Newer kernels have a patch that implements compatible functionality
directly. Adjust the attribute of our own patch in preparation for
dropping it later
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Enables support for Dynack feature.
When a remote station is far away, we need to compensate for the distance
by allowing more time for an ACK to arrive back before issueing a retransmission.
Currently, it needs to be set fixed to indicate the maximum distance the remote
station will ever be.
While this mostly works for static antennae, it introduces 2 issues:
- If the actual distance is less, speed is reduced due to a lot of wates wait-time
- If the distance becomes greater, retries start to occur and comms can get lost.
Allowing to set it dynamically using dynack ensures the best possible tradeoff
between speed vs distance.
This feature is currently only supported in ath9k.
it is also disabled by default.
Enabling it can be done in 2 ways:
- issue cmd: iw phy0 set distance auto
- sending the NL80211_ATTR_WIPHY_DYN_ACK flag to mac80211 driver using netlink
Disabling it can be done by providing a valid fixed value.
To give an idea of a practical example:
In my usecase, we have mesh wifi device installed on ships/platforms.
Currently, the coverage class is set at 12000m fixed.
When a vessel moved closer (ex. 1500m), the measured link capacity was a lot
lower compared to setting the coverage class fixed to 1500m
Dynack completely solved this, nearly providing double the bandwidth at closer range
compared to the fixed setting of 12000m being used.
Also when a vessel sailed to a distance greater than the fixed setting,
communication was lost as the ACK's never arrived within the max allowed timeframe.
Actual distance: 6010m
iperf 60s run avg
Fixed 12150m: 31 Mbit/s
Dynack: 58 Mbit/s
Fixed 6300m: 51 Mbit/s
Dynack: 59 Mbit/s
Fixed 3000m: 13 Mbit/s (lots of retries)
Dynack: 58 Mbit/s
Actual distance: 1504m
iperf 60s run avg
Fixed 12150m: 31 Mbit/s
Dynack: 86 Mbit/s
Fixed 6300m: 55 Mbit/s
Dynack: 87 Mbit/s
Fixed 3000m: 67 Mbit/s
Dynack: 87 Mbit/s
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Rereshed all patches
Reworked patches to match upstream:
335-v4.16-netfilter-nf_tables-add-single-table-list-for-all-fa.patch
Compile-tested on: cns3xxx, imx6, x86_64
Runtime-tested on: cns3xxx, imx6, x86_64
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
* device: print daddr not saddr in missing peer error
* receive: style
Debug messages now make sense again.
* wg-quick: android: support excluding applications
Android now supports excluding certain apps (uids) from the tunnel.
* selftest: ratelimiter: improve chance of success via retry
* qemu: bump default kernel version
* qemu: decide debug kernel based on KERNEL_VERSION
Some improvements to our testing infrastructure.
* receive: use NAPI on the receive path
This is a big change that should both improve preemption latency (by not
disabling it unconditionally) and vastly improve rx performance on most
systems by using NAPI. The main purpose of this snapshot is to test out this
technique.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Update to the latest version of iproute2; see https://lwn.net/Articles/756991/
for a full overview of the changes in 4.17.
Remove upstream patch 002-json_print-fix-hidden-64-bit-type-promotion.
Backport upstream patch 001-rdma-sync-some-IP-headers-with-glibc fixing
rdma compile issue.
At the same time re-organize patch numbering so the OpenWRT specific
patches start at 100.
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
Partition name is picked by a parser_trx_data_part_name(). It has to
get correct partition offset (taking care of bad blocks) to work
properly.
This fixes UBI support for devices that have kernel flashed on partition
with a bad block.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
The recent build failures on various platforms were apparently caused by
the fact that LTO build support in the configure script does not check
if it has a suitable version of gcc and simply assumes that gcc-ar is
available and can be used for intermediate files.
Since we really don't need to build nasm with LTO, simply disable it and
keep the whole build more portable
Signed-off-by: Felix Fietkau <nbd@nbd.name>
345bba0 dhcpv4: improve error checking in handle_dhcpv4()
c0f6390 odhcpd: Check if open the ioctl socket failed
Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
On Debian 9 nasm does not build when we force it to use ranlib, for
macOSX this is needed. Only force this on macOSX and not on any other
OS, this should fix the build of nasm on Linux systems. On my Debian
system the nasm configure script selects gcc-ranlib and gcc-ar instead.
