Marek Vasut [Wed, 21 Aug 2024 18:36:03 +0000 (20:36 +0200)]
wifi: wilc1000: Do not operate uninitialized hardware during suspend/resume
In case the hardware is not initialized, do not operate it during
suspend/resume cycle, the hardware is already off so there is no
reason to access it.
In fact, wilc_sdio_enable_interrupt() in the resume callback does
interfere with the same call when initializing the hardware after
resume and makes such initialization after resume fail. Fix this
by not operating uninitialized hardware during suspend/resume.
Frank Li [Tue, 20 Aug 2024 14:21:42 +0000 (10:21 -0400)]
dt-bindings: net: wireless: convert marvel-8xxx.txt to yaml format
Convert binding doc marvel-8xxx.txt to yaml format.
Additional change:
- Remove marvell,caldata_00_txpwrlimit_2g_cfg_set in example.
- Remove mmc related property in example.
- Add wakeup-source property.
- Remove vmmc-supply and mmc-pwrseq.
Fix below warning:
arch/arm64/boot/dts/freescale/imx8mp-beacon-kit.dtb: /soc@0/bus@30800000/mmc@30b40000/wifi@1:
failed to match any schema with compatible: ['marvell,sd8997']
wifi: iwlwifi: mvm: properly set the rates in link cmd
When a channel ctx is assigned to a link, we set the
LINK_CONTEXT_MODIFY_RATES_INFO to indicate that the rate fields are now
valid. But then we always take the rates of 2.4 GHz regardless of actual
used band.
This is because we are getting the band from bss_conf->chanctx_conf, but
this is assigned only after drv_assign_vif_chanctx returns, so we take
the bands of 2.4 GHz.
Fix it by taking the band from the iwl_mvm_link_info::phy_ctxt instead,
as this has already assigned in this point.
wifi: iwlwifi: mvm: tell the firmware about CSA with mode=1
When we de-activate a link because it started a CSA with mode=1, we want
to tell the firmware it can no longer transmit any frame for that link.
The firmware will do that on its own if the CSA indication (beacon /
action frame) was received on that same link, but with MLO, things got
more complex and the firmware can't track cross link CSA.
Tell the firmware if we de-activate a link because of CSA with mode=1 to
prevent it from transmitting, even if it is only an NDP PM=1 frame that
is part of the de-activation flow.
Johannes Berg [Sun, 1 Sep 2024 04:17:50 +0000 (07:17 +0300)]
wifi: iwlwifi: mvm: refactor scan channel description a bit
The channel number is at the same position across all versions
of the channel description struct, so move it out of the union
that versions it. Also add __packed annotations to all of the
sub-structs and the union so it's packed correctly, and fully
document the structure.
Issam Hamdi [Fri, 16 Aug 2024 14:24:18 +0000 (16:24 +0200)]
wifi: cfg80211: Set correct chandef when starting CAC
When starting CAC in a mode other than AP mode, it return a
"WARNING: CPU: 0 PID: 63 at cfg80211_chandef_dfs_usable+0x20/0xaf [cfg80211]"
caused by the chandef.chan being null at the end of CAC.
Solution: Ensure the channel definition is set for the different modes
when starting CAC to avoid getting a NULL 'chan' at the end of CAC.
Yan Zhen [Tue, 27 Aug 2024 10:30:12 +0000 (18:30 +0800)]
wifi: mac80211: scan: Use max macro
When the original file is guaranteed to contain the minmax.h header
file and compile correctly, using the real macro is usually
more intuitive and readable.
Johannes Berg [Tue, 27 Aug 2024 07:49:40 +0000 (09:49 +0200)]
wifi: mac80211: fix RCU list iterations
There are a number of places where RCU list iteration is
used, but that aren't (always) called with RCU held. Use
just list_for_each_entry() in most, and annotate iface
iteration with the required locks.
iwl_mvm_tx_skb_sta() and iwl_mvm_tx_mpdu() verify that the mvmvsta
pointer is not NULL.
It retrieves this pointer using iwl_mvm_sta_from_mac80211, which is
dereferencing the ieee80211_sta pointer.
