wifi: ath12k: Introduce the container for mac80211 hw
To support multi link operation, we need to combine all the link/pdev
under a single wiphy. This avoids the overhead of synchronization
across multiple hardware instances in both the cfg80211 and mac80211
layers. Currently, each link/pdev is registered as separate wiphy,
tightly coupled with link/pdev/radio (ar) structure. To enable single
wiphy registration within the chip, we decouple the wiphy data entity from
the link/pdev/radio (ar) structure and move it under the chip (ab)
structure with a new data container (ath12k_hw) structure. This approach
improves scalability for future multi link operation support.
mac80211 hw private data structure diagram
------------------------------------------
wifi: ath12k: Refactor the mac80211 hw access from link/radio
Currently, mac80211 hardware accesses link/radio structure directly in
multiple locations. Introduce helper function to avoid this direct
access, as this change will facilitate refactoring for Multi-link
operation support.
Baochen Qiang [Wed, 17 Jan 2024 12:05:39 +0000 (14:05 +0200)]
wifi: ath12k: change MAC buffer ring size to 2048
For WCN7850, there is a SRNG named MAC buffer ring, i.e.,
dp->rx_mac_buf_ring. During initialization, it is setup
by host and then under control of firmware. During RX
process, firmware fetches buffers from
dp->rx_refill_buf_ring to fill that MAC buffer ring,
and those buffers are taken by RXDMA to carry real WLAN
frames received from air.
Currently a low RX throughput is observed. Checking
firmware log, lots of errors are reported by MAC buffer
ring, complaining that it is running out of buffers,
which further indicates that RXDMA is suffering from
starvation. Currently the size of dp->rx_mac_buf_ring
is configured as 1024. After changing it to 2048, those
error messages are reduced, and a 6.4% increase is seen
in peak throughput. Note that 2048 is an empirical
value. It is chosen here because the RX throughput
meets our expectation after the change.
This change only applies to WCN7850 since other
chips don't have a MAC buffer ring.
Baochen Qiang [Wed, 17 Jan 2024 12:05:30 +0000 (14:05 +0200)]
wifi: ath12k: add support for BA1024
Currently the maximum block ACK window size supported is 256.
This results in that, when connected to an AP which supports
larger BA sizes like BA512 or BA1024, only BA256 is
established, leading to a lower peak throughput.
So add support for BA1024, this is done by allocating a larger
REO queue and advertising IEEE80211_MAX_AMPDU_BUF_EHT support
to MAC80211.
wifi: ath12k: add QMI PHY capability learn support
Currently, the number of PHY is learned from the firmware service ready
event message. However, on the QCN9274 platform, number of PHY is a
variable parameter. To enable MLO capability in the QMI host capability
request message, the driver needs the PHY count information earlier than
the firmware service ready event. Therefore, a new QMI message,
"PHY capability message", is introduced to retrieve this information.
This message allows the driver to fill in the MLO parameter in the QMI
host capability request message. If the new QMI PHY capability message
fails, the default configuration in the HW params will be used.
wifi: ath12k: refactor QMI MLO host capability helper function
Currently, QMI MLO host capability parameters are specific to the WCN7850
platform. To make use of this helper function across all the platforms,
move the platform specific MLO capability parameter to the HW param
configuration.
Jeff Johnson [Sun, 14 Jan 2024 15:02:42 +0000 (17:02 +0200)]
wifi: ath11k: document HAL_RX_BUF_RBM_SW4_BM
Commit 7636c9a6e7d7 ("wifi: ath11k: Add multi TX ring support for WCN6750")
added HAL_RX_BUF_RBM_SW4_BM to enum hal_rx_buf_return_buf_manager. However,
as flagged by the kernel-doc script, the documentation was not updated:
drivers/net/wireless/ath/ath11k/hal.h:689: warning: Enum value 'HAL_RX_BUF_RBM_SW4_BM' not described in enum 'hal_rx_buf_return_buf_manager'
So update the documentation. No functional changes, compile tested only.
wifi: ath12k: ath12k_start_vdev_delay(): convert to use ar
To support single wiphy abstraction, remove the mac80211 hw data
dependency from the start vdev delay function. This way, this
function can be extended to handle multiple link/radio in the future.
To support single wiphy abstraction, introduce link/radio specific helper
function in the mac80211 callback flush(). This way, the callback can be
extended to handle multiple link/radio in the future.
To support single wiphy abstraction, introduce link/radio specific helper
function in the mac80211 callback ampdu_action(). This way, the callback
can be extended to handle multiple link/radio in the future.
To support single wiphy abstraction, introduce link/radio specific helper
function in the mac80211 callback configure_filter(). This way, the
callback can be extended to handle multiple link/radio in the future.
To support single wiphy abstraction, introduce link/radio specific helper
function in the mac80211 callback update_vif_offload(). This way, the
callback can be extended to handle multiple link/radio in the future.
To support single wiphy abstraction, introduce link/radio specific helper
function in the mac80211 callback stop(). This way, the callback can be
extended to handle multiple link/radio in the future.
To support single wiphy abstraction, introduce link/radio specific helper
function in the mac80211 callback start(). This way, the callback can be
extended to handle multiple link/radio in the future.
To support single wiphy abstraction, introduce link/radio specific helper
function in the mac80211 callback conf_tx(). This way, the callback can
be extended to handle multiple link/radio in the future.
To support single wiphy abstraction, introduce link/radio specific helper
function in the mac80211 callback bss_info_change(). This way, the
callback can be extended to handle multiple link/radio in the future.
