Lingbo Kong [Mon, 13 Jan 2025 07:48:09 +0000 (15:48 +0800)]
wifi: ath12k: Adjust the timing to access ACPI table
Currently, the timing for accessing the ACPI table is inappropriate.
Due to special ACPI requirements, the ACPI table must be obtained before
downloading the board data file. Therefore, adjust the timing for accessing
the ACPI table accordingly.
This patch will not affect QCN9274, because only WCN7850 supports ACPI.
wifi: ath12k: handle ath12k_mac_ieee80211_sta_bw_to_wmi() for link sta
Currently ath12k_mac_ieee80211_sta_bw_to_wmi() handles the bandwidth from
sta's deflink member. This works only for non-ML station. Now that MLO
support is there, extend this function to use link sta instead of deflink.
Additionally, in ath12k_mac_handle_link_sta_state(), the link sta structure
is not accessible, making it difficult to fetch the bandwidth there.
However, ath12k_mac_station_assoc() does reference the link sta structure.
Therefore, move the initial assignment of the arsta bandwidth member to
ath12k_mac_station_assoc().
An upcoming change will invoke ath12k_mac_ieee80211_sta_bw_to_wmi() from a
line located above its current definition. Hence, relocate it to above
so that it can be invoked later on.
Avula Sri Charan [Fri, 24 Jan 2025 09:00:58 +0000 (14:30 +0530)]
wifi: ath12k: Avoid napi_sync() before napi_enable()
In case of MHI error a reset work will be queued which will try
napi_disable() after napi_synchronize().
As the napi will be only enabled after qmi_firmware_ready event,
trying napi_synchronize() before napi_enable() will result in
indefinite sleep in case of a firmware crash in QMI init sequence.
To avoid this, introduce napi_enabled flag to check if napi is enabled
or not before calling napi_synchronize().
wifi: ath12k: prevent CSA counter to reach 0 and hit WARN_ON_ONCE
Currently, when the driver receives a channel switch count WMI event from
the firmware with a count greater than 1, it calls
ieee80211_beacon_update_cntdwn(). If the beacon transmission fails, the
event will be received again with the previous count value. In this
scenario, the host decrements the mac80211 counter again, causing it to
move ahead of the firmware counter. Ultimately, when the firmware count
reaches 1, the mac80211 counter will reach zero, triggering a
WARN_ON_ONCE(). Therefore, there is a need to check the count value in the
event.
Hence to fix this, maintain the current ongoing counter in arvif. If the
count in the event does not match the expected value, silently discard the
event.
At present, the driver configures the firmware to send the Channel Switch
(CS) count event only when the count reaches zero during a Channel Switch
Announcement (CSA). For frames managed by the upper layer, where the driver
does not update the counter, the CS count in these frames remains unchanged
throughout the entire CSA period. This is because the upper layer is not
aware of the latest ongoing count. Indicating same count value throughout
the CSA time is wrong and could lead to connection instabilities.
Fix this by configuring firmware to send CS count event for every count and
then accordingly decrementing the count in mac80211.
wifi: ath12k: fix handling of CSA offsets in beacon template command
The driver is informed of the counter offsets in the beacon during CSA
through the ieee80211_mutable_offsets structure. According to the
documentation for the cntdwn_counter_offs member, "This array can contain
zero values which should be ignored." However, the current implementation
uses these values unconditionally, without checking for zeros.
Whenever CSA is active, these offsets are guaranteed to be set. Therefore,
add a check for CSA active status before setting the CSA switch count
offsets. This ensures that the offsets are only set when CSA is active,
preventing incorrect configurations.
wifi: ath12k: update beacon template function to use arvif structure
The current code has ath12k_wmi_bcn_tmpl() accepting separate ar and
vdev_id parameters. However, ath12k_link_vif structure can be used to
derive both of these.
Hence, simplify the function signature.
Later change needs arvif pointer access within the function hence it is
better if arvif is directly passed now.
Remi Pommarel [Wed, 29 Jan 2025 16:55:17 +0000 (17:55 +0100)]
wifi: ath12k: remove return for empty tx bitrate in mac_op_sta_statistics
Currently in ath12k_mac_op_sta_statistics() there is the following
logic:
if (!arsta->txrate.legacy && !arsta->txrate.nss)
return;
Because ath12k_sta_statistics is used to report many info to iw wlan0 link,
if it return for empty legacy and nss of arsta->txrate, then the other
stats after it will not be set.
To address this issue remove the return and instead invert the logic to set
the txrate logic if (arsta->txrate.legacy || arsta->txrate.nss).
The same was done also in both ath10k with commit 1cd6ba8ae33e ("ath10k:
remove return for NL80211_STA_INFO_TX_BITRATE") and ath11k as well with
commit 1d795645e1ee ("ath11k: remove return for empty tx bitrate in
mac_op_sta_statistics").
Since inception there is an obvious typo laying around in
ath12k_hal_tx_cmd_ext_desc_setup(). Instead of initializing + adding
flags to tcl_ext_cmd->info1, we initialize + override. This will be needed
in the future to make broadcast frames work with ethernet encapsulation.
Nicolas Escande [Fri, 17 Jan 2025 18:19:53 +0000 (19:19 +0100)]
wifi: ath11k: remove peer extra rssi update
Commit b205ce4c266c ("ath11k: support avg signal in station dump") added an
extra assignment of arsta->rssi_comb in ath11k_dp_rx_update_peer_stats()
when it added the average rssi support. So let's keep only one by removing
the legacy assignment so the two statements about rssi stay next to each
other. Compile tested only.
Dmitry Antipov [Thu, 23 Jan 2025 14:10:58 +0000 (17:10 +0300)]
wifi: ath9k: do not submit zero bytes to the entropy pool
In 'ath_cmn_process_fft()', it doesn't make too much sense to
add zero bytes in attempt to improve randomness. So swap calls
to 'memset()' and 'add_device_randomness()' to feed the pool
with actual FFT results rather than zeroes. Compile tested only.
