When a running wake_tx_queue() call is aborted due to a hw queue stop
the corresponding iTXQ is not always correctly marked for resumption:
wake_tx_push_queue() can stops the queue run without setting
@IEEE80211_TXQ_STOP_NETIF_TX.
Without the @IEEE80211_TXQ_STOP_NETIF_TX flag __ieee80211_wake_txqs()
will not schedule a new queue run and remaining frames in the queue get
stuck till another frame is queued to it.
Fix the issue for all drivers - also the ones with custom wake_tx_queue
callbacks - by moving the logic into ieee80211_tx_dequeue() and drop the
redundant @txqs_stopped.
@IEEE80211_TXQ_STOP_NETIF_TX is also renamed to @IEEE80211_TXQ_DIRTY to
better describe the flag.
Fixes: c850e31f79f0 ("wifi: mac80211: add internal handler for
wake_tx_queue")
Signed-off-by: Alexander Wetzel <alexander@wetzel-home.de>
Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
This updates mac80211 to version 6.1.110-1. This code is based on Linux
6.1.110 and contains all fixes included in the upstream wireless
subsystem from that kernel version. This includes many bugfixes and also
some security fixes.
The removed patches are already integrated in upstream Linux 6.1.110.
The following patches were integrated in upstream Linux:
subsys/311-v6.2-wifi-mac80211-fix-and-simplify-unencrypted-drop-chec.patch
subsys/312-v6.3-wifi-cfg80211-move-A-MSDU-check-in-ieee80211_data_to.patch
subsys/313-v6.3-wifi-cfg80211-factor-out-bridge-tunnel-RFC1042-heade.patch
subsys/314-v6.3-wifi-mac80211-remove-mesh-forwarding-congestion-chec.patch
subsys/315-v6.3-wifi-mac80211-fix-receiving-A-MSDU-frames-on-mesh-in.patch
subsys/316-v6.3-wifi-mac80211-add-a-workaround-for-receiving-non-sta.patch
subsys/321-mac80211-fix-mesh-forwarding.patch
subsys/322-wifi-mac80211-fix-mesh-path-discovery-based-on-unica.patch
subsys/329-wifi-mac80211-fix-receiving-mesh-packets-in-forwardi.patch
subsys/339-wifi-cfg80211-fix-receving-mesh-packets-without-RFC1.patch
subsys/350-v6.3-wifi-mac80211-Allow-NSS-change-only-up-to-capability.patch
subsys/351-v6.9-wifi-mac80211-track-capability-opmode-NSS-separately.patch
Link: https://github.com/openwrt/openwrt/pull/16368
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This fixes WARN_ONs when using AP_VLANs after station removal. The flush
call passed AP_VLAN vif to driver, but because these vifs are virtual and
not registered with drivers, we need to translate to the correct AP vif
first.
Fixes: openwrt#12420
Signed-off-by: Oldřich Jedlička <oldium.pro@gmail.com>
[Rename to 360-wifi-mac80211-do-not-pass-ap_vlan-vif-pointer-to-dri.patch]
(cherry picked from commit 3e738781a9)
Link: https://github.com/openwrt/openwrt/pull/15898
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This updates mac80211 to version 6.1.97-1. This code is based on Linux
6.1.97 and contains all fixes included in the upstream wireless
subsystem from that kernel version. This includes many bugfixes and also
some security fixes.
The removed patches are already integrated in upstream Linux 6.1.97 or
in backports.
The following patches were integrated in upstream Linux:
ath11k/0013-wifi-ath11k-synchronize-ath11k_mac_he_gi_to_nl80211_.patch
ath11k/0035-wifi-ath11k-Use-platform_get_irq-to-get-the-interrup.patch
ath11k/0036-wifi-ath11k-fix-SAC-bug-on-peer-addition-with-sta-ba.patch
ath11k/0047-wifi-ath11k-fix-deinitialization-of-firmware-resourc.patch
ath11k/0053-wifi-ath11k-fix-writing-to-unintended-memory-region.patch
ath11k/0060-wifi-ath11k-Ignore-frags-from-uninitialized-peer-in-.patch
ath11k/0065-wifi-ath11k-fix-tx-status-reporting-in-encap-offload.patch
ath11k/0067-wifi-ath11k-Fix-SKB-corruption-in-REO-destination-ri.patch
ath11k/0069-wifi-ath11k-fix-registration-of-6Ghz-only-phy-withou.patch
ath11k/0080-wifi-ath11k-add-support-default-regdb-while-searchin.patch
ath11k/0085-wifi-ath11k-fix-memory-leak-in-WMI-firmware-stats.patch
ath11k/0086-wifi-ath11k-Add-missing-check-for-ioremap.patch
ath11k/0096-wifi-ath11k-fix-boot-failure-with-one-MSI-vector.patch
subsys/337-wifi-mac80211-fix-race-condition-on-enabling-fast-xm.patch
The following patches were integrated in upstream backports:
ath11k/901-wifi-ath11k-pci-fix-compilation-in-5.16-and-older.patch
build/080-resv_start_op.patch
build/110-backport_napi_build_skb.patch
The following files are missing in backports, we do not have to remove
them any more. Some were already missing before some were removed in
this update:
include/linux/cordic.h
include/linux/crc8.h
include/linux/eeprom_93cx6.h
include/linux/wl12xx.h
include/net/ieee80211.h
backport-include/linux/bcm47xx_nvram.h
include/linux/ath9k_platform.h
include/net/bluetooth/
backports ships a dummy Mediatek wed header for older kernel versions.
