db3934d2f740 scripts/netifd-wireless.sh: properly fix WPA3 Enterprise support
Support the following values for the different WPA3 Enterprise modes:
- wpa3-mixed: WPA3 Enterprise transitional mode
This supports EAP with both SHA1 and SHA-256, with optional MFP
- wpa3: WPA3 Enterprise only mode
This supports only SHA256 with mandatory MFP
- wpa3-192: WPA3 Enterprise with mandatory 192 bit support
This uses only GCMP-256 ciphers
Disable 192 bit support and GCMP-256 ciphers for the regular "wpa3" mode.
It seems that even leaving in optional 192 bit support breaks auth on some
clients, including iOS devices.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 8c03dc962f)
WPA3 Enterprise-transitional requires optional MFP support and SHA1+SHA256
WPA3 Enterprise-only requires SHA1 support disabled and mandatory MFP.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit b63df6ce5d)
Use the SHA384 variant to account for longer keys with more security
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit f0d1349b52)
077e05f2b129 vlan/vlandev: pass through extra vlan information passed via hotplug
40fad91eb5be wireless: add network_vlan config attribute
1571e18e4a69 bridge: add support for configuring extra tagged vlans on member devices
b719f189f243 bridge: make hotplug-added vlans default to tagged
edf3aced9f9a bridge: add support for adding vlan ranges via hotplug
493e1589bc8b bridge: fix coverity false positive report
03a619947717 bridge: add support for configuring extra vlans for the bridge itself
4bea6d21a9ab wireless: fix changing reconf/serialize options in configuration
255b4d5c472e wireless: fix handling config reload with reconf=1
1ab992a74b43 wireless: fix another reconf issue
e94f7a81a039 bridge: fix config reload on 32 bit systems
8c2758b4fbbb wireless: add support for replacing data blobs at runtime
0ff22a6a68ce wireless: enable dynamic reconfiguration by default
4711f74479e2 netifd: fix disabling radio via config if reconf is being used
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 09fd59938b)
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>
For the ZTE MF287 series, a special recovery image is built. The Makefile
worked fine on snapshot, but created corrupt images on the 23.05 images.
By using the appropriate variable, this should be fixed.
Signed-off-by: Andreas Böhler <dev@aboehler.at>
(cherry picked from commit a9cc3708e0)
This commit removes the padded zeros in the date formatting.
The padded zeros from the date command causes the numbers
to be interpreted as an octal number by printf. Months, days,
and years with the number 08 or 09 raise an error in printf as an
"invalid octal number" and get interpreted as a zero.
Signed-off-by: Max Qian <public@maxqia.com>
(cherry picked from commit 794349a28a)
On some WLR-7100 routers, significant packet loss was observed. This is
fixed by configuring a delay on the GMAC0 RXD and RXDV lines.
The values used in this commit are copied from the values used by the
stock firmare (based on register dumping).
Out of four test routers, the problem was consistently observed on two.
It is unclear what the relevant difference is exactly (the two working
routers were v1 001 with AR1022 and v1 002 with AR9342, the two broken
routers were both v1 002 with AR1022). All PCB routing also seems
identical, so maybe there is some stray capacitance on some of these
that adds just enough delay or so...
With this change, the packet loss disappears on the broken routers,
without introducing new packet loss on the previously working routers.
Note that the PHY *also* has delays enabled (through
`qca,ar8327-initvals`) on both RX and TX lines, but apparently that is
not enough, or it is not effective (registers have been verified to be
written).
For detailed discussion of this issue and debug history, see
https://forum.openwrt.org/t/sitecom-wlr-7100-development-progress/79641
Signed-off-by: Matthijs Kooijman <matthijs@stdin.nl>
(cherry picked from commit d2ce3a61aa)
These suboptions (PLATFORM, FSL_MC and MLX5_VFIO_PCI)
may be prompted for when VFIO is enabled, regardless of
architecture.
These are not related to the main vfio use case
(passthrough of PCIe devices)
Signed-off-by: Mathew McBride <matt@traverse.com.au>
(5.15 version of abc536f547)
Arm platforms with the right hardware blocks (such as
GICv3.0+ interrupt controller and SMMU/IOMMU) are
able to use vfio-pci to pass through PCI devices
to a VM.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
(cherry picked from commit 9ac01aac0a)
The vfio module only exposes the enable_unsafe_noiommu_mode parameter
if CONFIG_VFIO_NOIOMMU is enabled. When it isn't, the module
will complain about an unknown parameter:
vfio: unknown parameter 'enable_unsafe_noiommu_mode' ignored
As CONFIG_VFIO_NOIOMMU is disabled by the module package,
we can remove the module loading parameter.