Signed-off-by: Dmitry Antipov <dmantipov@yandex.ru> Fixes: 2aa56cca3571 ("ath9k: Mix the received FFT bins to the random pool") Acked-by: Toke Høiland-Jørgensen <toke@toke.dk> Link: https://patch.msgid.link/20250123141058.1696502-1-dmantipov@yandex.ru Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Rosen Penev [Tue, 5 Nov 2024 22:23:26 +0000 (14:23 -0800)]
wifi: ath9k: return by of_get_mac_address
When using nvmem, ath9k could potentially be loaded before nvmem, which
loads after mtd. This is an issue if DT contains an nvmem mac address.
If nvmem is not ready in time for ath9k, -EPROBE_DEFER is returned. Pass
it to _probe so that ath9k can properly grab a potentially present MAC
address.
Aaradhana Sahu [Thu, 16 Jan 2025 03:28:35 +0000 (08:58 +0530)]
wifi: ath12k: Fetch regdb.bin file from board-2.bin
Currently, ath12k_core_fetch_regdb() finds regdb.bin file through
board id's but in board-2.bin file regdb.bin file is present with
default board id because of which regdb.bin is not fetched.
Add support to fetch regdb.bin file from board-2.bin through
default board id.
P Praneesh [Mon, 23 Dec 2024 06:01:31 +0000 (11:31 +0530)]
wifi: ath12k: Handle monitor drop TLVs scenario
During monitor destination ring back-pressure, hardware failed to send
HAL_RX_PPDU_END_STATUS_DONE TLV. But driver uses this TLV as a delimiter
to complete one PPDU worth of data parsing. This causes driver to overwrite
the existing PPDU information with the new PPDU information.
Fix it by recording the end reason which is provided under each buffer's
descriptor in skb->cb and uses it while parsing TLV tags to mark the PPDU
end delimiter.
P Praneesh [Mon, 23 Dec 2024 06:01:30 +0000 (11:31 +0530)]
wifi: ath12k: Avoid memory leak while enabling statistics
Driver uses monitor destination rings for extended statistics mode and
standalone monitor mode. In extended statistics mode, TLVs are parsed from
the buffer received from the monitor destination ring and assigned to the
ppdu_info structure to update per-packet statistics. In standalone monitor
mode, along with per-packet statistics, the packet data (payload) is
captured, and the driver updates per MSDU to mac80211.
When the AP interface is enabled, only extended statistics mode is
activated. As part of enabling monitor rings for collecting statistics,
the driver subscribes to HAL_RX_MPDU_START TLV in the filter
configuration. This TLV is received from the monitor destination ring, and
kzalloc for the mon_mpdu object occurs, which is not freed, leading to a
memory leak. The kzalloc for the mon_mpdu object is only required while
enabling the standalone monitor interface. This causes a memory leak while
enabling extended statistics mode in the driver.
Fix this memory leak by removing the kzalloc for the mon_mpdu object in
the HAL_RX_MPDU_START TLV handling. Additionally, remove the standalone
monitor mode handlings in the HAL_MON_BUF_ADDR and HAL_RX_MSDU_END TLVs.
These TLV tags will be handled properly when enabling standalone monitor
mode in the future.
P Praneesh [Mon, 23 Dec 2024 06:01:29 +0000 (11:31 +0530)]
wifi: ath12k: Handle PPDU spread across multiple buffers
Each PPDU contains numerous TLV tags. HAL_RX_PPDU_START marks the start
of the PPDU, and HAL_RX_PPDU_END_STATUS_DONE marks the end. From the
monitor destination rings, the driver retrieves skb containing these
TLV tags and their corresponding data. Sometimes, one PPDU’s information
spreads across multiple skbs. The current parsing logic uses memset on
struct hal_rx_mon_ppdu_info after parsing each skb, leading to information
loss if a PPDU spans multiple skbs. Fix this by setting the
ppdu_continuation flag when the driver fails to get
HAL_RX_PPDU_END_STATUS_DONE, and do memset on struct
hal_rx_mon_ppdu_info only when the ppdu_continuation flag is not set.
P Praneesh [Mon, 23 Dec 2024 06:01:28 +0000 (11:31 +0530)]
wifi: ath12k: Optimize NAPI budget by adjusting PPDU processing
In the current implementation, when PPDU spans multiple ring descriptors,
leading to inefficient use of the NAPI budget. The budget counter is
decremented for each ring descriptor, causing rapid depletion of the
budget even though the processing of a single PPDU might not be complete.
To address this issue, modify the code to decrement the budget counter
only when the driver receives HAL_MON_END_OF_PPDU as the end reason. This
change ensures that the budget is decremented only once per PPDU,
resulting in more efficient utilization of the NAPI budget and better
handling of monitor destination ring.
P Praneesh [Mon, 23 Dec 2024 06:01:27 +0000 (11:31 +0530)]
wifi: ath12k: Handle end reason for the monitor destination ring
Currently, the monitor destination ring's descriptor includes a 2-bit
field for the end reason. Out of all the end reason values, hardware uses
HAL_MON_FLUSH_DETECTED and HAL_MON_PPDU_TRUNCATED to indicate buffers that
should not be processed due to system level errors. Driver should not
process entries with these end reasons, as they contain junk values.
However, the current code lacks end reason-specific checks for the monitor
destination ring, leading to the processing of invalid buffers.
Fix this by adding checks for these two end reasons during the reaping
phase. Free the skb if either HAL_MON_FLUSH_DETECTED or
HAL_MON_PPDU_TRUNCATED is detected, preventing the driver from processing
invalid entries.
