mac80211 incorrectly processes A-MSDUs contained in A-MPDU frames. This
results in dropped packets and severely impacted throughput.
As a workaround, don't indicate support for A-MSDUs contained in
A-MPDUs. This improves throughput over mesh links by factor 10.
Ref: https://github.com/openwrt/mt76/issues/450
Signed-off-by: David Bauer <mail@david-bauer.net>
As also ramips/mt7621 now has a user of the ubnt-ledbar driver, make
the package available on all targets by removing the dependency on
@TARGET_mediatek_mt7622.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
This reverts commit 6e9613844c.
The patch was wrong in the first place as we base everything on
backports package and the compilation error was caused by an ath11k
present downstream. (will be needed later when backports package will be
updated but not now)
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
With kernel 5.15.61 the define IEEE80211_MAX_AMPDU_BUF got changed to
IEEE80211_MAX_AMPDU_BUF_HE. Add patch to fix compilation error on next
5.15 kernel versions.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
The "tx_burst" option which should control the value was
expecting more of a list and hence tx_queue_data2_burst
value wasn't updated.
Yes, it would make sense to have a list for this, the
existing code only updates tx_queue_data2_burst and
not the other tx_queue_data[0134]_burst values.
Signed-off-by: Alberto Martinez-Alvarez <amteza@gmail.com>
(formatted commit message, wrote extra information into commit,
moved tx_burst to existing json_get_vars)
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Prevents build errors by ensuring that it is only selected when a wext based
driver that needs it is also selected
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This package downloads raw files
which have names that are not corresponding to
the name and version of the package
as it is defined in the Makefile.
Use the option DL_SUBDIR to set the DL_DIR
to be a subdirectory named with
PKG_NAME and PKG_SOURCE_DATE
to better organize the downloads.
Define PKG_VERSION here
using PKG_SOURCE_DATE.
Signed-off-by: Michael Pratt <mcpratt@pm.me>
This matches the scheme used by other target packages and will avoid
confusion with any future version.
Signed-off-by: Andre Heider <a.heider@gmail.com>
Instead of always including the XHCI driver in the kernel on all
MediaTek boards, selectively include the kernel module only on boards
which actually make use of USB functionality.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
add Flow Queuing with Proportional Integral controller Enhanced (FQ-PIE) as an
optional kmod in network support and extract sched-pie from kmod-sched to
allow dependency on just kmod-sched-pie (PIE).
Signed-off-by: Kabuli Chana <newtownBuild@gmail.com>
Serge Vasilugin reports:
To improve mt7620 built-in wifi performance some changes:
1. Correct BW20/BW40 switching (see comments with mark (1))
2. Correct TX_SW_CFG1 MAC reg from v3 of vendor driver see
https://gitlab.com/dm38/padavan-ng/-/blob/master/trunk/proprietary/rt_wifi/rtpci/3.0.X.X/mt76x2/chips/rt6352.c#L531
3. Set bbp66 for all chains.
4. US_CYC_CNT init based on Programming guide, default value was 33 (pci),
set chipset bus clock with fallback to cpu clock/3.
5. Don't overwrite default values for mt7620.
6. Correct some typos.
7. Add support for external LNA:
a) RF and BBP regs never be corrected for this mode
b) eLNA is driven the same way as ePA with mt7620's pin PA
but vendor driver explicitly pin PA to gpio mode (for forrect calibration?)
so I'm not sure that request for pa_pin in dts-file will be enough
First 5 changes (really 2) improve performance for boards w/o eLNA/ePA.
Changes 7 add support for eLNA
Configuration w/o eLAN/ePA and with eLNA show results
tx/rx (from router point of view) for each stream:
35-40/30-35 Mbps for HT20
65-70/60-65 Mbps for HT40
Yes. Max results for 2T2R client is 140-145/135-140
with peaks 160/150, It correspond to mediatek driver results.
Boards with ePA untested.