Fixes: d3a7587eb9 ("toolchain/nasm: fix missing AR/RANLIB variables")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Update README to include Openwrt branding and improve wording.
Point at the Openwrt wiki in .github templates.
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
This commit adds support for the OCEDO Raccoon
SOC: Atheros AR9344
RAM: 128MB
FLASH: 16MiB
WLAN1: AR9344 2.4 GHz 802.11bgn 2x2
WLAN2: AR9382 5 GHz 802.11an 2x2
INPUT: RESET button
LED: Power, LAN, WiFi 2.4, WiFi 5
Serial: Header Next to Black metal shield
Pinout is 3.3V - GND - TX - RX (Arrow Pad is 3.3V)
The Serial setting is 115200-8-N-1.
NOTE: The U-Boot won't boot with the serial attached.
Boot the device without serial attached and attach it
after 3 seconds.
Tested and working:
- Ethernet
- 2.4 GHz WiFi
- 5 GHz WiFi
- TFTP boot from ramdisk image
- Installation via ramdisk image
- OpenWRT sysupgrade
- Buttons
- LEDs
Installation seems to be possible only through booting an OpenWRT
ramdisk image.
Hold down the reset button while powering on the device. It will load a
ramdisk image named 'raccoon-uImage-initramfs-lzma.bin' from 192.168.100.8.
Note: depending on the present software, the device might also try to
pull a file called 'raccoon-uimage-factory'. Only the name differs, it
is still used as a ramdisk image.
Wait for the ramdisk image to boot. OpenWRT can be written to the flash
via sysupgrade or mtd.
Due to the flip-flop bootloader which we not (yet) support, you need to
set the partition the bootloader is selecting. It is possible from the
initramfs image with
> fw_setenv bootcmd run bootcmd_1
Afterwards you can reboot the device.
Signed-off-by: David Bauer <mail@david-bauer.net>
CONFIG_USB_MTU3 is not visible for the mediatek target by default, but
only when CONFIG_USB_GADGET is set. This will config option will be
remove with when running "make kernel_oldconfig", move this option to
the generic config to prevent this.
This fixes the build of the mt7623 subtarget of the mediatek target.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Shrink the tiny kernel by moving all switch and ethernet phy drivers to
the generic kernel config instead of the target kernel config.
All boards in the tiny and nand target are either ar7240 or ar9331 based,
which don't support external xMII and therefore no external ethernet phy
can be connected. None of the boards uses a realtek switch either.
Signed-off-by: Lucian Cristian <lucian.cristian@gmail.com>
The Unifi AC-Mesh Pro has identical hardware to the Unifi AC-Pro except
USB support.
Furthermore for setting parameters like antenna gain it is helpful to
know the exact device variant.
Signed-off-by: Christoph Krapp <achterin@googlemail.com>
Add a template for safeloader images and include it instead of
overwriting variables defined in the common tp-link build commands.
Split the existing tp-link templates to proper implement the safeloader
template.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Drop the LOADER_TYPE variables in case no loader is used at all or move
the variable to devices which are using a loader.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Use the provided image build variables to point the kernel-bin build
command to the kernel we are interested in.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Use the LOADER_TYPE variable to specify that we need the elf preloader
and append the loader via the corresponding build recipe. It allows to
enable initramfs images again for mikrotik NAND images, which caused a
build error before.
Add the minor header only to the kernel of the sysupgrade images, as it
is only required for the bootloader to find the kernel on flash.
Signed-off-by: Mathias Kresin <dev@kresin.me>
I-O DATA WN-AX1167GR is a 2.4/5 GHz band 11ac router, based on
MediaTek MT7621A.
Specification:
- MT7621A (2-Cores, 4-Threads)
- 64 MB of RAM (DDR2)
- 16 MB of Flash (SPI)
- 2T2R 2.4/5 GHz
- 5x 10/100/1000 Mbps Ethernet
- 2x LEDs, 4x keys (2x buttons, 1x slide switch)
- UART header on PCB
- Vcc, GND, TX, RX from ethernet port side
- baudrate: 115200 bps (U-Boot, OpenWrt)
Stock firmware:
In the stock firmware, WN-AX1167GR has two os images each composed of
Linux kernel and rootfs.
These images are stored in "Kernel" and "app" partition of the
following partitions, respectively.