To support single wiphy abstraction, introduce link/radio specific helper
function in the mac80211 callback config(). This way, the callback can be
extended to handle multiple link/radio in the future.
wifi: ath12k: refactor ath12k_mac_register() and ath12k_mac_unregister()
Currently, the mac80211 hw registration procedure is tightly coupled with
the handling of link/radio (ar). Define a new helper function to separate
the link/radio handling from the mac80211 hw registration procedure for
improved code readability. Also, it can be easy to scale these
functionality to support single/multi link operation in the future.
Currently, the MAC setup helper function is accessing the mac80211 hw
data. In the future, to support single/multi link operation, need to
decouple the mac80211 hw data from this helper function so that it can be
easy to scale these functions to support single/multi link operations. So
remove the mac80211 hw access from the ath12k_mac_setup_channels_rates()
helper function.
wifi: ath12k: refactor ath12k_mac_allocate() and ath12k_mac_destroy()
Currently, the MAC allocation and destroy helper functions are tightly
coupled with the link/radio (ar) structure. In the future, to support
single/Multi link operations, need to refactor these helper functions
across the core and mac sub modules, so that it can be easy to scale
these functions to support single/Multi link operations.
Currently, the data path pdev pre alloc and mac allocate are called
separately from the core start procedure. The data path pdev pre alloc
can be called from the mac allocate procedure itself since initialization
related to pdev happens in the mac allocate procedure. So move the caller
of DP pdev pre alloc from the core start procedure to the mac allocate
procedure. This change helps in the future to easily decouple the mac
allocate procedure from core start handling in order to support MLO in
multi chip.
Jeff Johnson [Thu, 11 Jan 2024 18:05:32 +0000 (10:05 -0800)]
wifi: ath12k: Use initializers for QMI message buffers
Currently most of the QMI messaging functions use memset() to zero out
the QMI message buffers. Prefer to use a {} initializer to allow the
compiler to generate optimized code and avoid the function call
overhead.
Jeff Johnson [Thu, 11 Jan 2024 18:05:31 +0000 (10:05 -0800)]
wifi: ath12k: Add missing qmi_txn_cancel() calls
Per the QMI documentation "A client calling qmi_txn_init() must call
either qmi_txn_wait() or qmi_txn_cancel() to free up the allocated
resources."
Unfortunately, in most of the ath12k messaging functions, when
qmi_send_request() fails, the function returns without performing the
necessary cleanup. So update those functions to call qmi_txn_cancel()
when qmi_send_request() fails.
Currently most of the ath12k QMI messaging functions define their
struct qmi_txn variables with a {} initializer. However, all of these
functions subsequently call qmi_txn_init(), and the very first thing
that function does is zero the struct. Hence, the initializers are
unnecessary. Since these consume code space and cpu cycles, remove
them.
Wen Gong [Thu, 11 Jan 2024 13:56:59 +0000 (15:56 +0200)]
wifi: ath11k: use WMI_VDEV_SET_TPC_POWER_CMDID when EXT_TPC_REG_SUPPORT for 6 GHz
When station is connected to a 6 GHz AP, it has 2 ways to configure
the power limit to firmware. The first way is to send 2 WMI commands
WMI_PDEV_PARAM_TXPOWER_LIMIT2G/WMI_PDEV_PARAM_TXPOWER_LIMIT5G to
firmware, the second way is to send WMI_VDEV_SET_TPC_POWER_CMDID to
firmware which include more parameters for power control.
When firmware supports SERVICE_EXT_TPC_REG, it means firmware supports
WMI_VDEV_SET_TPC_POWER_CMDID, then ath11k selects the second way.
Wen Gong [Thu, 11 Jan 2024 13:56:58 +0000 (15:56 +0200)]
wifi: ath11k: add WMI_TLV_SERVICE_EXT_TPC_REG_SUPPORT service bit
Firmware advertises support for SERVICE_EXT_TPC_REG via a WMI service bit.
Add the definition of this service bit so that a subsequent patch can
check whether or not firmware supports this service.
Signed-off-by: Wen Gong <quic_wgong@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20231218085844.2658-11-quic_bqiang@quicinc.com
Signed-off-by: Wen Gong <quic_wgong@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20231218085844.2658-10-quic_bqiang@quicinc.com
Wen Gong [Thu, 11 Jan 2024 13:56:58 +0000 (15:56 +0200)]
wifi: ath11k: save max transmit power in vdev start response event from firmware
Save the max transmit power received in the vdev start response event from
firmware. A subsequent patch will use this to calculate the final power
value for WMI_VDEV_SET_TPC_POWER_CMDID.
Signed-off-by: Wen Gong <quic_wgong@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20231218085844.2658-9-quic_bqiang@quicinc.com
Signed-off-by: Wen Gong <quic_wgong@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20231218085844.2658-8-quic_bqiang@quicinc.com
Signed-off-by: Wen Gong <quic_wgong@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20231218085844.2658-7-quic_bqiang@quicinc.com
Wen Gong [Thu, 11 Jan 2024 13:56:58 +0000 (15:56 +0200)]
wifi: ath11k: update regulatory rules when connect to AP on 6 GHz band for station
When station connect to AP on 6 GHz band, it needs switch the regulatory
rules according to the regulatory info sub field in HE operation element.
Switch to the power type which AP used for station interface.
Signed-off-by: Wen Gong <quic_wgong@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20231218085844.2658-6-quic_bqiang@quicinc.com
Wen Gong [Thu, 11 Jan 2024 13:56:57 +0000 (15:56 +0200)]
wifi: ath11k: update regulatory rules when interface added
There are two power types for 6 GHz regulatory, one is AP, another
is client.
When firmware boots up, WMI_REG_CHAN_LIST_CC_EXT_EVENTID is sent from
firmware at an early stage, the interface mode is not decided at
this point, then ath11k select reg rules of AP type as default.