If sta is NULL, iwl_mvm_sta_from_mac80211 will dereference a NULL
pointer.
Fix this by checking the sta pointer before retrieving the mvmsta
from it. If sta is not NULL, then mvmsta isn't either.
Ping-Ke Shih [Thu, 22 Aug 2024 01:42:54 +0000 (09:42 +0800)]
wifi: mac80211: export ieee80211_purge_tx_queue() for drivers
Drivers need to purge TX SKB when stopping. Using skb_queue_purge() can't
report TX status to mac80211, causing ieee80211_free_ack_frame() warns
"Have pending ack frames!". Export ieee80211_purge_tx_queue() for drivers
to not have to reimplement it.
Zijun Hu [Sun, 11 Aug 2024 04:47:26 +0000 (12:47 +0800)]
wifi: rfkill: Correct parameter type for rfkill_set_hw_state_reason()
Change type of parameter @reason to enum rfkill_hard_block_reasons
for API rfkill_set_hw_state_reason() according to its comments, and
all kernel callers have invoked the API with enum type actually.
Hans de Goede [Sun, 25 Aug 2024 13:19:16 +0000 (15:19 +0200)]
net: rfkill: gpio: Do not load on Lenovo Yoga Tab 3 Pro YT3-X90
The Lenovo Yoga Tab 3 Pro YT3-X90 has a non functional "BCM4752" ACPI
device, which uses GPIO resources which are actually necessary / used
for the sound (codec, speaker amplifier) on the Lenovo Yoga Tab 3 Pro.
If the rfkill-gpio driver loads before the sound drivers do the sound
drivers fail to load because the GPIOs are already claimed.
Add a DMI based deny list with the Lenovo Yoga Tab 3 Pro on there and
make rfkill_gpio_probe() exit with -ENODEV for devices on the DMI based
deny list.
Johannes Berg [Thu, 8 Aug 2024 20:22:48 +0000 (23:22 +0300)]
wifi: iwlwifi: mvm: drop wrong STA selection in TX
This shouldn't happen at all, since in station mode all MMPDUs
go through the TXQ for the STA, and not this function. There
may or may not be a race in mac80211 through which this might
happen for some frames while a station is being added, but in
that case we can also just drop the frame and pretend the STA
didn't exist yet.
Also, the code is simply wrong since it uses deflink, and it's
not easy to fix it since the mvmvif->ap_sta pointer cannot be
used without the mutex, and perhaps the right link might not
even be known.
Just drop the frame at that point instead of trying to fix it
up.
This used to extract the timeout from the debug TLV
(FW_DBG_TRIGGER_TXQ_TIMERS), which is not in use.
Cleanup iwl_mvm_get_wd_timeout to not consider the debug TLV.
Miri Korenblit [Thu, 8 Aug 2024 20:22:45 +0000 (23:22 +0300)]
wifi: iwlwifi: use default command queue watchdog timeout
We used to have the opmode configuring it to the trans according to the
debug tlv value (FW_DBG_TRIGGER_TXQ_TIMERS).
But this debug is not used, so trans can just have the default value
hardcoded.
Ilan Peer [Thu, 8 Aug 2024 20:22:44 +0000 (23:22 +0300)]
wifi: iwlwifi: mvm: Stop processing MCC update if there was no change
When processing a MCC update notification from the firmware, if
the firmware indicates that the regulatory configuration didn't
change, abort the flow and do not notify higher layers.
Daniel Gabay [Thu, 8 Aug 2024 20:22:37 +0000 (23:22 +0300)]
wifi: iwlwifi: mvm: Offload RLC/SMPS functionality to firmware
Currently, the driver handles SMPS decisions by tracking AP
capabilities, BT coexistence changes, sending necessary SMPS
frames to the AP, and updating firmware with RX chain info
using the RLC_CONFIG_CMD.
Starting with version 3 of the RLC_CONFIG_CMD, the firmware
takes over this responsibility. It now tracks SMPS, sends
frames, and configures the RLC.
In this patch:
1. Stop sending RLC_CONFIG_CMD when firmware supports RLC
offload (version 3), as rlc.rx_chain_info is not needed by
firmware, and no other field in the cmd is used.