Reported-by: Serge Vasilugin <vasilugin@yandex.ru>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Prepare patches for sending upstream by adding patch descriptions
generated from the original OpenWrt commits adding each patch.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Enable MediaTek protocol in btusb module to support e.g. the Bluetooth
part of the MT7921K NGFF/M.2 module.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Devices using GPT usually have FAT filesystem on boot partition and
that's where the intermediary backup of system configuration is stored
on sysupgrade. Automatic restoring of OpenWrt configuration after
sysupgrade will be inhibited if the driver is not loaded and file system
type is not specified in mount command.
Signed-off-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Older MT7623 ARMv7 SoC as well as new Filogic platforms come with
inside-secure,safexcel-eip97 units. Enable them in DTS and select the
driver kernel module by default on those platforms.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
It is common for 802.11ax NICs to support more than just AP mode, which
results in there being a distinct set of HE capabilities for each mode. As
(bad) luck would have it, iw prints out info for each HE mode in sequential
order according to `enum nl80211_iftype`, and AP mode isn't always first.
As a result, the wrong set of HE capabilities can be parsed if an AP NIC
supports station (managed) mode or any other mode preceding AP mode, since
only the first set of HE capabilities printed by iw is parsed from awk's
output.
This has a noticeable impact on beamforming for example, since managed mode
usually doesn't have beamformer capabilities enabled, while AP mode does.
Hostapd won't be set up with the configs to enable beamformer capabilities
in this scenario, causing hostapd to disable beamforming to HE stations
even when it's supported by the AP.
Always parse the correct set of HE capabilities for AP mode to fix this.
This is achieved by trimming all of iw's output prior to the AP mode
capabilities, which ensures that the first set of HE capabilities are
always for AP mode.
Signed-off-by: Sultan Alsawaf <sultan@kerneltoast.com>
There is only one module in kmod-sched that depends on iptables. Move it
to its own kmod package so we can drop the kmod-ipt-core dependency from
kmod-sched. This makes it possible to disable all kmod-ipt-* packages
without having to disable kmod-sched. Since we now default to firewall4
and nftables, we should avoid iptables dependencies where we can.
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
This fixes the following build error:
Package kmod-hwmon-tps23861 is missing dependencies for the following libraries:
regmap-i2c.ko
Fixes: b664646db7 ("kernel: add kmod-hwmon-tps23861 support")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
While kmod-input-core was previously indirectly selected by
kmod-input-polldev, this is now only the case on Linux 5.10.
Select kmod-input-core as dependency independently of the kernel
version to fix build error:
Package kmod-input-gpio-keys-polled is missing dependencies for the following libraries:
input-core.ko
Fixes: 54878fbbdd ("kernel: kmod-input-polldev: Depend on kernel 5.10")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Add support for randomly generating a MAC address for a wifi-iface
instance by setting `macaddr` to `random`
When set to `random`, a new locally administered unicast MAC address
is generated and assigned to the iface everytime it is (re-)configured
Signed-off-by: Manas Sambhus <manas.sambhus+github@gmail.com>
The input-polldev.ko kernel module was removed from kernel 5.11. The
normal input implementation now supports polling.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The sps30.ko driver was split into a main sps30.ko driver and a
sps30_i2c.ko driver for the I2C interface with kernel 5.14. Add the
sps30_i2c.ko module to the package too.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The isdn4linux drivers and subsystem was removed in kernel 5.3, remove
the kernel package also from OpenWrt.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The rtc-pt7c4338.ko was never upstream under this name, the driver was
removed from OpenWrt some years ago, remove the kmod-rtc-pt7c4338
package too.
Fixes: 74d00a8c38 ("kernel: split patches folder up into backport, pending and hack folders")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This adds the kmod-wwan package. This provides the generic wwan driver
core which is needed for some existing packages.
Currently the drivers/net/wwan/wwan.ko driver is compiled into the
kernel when one of the wwan module is activated, better build it as a
kernel module.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The ulog iptables target was removed with kernel 3.17, remove the kernel
and also the iptables package in OpenWrt too.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The nft NAT packages for IPv4 and IPv6 were merged into the common
packages with kernel 5.1. The kmod-nft-nat6 package was empty in our
build, remove it.