After interface is created, it is exactly decided the interface
type such as AP/mesh point/station. Then ath11k need to update
reg rules to the exact power type matched to the interface type.
The client power type is used for station interface, and AP power
type is used for AP/mesh point interface.
Signed-off-by: Wen Gong <quic_wgong@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20231218085844.2658-5-quic_bqiang@quicinc.com
Baochen Qiang [Thu, 11 Jan 2024 13:56:57 +0000 (15:56 +0200)]
wifi: ath11k: fix a possible dead lock caused by ab->base_lock
spin_lock()/spin_unlock() are used in ath11k_reg_chan_list_event() to
acquire/release ab->base_lock. For now this is safe because that
function is only called in soft IRQ context.
But ath11k_reg_chan_list_event() will be called from process
context in an upcoming patch, and this can result in a deadlock if
ab->base_lock is acquired in process context and then soft IRQ occurs
on the same CPU and tries to acquire that lock.
Fix it by using spin_lock_bh() and spin_unlock_bh() instead.
Fixes: 69a0fcf8a9f2 ("ath11k: Avoid reg rules update during firmware recovery") Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com> Signed-off-by: Wen Gong <quic_wgong@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20231218085844.2658-4-quic_bqiang@quicinc.com
Wen Gong [Thu, 11 Jan 2024 13:56:57 +0000 (15:56 +0200)]
wifi: ath11k: store cur_regulatory_info for each radio
The regulatory info of WMI_REG_CHAN_LIST_CC_EXT_EVENTID is not saved
in ath11k now, the info should be saved in ath11k. Save the info for
each radio and support switch regulatory rules dynamically.
As mac.c will also call ath11k_reg_handle_chan_list() in next patches move the
function to reg.c.
Signed-off-by: Wen Gong <quic_wgong@quicinc.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Baochen Qiang <quic_bqiang@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20231218085844.2658-3-quic_bqiang@quicinc.com
Wen Gong [Thu, 11 Jan 2024 13:56:57 +0000 (15:56 +0200)]
wifi: ath11k: add support to select 6 GHz regulatory type
There are 3 types of regulatory rules for AP mode and 6 type for
station mode. Add wmi_vdev_type and ieee80211_ap_reg_power to
select the exact reg rules.
Since 'ath12k_wmi_tlv_parse_alloc()' always operates on 'skb->data,
skb->len' tuple, it may be simplified to pass the only 'skb' argument
instead (which implies refactoring of 'ath12k_pull_bcn_tx_status_ev()',
'ath12k_pull_chan_info_ev()' and 'ath12k_pull_pdev_temp_ev()' in the
same way). This is an ath12k counterpart of the recently submitted
ath11k patch and compile tested only as well.
Jeff Johnson [Sun, 17 Dec 2023 17:26:46 +0000 (09:26 -0800)]
wifi: ath12k: Make QMI message rules const
Commit ff6d365898d4 ("soc: qcom: qmi: use const for struct
qmi_elem_info") allows QMI message encoding/decoding rules
to be const, so do that for ath12k.
Baochen Qiang [Sun, 17 Dec 2023 11:29:04 +0000 (13:29 +0200)]
wifi: ath12k: support default regdb while searching board-2.bin for WCN7850
Sometimes board-2.bin does not have the regdb data which matches the
parameters such as vendor, device, subsystem-vendor, subsystem-device
etc. Add default regdb data with only 'bus=%s' into board-2.bin for
WCN7850, then ath12k uses 'bus=pci' to search regdb data in board-2.bin
for WCN7850.
[ 46.114895] ath12k_pci 0000:03:00.0: boot using board name 'bus=pci,vendor=17cb,device=1107,subsystem-vendor=17cb,subsystem-device=3378,qmi-chip-id=2,qmi-board-id=260'
[ 46.118167] ath12k_pci 0000:03:00.0: boot firmware request ath12k/WCN7850/hw2.0/board-2.bin size 380280
[ 46.118173] ath12k_pci 0000:03:00.0: board name
[ 46.118176] ath12k_pci 0000:03:00.0: 00000000: 62 75 73 3d 70 63 69 bus=pci
[ 46.118179] ath12k_pci 0000:03:00.0: failed to fetch regdb data for bus=pci,vendor=17cb,device=1107,subsystem-vendor=17cb,subsystem-device=3378,qmi-chip-id=2,qmi-board-id=260 from ath12k/WCN7850/hw2.0/board-2.bin
[ 46.118239] ath12k_pci 0000:03:00.0: boot using board name 'bus=pci'
[ 46.119842] ath12k_pci 0000:03:00.0: boot firmware request ath12k/WCN7850/hw2.0/board-2.bin size 380280
[ 46.119847] ath12k_pci 0000:03:00.0: board name
[ 46.119849] ath12k_pci 0000:03:00.0: 00000000: 62 75 73 3d 70 63 69 bus=pci
[ 46.119852] ath12k_pci 0000:03:00.0: boot found match regdb data for name 'bus=pci'
[ 46.119855] ath12k_pci 0000:03:00.0: boot found regdb data for 'bus=pci'
[ 46.119857] ath12k_pci 0000:03:00.0: fetched regdb
Wen Gong [Sun, 17 Dec 2023 11:29:04 +0000 (13:29 +0200)]
wifi: ath12k: add support to search regdb data in board-2.bin for WCN7850
Currently ath12k only downloads the same regdb.bin file for all WCN7850
chips, actually ath12k needs to distinguish all different WCN7850 chips.