2. Prevent the driver from forwarding any SMPS requests to
mac80211, i.e., the driver should not transmit SMPS frames
to the AP as firmware handles that.
3. Set NL80211_FEATURE_DYNAMIC_SMPS and NL80211_FEATURE_STATIC_SMPS
conditionally based on RLC version.
Emmanuel Grumbach [Thu, 8 Aug 2024 08:59:16 +0000 (11:59 +0300)]
wifi: mac80211: fix the comeback long retry times
When we had a comeback, we will never use the default timeout values
again because comeback is never cleared.
Clear comeback if we send another association request which will allow
to start a default timer after Tx status.
The problem was seen with iwlwifi where the tx_status on the association
request is handled before the association response frame (which is the
usual case).
1) Tx assoc request 1/3
2) Rx assoc response (comeback, timeout = 1 second)
3) wait 1 second
4) Tx assoc request 2/3
5) Set timer to IEEE80211_ASSOC_TIMEOUT_LONG = 500ms (1 second after
round_up)
6) tx_status on frame sent in 4) is ignored because comeback is still
true
7) AP does not reply with assoc response
8) wait 1s <= This is where the bug is felt
9) Tx assoc request 3/3
With this fix, in step 6 we will reset the timer to
IEEE80211_ASSOC_TIMEOUT_SHORT = 100ms and we will wait only 100ms in
step 8.
hhorace [Wed, 7 Aug 2024 08:22:05 +0000 (16:22 +0800)]
wifi: cfg80211: fix bug of mapping AF3x to incorrect User Priority
According to RFC8325 4.3, Multimedia Streaming: AF31(011010, 26),
AF32(011100, 28), AF33(011110, 30) maps to User Priority = 4
and AC_VI (Video).
However, the original code remain the default three Most Significant
Bits (MSBs) of the DSCP, which makes AF3x map to User Priority = 3
and AC_BE (Best Effort).
Rory Little [Tue, 6 Aug 2024 00:40:23 +0000 (17:40 -0700)]
wifi: mac80211: Add non-atomic station iterator
Drivers may at times want to iterate their stations with a function
which requires some non-atomic operations.
ieee80211_iterate_stations_mtx() introduces an API to iterate stations
while holding that wiphy's mutex. This allows the iterating function to
do non-atomic operations safely.
Christophe JAILLET [Mon, 5 Aug 2024 06:40:37 +0000 (08:40 +0200)]
wifi: lib80211: Handle const struct lib80211_crypto_ops in lib80211
lib80211_register_crypto_ops() and lib80211_unregister_crypto_ops() don't
modify their "struct lib80211_crypto_ops *ops" argument. So, it can be
declared as const.
Doing so, some adjustments are needed to also constify some date in
"struct lib80211_crypt_data", "struct lib80211_crypto_alg" and the
return value of lib80211_get_crypto_ops().
wifi: iwlwifi: mvm: add and improve EMLSR debug info
Add prints of the per-link MPDU counters, and change the other MPDU
counters related prints to use DL_INFO, which is already used for all
EMLSR tests anyway, instead of DL_STATS which pollutes the logs with all
the RX signal info.
wifi: iwlwifi: mvm: start to support the new BT profile notification
We have a new notification the BT profile. It contains almost nothing,
only the wifi loss information. Copy this into mvm. We still need to
iterate over the vifs / links to use this data.
The iwl_bt_coex_profile_notif is now being obsoleted. It'll be replaced
by a new notification coming with a non-legacy group. Rename it to avoid
confusion.
wifi: iwlwifi: mvm: increase the time between ranging measurements
The algo running in fw may take a little longer than 5 milliseconds,
(e.g. measurement on 80MHz while associated). Increase the minimum
time between measurements to 7 milliseconds.
wifi: iwlwifi: mvm: set the cipher for secured NDP ranging
The cipher pointer is not set, but is derefereced trying to set its
content, which leads to a NULL pointer dereference.
Fix it by pointing to the cipher parameter before dereferencing.