Multiple kernel configuration options were also removed, remove them
from our generic kernel configuration too.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The w1_ds2760.ko driver was merged into the ds2760_battery.ko driver.
The driver was removed and this package was never build any more.
This happened with kernel 4.19.
Remove this unused package.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The R8712U driver depends on cfg80211. cfg80211 is provided by mac80211
backports, we can not build any in kernel drivers which depend on
cfg80211 which is an out of tree module in OpenWrt.
The cfg80211 dependency was added with kernel 5.9.
We could add rtl8192su to backports and build it from there.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The malta target does not compile CONFIG_OF_MDIO into the kernel. On
malta the kmod-mdio-devres package depends on kmod-of-mdio.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The CONFIG_CACHEFILES configuration option makes the kernel build
cachefiles.ko, also package it. Build CONFIG_CACHEFILES as module and do
not try to build it into the kernel. This did not work because it
depended on CONFIG_FSCACHE which was already build as module.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The CONFIG_NLS_CODEPAGE_932 Config option builds the nls_cp932.ko and
the nls_euc-jp.ko kernel module, package both of them.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Add dwc2_pci kernel module into own kernel package.
The dwc2_pci.ko kernel module was always build when kmod-usb-dwc2
was selected, but it was not packaged.
Add the missing kmod-usb-phy-nop dependency to the kmod-usb-dwc2-pci
package too. The CONFIG_USB_DWC2_PLATFORM option is already gone for
some time.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The ar8xxx.ko kernel module uses the devm_mdiobus_alloc() function
provided by kmod-mdio-devres, add the missing dependency.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Both of these packages depend on CONFIG_WWAN
in the kernel.
Also fix the missing "wwan" subfolder in the path.
This fixes the missing devices in /dev after booting an MHI capable modem.
Fixes: 2519190fec ("kernel: package mhi wwan ctrl driver")
Fixes: 6af46796fa ("kernel: package mhi mbim driver")
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Backports upstream fix from 5.19
lockdep complains use of uninitialized spinlock at ieee80211_do_stop() [1],
for commit f856373e2f31ffd3 ("wifi: mac80211: do not wake queues on a vif
that is being stopped") guards clear_bit() using fq.lock even before
fq_init() from ieee80211_txq_setup_flows() initializes this spinlock.
According to discussion [2], Toke was not happy with expanding usage of
fq.lock. Since __ieee80211_wake_txqs() is called under RCU read lock, we
can instead use synchronize_rcu() for flushing ieee80211_wake_txqs().
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
This extracts kmod-sched-act-police to allow using it without adding all
the packages from the big kmod-sched package.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
Add the ipset/ip_set_hash_ipmac.ko file. The CONFIG_IP_SET_HASH_IPMAC
KConfig option is already set by the package.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
This adds the act_sample.ko and psample.ko kernel module which allows
traffic sampling.
Signed-off-by: Thomas Langer <tlanger@maxlinear.com>
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
Extract the kmod-sched-prio and kmod-sched-red kernel modules from the
big kmod-sched package. This allows adding the two kernel modules to
OpenWrt without adding the kmod-sched and all its depdnecy.
Signed-off-by: Thomas Langer <tlanger@maxlinear.com>
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
This puts the kmod-sched packages into an alphabetical order.
I kept the kmod-sched-core at the top as this is the main package.
No changes other than reordering were done.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
Change SCHED_MODULES_EXTRA to an explicit list of modules
instead of taking everything that is not filtered out.
This removes the need of updating the filter each time an extra
sch_*, act_* or similar is added with an own kmod definition.
Signed-off-by: Thomas Langer <tlanger@maxlinear.com>
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
The sch_fq_codel.ko and the sch_fifo.ko are always compiled into the
kernel, they are activated in the generic kernel configuration. There is
no need to activate the build of these kernel modules in the kmod-sched*
packages.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
"The MCP2221 is a USB-to-UART/I2C serial converter which enables
USB connectivity in applications that have a UART and I2C interfaces."