This is to re-use the string type which includes bus, chip id, board id,
vendor, device, subsystem-vendor, subsystem-device and variant for
WCN7850 to distinguish different regdb in board-2.bin.
ath12k will first load board-2.bin and search in it for the regdb data
with the above parameters. If matched with one regdb data, download it
to firmware. And if not, download regdb.bin instead.
Add enum value ATH12K_BD_IE_REGDB and enum type ath12k_bd_ie_regdb_type
to distinguish regdb data and board data since they are in the same file
board-2.bin.
Wen Gong [Sun, 17 Dec 2023 11:29:04 +0000 (13:29 +0200)]
wifi: ath12k: add fallback board name without variant while searching board-2.bin
Currently a variant value read from DT or SMBIOS is considered while
searching board-2.bin, this may fail because not all board-2.bin files
contains that symbol.
Add fallback board name which removes variant value and searches again
in board-2.bin when fails to increase boot up success rate.
Wen Gong [Sun, 17 Dec 2023 11:29:03 +0000 (13:29 +0200)]
wifi: ath12k: add string type to search board data in board-2.bin for WCN7850
Currently ath12k only supports string type with bus, chip id and board id
such as "bus=ahb,qmi-chip-id=1,qmi-board-id=4" for AHB bus chip and
"bus=pci,qmi-chip-id=0,qmi-board-id=255" for PCI bus chip in board-2.bin.
For WCN7850, it is not enough to distinguish all different chips.
Add a new string type which includes bus, chip id, board id, vendor,
device, subsystem-vendor and subsystem-device for WCN7850.
ath12k will first load board-2.bin and searches in it for the board data
with the above parameters. If matched with one board data, ath12k downloads
it to firmware. And if not, downloads board.bin instead.
Jeff Johnson [Sun, 17 Dec 2023 11:29:03 +0000 (13:29 +0200)]
wifi: ath10k: remove duplicate memset() in 10.4 TDLS peer update
In [1] it was identified that in ath10k_wmi_10_4_gen_tdls_peer_update()
the memset(skb->data, 0, sizeof(*cmd)) is unnecessary since function
ath10k_wmi_alloc_skb() already zeroes skb->data, so remove it.
Jeff Johnson [Sun, 17 Dec 2023 11:29:03 +0000 (13:29 +0200)]
wifi: ath10k: use flexible array in struct wmi_tdls_peer_capabilities
Currently struct wmi_tdls_peer_capabilities defines:
struct wmi_channel peer_chan_list[1];
Per the guidance in [1] this should be a flexible array, and at one
point Gustavo was trying to fix this [2], but had questions about the
correct behavior when the associated peer_chan_len is 0.
I have been unable to determine if firmware requires that at least one
record be present even if peer_chan_len is 0. But since that is the
current behavior, follow the example from [3] and replace the
one-element array with a union that contains both a flexible array and
a single instance of the array element. This results in a struct that
has the same footprint as the original, so no other driver changes are
required.
Jeff Johnson [Sun, 17 Dec 2023 11:29:03 +0000 (13:29 +0200)]
wifi: ath10k: remove unused template structs
Currently both the wmi_bcn_tmpl_cmd and wmi_prb_tmpl_cmd structs
define:
u8 data[1];
Per the guidance in [1] both instances of this should be flexible
arrays. However during conversion it was discovered that neither of
these structs are actually used, so just remove them.
Currently struct wmi_pdev_chanlist_update_event defines:
struct wmi_channel channel_list[1];
Per the guidance in [1] this should be a flexible array. However
during conversion it was discovered that this struct is not used, so
just remove the entire struct.
Jeff Johnson [Sun, 17 Dec 2023 11:29:03 +0000 (13:29 +0200)]
wifi: ath10k: use flexible arrays for WMI start scan TLVs
Currently ath10k defines the following struct:
struct wmi_start_scan_tlvs {
u8 tlvs[0];
} __packed;
Per the guidance in [1] this should be a flexible array. However, a
direct replace to u8 tlvs[] results in the compilation error:
flexible array member in a struct with no named members
This is because C99 6.7.2.1 (16) requires that a structure containing
a flexible array member must have more than one named member.
So rather than defining a separate struct wmi_start_scan_tlvs which
contains the flexible tlvs[] array, just define the tlvs[] array where
struct wmi_start_scan_tlvs is being used.
Jeff Johnson [Sun, 17 Dec 2023 11:29:03 +0000 (13:29 +0200)]
wifi: ath10k: use flexible array in struct wmi_host_mem_chunks
Currently struct wmi_host_mem_chunks defines:
struct host_memory_chunk items[1];
Per the guidance in [1] this should be a flexible array. However there
is a documented requirement:
some fw revisions require at least 1 chunk regardless of count
To satisfy this requirement, follow the guidance from [2] and wrap the
array in a union which contains both the flexible array and a single
instance of the underlying struct. Since the footprint of the struct
is unchanged, no additional driver changes are required.
Dmitry Antipov [Sun, 17 Dec 2023 11:29:03 +0000 (13:29 +0200)]
wifi: ath9k: fix LNA selection in ath_ant_try_scan()
In 'ath_ant_try_scan()', (most likely) the 2nd LNA's signal
strength should be used in comparison against RSSI when
selecting first LNA as the main one. Compile tested only.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Uwe Kleine-König [Sun, 17 Dec 2023 11:29:03 +0000 (13:29 +0200)]
wifi: ath9k: Convert to platform remove callback returning void
The .remove() callback for a platform driver returns an int which makes
many driver authors wrongly assume it's possible to do error handling by
returning an error code. However the value returned is ignored (apart
from emitting a warning) and this typically results in resource leaks.