(excerpt from dmesg):
MX25L12805D(c2 2018c220) (16384 Kbytes)
mtd .name = raspi, .size = 0x01000000 (16M) .erasesize = 0x00010000 (64K) .numeraseregions = 0
Creating 10 MTD partitions on "raspi":
0x000000000000-0x000001000000 : "ALL"
0x000000000000-0x000000030000 : "Bootloader"
0x000000030000-0x000000040000 : "Config "
0x000000040000-0x000000050000 : "Factory"
0x000000050000-0x000000060000 : "iNIC_rf"
0x000000060000-0x0000007e0000 : "Kernel"
0x000000800000-0x000000f80000 : "app"
0x000000f90000-0x000000fa0000 : "Key"
0x000000fa0000-0x000000fb0000 : "backup"
0x000000fb0000-0x000001000000 : "storage"
The flag for boot partition is stored in "Key" partition, and U-Boot
reads this and determines the partition to boot.
If the image that U-Boot first reads according to the flag is
"Bad Magic Number", U-Boot then tries to boot from the other image.
If the second image is correct, change the flag to the number
corresponding to that image and boot from that image.
(example):
## Booting image at bc800000 ...
Bad Magic Number,FFFFFFFF
Boot from KERNEL 1 !!
## Booting image at bc060000 ...
Image Name: MIPS OpenWrt Linux-4.14.50
Image Type: MIPS Linux kernel Image (lzma compressed)
Data Size: 1865917 Bytes = 1.8 MB
Load Address: 80001000
Entry Point: 80001000
Verifying Checksum ... OK
Uncompressing Kernel Image ... OK
raspi_erase_write: offs:f90000, count:34
.
.
Done!
Starting kernel ...
Flash instruction using factory image:
1. Connect the computer to the LAN port of WN-AX1167GR
2. Connect power cable to WN-AX1167GR and turn on it
3. Access to "192.168.0.1" on the web browser and open firmware
update page ("ファームウェア")
4. Select the OpenWrt factory image and perform firmware update
5. On the initramfs image, execute "mtd erase firmware" to erase stock
firmware and execute sysupgrade with sysupgrade image for WN-AX1167GR
6. Wait ~180 seconds to complete flasing
Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
* change mx6qsabresd to mx6qsabres to match defconfig name
* merge wanboard profiles since there is only one defconfig for the target device
* move wanboard options from wandboard.h to defconfig
* remove legacy patches
Signed-off-by: Vladimir Vid <vladimir.vid@sartura.hr>
Most of the implementations behind cfg80211_get_station will not initialize
sinfo to zero before manipulating it. For example, the member "filled",
which indicates the filled in parts of this struct, is often only modified
by enabling certain bits in the bitfield while keeping the remaining bits
in their original state. A caller without a preinitialized sinfo.filled can
then no longer decide which parts of sinfo were filled in by
cfg80211_get_station (or actually the underlying implementations).
cfg80211_get_station must therefore take care that sinfo is initialized to
zero. Otherwise, the caller may tries to read information which was not
filled in and which must therefore also be considered uninitialized. In
batadv_v_elp_get_throughput's case, an invalid "random" expected throughput
may be stored for this neighbor and thus the B.A.T.M.A.N V algorithm may
switch to non-optimal neighbors for certain destinations.
Signed-off-by: Sven Eckelmann <sven.eckelmann@openmesh.com>
Board Data File (BDF) is loaded upon driver boot-up procedure. The right
board data file is identified on QCA4019 using bus, bmi-chip-id and
bmi-board-id.
The problem, however, can occur when the (default) board data file cannot
fulfill the vendor requirements and it is necessary to use a different
board data file.
This problem was solved for SMBIOS by adding a special SMBIOS type 0xF8.
Something similar has to be provided for systems without SMBIOS but with
device trees. No solution was specified by QCA and therefore a new one has
to be found for ath10k.
The device tree requires addition strings to define the variant name
wifi@a000000 {
status = "okay";
qcom,ath10k-calibration-variant = "RT-AC58U";
};
wifi@a800000 {
status = "okay";
qcom,ath10k-calibration-variant = "RT-AC58U";
};
This would create the boarddata identifiers for the board-2.bin search
* bus=ahb,bmi-chip-id=0,bmi-board-id=16,variant=RT-AC58U
* bus=ahb,bmi-chip-id=0,bmi-board-id=17,variant=RT-AC58U
Signed-off-by: Sven Eckelmann <sven.eckelmann@open-mesh.com>
In order for monitoring tools such as atop and htop to track and report
i/o data, kernel support for task statistics and io accounting is
required.
