Now some chips which support 'support_dual_stations' will enable DBS,
but will disable DFS. Restructure the ath11k_mac_setup_iface_combinations
function to support DBS and DFS compatibility.
About 'support_dual_station' feature can refer:
https://msgid.link/20230714023801.2621802-2-quic_cjhuang@quicinc.com
Add a ieee80211_iface_combination to support DBS and DFS compatibility,
one combination can support DFS(same with non dual sta), another
combination can support DBS. When running different scenarios that will
use different ieee80211_iface_combination due to mac80211 will go through
all of possible interface combinations.
In addition, maximum number of interfaces of these types should be total
allowed in this group.
The chips affected are:
QCA6390 hw2.0
WCN6855 hw2.0
WCN6855 hw2.1
Other chips are not affected.
Tested-on: IPQ8074 hw2.0 AHB WLAN.HK.2.1.0.1-00410-QCAHKSWPL_SILICONZ-2
Tested-on: IPQ6018 hw1.0 AHB WLAN.HK.2.1.0.1-01161-QCAHKSWPL_SILICONZ-1
Tested-on: WCN6855 hw2.1 PCI WLAN.HSP.1.1-04402-QCAHSPSWPL_V1_V2_SILICONZ_IOE-1.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-05266-QCAHSTSWPLZ_V2_TO_X86-1
Signed-off-by: Yu Zhang (Yuriy) <quic_yuzha@quicinc.com>
Reviewed-by: Baochen Qiang <quic_bqiang@quicinc.com>
Reviewed-by: Vasanthakumar Thiagarajan <vasanthakumar.thiagarajan@oss.qualcomm.com>
Acked-by: Jeff Johnson <quic_jjohnson@quicinc.com>
Tested-by: Mihai Moldovan <ionic@ionic.de>
Link: https://patch.msgid.link/20241127022742.4016870-1-quic_yuzha@quicinc.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Add the missing memory barrier to make sure that the REO dest ring
descriptor is read after the head pointer to avoid using stale data on
weakly ordered architectures like aarch64.
This may fix the ring-buffer corruption worked around by commit
f9fff67d2d ("wifi: ath11k: Fix SKB corruption in REO destination
ring") by silently discarding data, and may possibly also address user
reported errors like:
ath11k_pci 0006:01:00.0: msdu_done bit in attention is not set
Tested-on: WCN6855 hw2.1 WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.41
Fixes: d5c65159f2 ("ath11k: driver for Qualcomm IEEE 802.11ax devices")
Cc: stable@vger.kernel.org # 5.6
Link: https://bugzilla.kernel.org/show_bug.cgi?id=218005
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Tested-by: Clayton Craft <clayton@craftyguy.net>
Link: https://patch.msgid.link/20250321145302.4775-1-johan+linaro@kernel.org
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Users of the Lenovo ThinkPad X13s have reported that Wi-Fi sometimes
breaks and the log fills up with errors like:
ath11k_pci 0006:01:00.0: HTC Rx: insufficient length, got 1484, expected 1492
ath11k_pci 0006:01:00.0: HTC Rx: insufficient length, got 1460, expected 1484
which based on a quick look at the driver seemed to indicate some kind
of ring-buffer corruption.
Miaoqing Pan tracked it down to the host seeing the updated destination
ring head pointer before the updated descriptor, and the error handling
for that in turn leaves the ring buffer in an inconsistent state.
Add the missing memory barrier to make sure that the descriptor is read
after the head pointer to address the root cause of the corruption while
fixing up the error handling in case there are ever any (ordering) bugs
on the device side.
Note that the READ_ONCE() are only needed to avoid compiler mischief in
case the ring-buffer helpers are ever inlined.
Tested-on: WCN6855 hw2.1 WLAN.HSP.1.1-03125-QCAHSPSWPL_V1_V2_SILICONZ_LITE-3.6510.41
Fixes: d5c65159f2 ("ath11k: driver for Qualcomm IEEE 802.11ax devices")
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218623
Link: https://lore.kernel.org/20250310010217.3845141-3-quic_miaoqing@quicinc.com
Cc: Miaoqing Pan <quic_miaoqing@quicinc.com>
Cc: stable@vger.kernel.org # 5.6
Signed-off-by: Johan Hovold <johan+linaro@kernel.org>
Reviewed-by: Miaoqing Pan <quic_miaoqing@quicinc.com>
Tested-by: Steev Klimaszewski <steev@kali.org>
Tested-by: Jens Glathe <jens.glathe@oldschoolsolutions.biz>
Tested-by: Clayton Craft <clayton@craftyguy.net>
Link: https://patch.msgid.link/20250321094916.19098-1-johan+linaro@kernel.org
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Firmware requests 2 segments at first. The first segment is of 6799360
whose allocation fails due to dma remapping not available. The success
is returned to firmware. Then firmware asks for 22 smaller segments
instead of 2 big ones. Those get allocated successfully. At suspend/
hibernation time, these segments aren't freed as they will be reused
by firmware after resuming.