To improve here there is a quest to make the remove callback return
void. In the first step of this quest all drivers are converted to
.remove_new(), which already returns void. Eventually after all drivers
are converted, .remove_new() will be renamed to .remove().
Trivially convert this driver from always returning zero in the remove
callback to the void returning variant.
Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> Reviewed-by: Jeff Johnson <quic_jjohnson@quicinc.com> Acked-by: Toke Høiland-Jørgensen <toke@toke.dk> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20231117093056.873834-11-u.kleine-koenig@pengutronix.de
Since 'ath11k_wmi_tlv_parse_alloc()' always operates on
'skb->data, skb->len' tuple, it may be simplified to pass
the only 'skb' argument instead (which also implies
refactoring of 'ath11k_pull_bcn_tx_status_ev()' and
'ath11k_pull_chan_info_ev()' in the same way). Compile
tested only.
Xingyuan Mo [Sun, 17 Dec 2023 11:29:01 +0000 (13:29 +0200)]
wifi: ath10k: fix NULL pointer dereference in ath10k_wmi_tlv_op_pull_mgmt_tx_compl_ev()
We should check whether the WMI_TLV_TAG_STRUCT_MGMT_TX_COMPL_EVENT tlv is
present before accessing it, otherwise a null pointer deference error will
occur.
Fixes: dc405152bb64 ("ath10k: handle mgmt tx completion event") Signed-off-by: Xingyuan Mo <hdthky0@gmail.com> Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com> Signed-off-by: Kalle Valo <quic_kvalo@quicinc.com> Link: https://msgid.link/20231208043433.271449-1-hdthky0@gmail.com
wifi: mt76: mt7996: Use DECLARE_FLEX_ARRAY() and fix -Warray-bounds warnings
Transform zero-length arrays `rate`, `adm_stat` and `msdu_cnt` into
proper flexible-array members in anonymous union in `struct
mt7996_mcu_all_sta_info_event` via the DECLARE_FLEX_ARRAY()
helper; and fix multiple -Warray-bounds warnings:
drivers/net/wireless/mediatek/mt76/mt7996/mcu.c:544:61: warning: array subscript <unknown> is outside array bounds of 'struct <anonymous>[0]' [-Warray-bounds=]
drivers/net/wireless/mediatek/mt76/mt7996/mcu.c:551:58: warning: array subscript <unknown> is outside array bounds of 'struct <anonymous>[0]' [-Warray-bounds=]
drivers/net/wireless/mediatek/mt76/mt7996/mcu.c:553:58: warning: array subscript <unknown> is outside array bounds of 'struct <anonymous>[0]' [-Warray-bounds=]
drivers/net/wireless/mediatek/mt76/mt7996/mcu.c:530:61: warning: array subscript <unknown> is outside array bounds of 'struct <anonymous>[0]' [-Warray-bounds=]
drivers/net/wireless/mediatek/mt76/mt7996/mcu.c:538:66: warning: array subscript <unknown> is outside array bounds of 'struct <anonymous>[0]' [-Warray-bounds=]
drivers/net/wireless/mediatek/mt76/mt7996/mcu.c:540:66: warning: array subscript <unknown> is outside array bounds of 'struct <anonymous>[0]' [-Warray-bounds=]
drivers/net/wireless/mediatek/mt76/mt7996/mcu.c:520:57: warning: array subscript <unknown> is outside array bounds of 'struct all_sta_trx_rate[0]' [-Warray-bounds=]
drivers/net/wireless/mediatek/mt76/mt7996/mcu.c:526:76: warning: array subscript <unknown> is outside array bounds of 'struct all_sta_trx_rate[0]' [-Warray-bounds=]
drivers/net/wireless/mediatek/mt76/mt7996/mcu.c:526:76: warning: array subscript <unknown> is outside array bounds of 'struct all_sta_trx_rate[0]' [-Warray-bounds=]
drivers/net/wireless/mediatek/mt76/mt7996/mcu.c:526:76: warning: array subscript <unknown> is outside array bounds of 'struct all_sta_trx_rate[0]' [-Warray-bounds=]
drivers/net/wireless/mediatek/mt76/mt7996/mcu.c:526:76: warning: array subscript <unknown> is outside array bounds of 'struct all_sta_trx_rate[0]' [-Warray-bounds=]
This results in no differences in binary output, helps with the ongoing
efforts to globally enable -Warray-bounds.
Reviewed-by: Kees Cook <keescook@chromium.org> Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://msgid.link/ZXiU9ayVCslt3qiI@work
Kalle Valo [Thu, 14 Dec 2023 16:17:40 +0000 (18:17 +0200)]
wifi: ath11k: workaround too long expansion sparse warnings
In v6.7-rc1 sparse warns:
drivers/net/wireless/ath/ath11k/mac.c:4702:15: error: too long token expansion
drivers/net/wireless/ath/ath11k/mac.c:4702:15: error: too long token expansion
drivers/net/wireless/ath/ath11k/mac.c:8393:23: error: too long token expansion
drivers/net/wireless/ath/ath11k/mac.c:8393:23: error: too long token expansion
Workaround the warnings by refactoring the code to a new function, which also
reduces code duplication. And in the new function use max3() to make the code
more readable.
Revert "wifi: ath12k: use ATH12K_PCI_IRQ_DP_OFFSET for DP IRQ"
This reverts commit 1f1f7d548a00ebe50808cb1f580df9693e194a7c. The commit
caused bootup failure on QCN9274 hw2.0 platform. Incorrect hardcode DP
irq offset overwrite the CE irq, which caused the driver to miss the
mandatory bootup message from the firmware through the CE interrupt. This
occurs because the CE count differs between platforms. The revert has no
impact since the original change was based on an incorrect assumption.