P Praneesh [Mon, 23 Dec 2024 06:01:26 +0000 (11:31 +0530)]
wifi: ath12k: Add drop descriptor handling for monitor ring
When monitor block in Hardware experiences internal backpressure, a ring
entry with the EMPTY_DESC reason is received in the monitor destination
ring descriptor. The cookie field for this corresponding entry is invalid.
Currently driver attempts to process this cookie resulting in invalid
buf_id warning logs flooding the console. To fix this, skip processing the
destination descriptor when the EMPTY_DESCRIPTOR bit is set to true in ring
descriptor, thereby avoiding the processing of junk cookies.
P Praneesh [Mon, 23 Dec 2024 06:01:25 +0000 (11:31 +0530)]
wifi: ath12k: Fix end offset bit definition in monitor ring descriptor
End offset for the monitor destination ring descriptor is defined as
16 bits, while the firmware definition specifies only 12 bits.
The remaining bits (bit 12 to bit 15) are reserved and may contain
junk values, leading to invalid information retrieval. Fix this issue
by updating the correct genmask values.
P Praneesh [Mon, 23 Dec 2024 06:01:24 +0000 (11:31 +0530)]
wifi: ath12k: Fix invalid entry fetch in ath12k_dp_mon_srng_process
Currently, ath12k_dp_mon_srng_process uses ath12k_hal_srng_src_get_next_entry
to fetch the next entry from the destination ring. This is incorrect because
ath12k_hal_srng_src_get_next_entry is intended for source rings, not destination
rings. This leads to invalid entry fetches, causing potential data corruption or
crashes due to accessing incorrect memory locations. This happens because the
source ring and destination ring have different handling mechanisms and using
the wrong function results in incorrect pointer arithmetic and ring management.
To fix this issue, replace the call to ath12k_hal_srng_src_get_next_entry with
ath12k_hal_srng_dst_get_next_entry in ath12k_dp_mon_srng_process. This ensures
that the correct function is used for fetching entries from the destination
ring, preventing invalid memory accesses.
P Praneesh [Mon, 23 Dec 2024 06:01:23 +0000 (11:31 +0530)]
wifi: ath12k: Restructure the code for monitor ring processing
Currently, monitor ring reaping and processing occur in the same loop,
which requires holding ring locks until skb processing is complete.
However, only the ring reaping part requires the ring lock; the skb
processing part does not need it. This approach is problematic because it
unnecessarily extends the duration for which the ring locks are held,
leading to increased contention and potential backpressure issues.
Fix it by holding ring locks only during the reaping phase, as skb
processing does not require them. First, reap the monitor destination ring
with the ring lock and queue the skbs into an skb list. Then, process the
skbs in this list in a separate loop without holding the ring lock.
P Praneesh [Mon, 23 Dec 2024 06:01:22 +0000 (11:31 +0530)]
wifi: ath12k: Avoid code duplication in monitor ring processing
The current implementation processes the monitor destination ring using
two separate functions, ath12k_dp_mon_srng_process() for standalone
monitor mode and ath12k_dp_mon_rx_process_stats() for statistics.
However, both functions contain same code which performs monitor ring
reaping and skb data processing. To eliminate redundancy, remove the
duplicate code and use a single function to handle both the cases.
P Praneesh [Mon, 23 Dec 2024 06:01:21 +0000 (11:31 +0530)]
wifi: ath12k: Avoid multiple times configuring monitor filter
ath12k_mac_op_configure_filter() gets called multiple times during
interface bringup. Applying filter configuration from this function leads
to writing same filter configurations multiple times.
Resolve this issue by relocating the filter configuration to
ath12k_mac_config_mon_status_default(), which is invoked by both
ath12k_mac_op_start() and ath12k_mac_op_stop(). Additionally, set the
rxmon_disable flag to true when called from ath12k_mac_op_stop() to
disable the monitor destination ring.
P Praneesh [Mon, 23 Dec 2024 06:01:20 +0000 (11:31 +0530)]
wifi: ath12k: Enable filter config for monitor destination ring
Add provision to configure monitor filter for the destination
ring. These filters are used for requesting statistics or monitor mode
through the monitor destination ring.
P Praneesh [Mon, 23 Dec 2024 06:01:19 +0000 (11:31 +0530)]
wifi: ath12k: Add HTT source ring ID for monitor rings
Add source buffer ring and destination buffer ring ID for monitor rings.
These IDs are used for ring configuration during initial ring setup.
Since monitor rings are enabled based on the rxdma1_enable flag, enable it
in the hardware param for the QCN9274 version 2 hardware and increase the
destination ring size to handle MSDU data buffers.
Rameshkumar Sundaram [Thu, 2 Jan 2025 04:30:48 +0000 (10:00 +0530)]
wifi: ath12k: Fix pdev lookup in WBM error processing
Currently in ath12k_dp_rx_process_wbm_err(), when processing packets
received on the WBM error ring, pdev validation is done based upon the
hw_link_id. But hw_link_id corresponds to link id of a given partner pdev
in a MLO hardware group, and is not the correct index to use to lookup a
pdev in an SoC(ab). As a result, pdev validation fails, and the reaped
packets are dropped instead of being processed.
The correct index to use is the pdev_id, which is already derived in the
function. So update the logic to validate the pdev based upon the pdev_id
instead of the hw_link_id. This matches the logic used in other Rx ring
processing functions.