After resuming, the firmware asks for the 2 segments again with the
first segment of 6799360 size. Since chunk->vaddr is not NULL, the
type and size are compared with the previous type and size to know if
it can be reused or not. Unfortunately, it is detected that it cannot
be reused and this first smaller segment is freed. Then we continue to
allocate 6799360 size memory which fails and ath11k_qmi_free_target_mem_chunk()
is called which frees the second smaller segment as well. Later success
is returned to firmware which asks for 22 smaller segments again. But
as we had freed 2 segments already, we'll allocate the first 2 new
smaller segments again and reuse the remaining 20. Hence 20 small
segments are being reused instead of 22.
Add skip logic when vaddr is set, but size/type don't match. Use the
same skip and success logic as used when dma_alloc_coherent() fails.
By skipping, the possibility of resume failure due to kernel failing to
allocate memory for QMI can be avoided.
kernel: ath11k_pci 0000:03:00.0: failed to allocate dma memory for qmi (524288 B type 1)
ath11k_pci 0000:03:00.0: failed to allocate qmi target memory: -22
Tested-on: WCN6855 WLAN.HSP.1.1-03926.13-QCAHSPSWPL_V2_SILICONZ_CE-2.52297.6
Signed-off-by: Muhammad Usama Anjum <usama.anjum@collabora.com>
Reviewed-by: Baochen Qiang <quic_bqiang@quicinc.com>
Link: https://patch.msgid.link/20250428080242.466901-1-usama.anjum@collabora.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Simplifies probe slightly and adds extra error codes.
Switching from devm_ioremap to the platform variant ends up calling
devm_request_mem_region, which reserves the memory region for the
various wmacs. Per board, there is only one wmac and after some fairly
thorough analysis, there are no overlapping memory regions between wmacs
and other devices on the ahb.
Tested on a TP-Link Archer C7v2.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Acked-by: Toke Høiland-Jørgensen <toke@toke.dk>
Link: https://patch.msgid.link/20250421040044.44887-1-rosenp@gmail.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
According to the code documentation in ath12k_mac_op_hw_scan(), "if no
links of an ML VIF are already active on the radio corresponding to the
given scan frequency, the scan link (link ID 15) should be used". This rule
should apply to non-ML interfaces as well to maintain uniformity across the
driver. However, currently, link 0 is selected as the scan link during
non-ML operations.
Update the code to use scan link ID 15 in all cases.
Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.4.1-00199-QCAHKSWPL_SILICONZ-1
Reviewed-by: Vasanthakumar Thiagarajan <vasanthakumar.thiagarajan@oss.qualcomm.com>
Signed-off-by: Aditya Kumar Singh <aditya.kumar.singh@oss.qualcomm.com>
Link: https://patch.msgid.link/20250417-fix_scan_vdev_handling-v3-2-9ec42513d26b@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
For scanning purposes, the driver can use link ID 15 (scan link). A future
change will make non-ML interfaces to select the scan link. In this
scenario, arvif->link_id will be used to retrieve the link configuration in
ath12k_mac_vdev_create(). However, link ID 15 is not recognized as a valid
link ID in the upper kernel, which will result in a failure to fetch
link_conf and subsequently cause the scan to fail.
To avoid this issue, ensure link_conf is fetched only when the link ID is
within the valid range. Since link_conf cannot be retrieved using the scan
link, use vif->addr as the Ethernet address for creating the scan vdev.
This address will serve as the source address (address 2) in the probe
request frames during scanning. Additionally, use the automatic Tx power
value for the vdev. As this is a scan vdev, these values do not affect the
scan functionality.
Note that vif->addr will only be taken when a valid link_conf can not be
fetched. Otherwise, link_conf's address will be taken as address 2.
Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.4.1-00199-QCAHKSWPL_SILICONZ-1
Reviewed-by: Vasanthakumar Thiagarajan <vasanthakumar.thiagarajan@oss.qualcomm.com>
Signed-off-by: Aditya Kumar Singh <aditya.kumar.singh@oss.qualcomm.com>
Link: https://patch.msgid.link/20250417-fix_scan_vdev_handling-v3-1-9ec42513d26b@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Read the country code from SMBIOS and send it to the firmware. The
firmware will then indicate the regulatory domain information for
the country code, which ath12k will use.
dmesg:
[ 1242.637253] ath12k_pci 0000:02:00.0: worldwide regdomain setting from SMBIOS
[ 1242.637259] ath12k_pci 0000:02:00.0: bdf variant name not found.
[ 1242.637261] ath12k_pci 0000:02:00.0: SMBIOS bdf variant name not set.
[ 1242.927543] ath12k_pci 0000:02:00.0: set current country pdev id 0 alpha2 00
Tested-on: WCN7850 hw2.0 PCI WLAN.HMT.1.0-02582-QCAHMTSWPL_V1.0_V2.0_SILICONZ-1
Signed-off-by: Wen Gong <quic_wgong@quicinc.com>
Signed-off-by: Kang Yang <kang.yang@oss.qualcomm.com>
Reviewed-by: Baochen Qiang <quic_bqiang@quicinc.com>
Link: https://patch.msgid.link/20250417024227.1712-1-kang.yang@oss.qualcomm.com
Signed-off-by: Jeff Johnson <jeff.johnson@oss.qualcomm.com>
Ping-Ke Shih says:
==================
rtw-next patches for v6.16
Some fixes and refinements across drivers, and regular development of
MLO and STA + P2P concurrency. Major changes are listed below.
rtw88:
* improve throughput for RTL8814AU
rtw89:
* support MLO
* improve user experience for STA + P2P concurrency
* dynamic antenna gain (DAG) with different power by antenna
* load SAR tables from ACPI
==================
Link: https://patch.msgid.link/17e74675-70cc-43d7-a797-afb937030d34@RTEXMBS04.realtek.com.tw/
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
MCC will track the changes of beacon offset, and trigger a recalculation
when the difference is larger than the tolerance. It means that a better
pattern is expected after recalculating. However, in the cases which get
a worse beacon offset, there is no chance to improve the pattern even if
recalculating. So, bypass them.
Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20250511035217.10410-7-pkshih@realtek.com
Originally, MCC just took periodic NoA into account. When the connected GO
announces non-periodic NoA and GC side is during MCC, sometimes GC cannot
receive beacons well if the MCC scheduling conflicts with the non-periodic
NoA planning. After the loss exceeds the tolerable amount, beacon filter
will report connection loss. However, in this case, the loss is acceptable.
So now, MCC will calculate the range of non-periodic NoA. And then, don't
care beacon loss during the range.
Besides, rtw89_mcc_fill_role_limit() only makes sense for GC. Remove the
redundant check of GO.
Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20250511035217.10410-6-pkshih@realtek.com
In the cases that two MCC roles' TBTTs are too close or too far, original
MCC pattern calculation logic will lead to a result that both roles might
not cover its TBTT with sufficient time. Introduce a new calculation logic
called anchor pattern for these corner cases. It allows to choose one role
as anchor to put its TBTT in the middle of its duration directly. For now,
a P2P role has a higher priority to be chosen as an anchor. Then, if able,
another role might need to depend on courtesy mechanism to take time from
anchor.
Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20250511035217.10410-5-pkshih@realtek.com
In one enablement of courtesy mechanism, there is one provider and
one receiver. And, receiver can use the provider's time in a given
period. But, to make P2P NoA protocol work as expected as possible,
GO should be present at the time it doesn't announce absent, and GC
should not use the time when GO announces absent. So, don't enable
courtesy mechanism if provider is GO or receiver is GC.
Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20250511035217.10410-4-pkshih@realtek.com
When MCC is about to stop, there may be some chanctx changes which are
queued for work but have not yet been run. To avoid these changes from
being processed in a wrong state (e.g. next new MCC instance), cancel
the queued work and drop queued changes.