Ilan Peer [Sun, 25 Aug 2024 05:56:37 +0000 (08:56 +0300)]
wifi: iwlwifi: mvm: Fix a race in scan abort flow
When the upper layer requests to cancel an ongoing scan, a race
is possible in which by the time the driver starts to handle the
upper layers scan cancel flow, the FW already completed handling
the scan request and the driver received the scan complete
notification but still did not handle the notification. In such a
case the FW will simply ignore the scan abort request coming from
the driver, no notification would arrive from the FW and the entire
abort flow would be considered a failure.
To better handle this, check the status code returned by the FW for
the scan abort command. In case the status indicates that
no scan was aborted, complete the scan abort flow with success, i.e.,
the scan was aborted, as the flow is expected to consume the scan
complete notification.
SAP version 4 uses larger Host to MEI notification queue.
Since it is unknown which SAP version is used by the CSME firmware
when the driver loads, try version 4 first. In case the CSME firmware
uses version 3, the memory allocation will fail. In this case the
driver will try again to allocate the memory for version 3.
Johannes Berg [Mon, 29 Jul 2024 17:20:07 +0000 (20:20 +0300)]
wifi: iwlwifi: config: label 'gl' devices as discrete
The 'gl' devices are in the bz family, but they're not,
integrated, so should have their own trans config struct.
Fix that, also necessitating the removal of LTR config,
and while at it remove 0x2727 and 0x272D IDs that were
only used for test chips.
Johannes Berg [Mon, 29 Jul 2024 17:20:06 +0000 (20:20 +0300)]
wifi: iwlwifi: pcie: print function scratch before writing
We may want to know the value of this register for certain
debug scenarios before we overwrite it, it indicates which
parts of the chip have recently been reset.
Johannes Berg [Mon, 29 Jul 2024 17:20:05 +0000 (20:20 +0300)]
wifi: iwlwifi: mvm: use correct key iteration
In the cases changed here, key iteration isn't done from
an RCU critical section, but rather using the wiphy lock
as protection. Therefore, just use ieee80211_iter_keys().
The link switch case can therefore also use sync commands.
wifi: iwlwifi: mvm: prepare the introduction of V9 of REDUCED_TX_POWER
* Rename iwl_dev_tx_power_cmd to iwl_dev_tx_power_cmd_v3_v8
* struct iwl_dev_tx_power_common needs to be packed. It was always the
case, but now that its size is not a multiple of 4, it becomes
meaningful.
* Move per_band data out of the common structure since it won't be
present in the new versions of the command.
wifi: mac80211: don't use rate mask for offchannel TX either
Like the commit ab9177d83c04 ("wifi: mac80211: don't use rate mask for
scanning"), ignore incorrect settings to avoid no supported rate warning
reported by syzbot.
The syzbot did bisect and found cause is commit 9df66d5b9f45 ("cfg80211:
fix default HE tx bitrate mask in 2G band"), which however corrects
bitmask of HE MCS and recognizes correctly settings of empty legacy rate
plus HE MCS rate instead of returning -EINVAL.
As suggestions [1], follow the change of SCAN TX to consider this case of
offchannel TX as well.
Introduce 'ieee80211_mgmt_ba()' to avoid code duplication between
'ieee80211_send_addba_resp()', 'ieee80211_send_addba_request()',
and 'ieee80211_send_delba()', ensure that all related addresses
are '__aligned(2)', and prefer convenient 'ether_addr_copy()'
over generic 'memcpy()'. No functional changes expected.
Zong-Zhe Yang [Fri, 12 Jul 2024 07:49:38 +0000 (15:49 +0800)]
wifi: mac80211_hwsim: correct MODULE_PARM_DESC of multi_radio
Correct the name field in multi_radio's MODULE_PARM_DESC.
Fixes: d2601e34a102 ("wifi: mac80211_hwsim: add support for multi-radio wiphy") Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com> Acked-by: Felix Fietkau <nbd@nbd.name> Link: https://patch.msgid.link/20240712074938.26437-1-kevin_yang@realtek.com Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Sascha Hauer [Fri, 16 Aug 2024 06:56:10 +0000 (08:56 +0200)]
wifi: mwifiex: remove unnecessary checks for valid priv
The pointers in adapter->priv[] are allocated in mwifiex_register().
