crashed.
Add a simple print in the debugfs hook to ease the debug.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512075822.e2558de222dc.Idd81777c47264e6f557b086625895c1dc2f667f3@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: disable dynamic EMLSR when AUTO_EML is false

When AUTO_EML is set to false, this change ignores the exit criteria,
for testing purposes. Currently, if AUTO_EML is disabled, the
driver will not select a link or enter EMLSR, but will still exit if one
of the criteria is not fulfilled.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Link: https://msgid.link/20240512072733.10a0b5da6ec2.I46fd578a3ef6cdbf14fdc4dfa97b4be008fe68e3@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: use vif P2P type helper

Use ieee80211_vif_type_p2p instead of checking for IFTYPE_AP in
combination with vif->p2p.

Signed-off-by: Benjamin Berg <benjamin.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512072733.15a00a812c1b.I5ffee795d960c9beda46a934f5b4c70acde112f9@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: don't track used links separately

We track which link is using which FW link ID, so there
really isn't a need to separately track which link IDs
are in use. Remove that code and check the table when
looking for a new link ID to use.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512072733.1a67d8af815f.Ie642c12dce3ab55c688abd9a25918569e83e558a@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: always print the firmware version in hex

All our firmware versions should now be printed in hex.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512072733.cae4c562b2c3.I5043383bd0e81f872c55046afd2c6560f482c8e0@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: fw: avoid bad FW config on RXQ DMA failure

In practice, iwl_trans_get_rxq_dma_data() will not fail,
it only can do that if called with the wrong arguments.
But it does have an error value and doesn't initialize
the argument when it fails, so don't use the value then.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512072733.faf3a1b87589.I77c950173cb914676fbd28483e5ef420bb9f1bc9@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: fw: api: Add new timepoint for scan failure

Add new time point to get the details of scan failure
While on it added other missing enum to match with FW.

Signed-off-by: Mukesh Sisodiya <mukesh.sisodiya@intel.com>
Reviewed-by: Pagadala Yesu Anjaneyulu <pagadala.yesu.anjaneyulu@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512072733.6781e76990c4.Iadfb3f6dd63a98fc35019772266b5ebddc5b5270@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mei: unify iwl_mei_set_power_limit() prototype

The two versions of iwl_mei_set_power_limit() are a bit different,
and while really the const isn't all that necessary on the inline,
it's still better to have it be the same.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512072733.854427753c92.I557716085cb1f6a35d1f97640388fa421f42a56a@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: mark bad no-data RX as having bad PLCP

Just like with any other frame, mac80211 will do sanity checks
on no-data RX (representing e.g. sounding PPDUs), and if e.g.
the NSS is wrong, it will warn. This isn't a good idea if we
already know the frame wasn't received well, e.g. has bad PLCP.
Unless the firmware reports "no error", set the bad PLCP flag
to skip checks in mac80211.

Also, since we're now extracting two different values from the
info field, use le32_get_bits() for both.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Reviewed-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512072733.e6adcb9a6ece.Ic14c2e8ed5e80d48af78b2f04e9f08beeb62d68e@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: bump FW API to 91 for BZ/SC devices

Start supporting API version 91 for new devices.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512072733.27c00099e8b7.I99b4f24bba073414dd04a6e04a359c7fbba52990@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: add a of print of a few commands

This is needed for debugging.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512072733.65fc46e2ffbc.I19d4ef2cc06bfee7d9644a376d39399f0f6eaa15@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: Use the SMPS cfg of the correct link

The exiting code is checking the SMPS mode of deflink.
Use the SMPS mode of the link_sta function parameter.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512072733.f3e2834e78f5.Ie24b431a6ad0dbf977afbbd4d2116bfe2426a0d5@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: don't always set antenna in beacon template cmd

This needs to be set only if the FW is not capable of selecting an
antenna on its own.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240512072733.73d790911268.I158dc1dbc9c2e199c69d8213e5496d2de72da28b@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: remove redundant reading from NVM file

The driver reads xtal_calib from NVM file, but actually never uses it.
This is only used in dvm driver.

Signed-off-by: Chaya Rachel Ivgi <chaya.rachel.ivgi@intel.com>
Reviewed-by: Luciano Coelho <luciano.coelho@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240510170500.c4d1a4571049.I7d7b73dccb793e220f023e0d049b082b043ca95e@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: remove fw_running op

fw_running assumes that memory can be retrieved only after alive.
This assumption is no longer true as we support dump before alive.
To avoid invalid access to the NIC, check that STATUS_DEVICE_ENABLED
bit in trans status is set before dumping instead of the prior check.

Signed-off-by: Shahar S Matityahu <shahar.s.matityahu@intel.com>
Reviewed-by: Luciano Coelho <luciano.coelho@intel.com>
Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240510170500.ca07138cedeb.I090e31d3eaeb4ba19f5f84aba997ccd36927e9ac@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: Don't set NO_HT40+/- flags on 6 GHz band

These flags are not needed on 6 GHz channels, and anyway they were set
incorrectly.
This in turn resulted in alternating channel flags, preventing reg domain
rules to be merged together, so cfg80211 couldn't even send them to the
user space, as the regulatory domain was too large.
Fix it.

Signed-off-by: Andrei Otcheretianski <andrei.otcheretianski@intel.com>
Reviewed-by: Luciano Coelho <luciano.coelho@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240510170500.b0d51b2229f8.I092e21cde43320ffc2eff17f5748ff9c87c87fcf@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>

wifi: iwlwifi: mvm: call ieee80211_sta_recalc_aggregates on A-MSDU size update

The driver needs to call this function when the A-MSDU size changes.

Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://msgid.link/20240510170500.632ee6612a79.Ice3a536a4cbfe60d0edfa231fcb79ee7ab9495dd@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>