Signed-off-by: Zong-Zhe Yang <kevin_yang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20250511035217.10410-3-pkshih@realtek.com
RTL8811AU stops responding during the firmware download on some systems:
[ 809.256440] rtw_8821au 5-2.1:1.0: Firmware version 42.4.0, H2C version 0
[ 812.759142] rtw_8821au 5-2.1:1.0 wlp48s0f4u2u1: renamed from wlan0
[ 837.315388] rtw_8821au 1-4:1.0: write register 0x1ef4 failed with -110
[ 867.524259] rtw_8821au 1-4:1.0: write register 0x1ef8 failed with -110
[ 868.930976] rtw_8821au 5-2.1:1.0 wlp48s0f4u2u1: entered promiscuous mode
[ 897.730952] rtw_8821au 1-4:1.0: write register 0x1efc failed with -110
Maybe it takes too long when writing the firmware 4 bytes at a time.
Write 196 bytes at a time for RTL8821AU, RTL8811AU, and RTL8812AU,
and 254 bytes at a time for RTL8723DU. These are the sizes used in
their official drivers. Tested with all these chips.
Cc: stable@vger.kernel.org
Link: https://github.com/lwfinger/rtw88/issues/344
Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/43f1daad-3ec0-4a3b-a50c-9cd9eb2c2f52@gmail.com
RTL8811AU stops responding during the firmware download on some systems:
[ 809.256440] rtw_8821au 5-2.1:1.0: Firmware version 42.4.0, H2C version 0
[ 812.759142] rtw_8821au 5-2.1:1.0 wlp48s0f4u2u1: renamed from wlan0
[ 837.315388] rtw_8821au 1-4:1.0: write register 0x1ef4 failed with -110
[ 867.524259] rtw_8821au 1-4:1.0: write register 0x1ef8 failed with -110
[ 868.930976] rtw_8821au 5-2.1:1.0 wlp48s0f4u2u1: entered promiscuous mode
[ 897.730952] rtw_8821au 1-4:1.0: write register 0x1efc failed with -110
Each write takes 30 seconds to fail because that's the timeout currently
used for control messages in rtw_usb_write().
In this scenario the firmware download takes at least 2000 seconds.
Because this is done from the USB probe function, the long delay makes
other things in the system hang.
Reduce the timeout to 500 ms. This is the value used by the official USB
wifi drivers from Realtek.
Of course this only makes things hang for ~30 seconds instead of ~30
minutes. It doesn't fix the firmware download.
Tested with RTL8822CU, RTL8812BU, RTL8811CU, RTL8814AU, RTL8811AU,
RTL8812AU, RTL8821AU, RTL8723DU.
Cc: stable@vger.kernel.org
Fixes: a82dfd33d1 ("wifi: rtw88: Add common USB chip support")
Link: https://github.com/lwfinger/rtw88/issues/344
Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/1e35dd26-3f10-40b1-b2b4-f72184a26611@gmail.com
Context info was introduced in 22000, and was significantly changed in
ax210. The new version of context info was called 'gen3',
probably because in 22000, the gen2 transport was added.
But this name is just wrong:
- if 'gen' enumerates transports, there was not a gen3 transport, just a
few modifications to gen1/2 transports needed for ax210.
- if 'gen' enumerates devices, then we can just use the device names.
Also, context info will soon become a lib, agnostic of the transport
generations.
Simply replace 'gen3' with 'v2'.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://patch.msgid.link/20250511195137.a580bd8d4f74.Ie413a02233f1a5ad538e13071c09760b9d97be3b@changeid
We have iwl_tx_cmd for devices older than 22000, iwl_tx_cmd_gen2 for
22000 devices, and iwl_tx_cmd_gen3 ax210 and up.
But the convention for all other APIs is to have the latest version
without any prefix and the older ones - with a _vX prefix,
where X is the highest version that this struct support.
The term 'gen' was introduced as the name of the (back then) new
transport, and should not be used as a device name (for that we have the
actual names: 22000, ax210, etc.)
Now as a new transport, called 'gen3', is going to be written and it can
be confused with this API.
Move iwl_tx_cmd to use the regular versioning convention.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://patch.msgid.link/20250511195137.806e40c8f767.Ibc0e95e43a6fa6d47f72823bf804314d5db84618@changeid
We don't need the CORES() match nor jacket (which really doesn't
even make sense to match to the RF anyway), and since the subdevice
masks we care about are contiguous, we can encode them as highest
and lowest bit set (automatically.) By encoding whether to match or
not as separate flags and taking advantage of the limited range of
the RF type, step and ID we can reduce the amount of memory needed
for the table, while also making the logic (apart perhaps from the
subdevice mask) easier to understand.
This reduces the size of the module by about 1.5KiB on x86-64.
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://patch.msgid.link/20250511195137.38a805a7c96f.Ieece00476cea6054b0827cd075eb8ba5943373df@changeid