HTT cookies queued 0
HTT cookies disp. 0
MSDU queued 0
MPDU queued 0
MSDUs dropped 0
Local enqued 0
Local freed 0
HW queued 0
PPDUs reaped 0
Num underruns 0
PPDUs cleaned 0
MPDUs requeued 0
Excessive retries 0
HW rate 0
Sched self triggers 0
Dropped due to SW retries 0
Illegal rate phy errors 0
PDEV continuous xretry 0
TX timeout 9
PDEV resets 0
Stateless TIDs alloc failures 0
PHY underrun 0
MPDU is more than txop limit 0
ath12k PDEV RX stats
====================
Mid PPDU route change 0
Tot. number of statuses 0
Extra frags on rings 0 0
Extra frags on rings 1 0
Extra frags on rings 2 0
Extra frags on rings 3 0
MSDUs delivered to HTT 0
MPDUs delivered to HTT 0
MSDUs delivered to stack 0
MPDUs delivered to stack 0
Oversized AMSUs 0
PHY errors 0
PHY errors drops 0
MPDU errors (FCS, MIC, ENC) 0
VDEV ID 0
VDEV MAC address 00:03:7f:6c:9c:1a
beacon snr 96
data snr 255
num rx frames 0
num rts fail 0
num rts success 0
num rx err 0
num rx discard 0
num tx not acked 0
num tx frames [00] 0
num tx frames [01] 0
num tx frames [02] 0
num tx frames [03] 2
num tx frames retries [00] 0
num tx frames retries [01] 0
num tx frames retries [02] 0
num tx frames retries [03] 0
num tx frames failures [00] 0
num tx frames failures [01] 0
num tx frames failures [02] 0
num tx frames failures [03] 0
tx rate history [00] 0x00000000
tx rate history [01] 0x00000000
tx rate history [02] 0x00000000
tx rate history [03] 0x00000000
tx rate history [04] 0x00000000
tx rate history [05] 0x00000000
tx rate history [06] 0x00000000
tx rate history [07] 0x00000000
tx rate history [08] 0x00000000
tx rate history [09] 0x00000000
beacon rssi history [00] 0
beacon rssi history [01] 0
beacon rssi history [02] 0
beacon rssi history [03] 0
beacon rssi history [04] 0
beacon rssi history [05] 0
beacon rssi history [06] 0
beacon rssi history [07] 0
beacon rssi history [08] 0
beacon rssi history [09] 0
The firmware memory was allocated in ath12k_pci_probe(), but not
freed in ath12k_pci_remove() in case ATH12K_FLAG_QMI_FAIL bit is
set. So call ath12k_fw_unmap() to free the memory.
Sathishkumar Muruganandam [Tue, 7 Jan 2025 04:01:39 +0000 (09:31 +0530)]
wifi: ath12k: encode max Tx power in scan channel list command
Currently, when sending the scan channel list command to the firmware, the
maximum Tx power is not encoded in the reg2 member. This omission causes
the firmware to be unaware of the host's maximum Tx power, leading to
incorrect Tx power derivation at firmware level.
To resolve this issue, encode the maximum Tx power in the scan channel list
command before sending it to firmware.
Miaoqing Pan [Thu, 23 Jan 2025 08:49:48 +0000 (16:49 +0800)]
wifi: ath11k: fix memory leak in ath11k_xxx_remove()
The firmware memory was allocated in ath11k_pci_probe() or
ath11k_ahb_probe(), but not freed in ath11k_xxx_remove() in case
ATH11K_FLAG_QMI_FAIL bit is set. So call ath11k_fw_destroy() to
free the memory.
Found while fixing the same problem in ath12k:
https://lore.kernel.org/linux-wireless/20240314012746.2729101-1-quic_miaoqing@quicinc.com
Lingbo Kong [Wed, 15 Jan 2025 06:35:37 +0000 (14:35 +0800)]
wifi: ath12k: report station mode signal strength
Currently, the signal strength of "iw dev xxx station dump" always show an
invalid value.
This is because signal strength is only set in ath12k_mgmt_rx_event()
function, and not set for received data packet. So, change to get signal
from firmware and report to mac80211.
After that, "iw dev xxx station dump" show the correct signal strength.
Such as:
Station 00:03:7f:12:03:03 (on wlo1)
inactive time: 36 ms
rx bytes: 61571
rx packets: 336
tx bytes: 28204
tx packets: 205
tx retries: 49
tx failed: 0
beacon loss: 0
beacon rx: 83
rx drop misc: 66
signal: -24 dBm
beacon signal avg: -22 dBm
For WCN7850, the firmware supports db2dbm, so not need to add noise floor.
For QCN9274, the firmware not support db2dbm, so need to add noise floor.
This patch affects the station mode of WCN7850 and QCN9274.
Nicolas Escande [Wed, 22 Jan 2025 16:01:12 +0000 (17:01 +0100)]
wifi: ath12k: fix skb_ext_desc leak in ath12k_dp_tx() error path
When vlan support was added, we missed that when
ath12k_dp_prepare_htt_metadata() returns an error we also need to free
the skb holding the metadata before going on with the cleanup process.
Compile tested only.
Fixes: 26dd8ccdba4d ("wifi: ath12k: dynamic VLAN support") Signed-off-by: Nicolas Escande <nico.escande@gmail.com> Reviewed-by: Aditya Kumar Singh <aditya.kumar.singh@oss.qualcomm.com> Link: https://patch.msgid.link/20250122160112.3234558-1-nico.escande@gmail.com Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Aaradhana Sahu [Sun, 19 Jan 2025 08:36:56 +0000 (14:06 +0530)]
wifi: ath12k: add factory test mode support
Add support to process factory test mode commands(FTM) for calibration.
By default firmware start with MISSION mode and to process the FTM commands
firmware needs to be restarted in FTM mode using module parameter ftm_mode.
The pre-request is all the radios should be down before starting the test.
All ath12k test mode interface related commands specified in enum
ath_tm_cmd.
When start command ATH_TM_CMD_TESTMODE_START is received, ar state
is set to test Mode and FTM daemon sends test mode command to wifi
driver via cfg80211. Wifi driver sends these command to firmware as
wmi events. If it is segmented commands it will be broken down into
multiple segments and encoded with TLV header else it is sent to
firmware as it is.