Log:
ath12k_pci 0000:06:00.0: fw_version 0x1011001d fw_build_timestamp 2022-12-02 01:16 fw_build_id QC_IMAGE_VERSION_STRING=WLAN.WBE.1.0.1-00029-QCAHKSWPL_SILICONZ-1
ath12k_pci 0000:06:00.0: failed to receive control response completion, polling..
ath12k_pci 0000:06:00.0: Service connect timeout
ath12k_pci 0000:06:00.0: failed to connect to HTT: -110
ath12k_pci 0000:06:00.0: failed to start core: -110
Dmitry Antipov [Wed, 13 Dec 2023 05:14:43 +0000 (08:14 +0300)]
wifi: rt2x00: remove useless code in rt2x00queue_create_tx_descriptor()
In 'rt2x00queue_create_tx_descriptor()', there is no need to call
'ieee80211_get_rts_cts_rate()' while checking for RTS/CTS frame
since this function returns NULL or pointer to internal bitrate
table entry, and the return value is not actually used. Compile
tested only.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Ping-Ke Shih [Mon, 11 Dec 2023 08:33:39 +0000 (16:33 +0800)]
wifi: rtw89: mac: add suffix _ax to MAC functions
Many existing MAC access functions are used by WiFi 6 chips only, so add
suffix _ax to be clearer. Some are common and can be used by WiFi 7, so
export this kind of functions. This patch doesn't change logic at all.
Ping-Ke Shih [Mon, 11 Dec 2023 08:33:38 +0000 (16:33 +0800)]
wifi: rtw89: mac: add flags to check if CMAC and DMAC are enabled
Before accessing CMAC and DMAC registers, we should ensure they have been
powered on, so add flag to determine the state. For old chips, we read
registers and check corresponding bit, but it takes extra cost to read.
Ping-Ke Shih [Mon, 11 Dec 2023 08:33:36 +0000 (16:33 +0800)]
wifi: rtw89: add XTAL SI for WiFi 7 chips
The XTAL SI is a serial interface to indirectly access registers of
analog hardware circuit. Since WiFi 7 chips use different registers, add
a ops to access them via common functions. This patch doesn't change logic
for existing chips.
Ping-Ke Shih [Wed, 13 Dec 2023 00:50:54 +0000 (08:50 +0800)]
wifi: rtw89: phy: print out RFK log with formatted string
With formatted string loaded from firmware file, we can use the formatted
string ID and get corresponding string, and then use regular rtw89_debug()
to show the message if debug mask of RFK is enabled.
If the string ID doesn't present, fallback to print plain hexadecimal.
Ping-Ke Shih [Wed, 13 Dec 2023 00:50:53 +0000 (08:50 +0800)]
wifi: rtw89: parse and print out RFK log from C2H events
RFK log events contains two types. One called RUN log is to reflect state
during RFK is running, and it replies on formatted string loaded from
firmware file, but print this type as plain hexadecimal only in this patch.
The other is REPORT log that reflects the final result of a RFK, and
each calibration has its own struct to carry many specific information.
Ping-Ke Shih [Wed, 13 Dec 2023 00:50:52 +0000 (08:50 +0800)]
wifi: rtw89: add C2H event handlers of RFK log and report
Trigger a RFK (RF calibration) in firmware by a H2C command, and in
progress it reports log and a result finally by C2H events. Firstly, add
prototype of the C2H event handlers to have a simple picture of framework.
The callers who trigger H2C will wait until a C2H event is received,
so we must process these C2H events in receiving process. Thus, mark this
kind of C2H events as atomic. Also, timestamp is also useful for
debugging, mark C2H events carrying RFK log as atomic as well.
Ping-Ke Shih [Wed, 13 Dec 2023 00:50:51 +0000 (08:50 +0800)]
wifi: rtw89: load RFK log format string from firmware file
To debug RFK (RF calibration) in firmware, it sends log via firmware C2H
events to driver with string format ID and four arguments. Load formatted
string from firmware file, and the string ID can get back its string. Then,
use regular print format to show the message.
This firmware element layout looks like
+============================================+
| elm ID | elm size | version | |
+----------+----------+----------+-----------+
| | nr |rsvd |rfk_id|rsvd|
+--------------------------------------------+
| offset[] (__le16 * nr) |
| ... |
+--------------------------------------------+
| formatted string with null termintor (*nr) |
| ... |
+============================================+
* a firmware file can contains more than one elements with this element ID
named RTW89_FW_ELEMENT_ID_RFKLOG_FMT (19), because many RFK needs its
own formatted strings, so add 'rfk_id' to know it belongs to which RFK.
* the 'formatted string' just follow 'offset[]' without padding to align
32bits.
Ping-Ke Shih [Wed, 13 Dec 2023 00:50:50 +0000 (08:50 +0800)]
wifi: rtw89: fw: add version field to BB MCU firmware element
8922AE has more than one hardware version, and they use different BB MCU
firmware, so occupy a byte from element priv[] to annotate version. Since
there are more than one firmware and only matched version is adopted,
return 1 to ignore not matched firmware.