Signed-off-by: Mathew McBride <matt@traverse.com.au>
(cherry picked from commit 7ad8612a23)
vhost-net is used to accelerate traffic to virtualisation
guests that use the virtio-net network card in QEMU.
Generally it is invoked by specifying "vhost=on" to a
QEMU -netdev device:
qemu-system-aarch64 -nographic -M virt -cpu host \
--enable-kvm -bios u-boot.bin -smp 1 -m 2048 \
-drive file=openwrt-armsr-armv8.img,format=raw,index=0,media=disk \
-device "virtio-net,netdev=landev,disable-legacy=off,disable-modern=off" \
-netdev "tap,id=landev,helper=/usr/lib/qemu-bridge-helper --br=br-lan,vhost=on"
Signed-off-by: Mathew McBride <matt@traverse.com.au>
(cherry picked from commit d188db8aed)
There's a typo in here: board_name is a function, not a variable. This
issue was pointed out on the OpenWrt forum.
Closes: #13409
Reviewed-by: Robert Marko <robimarko@gmail.com>
Signed-off-by: Brian Norris <computersforpeace@gmail.com>
(cherry picked from commit 7b78a19e6a)
When the membase and pci_dev pointer were moved to a new struct in priv,
the actual membase users were left untouched, and they started reading
out arbitrary memory behind the struct instead of registers. This
unfortunately turned the RNG into a constant number generator, depending
on the content of what was at that offset.
To fix this, update geode_rng_data_{read,present}() to also get the
membase via amd_geode_priv, and properly read from the right addresses
again.
Closes#13417.
Reported-by: Timur I. Davletshin <timur.davletshin@gmail.com>
Tested-by: Timur I. Davletshin <timur.davletshin@gmail.com>
Suggested-by: Jo-Philipp Wich <jo@mein.io>
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
(cherry picked from commit 09d13cd8d8)
Add 2 seconds sleep after each forcibly killed/tried-to-kill process
in the final process termination loop in sysupgrade stage2.
This is needed especially for qualcommax/ipq807x, where ath11k
wireless driver may have a long 10-20 seconds delay after termination
before actually getting killed. This often breaks sysupgrade.
The current KILL loop in kill_remaining does all 10 kill attempts
consecutively without any delay, as evidenced here in a failing sysupgrade.
It does not allow any time for the process to finalize its internal
termination.
Sat Sep 2 19:05:56 EEST 2023 upgrade: Sending TERM to remaining processes ...
Sat Sep 2 19:05:56 EEST 2023 upgrade: Sending signal TERM to hostapd (2122)
Sat Sep 2 19:05:56 EEST 2023 upgrade: Sending signal TERM to hostapd (2138)
Sat Sep 2 19:06:00 EEST 2023 upgrade: Sending KILL to remaining processes ...
Sat Sep 2 19:06:00 EEST 2023 upgrade: Sending signal KILL to hostapd (2122)
Sat Sep 2 19:06:00 EEST 2023 upgrade: Sending signal KILL to hostapd (2138)
Sat Sep 2 19:06:00 EEST 2023 upgrade: Sending signal KILL to hostapd (2138)
Sat Sep 2 19:06:00 EEST 2023 upgrade: Sending signal KILL to hostapd (2138)
Sat Sep 2 19:06:00 EEST 2023 upgrade: Sending signal KILL to hostapd (2138)
Sat Sep 2 19:06:00 EEST 2023 upgrade: Sending signal KILL to hostapd (2138)
Sat Sep 2 19:06:00 EEST 2023 upgrade: Sending signal KILL to hostapd (2138)
Sat Sep 2 19:06:00 EEST 2023 upgrade: Sending signal KILL to hostapd (2138)
Sat Sep 2 19:06:00 EEST 2023 upgrade: Sending signal KILL to hostapd (2138)
Sat Sep 2 19:06:00 EEST 2023 upgrade: Sending signal KILL to hostapd (2138)
Sat Sep 2 19:06:00 EEST 2023 upgrade: Sending signal KILL to hostapd (2138)
Sat Sep 2 19:06:00 EEST 2023 upgrade: Failed to kill all processes.
sysupgrade aborted with return code: 256
The change in this commit adds a 2 seconds delay after each kill attempt
in order to allow some processes to more gracefully handle their
internal termination.
The result is like this:
Sun Sep 3 11:15:10 EEST 2023 upgrade: Sending TERM to remaining processes ...