With an allocation failed the function will return an error and
driver initialization is aborted. This makes all checks for valid
priv pointers unnecessary throughout the driver. In many places
the pointers are assumed to be valid without checks, this patch
removes the remaining checks.
-Wflex-array-member-not-at-end was introduced in GCC-14, and we are
getting ready to enable it, globally.
So, in order to avoid ending up with a flexible-array member in the
middle of multiple other structs, we use the `__struct_group()`
helper to create new tagged structures with the suffix `_hdr`.
These structures group together all the members of the original
flexible structures except the flexible arrays.
As a result, the arrays are effectively separated from the rest of the
members without modifying the memory layout of the flexible structures.
We then change the type of the middle struct members currently causing
trouble from the original flex struct to the newly created structs with
suffix `_hdr`.
We also want to ensure that when new members need to be added to the
flexible structures, they are always included within the newly created
tagged structs. For this, we use `static_assert()`. This ensures that the
memory layout for both the flexible structure and the new tagged struct
is the same after any changes.
This approach avoids having to implement the `_hdr` structures as
completely separate structures, thus preventing having to maintain
two independent but basically identical structures, closing the door
to potential bugs in the future.
We also use `container_of()` whenever we need to retrieve a pointer to
the flexible structure, through which we can access the flexible-array
member, if necessary.
Also, remove a couple of unused zero-length arrays and flexible-array
members.
So, with these changes, fix the following warnings:
drivers/net/wireless/intel/iwlegacy/commands.h:1196:38: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/iwlegacy/commands.h:1197:36: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/iwlegacy/commands.h:2505:30: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/iwlegacy/commands.h:2549:26: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/iwlegacy/commands.h:2654:31: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/iwlegacy/commands.h:2665:30: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/iwlegacy/commands.h:2673:26: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/iwlegacy/commands.h:3349:30: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
Sascha Hauer [Fri, 9 Aug 2024 09:51:48 +0000 (11:51 +0200)]
wifi: mwifiex: keep mwifiex_cfg80211_ops constant
With host_mlme support being added mwifiex_cfg80211_ops is no longer
constant, but supplemented with the host_mlme related ops when host_mlme
support is enabled. This doesn't work with multiple adapters when only
few of then have host_mlme support. Duplicate mwifiex_cfg80211_ops
before using it and keep the original constant.
While at it mark mwifiex_cfg80211_ops const to prevent people from
changing it again during runtime.
Kalle Valo [Tue, 13 Aug 2024 09:58:32 +0000 (12:58 +0300)]
Merge tag 'ath-next-20240812' of git://git.kernel.org/pub/scm/linux/kernel/git/ath/ath
ath.git patches for v6.12
This is a fairly light pull request since ath12k is still working on
MLO-related changes, and the other drivers are mostly in maintenance
mode with a few cleanups and bug fixes.
Major changes:
ath12k
* DebugFS support for transmit DE stats
* Make ASPM support hardware-dependent
* Align BSS Channel information command and message with firmware
ath11k
* Use work queue for beacon tx events
ath9k
* Use devm for gpio_request_one
* Use unmanaged PCI functions in ath9k_pci_owl_loader()
Gustavo A. R. Silva [Thu, 8 Aug 2024 22:11:15 +0000 (16:11 -0600)]
wifi: mwl8k: Use static_assert() to check struct sizes
Commit 5c4250092fad ("wifi: mwl8k: Avoid -Wflex-array-member-not-at-end
warnings") introduced tagged `struct mwl8k_cmd_pkt_hdr`. We want to
ensure that when new members need to be added to the flexible structure,
they are always included within this tagged struct.
We use `static_assert()` to ensure that the memory layout for both
the flexible structure and the tagged struct is the same after any
changes.
Kalle Valo [Sat, 10 Aug 2024 07:42:26 +0000 (10:42 +0300)]
Merge tag 'rtw-next-2024-08-09' of https://github.com/pkshih/rtw
rtw-next patches for v6.12
A pull-request for v6.12 containing fixes and features, and list major
changes below:
rtw88:
- improve USB performance by aggregation
- fix RX bandwidth for rx_status
rtw89:
- support new WiFi 6 chip RTL8852BE-VT
- support WoWLAN net-detect feature
- support hardware encryption in unicast management frames
- support hardware rfkill
-Wflex-array-member-not-at-end was introduced in GCC-14, and we are
getting ready to enable it, globally.