Add a config option to enable building this support in the kernel.
Signed-off-by: Jeremiah McConnell <miah@miah.com>
With deterministic ECDSA the value k needed for the ECDSA signature is
not randomly generated any more, but generated from a hash over the
private key and the message to sign. If the value k used in a ECDSA
signature or the relationship between the two values k used in two
different ECDSA signatures over the same content is know to an attacker
he can derive the private key pretty easily. Using deterministic ECDSA
as defined in the RFC6979 removes this problem by deriving the value k
deterministically from the private key and the content which gets
signed.
The resulting signature is still compatible to signatures generated not
deterministic.
This increases the size of the ipk on mips 24Kc by about 2 KByte.
old:
166.240 libmbedtls_2.11.0-1_mips_24kc.ipk
new:
167.811 libmbedtls_2.11.0-1_mips_24kc.ipk
This does not change the ECDSA performance in a measurable way.
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Disable MBEDTLS_SHA256_SMALLER implementation, not enabled by default in
upstream and reduces performance by quite a bit.
Source: include/mbedtls/config.h
Enable an implementation of SHA-256 that has lower ROM footprint but also
lower performance.
The default implementation is meant to be a reasonnable compromise between
performance and size. This version optimizes more aggressively for size at
the expense of performance. Eg on Cortex-M4 it reduces the size of
mbedtls_sha256_process() from ~2KB to ~0.5KB for a performance hit of
about 30%.
The size of mbedtls increased a little bit:
ipkg for mips_24kc before:
164.382 Bytes
ipkg for mips_24kc after:
166.240 Bytes
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
Update mbed TLS to 2.11.0
Disable OFB block mode and XTS block cipher mode, added in 2.11.0.
The soVersion of mbedtls changed, bump PKG_RELEASE for packages that use mbedTLS
This is to avoid having a mismatch between packages when upgrading.
The size of mbedtls increased a little bit:
ipkg for mips_24kc before:
163.846 Bytes
ipkg for mips_24kc after:
164.382 Bytes
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
USB storage support is however SCSI Disk block device support isn't
meaning that connected devices wont enumerate.
Enable CONFIG_BLK_DEV_SD by default to fix it.
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
BusyBox's `tar` command does not support the `--directory` directive, which
is essentially `-C` in short-form option.
BusyBox's `tar` command supports `-C`.
Signed-off-by: Alexandru Ardelean <ardeleanalex@gmail.com>
These options are handled by generic configuration
Targets that need these options should select KERNEL_DEVMEM
and/or KERNEL_DEVKMEM options on OpenWRT's config
Signed-off-by: Luis Araneda <luaraneda@gmail.com>
The WDR4900v1 uses the P1040 SoC, so the device tree pulls in the
definition for the related P1010 SoC. However, the P1040 lacks the
CAAM/SEC4 hardware crypto accelerator which the P1010 device tree
defines. If left defined, this causes the CAAM drivers (if present) to
attempt to use the non-existent device, making various crypto-related
operations (e.g. macsec and ipsec) fail.
This commit overrides the incorrect dt node definition in the included
file.
See also:
- https://bugs.openwrt.org/index.php?do=details&task_id=1262
- https://community.nxp.com/thread/338432#comment-474107
Signed-off-by: Tim Small <tim@seoss.co.uk>
OpenSSL defaults X509_CERT_FILE to /etc/ssl/cert.pem. This change is
needed for wget-ssl and possibly others to work seamlessly with fresh
ca-bundle installation
Fixesopenwrt/packages#6152
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
Add two patches submitted for upstream review that significantly improve
the dwc2 driver on openwrt from kernel stability and performance
perspectives.
Fixes: FS#1367
Signed-off-by: Antti Seppälä <a.seppala@gmail.com>
Felix Fietkau
Koen Vandeputte
Moritz Warning
Jason A. Donenfeld
Hans Dedecker
Rafał Miłecki
Hauke Mehrtens
Kevin Darbyshire-Bryant
David Bauer
Ted Hess
Lucian Cristian
Christoph Krapp
Ademar Arvati Filho
Mathias Kresin
INAGAKI Hiroshi
Konstantin Demin
Vladimir Vid
Sven Eckelmann
Sven Eckelmann
Jeremiah McConnell
Daniel Engberg
Alexandru Ardelean
Enrico Mioso
Luis Araneda
Tim Small
Yousong Zhou
Antti Seppälä
Florian Eckert