Sun Sep 3 11:15:10 EEST 2023 upgrade: Sending signal TERM to hostapd (2309)
Sun Sep 3 11:15:10 EEST 2023 upgrade: Sending signal TERM to hostapd (2324)
Sun Sep 3 11:15:14 EEST 2023 upgrade: Sending KILL to remaining processes ...
Sun Sep 3 11:15:14 EEST 2023 upgrade: Sending signal KILL to hostapd (2309)
[ 699.827521] br-lan: port 7(hn5wpa2r) entered disabled state
[ 699.908673] device hn5wpa2r left promiscuous mode
[ 699.908721] br-lan: port 7(hn5wpa2r) entered disabled state
[ 701.038029] br-lan: port 6(hn5wpa3) entered disabled state
Sun Sep 3 11:15:16 EEST 2023 upgrade: Sending signal KILL to hostapd (2324)
[ 702.058256] br-lan: port 5(hn2wlan) entered disabled state
[ 709.250063] stage2 (8237): drop_caches: 3
Sun Sep 3 11:15:25 EEST 2023 upgrade: Switching to ramdisk...
The delay introduced here only kicks in if there is some process that
does not get terminated by the first TERM call. Then there is at least
one 2 sec wait after the first KILL loop round.
This commit is related to discussion in PRs #12235 and #12632
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
Reviewed-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 47d56ae546)
ALFA Network AX1800RM (FCC ID: 2AB877621) is a dual-band Wi-Fi 6
(AX1800) router, based on MediaTek MT7621A + MT79x5D platform.
Specifications:
- SOC: MT7621A (880 MHz)
- DRAM: DDR3 256 MiB (Nanya NT5CC128M16JR-EK)
- Flash: 16 MiB SPI NOR (EN25QH128A-104HIP)
- Ethernet: 4x 10/100/1000 Mbps (SOC's built-in switch)
- Wi-Fi: 2x2:2 2.4/5 GHz (MT7905DAN + MT7975DN)
(MT7905DAN doesn't support background DFS scan/BT)
- LED: 6x green, 1x green/red
- Buttons: 2x (reset, WPS)
- Antenna: 4x external, non-detachable omnidirectional
- UART: 1x 4-pin (2.54 mm pitch, J4, not populated)
- Power: 12 V DC/1 A (DC jack)
MAC addresses:
LAN: 00:c0:ca:xx:xx:4e (factory 0x4, +2)
WAN: 00:c0:ca:xx:xx:4f (factory 0x4, +3)
2.4 GHz: 00:c0:ca:xx:xx:4c (factory 0x4, device's label)
5 GHz: 00:c0:ca:xx:xx:4c (factory 0xa)
Flash instructions for web-based U-Boot recovery:
1. Power the device with WPS button pressed and wait around 10 seconds.
2. Setup static IP 192.168.1.2/24 on your PC.
3. Go to 192.168.1.1 in browser and upload 'recovery' image.
The device runs LEDE 17.01 (kernel 4.4.x) based firmware with 'failsafe'
mode available which allows alternative upgrade method:
1. Run device in 'failsafe' mode and change password for default user.
2. SSH to the device, transfer 'sysupgrade' image and perform upgrade
in forced mode, without preserving settings: 'sysupgrade -n -F ...'.
Other notes:
If you own early version of this device, the vendor firmware might
refuse OpenWrt image because of missing custom header. In that case,
ask vendor's customer support for stock firmware without custom header
support/requirement.
Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
(backported from commit f1aaa267f0)
With recent updates of TF-A the previously already fixed bug slipped
back into the source tree. Again, reorder bl2 init for MT7622 and
initialize WDT only after DRAM init has completed to avoid the
notorious hang.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 948ad2ec7a)
Sync SPI-NAND/ECC controller driver for MT7622, MT7981, MT7986 and MT7988:
* Platform data for MT7981 was actually missing and is now added.
* Add support for Winbond W25N01KV 1Gbit chip.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 9725524235)
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)
MT7620 wireless radio needs change the pin group function between
"gpio" and "pa" during the calibration process. However, ralink
pinctrl driver doesn't support requesting different functions for
the same group. This patch enables pinctrl consumers to perform
such operations.
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
(cherry picked from commit b4ea49ad44)
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)
Relocating the device tree is required for being apply to apply
device tree overylay at boot.
Fixes: 34bb33094a ("mediatek: use updated device tree overlay mechanism for BPi-R64")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit a8cbee8e2d)
Always enable built-in 2.5G PHY on MT7988 for now, so that it can be
used. In future it would be nice to be able to switch power and MDIO
access via address 0 at run-time in Linux, both, to be able to use
external PHYs at address 0 and to reduce power consumption on systems
not using the built-in 2.5G PHY.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 33046d2a47)
The U-Boot build for the MT7988 reference board booting from SD card
wrongly depended on the 'ddr4' variant of the ARM TrustedFirmware-A build
even though the 'comb' variant is used. Fix that dependency.