So, in order to avoid ending up with a flexible-array member in the
middle of multiple other structs, we use the `__struct_group()`
helper to create a new tagged `struct libipw_hdr_3addr_hdr`.
This structure groups together all the members of the flexible
`struct libipw_hdr_3addr` except the flexible array.
As a result, the array is effectively separated from the rest of the
members without modifying the memory layout of the flexible structure.
We then change the type of the middle struct members currently causing
trouble from `struct libipw_hdr_3addr` to `struct libipw_hdr_3addr_hdr`.
We also want to ensure that when new members need to be added to the
flexible structure, they are always included within the newly created
tagged struct. For this, we use `static_assert()`. This ensures that the
memory layout for both the flexible structure and the new tagged struct
is the same after any changes.
This approach avoids having to implement `struct libipw_hdr_3addr_hdr`
as a completely separate structure, thus preventing having to maintain
two independent but basically identical structures, closing the door
to potential bugs in the future.
Also, remove a couple of unused structures `struct libipw_ibss_dfs` and
`struct libipw_assoc_request`.
So, with these changes, fix the following warnings:
drivers/net/wireless/intel/ipw2x00/libipw.h:403:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/ipw2x00/libipw.h:420:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/ipw2x00/libipw.h:433:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/ipw2x00/libipw.h:441:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/ipw2x00/libipw.h:447:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/ipw2x00/libipw.h:460:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/ipw2x00/libipw.h:468:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/ipw2x00/libipw.h:476:33: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
drivers/net/wireless/intel/ipw2x00/libipw.h:592:36: warning: structure containing a flexible array member is not at the end of another structure [-Wflex-array-member-not-at-end]
Sascha Hauer [Mon, 5 Aug 2024 06:39:15 +0000 (08:39 +0200)]
wifi: mwifiex: add support for WPA-PSK-SHA256
This adds support for the WPA-PSK AKM suite with SHA256 as hashing
method (WPA-PSK-SHA256). Tested with a wpa_supplicant provided AP
using key_mgmt=WPA-PSK-SHA256.
Sascha Hauer [Mon, 5 Aug 2024 06:39:14 +0000 (08:39 +0200)]
wifi: mwifiex: fix key_mgmt setting
bss_config->key_mgmt specifies the AKM suites that are usable in
hardware. This variable is set to specific values while iterating
over the advertised AKM suites. This means the final value of the
variable depends on the order of the entries in the AKM suites
array. Instead of setting the variable, just set the relevant bits
in the key_mgmt bit field to make us independent of the order of
entries.
This behaviour is derived from the downstream driver that does the
same.
Also, set bss_config->key_mgmt to zero explicitly right before the
loop. bss_config has been zero allocated by the caller already, but
do so again to save the reader from following the code path.
Sascha Hauer [Mon, 5 Aug 2024 06:39:13 +0000 (08:39 +0200)]
wifi: mwifiex: simplify WPA flags setting
The WPA flags setting only depends on the wpa_versions bitfield and not
on the AKM suite, so move it out of the switch/case to simplify the code
a bit. Also set bss_config->protocol to zero explicitly. This is done
only to make the code clearer, bss_config has been zero alloced by the
caller, so should be zero already. No functional change intended.
Felix pointed out that moving to devm for request_irq() can lead to a use after
free, and that avoiding that means having explicit frees that makes the devm
thing pretty pointless. So let's just revert the patch.
Bitterblue Smith [Wed, 7 Aug 2024 22:23:06 +0000 (01:23 +0300)]
wifi: rtw88: Enable USB RX aggregation for 8822c/8822b/8821c
Enable USB RX aggregation when there is at least 1 Mbps RX or TX
traffic, otherwise disable it.
USB RX aggregation improves the RX speed of RTL8811CU on certain ARM
systems, like the NanoPi NEO Core2. Before: 28 Mbps, after: 231 Mbps.