We backported the feature in our kernel, remove the dummy header:
backport-include/linux/soc/mediatek/mtk_wed.h
Remove header files for subsystems used form the mainline kernel:
include/trace/events/qrtr.h
include/net/rsi_91x.h
backport-include/linux/platform_data/brcmnand.h
Link: https://github.com/openwrt/openwrt/pull/15827
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
It's needed by some drivers, e.g. mt7925, see:
mt7925/main.c:264:9: error: implicit declaration of function '_ieee80211_set_sband_iftype_data' [-Werror=implicit-function-declaration]
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
It's needed by some drivers, e.g. mt7925, see:
mt7925/mcu.c:2181:23: error: implicit declaration of function 'ieee80211_vif_is_mld' [-Werror=implicit-function-declaration]
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Synchronize the ath11k backports with upstream linux.
Most of them are changes in kernel 6.5, the rest are
fixes for the ath11k_pci. The most important one is
"Revert 'wifi: ath11k: Enable threaded NAPI'", which
fixes the problem that QCN9074 cannot be used after
restarting on the x86 platform.
[ 23.462718] ath11k_pci 0000:02:00.0: failed to vdev 0 create peer for AP: -110
[ 28.503020] ath11k_pci 0000:02:00.0: Timeout in receiving vdev delete response
Changes to ipq8074 coldboot part pick from commit
b33bfcf ("mac80211: ath11k: sync with ath-next").
Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
With mac80211_hwsim I have seen such entries in OpenWrt 22.03:
HE Iftypes: managed, AP
The mac80211.sh script did not detect the entry and failed. Allow
arbitrary other entries before to fix this problem.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 5df7a78e82)
Frequent crashes have been observed on MT7916 based platforms. While the
root of these crashes are currently unknown, they happen when decoding
rate information of connected STAs in AP mode. The rate-information is
associated with a band which is not available on the PHY.
Check for this condition in order to avoid crashing the whole system.
This patch should be removed once the roout cause has been found and
fixed.
Link: https://github.com/freifunk-gluon/gluon/issues/2980
Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry picked from commit 1278d47bea)
This prevents min_tx_power from functioning properly in some circumstances.
Add the missing newline.
Signed-off-by: Rany Hany <rany_hany@riseup.net>
(cherry picked from commit 6ca8752a9c)
fast-xmit must only be enabled after the sta has been uploaded to the driver,
otherwise it could end up passing the not-yet-uploaded sta via drv_tx calls
to the driver, leading to potential crashes because of uninitialized drv_priv
data.
Add a missing sta->uploaded check and re-check fast xmit after inserting a sta.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 438a97fab6)
noscan option was changed to hostapd_noscan but the entry in
wpa_supplicant was never updated resulting in the noscan option actually
never set.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 1070fbce6e)
Allow grace period for DFS available after shutting down beacons on the channel
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 16889141d9)
This significantly improves config reload behavior and also fixes some
corner cases related to running AP + mesh interfaces at the same time.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
During the channel scanning process, the driver will continuously
switch channels. It seems that the full RF calibration step in
rt2800_config_channel() caused the channel scanning function to
timeout. To fix it, move the RF calibration to rt2800_enable_radio()
so that it is only executed once. This commit also includes some
coding format adjustments to follow the Linux recommended style.
Fixes: 2824fa6963 ("mac80211: rework MT7620 PA/LNA RF calibration")
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
(cherry picked from commit ce6ad123e7)
This patch makes some improvements to the MT7620 RF calibration.
1. Move MT7620 PA/LNA calibration code to dedicated functions.
2. Restore RF and BBP registers before R-Calibration.
3. Do Rx DCOC calibration again before RXIQ calibration.
4. Use SoC specific AGC initial LNA value.
5. Correct MAC_RX_EN mask in rt2800_r_calibration()[1].
[1] This change may fix the "BBP/RF register access failed" error:
ieee80211 phy0: rt2800_wait_bbp_rf_ready: Error - BBP/RF register access failed, aborting
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
(cherry picked from commit 2824fa6963)
This patch will only force mac80211 loss detection upon ath10k by
masking the driver-specific loss-detection bit.
Ref: commit ed816f6ba8 ("mac80211: always use mac80211 loss detection")
Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry picked from commit d9070f8d23)
[felix.bau@gmx.de: replace path 6.2 with 5.15, refresh patches]
Signed-off-by: Felix Baumann <felix.bau@gmx.de>
ath10k does not report excessive loss in case of broken block-ack
sessions. The loss is communicated to the host-os, but ath10k does not
trigger a low-ack events by itself.