Fixes: 572ea68070 ("uboot-mediatek: add patches for MT7988 and builds for RFB")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 78e3adcaf9)
When adapting the network configuration for MT7988 RFB a stray quote
was left in a script. Remove it to fix generating the default network
configuration.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 8f5986355c)
Import commits from upstream Linux replacing some downstream patches.
Move accepted patches from pending-{5.15,6.1} to backport-{5.15,6.1}.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit f631c7bbb1)
Some recent models of the Ubiquiti Networks UniFi 6 LR access point
come with a RealTek RTL8211FS 1000M/100M/10M PHY instead of the
Aquantia AQR112 2500M/1000M/100M/10M PHY used in both v1 and v2. Add
build for this variant so we can support Ethernet with the PHY.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit a0f4eadf6a)
While the v2 is nearly identical to v1, v3 uses a different PHY and
needs a different build for Ethernet to work in U-Boot.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit fff4441135)
Switch to OpenWrt uImage.FIT bootmethod and include various bootloader
artifacts with the generated binaries.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 035a88ae55)
Select many potentially useful options for the MT7988 RFB U-Boot builds.
The resulting loader is intended as a development tool and intends to be
generic. It does *not* have a default bootcmd set, but allows to boot
pretty much everything, including EFI executables.
To install this U-Boot build to the eMMC:
opkg install mmc-utils partx-utils
mmc bootpart enable 1 1 /dev/mmcblk0
echo 0 > /sys/block/mmcblk0boot0/force_ro
dd if=*mediatek_mt7988a-rfb-nand-emmc-preloader.bin of=/dev/mmcblk0boot0
dd if=*mediatek_mt7988a-rfb-nand-emmc-gpt.bin of=/dev/mmcblk0
partx -a /dev/mmcblk0
dd if=*mediatek_mt7988a-rfb-nand-emmc-bl31-uboot.fip of=/dev/mmcblk0p3
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 1d5778e18d)
* re-factor WED components to boot fine also on limited loaders
* add LEDs of integrated GE PHY
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 3ef8760e87)
Set correct pull-type data and add additional uart groups for MT7981.
Assign functions to configure pin bias for MT7986.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 9f8fde216d)
Using the I2C host controller on the MT7981 SoC requires 4 clocks to
be enabled. One of them, the pmic clk, is only enabled in case
'mediatek,have-pmic' is also set which has other consequences which
are not desired in this case.
Allow defining a pmic clk even in case the 'mediatek,have-pmic' propterty
is not present and the bus is not used to connect to a pmic, but may
still require to enable the pmic clock.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 2544dc34f2)
Unfortunately some device tree properties have slipped under the table
when switching from our downstream device tree.
Bring back 3W power for SFP cages and restore thermal trip points to
make sense again.
Fixes: 7a0ec001ff ("mediatek: sync MT7986 device trees with upstream")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 06a150aed7)
Backport initial LEDs hw control support. Currently this is limited to
only rx/tx and link events for the netdev trigger but the API got
accepted and the additional modes are working on and will be backported
later.
Refresh every patch and add the additional config flag for QCA8K new
LEDs support.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 0a4b309f41)
This allows supporting a mix of devices with or without hw offloading support
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit c5b7be8316)
7aefb47 jitterentropy-rngd: update to the v1.2.0
What's interesting about jitterentropy-rngd v1.2.0 release is that it
bumps its copy of jitterentropy-library from v2.2.0 to the v3.0.0. That
bump includes a relevant commit 3130cd9 ("replace LSFR with SHA-3 256").
When initializing entropy jent calculates time delta. Time values are
obtained using clock_gettime() + CLOCK_REALTIME. There is no guarantee
from CLOCK_REALTIME of unique values and slow devices often return
duplicated ones.
A switch from jent_lfsr_time() to jent_hash_time() resulted in many less
cases of zero delta and avoids ECOARSETIME.
Long story short: on some system this fixes:
[ 6.722725] urngd: jent-rng init failed, err: 2
This is important change for BCM53573 which doesn't include hwrng and
seems to have arch_timer running at 36,8 Hz.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
(cherry picked from commit c74b5e09e6)
This reverts commit 23953cfa5a.
01837 seems to be having multiple issues, one being broken multicast so
lets revert to the last version that was better for the stable branch.
Signed-off-by: Robert Marko <robimarko@gmail.com>