It also improves the RX speed of RTL8822CU on some x86_64 systems.
Before: ~200 Mbps, after: ~300 Mbps.
The official drivers for these chips use the same logic for SDIO, but
for some reason the SDIO driver in rtw88 always enables RX aggregation,
so this patch only toggles aggregation for USB devices.
RTL8703B is likely not found in USB devices, and RTL8723DU doesn't like
aggregation.
Bitterblue Smith [Wed, 7 Aug 2024 22:21:36 +0000 (01:21 +0300)]
wifi: rtw88: usb: Support RX aggregation
The chips can be configured to aggregate several frames into a single
USB transfer. Modify rtw_usb_rx_handler() to support this case.
RX aggregation improves the RX speed of RTL8811CU on certain ARM
systems, like the NanoPi NEO Core2. It also improves the RX speed of
RTL8822CU on some x86_64 systems.
Currently none of the chips are configured to aggregate frames.
Peter Robinson [Tue, 6 Aug 2024 08:22:31 +0000 (09:22 +0100)]
wifi: rtl8xxxu: add missing rtl8192cu USB IDs
The rtl8xxxu has all the rtl8192cu USB IDs except for the
following 3 so add them to the untested section so they
can be used with the rtl8xxxu as the rtl8192cu are well
supported.
Peter Robinson [Mon, 5 Aug 2024 22:19:04 +0000 (23:19 +0100)]
wifi: rtl8xxxu: drop reference to staging drivers
The Kconfig notes mention staging drivers that have
since been removed so update the driver's description
so it no longer references the staging rtl8723au and
rtl8192u drivers which have now been deleted.
Chin-Yen Lee [Mon, 5 Aug 2024 09:00:27 +0000 (17:00 +0800)]
wifi: rtw89: wow: add delay option for net-detect
The delay option is the period in unit of second for WoWLAN firmware
to wait before the first scan. We get the option from cfg80211 and
practice it. Another, in some platform, WoWLAN firmware may found
configured network and then trigger resume process, before suspend
process is completed, lead to the wakeup function failed.
So the default value is set one to avoid the issue.
Chin-Yen Lee [Mon, 5 Aug 2024 09:00:26 +0000 (17:00 +0800)]
wifi: rtw89: wow: add WoWLAN net-detect support
Net-detect is an option of WoWLAN to allow the device to
be woken up from suspend mode when configured network is detected.
When user enables net-detect and lets the device enter suspend
state, WoWLAN firmware will periodically scan until beacon or
probe response of configured networks are received. If configured
networks are detected, WoWLAN firmware will trigger resume process.
Chin-Yen Lee [Mon, 5 Aug 2024 09:00:25 +0000 (17:00 +0800)]
wifi: rtw89: wow: implement PS mode for net-detect
When net-detect is enabled, WoWLAN firmware will periodically
scan until beacon or probe response of configured networks are
received. To reduce power consumption, the FW-IPS mode is
implemented to keep WiFi chip in idle mode between each scan.
The FW-IPS is controlled by WoWLAN firmware to turn of some critical
electrical components, and is different from the original IPS mode
which most electrical components are turned off.
Ping-Ke Shih [Mon, 5 Aug 2024 09:00:24 +0000 (17:00 +0800)]
wifi: rtw89: 8852c: support firmware format up to v1
Driver has supported different WoWLAN reason code by commit 0e5210217768
("wifi: rtw89: wow: update WoWLAN reason register for different FW")
since firmware version 0.27.80.0. The old driver can't support two
kinds of WoWLAN reason, so increase firmware format to v1.
Also driver tables of BB and RF registers and power values will be added
into v1 format.
Fixes: d889913205cf ("wifi: ath12k: driver for Qualcomm Wi-Fi 7 devices") Signed-off-by: P Praneesh <quic_ppranees@quicinc.com> Signed-off-by: Karthikeyan Kathirvel <quic_kathirve@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://patch.msgid.link/20240331183232.2158756-3-quic_kathirve@quicinc.com
P Praneesh [Sun, 31 Mar 2024 18:32:31 +0000 (00:02 +0530)]
wifi: ath12k: fix BSS chan info request WMI command
Currently, the firmware returns incorrect pdev_id information in
WMI_PDEV_BSS_CHAN_INFO_EVENTID, leading to incorrect filling of
the pdev's survey information.