The mac80211 framework for loss detection however detects this
circumstance well in case of ath10k. So use it regardless of ath10k's
own loss detection mechanism.
Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry picked from commit ed816f6ba8)
Synchronize the ath11k backports with the current ath-next tree.
This introduces support for MBSSID and EMA, adds factory test mode and
some new HTT stats.
Tested-by: Francisco G Luna <frangonlun@gmail.com>
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit acde5271a6)
Backport EMA beacon support from kernel 6.4.
It is required for MBSSID/EMA suport in ath11k that will follow.
Tested-by: Francisco G Luna <frangonlun@gmail.com>
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 84b5735b4c)
That was a workaround for OpenWrt generation of config files. This patch
was used to postpone returning from probe function until loading
firmware and calling register_wiphy().
All of that is not needed anymore thanks to the ieee80211 hotplug.d
script introduced in the commit 5f8f8a3661 ("base-files, mac80211,
broadcom-wl: wifi detection and configuration"). That takes care of
generating /etc/config/wireless entries even if wireless device appears
late in the booting process.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
(cherry picked from commit bd26266314)
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
If CPTCFG_BRCMFMAC_SDIO is not used brcmf_of_fwnames is defined but not
used in the case of CONFIG_OF not compiled.
Restrict brcmf_of_fwnames only with CPTCFG_BRCMFMAC_SDIO selected.
Fix compilation warning:
/__w/openwrt/openwrt/openwrt/build_dir/target-x86_64-openwrt-linux-musl_musl/linux-x86_64/backports-6.1.24/drivers/net/wireless/broadcom/brcm80211/brcmfmac/of.h:16:1: error: 'brcmf_of_fwnames' defined but not used [-Werror=unused-function]
16 | brcmf_of_fwnames(struct device *dev, u32 *map_count)
| ^~~~~~~~~~~~~~~~
cc1: all warnings being treated as errors
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Handle error from copy_to_user in read_file_eeprom.
Fix compilarion warning:
/__w/openwrt/openwrt/openwrt/build_dir/target-x86_64-openwrt-linux-musl_musl/linux-x86_64/backports-6.1.24/drivers/net/wireless/ath/ath9k/debug.c: In function 'read_file_eeprom':
/__w/openwrt/openwrt/openwrt/build_dir/target-x86_64-openwrt-linux-musl_musl/linux-x86_64/backports-6.1.24/drivers/net/wireless/ath/ath9k/debug.c:1451:17: error: ignoring return value of 'copy_to_user' declared with attribute 'warn_unused_result' [-Werror=unused-result]
1451 | copy_to_user(user_buf, from, bytes);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
cc1: all warnings being treated as errors
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Synchronize the ath11k backports with the current ath-next tree.
This replaces the 160MHz with the upstreamed one, fixes 6GHz only WIPHY
registration, allows SAR usage on WCN6750 and plenty of REO fixes.
Signed-off-by: Robert Marko <robimarko@gmail.com>
By default both kmod-bcma and kmod-ssb are selected by kmod-b43.
However, only one of both modules is needed for bmips subtargets:
- bcma: bcm6318, bcm6328, bcm6362, bcm63268
- ssb: bcm6358, bcm6368
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
QCA has finally sent a proper fixup for the 160MHz regression upstream,
so lets use the pending fix which also properly sets center frequency 2
in case 80+80 MHz is used.
Signed-off-by: Robert Marko <robimarko@gmail.com>
Synchronize the ath11k backports with the current ath-next tree.
This replaces the management TLV pending fix with the upstreamed one,
fixes traffic flooding when AP and monitor modes are used at the same time,
fixes QCN9074 always showing -95 dBm for station RSSI in dumps,
fixes potential crash on boot if spectral scan is enabled due to writing to
unitialized memory and adds 11d scan offloading for WCN6750 and WCN6855.
Signed-off-by: Robert Marko <robimarko@gmail.com>
Currently, during initialization ath11k will receive a regulatory event
from the firmware in which it will receive the default regulatory domain
code and accompanying rules list and report those to the kernel.
Then if you try to change the regulatory domain to a different country code
it will do a weird thing in which it will send that to the FW and after
receiving the appropriate regulatory event it will parse the rules.
However, while its parsing there is a weird thing being done, and that is
that new raw rules from FW get intersected with the rules from the default
domain.
This is creating a big issue as the default domain is almost always set to
"US" or just "00" aka world so ath11k will unfairly limit you to the most
restrictive combination of rules based on the default domain and your
desired domain.
For example, in ETSI countries this is causing channels 12 and 13 on 2.4GHz
to not be usable since "US" limits 2.4GHz to 2472MHz instead of 2482MHz
like ETSI countries do.
So, lets do what TIP and even QCA do in their ath11k downstream tree and
completely get rid of the interesection code in ath11k.
Signed-off-by: Robert Marko <robimarko@gmail.com>