+===========================================+
| elm ID | elm size | version | |
+----------+----------+----------+----------+
| | element_priv[] |
+-------------------------------------------+
change to |
v
+===========================================+
| elm ID | elm size | version | |
+----------+----------+----------+----------+
| | cv | element_rsvd[] |
+-------------------------------------------+
Ping-Ke Shih [Wed, 13 Dec 2023 00:50:49 +0000 (08:50 +0800)]
wifi: rtw89: fw: load TX power track tables from fw_element
The TX power track tables are used to define compensation power reflected
to thermal value. Currently, we have 16 (2 * 4 * 2) tables made by
combinations of
{negative/positive thermal value, 2GHz/2GHz-CCK/5GHz/6GHz, path A/B}
David Lin [Fri, 15 Dec 2023 00:51:18 +0000 (08:51 +0800)]
wifi: mwifiex: configure BSSID consistently when starting AP
AP BSSID configuration is missing at AP start. Without this fix, FW returns
STA interface MAC address after first init. When hostapd restarts, it gets MAC
address from netdev before driver sets STA MAC to netdev again. Now MAC address
between hostapd and net interface are different causes STA cannot connect to
AP. After that MAC address of uap0 mlan0 become the same. And issue disappears
after following hostapd restart (another issue is AP/STA MAC address become the
same).
This patch fixes the issue cleanly.
Signed-off-by: David Lin <yu-hao.lin@nxp.com> Fixes: 12190c5d80bd ("mwifiex: add cfg80211 start_ap and stop_ap handlers") Cc: stable@vger.kernel.org Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com> Tested-by: Rafael Beims <rafael.beims@toradex.com> # Verdin iMX8MP/SD8997 SD Acked-by: Brian Norris <briannorris@chromium.org> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://msgid.link/20231215005118.17031-1-yu-hao.lin@nxp.com
David Lin [Fri, 8 Dec 2023 23:40:29 +0000 (07:40 +0800)]
wifi: mwifiex: add extra delay for firmware ready
For SDIO IW416, due to a bug, FW may return ready before complete full
initialization. Command timeout may occur at driver load after reboot.
Workaround by adding 100ms delay at checking FW status.
Signed-off-by: David Lin <yu-hao.lin@nxp.com> Cc: stable@vger.kernel.org Reviewed-by: Francesco Dolcini <francesco.dolcini@toradex.com> Acked-by: Brian Norris <briannorris@chromium.org> Tested-by: Marcel Ziswiler <marcel.ziswiler@toradex.com> # Verdin AM62 (IW416) Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://msgid.link/20231208234029.2197-1-yu-hao.lin@nxp.com
Randy Dunlap [Wed, 13 Dec 2023 05:48:09 +0000 (21:48 -0800)]
wifi: mac80211: sta_info.c: fix sentence grammar
Correct run-on sentences by changing "," to ";".
Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Kalle Valo <kvalo@kernel.org> Cc: linux-wireless@vger.kernel.org Cc: "David S. Miller" <davem@davemloft.net> Cc: Eric Dumazet <edumazet@google.com> Cc: Jakub Kicinski <kuba@kernel.org> Cc: Paolo Abeni <pabeni@redhat.com> Link: https://msgid.link/20231213054809.23475-1-rdunlap@infradead.org Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Randy Dunlap [Wed, 13 Dec 2023 05:48:00 +0000 (21:48 -0800)]
wifi: mac80211: rx.c: fix sentence grammar
Correct a run-on sentence by changing "," to ";".
Add a subject in one sentence.
Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: Kalle Valo <kvalo@kernel.org> Cc: linux-wireless@vger.kernel.org Cc: "David S. Miller" <davem@davemloft.net> Cc: Eric Dumazet <edumazet@google.com> Cc: Jakub Kicinski <kuba@kernel.org> Cc: Paolo Abeni <pabeni@redhat.com> Link: https://msgid.link/20231213054800.22561-1-rdunlap@infradead.org Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Randy Dunlap [Wed, 13 Dec 2023 04:43:15 +0000 (20:43 -0800)]
wifi: nl80211: fix grammar & spellos
Correct spelling as reported by codespell.
Correct run-on sentences and other grammar issues.
Add hyphenation of adjectives.
Correct some punctuation.
Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: linux-wireless@vger.kernel.org Cc: Kalle Valo <kvalo@kernel.org> Cc: "David S. Miller" <davem@davemloft.net> Cc: Eric Dumazet <edumazet@google.com> Cc: Jakub Kicinski <kuba@kernel.org> Cc: Paolo Abeni <pabeni@redhat.com> Link: https://msgid.link/20231213044315.19459-1-rdunlap@infradead.org Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Randy Dunlap [Wed, 13 Dec 2023 04:35:58 +0000 (20:35 -0800)]
wifi: cfg80211: fix spelling & punctutation
Correct spelling and run-on sentences.
Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Cc: Johannes Berg <johannes@sipsolutions.net> Cc: linux-wireless@vger.kernel.org Cc: Kalle Valo <kvalo@kernel.org> Cc: "David S. Miller" <davem@davemloft.net> Cc: Eric Dumazet <edumazet@google.com> Cc: Jakub Kicinski <kuba@kernel.org> Cc: Paolo Abeni <pabeni@redhat.com> Link: https://msgid.link/20231213043558.10409-1-rdunlap@infradead.org Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Hancheng Yang [Tue, 5 Dec 2023 17:06:23 +0000 (18:06 +0100)]
wifi: ath9k: reset survey of current channel after a scan started
In the `ath_set_channel()` function, we only reset surveys that are not
from the current channel. This leads to the accumulation of survey data for
the current channel indefinitely. This may not be the most optimal
approach, as we want the ACS to rely on the most recent survey.
So reset the survey data for the current channel at the start of each scan.
Lingbo Kong [Wed, 6 Dec 2023 14:17:59 +0000 (22:17 +0800)]
wifi: ath12k: fix the issue that the multicast/broadcast indicator is not read correctly for WCN7850
We observe some packets are discarded in ieee80211_rx_handlers_result
function for WCN7850. This is because the way to get multicast/broadcast
indicator with RX_MSDU_END_INFO5_DA_IS_MCBC & info5 is incorrect. It should
use RX_MSDU_END_INFO13_MCAST_BCAST & info13 to get multicast/broadcast
indicator.