To prevent this issue, when requesting BSS channel information
through WMI_PDEV_BSS_CHAN_INFO_REQUEST_CMDID, firmware expects
pdev_id as one of the arguments in this WMI command.
Add pdev_id to the struct wmi_pdev_bss_chan_info_req_cmd and fill it
during ath12k_wmi_pdev_bss_chan_info_request(). This resolves the
issue of sending the correct pdev_id in WMI_PDEV_BSS_CHAN_INFO_EVENTID.
Fixes: d889913205cf ("wifi: ath12k: driver for Qualcomm Wi-Fi 7 devices") Signed-off-by: P Praneesh <quic_ppranees@quicinc.com> Signed-off-by: Karthikeyan Kathirvel <quic_kathirve@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://patch.msgid.link/20240331183232.2158756-2-quic_kathirve@quicinc.com
wifi: ath12k: restore ASPM for supported hardwares only
During PCI based hardware device start up, ASPM is disabled for all. And
once firmware is ready, it is restored back. However, not all hardwares
(for example QCN9274) supports ASPM. Hence there is a need to conditionally
restore ASPM back. Or else, for such hardwares, issue can be seen during
sending and receiving packets.
Introduce a new hardware param supports_aspm which identifies whether a
given hardware supports ASPM or not and then accordingly restore it.
Toke Høiland-Jørgensen [Mon, 5 Aug 2024 11:02:22 +0000 (13:02 +0200)]
wifi: ath9k: Remove error checks when creating debugfs entries
We should not be checking the return values from debugfs creation at all: the
debugfs functions are designed to handle errors of previously called functions
and just transparently abort the creation of debugfs entries when debugfs is
disabled. If we check the return value and abort driver initialisation, we break
the driver if debugfs is disabled (such as when booting with debugfs=off).
Earlier versions of ath9k accidentally did the right thing by checking the
return value, but only for NULL, not for IS_ERR(). This was "fixed" by the two
commits referenced below, breaking ath9k with debugfs=off starting from the 6.6
kernel (as reported in the Bugzilla linked below).
Restore functionality by just getting rid of the return value check entirely.
Link: https://bugzilla.kernel.org/show_bug.cgi?id=219122 Fixes: 1e4134610d93 ("wifi: ath9k: use IS_ERR() with debugfs_create_dir()") Fixes: 6edb4ba6fb5b ("wifi: ath9k: fix parameter check in ath9k_init_debug()") Reported-by: Daniel Tobias <dan.g.tob@gmail.com> Tested-by: Daniel Tobias <dan.g.tob@gmail.com> Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://patch.msgid.link/20240805110225.19690-1-toke@toke.dk
wifi: rtw89: add support for HW encryption in unicast management frames
Add hardware encryption support for unicast management frames for
8922AE and 8852CE. Other chips will continue to use software
encryption for unicast management frames.
wifi: rtw89: avoid to add interface to list twice when SER
If SER L2 occurs during the WoWLAN resume flow, the add interface flow
is triggered by ieee80211_reconfig(). However, due to
rtw89_wow_resume() return failure, it will cause the add interface flow
to be executed again, resulting in a double add list and causing a kernel
panic. Therefore, we have added a check to prevent double adding of the
list.
wifi: rtw89: 8922a: Add new fields for scan offload H2C command
Update scan offload H2C format to fit firmware version 35.21.
The new fields indicate lengths of variable length members, so
when driver and firmware are using mismatch version, FW could
handle the parsing better.
Dmitry Kandybka [Thu, 1 Aug 2024 10:15:31 +0000 (13:15 +0300)]
wifi: brcmsmac: clean up unnecessary current_ampdu_cnt and related checks
In 'brcms_c_ffpld_check_txfunfl()', 'current_ampdu_cnt' is hardcoded to
zero, so 'txunfl_ratio' is always zero as well and some dead code can be
removed. Compile tested only.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
projects / users / jedix / linux-maple.git / log