Firmware response via UTF events, wifi driver creates skb and send
to cfg80211, cfg80211 sends firmware response to FTM daemon via
netlink message.
Command to boot in ftm mode
insmod ath12k ftm_mode=1
Aaradhana Sahu [Sun, 19 Jan 2025 08:36:54 +0000 (14:06 +0530)]
wifi: ath: create common testmode_i.h file for ath drivers
User space application requires that the testmode interface
is exactly same between ath drivers. Move testmode_i.h file
in ath directory to ensure that all ath driver uses same testmode
interface instead of duplicating testmode_i.h for each ath drivers.
P Praneesh [Sun, 19 Jan 2025 16:42:19 +0000 (22:12 +0530)]
wifi: ath11k: Use dma_alloc_noncoherent for rx_tid buffer allocation
Currently, the driver allocates cacheable DMA buffers for the rx_tid
structure using kzalloc() and dma_map_single(). These buffers are
long-lived and can persist for the lifetime of the peer, which is not
advisable. Instead of using kzalloc() and dma_map_single() for allocating
cacheable DMA buffers, utilize the dma_alloc_noncoherent() helper for the
allocation of long-lived cacheable DMA buffers, such as the peer's rx_tid.
Since dma_alloc_noncoherent() returns unaligned physical and virtual
addresses, align them internally before use within the driver. This
ensures proper allocation of non-coherent memory through the kernel
helper.
P Praneesh [Sun, 19 Jan 2025 16:42:18 +0000 (22:12 +0530)]
wifi: ath11k: Fix DMA buffer allocation to resolve SWIOTLB issues
Currently, the driver allocates cacheable DMA buffers for rings like
HAL_REO_DST and HAL_WBM2SW_RELEASE. The buffers for HAL_WBM2SW_RELEASE
are large (1024 KiB), exceeding the SWIOTLB slot size of 256 KiB. This
leads to "swiotlb buffer is full" error messages on systems without an
IOMMU that use SWIOTLB, causing driver initialization failures. The driver
calls dma_map_single() with these large buffers obtained from kzalloc(),
resulting in ring initialization errors on systems without an IOMMU that
use SWIOTLB.
To address these issues, replace the flawed buffer allocation mechanism
with the appropriate DMA API. Specifically, use dma_alloc_noncoherent()
for cacheable DMA buffers, ensuring proper freeing of buffers with
dma_free_noncoherent().
Error log:
[ 10.194343] ath11k_pci 0000:04:00.0: swiotlb buffer is full (sz:1048583 bytes), total 32768 (slots), used 2529 (slots)
[ 10.194406] ath11k_pci 0000:04:00.0: failed to set up tcl_comp ring (0) :-12
[ 10.194781] ath11k_pci 0000:04:00.0: failed to init DP: -12
Reported-by: Tim Harvey <tharvey@gateworks.com> Closes: https://lore.kernel.org/all/20241210041133.GA17116@lst.de/ Signed-off-by: P Praneesh <quic_ppranees@quicinc.com> Tested-by: Tim Harvey <tharvey@gateworks.com> Link: https://patch.msgid.link/20250119164219.647059-2-quic_ppranees@quicinc.com Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Karthikeyan Periyasamy [Tue, 24 Dec 2024 14:36:13 +0000 (20:06 +0530)]
wifi: ath12k: Refactor Rx status TLV parsing procedure argument
Currently, ath12k_dp_mon_rx_parse_status_tlv() takes the TLV tag, TLV data
and TLV userid as separate arguments from the caller. In the future, the
TLV length will be needed for parsing the EHT TLV tag. Therefore, instead
of increasing the number of arguments, pass the TLV header and retrieve the
necessary fields from the TLV header itself.
Karthikeyan Periyasamy [Tue, 24 Dec 2024 14:36:12 +0000 (20:06 +0530)]
wifi: ath12k: Refactor the monitor Tx/RX handler procedure arguments
Currently, the pdev handle is given along with the mac id to all the
monitor Tx/Rx handler procedure arguments. The mac id information is
derived from the pdev handle itself. Therefore, remove the unnecessary
mac id argument from the handler.
Karthikeyan Periyasamy [Tue, 24 Dec 2024 14:36:11 +0000 (20:06 +0530)]
wifi: ath12k: Refactor the monitor Rx parser handler argument
Currently, the monitor Rx parser handlers
ath12k_dp_mon_rx_parse_status_tlv() and ath12k_dp_mon_parse_rx_dest()
take the device handle from the caller. However, these handlers
functionality is technically pdev specific. Additionally, the device handle
can be retrieved from the pdev handle. Therefore, for better code
understanding, change the monitor Rx parser handlers argument from the
device handle to the pdev handle.
Dmitry Antipov [Wed, 15 Jan 2025 17:17:50 +0000 (20:17 +0300)]
wifi: ath9k: use unsigned long for activity check timestamp
Since 'rx_active_check_time' of 'struct ath_softc' is in jiffies,
prefer 'unsigned long' over 'u32' to avoid possible truncation in
'ath_hw_rx_inactive_check()'. Found with clang's -Wshorten-64-to-32,
compile tested only.
Dmitry Antipov [Wed, 15 Jan 2025 17:17:49 +0000 (20:17 +0300)]
wifi: ath9k: cleanup struct ath_tx_control and ath_tx_prepare()
After switching to mac80211 software queues, pointer to 'struct ath_node'
in 'struct ath_tx_control' is still assigned but not actually used. So drop
it and cleanup related things in 'ath_tx_prepare()'. Compile tested only.
Since brcmf_get_ip() can and actually will return NULL in this case the
call to brcmf_txfinalize() will result in a NULL pointer dereference inside
brcmf_txfinalize() when trying to update ifp->ndev->stats.tx_errors.
This will only happen if a flowring still has an skb.