<https://www.microchip.com/en-us/product/MCP2221>
Signed-off-by: Leo Soares <leo@hyper.ag>
(replaced GPIOLIB KConfig with @GPIO_SUPPORT)
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
all other drivers depend on @GPIO_SUPPORT rather than
forcing CONFIG_GPIOLIB=y.
(I wonder what would happen if someone decides to try
UML with USBIP?)
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
This updates mac80211 to version 5.15.58-1 which is based on kernel
5.15.58.
The removed patches were applied upstream.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Add driver for NVM Express block devices, ie. PCIe connected SSDs.
Targets which allow booting from NVMe (x86, maybe some mvebu boards come
to mind) should have it built-in, so rootfs can be mounted from there.
For targets without NVMe support in bootloader or BIOS/firmware it's
sufficient to provide the kernel module package.
On targets having the NVMe driver built-in the resulting kmod package
is an empty dummy. In any case, depending on or installing kmod-nvme
results in driver support being available (either because it was already
built-in or because the relevant kernel modules are added and loaded).
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Accessing the console on many devices is difficult.
netconsole eases debugging on devices that crash
after the network is up.
Reference to the netconsole documentation in upstream Linux:
<https://www.kernel.org/doc/html/latest/networking/netconsole.html>
|
|netconsole=[+][src-port]@[src-ip]/[<dev>],[tgt-port]@<tgt-ip>/[tgt-macaddr]
|
| where
| + if present, enable extended console support
| src-port source for UDP packets (defaults to 6665)
| src-ip source IP to use (interface address)
| dev network interface (eth0)
| tgt-port port for logging agent (6666)
| tgt-ip IP address for logging agent
| tgt-macaddr ethernet MAC address for logging agent (broadcast)
OpenWrt specific notes:
OpenWrt's device userspace scripts are attaching the network
interface (i.e. eth0) to a (virtual) bridge (br-lan) device.
This will cause netconsole to report:
|network logging stopped on interface eth0 as it is joining a master device
(and unfortunately the traffic/logs to stop at this point)
As a workaround, the netconsole module can be manually loaded
again after the bridge has been setup with:
insmod netconsole netconsole=@/br-lan,@192.168.1.x/MA:C...
One way of catching errors before the handoff, try to
append the /etc/modules.conf file with the following extra line:
options netconsole netconsole=@/eth0,@192.168.1.x/MA:C...
and install the kmod-netconsole (=y) into the base image.
Signed-off-by: Catalin Toda <catalinii@yahoo.com>
(Added commit message from PR, added links to documentation)
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
the legacy driver was dropped in linux 5.14-rc3:
commit d249ff28b1d8 ("intersil: remove obsolete prism54 wireless driver")
Quoting Lukas Bulwahn:
"p54 replaces prism54 so users should be unaffected."
Reported-by: Marius Dinu <m95d+git@psihoexpert.ro>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
TechNexion PICO-PI-IMX7D uses BCM4339 Wi-Fi interface in SDIO mode.
Enable SDIO support for imx/cortexa7 to fully support it in images.
Signed-off-by: Lech Perczak <lech.perczak@gmail.com>
Add package supporting Bluetooth HCI interfaces connected over SDIO.
Signed-off-by: Lech Perczak <lech.perczak@gmail.com>
[pepe2k@gmail.com: dropped rfkill dependency, other minor text fixes]
Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
Backport upstream fix to build on kernel 5.15.52 or later since kernel
devs backported newer functionality to older kernels.
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
The mac80211-hwsim and the Intel iwlwifi driver support ieee80211ax, add
the missing DRIVER_11AX_SUPPORT dependency too.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
c07f45927839 firmware: update mt7622 firmware to version 20220630
af406a2d1c36 mt76: do not use skb_set_queue_mapping for internal purposes
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This external switch driver should be loaded on boot for network
support in failsafe mode.
Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
This external switch driver should be loaded on boot for network
support in failsafe mode.
Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