Jeff Johnson [Tue, 5 Dec 2023 21:00:17 +0000 (13:00 -0800)]
wifi: ath11k: Fix ath11k_htc_record flexible record
Transform the zero-length ath11k_htc_record::credit_report array into
a proper flexible array. Since this is the only array in
ath11k_htc_record, remove the unnecessary union.
Chris Morgan [Fri, 8 Dec 2023 15:07:39 +0000 (09:07 -0600)]
wifi: rtw88: Use random MAC when efuse MAC invalid
When the MAC address read from the efuse data is invalid, warn the
user and use a random MAC address instead.
On a device I am currently using (Anbernic RG-ARC) with a rtw8821cs
the efuse appears to be incompletely/improperly programmed. The MAC
address reads as ff:ff:ff:ff:ff:ff. When networkmanager attempts to
initiate a connection (and I haven't hard-coded a MAC address or
set it to random) it fails to establish a connection.
Dmitry Antipov [Mon, 4 Dec 2023 17:11:28 +0000 (20:11 +0300)]
wifi: wfx: fix possible NULL pointer dereference in wfx_set_mfp_ap()
Since 'ieee80211_beacon_get()' can return NULL, 'wfx_set_mfp_ap()'
should check the return value before examining skb data. So convert
the latter to return an appropriate error code and propagate it to
return from 'wfx_start_ap()' as well. Compile tested only.
Signed-off-by: Dmitry Antipov <dmantipov@yandex.ru> Tested-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Acked-by: Jérôme Pouiller <jerome.pouiller@silabs.com> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://lore.kernel.org/r/20231204171130.141394-1-dmantipov@yandex.ru
Arnd Bergmann [Mon, 4 Dec 2023 07:30:13 +0000 (08:30 +0100)]
wifi: rtw89: avoid stringop-overflow warning
After -Wstringop-overflow got enabled, the rtw89 driver produced
two odd warnings with gcc-13:
drivers/net/wireless/realtek/rtw89/coex.c: In function 'rtw89_btc_ntfy_scan_start':
drivers/net/wireless/realtek/rtw89/coex.c:5362:50: error: writing 1 byte into a region of size 0 [-Werror=stringop-overflow=]
5362 | wl->dbcc_info.scan_band[phy_idx] = band;
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~
In file included from drivers/net/wireless/realtek/rtw89/coex.h:8,
from drivers/net/wireless/realtek/rtw89/coex.c:5:
drivers/net/wireless/realtek/rtw89/core.h:1441:12: note: at offset [64, 255] into destination object 'scan_band' of size 2
1441 | u8 scan_band[RTW89_PHY_MAX]; /* scan band in each phy */
| ^~~~~~~~~
drivers/net/wireless/realtek/rtw89/coex.c: In function 'rtw89_btc_ntfy_switch_band':
drivers/net/wireless/realtek/rtw89/coex.c:5406:50: error: writing 1 byte into a region of size 0 [-Werror=stringop-overflow=]
5406 | wl->dbcc_info.scan_band[phy_idx] = band;
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~
drivers/net/wireless/realtek/rtw89/core.h:1441:12: note: at offset [64, 255] into destination object 'scan_band' of size 2
1441 | u8 scan_band[RTW89_PHY_MAX]; /* scan band in each phy */
| ^~~~~~~~~
I don't know what happened here, but adding an explicit range check
shuts up the output.
Dmitry Antipov [Fri, 8 Dec 2023 15:31:25 +0000 (18:31 +0300)]
wifi: mac80211: drop spurious WARN_ON() in ieee80211_ibss_csa_beacon()
The WARN_ON() in subject was actually seen only once, with 5.10.200
under syzkaller. It looks like a weird artifact of (ab?)using the
syzkaller itself [1], and hopefully may be safely removed.
Johannes Berg [Mon, 11 Dec 2023 07:05:29 +0000 (09:05 +0200)]
wifi: mac80211: don't set ESS capab bit in assoc request
The ESS capability bit is reserved in frames transmitted by
the client, so we shouldn't set it. Since we've set it for
decades, keep that old behaviour unless we're connection to
a new EHT AP.
Benjamin Berg [Mon, 11 Dec 2023 07:05:28 +0000 (09:05 +0200)]
wifi: cfg80211: consume both probe response and beacon IEs
When doing a channel switch, cfg80211_update_known_bss may be called
with a BSS where both proberesp_ies and beacon_ies is set. If that
happens, both need to be consumed.
Benjamin Berg [Mon, 11 Dec 2023 07:05:27 +0000 (09:05 +0200)]
wifi: cfg80211: generate an ML element for per-STA profiles
The specification says that this information should not be explicitly
included in the per-STA profile. However, we need this information
readily available in the BSS for userspace and also internally when
associating. As such, append the appropriate element before
adding/updating the BSS.
Ilan Peer [Mon, 11 Dec 2023 07:05:22 +0000 (09:05 +0200)]
wifi: cfg80211: Add support for setting TID to link mapping
Add support for setting the TID to link mapping for a non-AP MLD
station.
This is useful in cases user space needs to restrict the possible
set of active links, e.g., since it got a BSS Transition Management
request forcing to use only a subset of the valid links etc.
Johannes Berg [Mon, 11 Dec 2023 07:05:21 +0000 (09:05 +0200)]
wifi: mac80211: update some locking documentation
With the locking rework, more functions need to be called
with the wiphy mutex held. Document that, and for that use
the "Context" description that shows up more nicely in the
generated documentation.
projects / thirdparty / linux.git / log