Although the NULL pointer dereference has only been seen when trying to
update the tx statistic, all other uses of the ifp pointer have been
guarded as well with an early return if ifp is NULL.
Ping-Ke Shih [Thu, 16 Jan 2025 12:04:24 +0000 (20:04 +0800)]
wifi: rtw88: add RTW88_LEDS depends on LEDS_CLASS to Kconfig
When using allmodconfig, .config has CONFIG_LEDS_CLASS=m but
autoconf.h has CONFIG_LEDS_CLASS_MODULE (additional suffix _MODULE)
instead of CONFIG_LEDS_CLASS, which condition CONFIG_LEDS_CLASS in
rtw88/led.h can't work properly.
Add RTW88_LEDS to Kconfig, and use it as condition to fix this problem.
drivers/net/wireless/realtek/rtw88/led.c:19:6: error: redefinition of 'rtw_led_init'
19 | void rtw_led_init(struct rtw_dev *rtwdev)
| ^~~~~~~~~~~~
In file included from drivers/net/wireless/realtek/rtw88/led.c:7:
drivers/net/wireless/realtek/rtw88/led.h:15:20: note: previous definition of 'rtw_led_init' with type 'void(struct rtw_dev *)'
15 | static inline void rtw_led_init(struct rtw_dev *rtwdev)
| ^~~~~~~~~~~~
drivers/net/wireless/realtek/rtw88/led.c:64:6: error: redefinition of 'rtw_led_deinit'
64 | void rtw_led_deinit(struct rtw_dev *rtwdev)
| ^~~~~~~~~~~~~~
drivers/net/wireless/realtek/rtw88/led.h:19:20: note: previous definition of 'rtw_led_deinit' with type 'void(struct rtw_dev *)'
19 | static inline void rtw_led_deinit(struct rtw_dev *rtwdev)
| ^~~~~~~~~~~~~~
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au> Closes: https://lore.kernel.org/linux-wireless/e19a87ad9cd54bfa9907f3a043b25d30@realtek.com/T/#me407832de1040ce22e53517bcb18e322ad0e2260 Fixes: 4b6652bc6d8d ("wifi: rtw88: Add support for LED blinking") Cc: Bitterblue Smith <rtl8821cerfe2@gmail.com> Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://patch.msgid.link/20250116120424.13174-1-pkshih@realtek.com
Kalle Valo [Wed, 15 Jan 2025 16:51:56 +0000 (18:51 +0200)]
Merge tag 'ath-next-20250114' of git://git.kernel.org/pub/scm/linux/kernel/git/ath/ath
ath.git patches for v6.14
This development cycle again featured multiple patchsets to ath12k to
support the new 802.11be MLO feature, this time including the device
grouping infrastructure, and the advertisement of MLO support to the
wireless core. However the MLO feature is still considered to be
incomplete.
In addition, there was the usual set of bug fixes and cleanups, mostly
in ath12k, but also in ath9k.
Alexis Lothoré [Tue, 14 Jan 2025 10:45:34 +0000 (11:45 +0100)]
wifi: wilc1000: unregister wiphy only after netdev registration
wiphy_unregister()/wiphy_free() has been recently decoupled from
wilc_netdev_cleanup() to fix a faulty error path in sdio/spi probe
functions. However this change introduced a new failure when simply
loading then unloading the driver:
$ modprobe wilc1000-sdio; modprobe -r wilc1000-sdio
WARNING: CPU: 0 PID: 115 at net/wireless/core.c:1145 wiphy_unregister+0x904/0xc40 [cfg80211]
Modules linked in: wilc1000_sdio(-) wilc1000 cfg80211 bluetooth ecdh_generic ecc
CPU: 0 UID: 0 PID: 115 Comm: modprobe Not tainted 6.13.0-rc6+ #45
Hardware name: Atmel SAMA5
Call trace:
unwind_backtrace from show_stack+0x18/0x1c
show_stack from dump_stack_lvl+0x44/0x70
dump_stack_lvl from __warn+0x118/0x27c
__warn from warn_slowpath_fmt+0xcc/0x140
warn_slowpath_fmt from wiphy_unregister+0x904/0xc40 [cfg80211]
wiphy_unregister [cfg80211] from wilc_sdio_remove+0xb0/0x15c [wilc1000_sdio]
wilc_sdio_remove [wilc1000_sdio] from sdio_bus_remove+0x104/0x3f0
sdio_bus_remove from device_release_driver_internal+0x424/0x5dc
device_release_driver_internal from driver_detach+0x120/0x224
driver_detach from bus_remove_driver+0x17c/0x314
bus_remove_driver from sys_delete_module+0x310/0x46c
sys_delete_module from ret_fast_syscall+0x0/0x1c
Exception stack(0xd0acbfa8 to 0xd0acbff0)
bfa0: 0044b2100044b2100044b24c000008000000000000000000
bfc0: 0044b2100044b2100000000000000081000000000044b2100000000000000000
bfe0: 00448e24b6af99c40043ea0daea2e12c
irq event stamp: 0
hardirqs last enabled at (0): [<00000000>] 0x0
hardirqs last disabled at (0): [<c01588f0>] copy_process+0x1c4c/0x7bec
softirqs last enabled at (0): [<c0158944>] copy_process+0x1ca0/0x7bec
softirqs last disabled at (0): [<00000000>] 0x0
The warning is triggered by the fact that there is still a
wireless_device linked to the wiphy we are unregistering, due to
wiphy_unregister() now being called after net device unregister (performed
in wilc_netdev_cleanup()). Fix this warning by moving wiphy_unregister()
after wilc_netdev_cleanup() is nominal paths (ie: driver removal).
wilc_netdev_cleanup() ordering is left untouched in error paths in probe
function because net device is not registered in those paths (so the
warning can not trigger), yet the wiphy can still be registered, and we
still some cleanup steps from wilc_netdev_cleanup().
Dmitry Antipov [Mon, 13 Jan 2025 15:54:17 +0000 (18:54 +0300)]
wifi: cfg80211: adjust allocation of colocated AP data
In 'cfg80211_scan_6ghz()', an instances of 'struct cfg80211_colocated_ap'
are allocated as if they would have 'ssid' as trailing VLA member. Since
this is not so, extra IEEE80211_MAX_SSID_LEN bytes are not needed.
Briefly tested with KUnit.
Currently, an interface is created in the driver during channel assignment.
If mac80211 attempts to set a key for an interface before this assignment,
the driver caches the key. Once the interface is created, the driver
installs the cached key to the hardware. This sequence is exemplified in
mesh mode operation where the group key is set before channel assignment.
However, in ath12k_mac_update_key_cache(), after caching the key, due to
incorrect logic, it is deleted from the cache during the subsequent loop
iteration. As a result, after the interface is created, the driver does not
find any cached key, and the key is not installed to the hardware which is
wrong. This leads to issue in mesh, where broadcast traffic is not
encrypted over the air.
Fix this issue by adjusting the logic of ath12k_mac_update_key_cache()
properly.
Fixes: 25e18b9d6b4b ("wifi: ath12k: modify ath12k_mac_op_set_key() for MLO") Signed-off-by: Aditya Kumar Singh <quic_adisi@quicinc.com> Acked-by: Kalle Valo <kvalo@kernel.org> Tested-by: Nicolas Escande <nico.escande@gmail.com> Link: https://patch.msgid.link/20250112-fix_key_cache_handling-v2-1-70e142c6153e@quicinc.com Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Karthikeyan Periyasamy [Sun, 12 Jan 2025 07:16:30 +0000 (12:46 +0530)]
wifi: ath12k: Fix uninitialized variable access in ath12k_mac_allocate() function
Currently, the uninitialized variable 'ab' is accessed in the
ath12k_mac_allocate() function. Initialize 'ab' with the first radio device
present in the hardware abstraction handle (ah). Additionally, move the
default setting procedure from the pdev mapping iteration to the total
radio calculating iteration for better code readability. Perform the
maximum radio validation check for total_radio to ensure that both num_hw
and radio_per_hw are validated indirectly, as these variables are derived
from total_radio. This also fixes the below Smatch static checker warning.
Smatch warning:
ath12k_mac_allocate() error: uninitialized symbol 'ab'
Fixes: a343d97f27f5 ("wifi: ath12k: move struct ath12k_hw from per device to group") Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com> Acked-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com> Acked-by: Kalle Valo <kvalo@kernel.org> Link: https://patch.msgid.link/20250112071630.4059410-5-quic_periyasa@quicinc.com Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Karthikeyan Periyasamy [Sun, 12 Jan 2025 07:16:29 +0000 (12:46 +0530)]
wifi: ath12k: Remove ath12k_get_num_hw() helper function
Currently, the ath12k_get_num_hw() helper function takes the device handle
as an argument. Here, the number of hardware is retrieved from the group
handle. Demanding the device handle from the caller is unnecessary since
in some cases the group handle is already available. Additionally, there
is no longer a need for multiple indirections to get the number of
hardware. Therefore, remove this helper function and directly use
ag->num_hw. This change also fixes the below Smatch static checker
warning.
Smatch warning:
ath12k_mac_destroy() error: we previously assumed 'ab' could be null
Reported-by: Dan Carpenter <dan.carpenter@linaro.org> Closes: https://lore.kernel.org/ath12k/3e705de0-67d1-4437-97ff-4828d83ae2af@stanley.mountain/ Closes: https://scan7.scan.coverity.com/#/project-view/52682/11354?selectedIssue=1602340 Fixes: a343d97f27f5 ("wifi: ath12k: move struct ath12k_hw from per device to group") Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com> Acked-by: Kalle Valo <kvalo@kernel.org> Acked-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com> Link: https://patch.msgid.link/20250112071630.4059410-4-quic_periyasa@quicinc.com Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Karthikeyan Periyasamy [Sun, 12 Jan 2025 07:16:28 +0000 (12:46 +0530)]
wifi: ath12k: Refactor the ath12k_hw get helper function argument
Currently, ath12k_hw is placed inside the ath12k_hw_group. However, the
ath12k_hw get helper function takes the device handle and the index as
parameters. Here, the index parameter is specific to the group handle.
Therefore, change this helper function argument from the device handle
to the group handle.
Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com> Acked-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com> Acked-by: Kalle Valo <kvalo@kernel.org> Link: https://patch.msgid.link/20250112071630.4059410-3-quic_periyasa@quicinc.com Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Karthikeyan Periyasamy [Sun, 12 Jan 2025 07:16:27 +0000 (12:46 +0530)]
wifi: ath12k: Refactor ath12k_hw set helper function argument
Currently, ath12k_hw is placed inside the ath12k_hw_group. However, the
ath12k_hw set helper function takes the device handle and the index as
parameters. Here, the index parameter is specific to the group handle.
Therefore, change this helper function argument from the device handle to
the group handle.
Signed-off-by: Karthikeyan Periyasamy <quic_periyasa@quicinc.com> Acked-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com> Acked-by: Kalle Valo <kvalo@kernel.org> Link: https://patch.msgid.link/20250112071630.4059410-2-quic_periyasa@quicinc.com Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Howard Hsu [Tue, 14 Jan 2025 10:10:26 +0000 (18:10 +0800)]
wifi: mt76: mt7996: add implicit beamforming support for mt7992
The configuration of mt7992 iBF has some differences to mt7996, rework
related code to make it work.
Also, most of our chipsets support iBF only on HT and VHT mode, except
mt7992 variant that has 5T on 5GHz band also supports iBF on HE and EHT
mode.
Michael-CY Lee [Tue, 14 Jan 2025 10:10:25 +0000 (18:10 +0800)]
wifi: mt76: mt7996: fix beacon command during disabling
When stopping AP, mac80211 frees beacon template before it calls
driver's stop_ap operation. In other words, on the path of stopping
AP, ieee80211_beacon_get_template() must returns NULL in
mt7996_mcu_add_beacon(). In such case mt7996 immediately returns
-EINVAL without telling FW to disable the beacon.
This commit refactors mt7996_mcu_add_beacon() so that FW can be
correctly informed when disabling AP interface.
Peter Chiu [Tue, 14 Jan 2025 10:10:24 +0000 (18:10 +0800)]
wifi: mt76: mt7996: fix ldpc setting
The non-AP interfaces would not use conf->vht_ldpc so they never set
STA_CAP_VHT_LDPC even if peer-station support LDPC.
Check conf->vht_ldpc only for AP interface.
Without this patch, station only uses BCC to transmit packet in VHT mode.
Benjamin Lin [Tue, 14 Jan 2025 10:10:23 +0000 (18:10 +0800)]
wifi: mt76: mt7996: fix definition of tx descriptor
For mt7992 chipsets, the definition of TXD.DW6.BIT10~15 has different
interpretations on different frame types. Driver only needs to fill
MSDU_CNT for non-management frames.
Fixes: 408566db8cad ("wifi: mt76: connac: add new definition of tx descriptor") Co-developed-by: Shayne Chen <shayne.chen@mediatek.com> Signed-off-by: Shayne Chen <shayne.chen@mediatek.com> Signed-off-by: Benjamin Lin <benjamin-jw.lin@mediatek.com> Link: https://patch.msgid.link/20250114101026.3587702-6-shayne.chen@mediatek.com Signed-off-by: Felix Fietkau <nbd@nbd.name>
Howard Hsu [Tue, 14 Jan 2025 10:10:19 +0000 (18:10 +0800)]
wifi: mt76: mt7996: fix the capability of reception of EHT MU PPDU
This commit includes two changes. First, enable "EHT MU PPDU With 4x
EHT-LTF And 0.8us GI" in EHT Phy capabilities element since hardware
can support. Second, fix the value of "Maximum number of supported
EHT LTFs" in the same element, where the previous setting of 3 in
Bit 3-4 was incorrect.
Peter Chiu [Tue, 14 Jan 2025 10:10:18 +0000 (18:10 +0800)]
wifi: mt76: mt7996: add max mpdu len capability
Set max mpdu len to 11454 according to hardware capability.
Without this patch, the max ampdu length would be 3895 and count not get
expected performance.
Leon Yen [Fri, 12 Apr 2024 08:53:57 +0000 (16:53 +0800)]
wifi: mt76: mt7921: avoid undesired changes of the preset regulatory domain
Some countries have strict RF restrictions where changing the regulatory
domain dynamically based on the connected AP is not acceptable.
This patch disables Beacon country IE hinting when a valid country code
is set from usersland (e.g., by system using iw or CRDA).
Quan Zhou [Fri, 3 Jan 2025 13:53:50 +0000 (21:53 +0800)]
wifi: mt76: mt7925e: fix too long of wifi resume time
The mt7925e takes a long time to resume in platform resume
scenarios. This is due to the mt7925 reconfiguring the tx
power for each wifi channel during resume. By streamlining
the process and removing redundant power configurations,
the resume time for the mt7925 can be shortened.
Quan Zhou [Fri, 3 Jan 2025 13:44:54 +0000 (21:44 +0800)]
wifi: mt76: mt7925: add handler to hif suspend/resume event
When the system suspend or resume, the WiFi driver sends
an hif_ctrl command to the firmware and waits for an event.
Due to changes in the event format reported by the chip, the
current mt7925's driver does not account for these changes,
resulting in command timeout. Add flow to handle hif_ctrl
event to avoid command timeout. We also exented API
mt76_connac_mcu_set_hif_suspend for connac3 this time.
Quan Zhou [Fri, 3 Jan 2025 13:39:43 +0000 (21:39 +0800)]
wifi: mt76: mt7925: fix CLC command timeout when suspend/resume
When enter suspend/resume while in a connected state, the upper layer
will trigger disconnection before entering suspend, and at the same time,
it will trigger regd_notifier() and update CLC, causing the CLC event to
not be received due to suspend, resulting in a command timeout.
Therefore, the update of CLC is postponed until resume, to ensure data
consistency and avoid the occurrence of command timeout.
Quan Zhou [Fri, 3 Jan 2025 13:35:59 +0000 (21:35 +0800)]
wifi: mt76: mt7925: fix the unfinished command of regd_notifier before suspend
Before entering suspend, we need to ensure that all MCU command are
completed. In some cases, such as with regd_notifier, there is a
chance that CLC commands, will be executed before suspend.
Peter Chiu [Thu, 9 Jan 2025 11:04:36 +0000 (19:04 +0800)]
wifi: mt76: mt7996: fix register mapping
Bypass the entry when ofs is equal to dev->reg.map[i].size.
Without this patch, it would get incorrect register mapping when the CR
address is located at the boundary of an entry.
Peter Chiu [Thu, 9 Jan 2025 11:04:35 +0000 (19:04 +0800)]
wifi: mt76: mt7915: fix register mapping
Bypass the entry when ofs is equal to dev->reg.map[i].size.
Without this patch, it would get incorrect register mapping when the CR
address is located at the boundary of an entry.
Felix Fietkau [Thu, 2 Jan 2025 16:35:07 +0000 (17:35 +0100)]
wifi: mt76: mt7996: switch to single multi-radio wiphy
Use generic mt76 chanctx functions and look up phy from vif link.
The driver now uses only a single phy to handle multiple interfaces on
different channels. This is preparation for full MLO support.
projects / users / jedix / linux-maple.git / log