The DTS_DIR variable is not a device variable, thus it should not
be set inside Device/Default but globally.
Fixes: c640370939 ("lantiq: use soc_vendor_device scheme on DTS file")
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The number 3 was accidentally removed from the name during split
of DEVICE_TITLE.
Fixes: fd66687058 ("lantiq: split up DEVICE_TITLE")
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
In ramips, all devices in mt7621, mt76x8 and rt288x subtarget have
the same value set to the SOC variable for each device individually.
This patch introduces a non-device-dependent variable DEFAULT_SOC,
which is used if no specific SOC is set for a device, and thus reduces
the number of redundant definitions drastically.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
* Makefile: remove pwd from compile output
* Makefile: add standard 'all' target
* Makefile: evaluate git version lazily
Quality of life improvements for packagers.
* ipc: simplify inflatable buffer and add fuzzer
* fuzz: add generic command argument fuzzer
* fuzz: add set and setconf fuzzers
More fuzzers and a slicker string list implementation. These fuzzers now find
themselves configuring wireguard interfaces from scratch after several million
mutations, which is fun to watch.
* netlink: make sure to clear return value when trying again
Prior, if a dump was interrupted by a concurrent set operation, we'd try
again, but forget to reset an error flag, so we'd keep trying again forever.
Now we do the right thing and succeed when we succeed.
* Makefile: sort inputs to linker so that build is reproducible
Earlier versions of make(1) passed GLOB_NOSORT to glob(3), resulting in the
linker receiving its inputs in a filesystem-dependent order. This screwed up
reproducible builds.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
* Makefile: strip prefixed v from version.h
This fixes a mistake in dmesg output and when parsing the sysfs entry in the
filesystem.
* device: skb_list_walk_safe moved upstream
This is a 5.6 change, which we won't support here, but it does make the code
cleaner, so we make this change to keep things in sync.
* curve25519: x86_64: replace with formally verified implementation
This comes from INRIA's HACL*/Vale. It implements the same algorithm and
implementation strategy as the code it replaces, only this code has been
formally verified, sans the base point multiplication, which uses code
similar to prior, only it uses the formally verified field arithmetic
alongside reproducable ladder generation steps. This doesn't have a
pure-bmi2 version, which means haswell no longer benefits, but the
increased (doubled) code complexity is not worth it for a single
generation of chips that's already old.
Performance-wise, this is around 1% slower on older microarchitectures,
and slightly faster on newer microarchitectures, mainly 10nm ones or
backports of 10nm to 14nm. This implementation is "everest" below:
Xeon E5-2680 v4 (Broadwell)
armfazh: 133340 cycles per call
everest: 133436 cycles per call
Xeon Gold 5120 (Sky Lake Server)
armfazh: 112636 cycles per call
everest: 113906 cycles per call
Core i5-6300U (Sky Lake Client)
armfazh: 116810 cycles per call
everest: 117916 cycles per call
Core i7-7600U (Kaby Lake)
armfazh: 119523 cycles per call
everest: 119040 cycles per call
Core i7-8750H (Coffee Lake)
armfazh: 113914 cycles per call
everest: 113650 cycles per call
Core i9-9880H (Coffee Lake Refresh)
armfazh: 112616 cycles per call
everest: 114082 cycles per call
Core i3-8121U (Cannon Lake)
armfazh: 113202 cycles per call
everest: 111382 cycles per call
Core i7-8265U (Whiskey Lake)
armfazh: 127307 cycles per call
everest: 127697 cycles per call
Core i7-8550U (Kaby Lake Refresh)
armfazh: 127522 cycles per call
everest: 127083 cycles per call
Xeon Platinum 8275CL (Cascade Lake)
armfazh: 114380 cycles per call
everest: 114656 cycles per call
Achieving these kind of results with formally verified code is quite
remarkable, especialy considering that performance is favorable for
newer chips.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Based on a script for comparison, this fixes (hopefully) all errors
in SUPPORTED_DEVICES for ar71xx->ath79 upgrade.
Devices where old string is removed as the device does not exist
in ar71xx:
- dlink_dir-859-a1
- tplink_archer-a7-v5
- tplink_cpe510-v3
Devices where string is changed because it did not match the board
name in ar71xx:
- tplink_tl-mr3220-v1
- tplink_tl-mr3420-v1
- tplink_tl-wr2543-v1
- tplink_tl-wr741nd-v4
- tplink_tl-wr841-v7
- ubnt_unifiac-mesh
- ubnt_unifiac-mesh-pro
- ubnt_unifiac-pro
For this device, the correct string could not be found, but we could
not determine the correct one. Thus, the string is removed for now:
- tplink_tl-wr740n-v4
The script for checking this is quite simple (note that newer
entries, i.e. ath79->ath79 upgrade, are displayed as missing):
newpath=target/linux/ath79/image/
oldpath=target/linux/ar71xx/base-files/lib/ar71xx.sh
for s in $(grep -roh "SUPPORTED_DEVICES.*" $newpath | sed 's/SUPPORTED_DEVICES *.= *//'); do
found="Missing"
grep -q -r "\"$s\"" $oldpath && found="Found"
echo "$s: $found."
done
The errors might be filtered by appending 'grep "Missing"' to the script.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
In certain cases, the uncompressed initramfs image will overwrite
the lzma-loader, which is currently only 10 MB away from kernel image
start. To prevent this, change LZMA_TEXT_START to 24 MB, so loader
and compressed image have 8 MB at the end of RAM and uncompressed
image has 24 MB available.
This is only enabled for ath79 at the moment, as there we can be sure
that all devices have 32+ MB RAM and TARGET_INITRAMFS_COMPRESSION_LZMA
is not enabled there.
Despite, since lzma-loader is currently build specifically for ath79
anyway, there is no need to re-specify LOADADDR and LZMA_TEXT_START
in image/Makefile, so the values are set directly in
image/lzma-loader/Makefile and the overwrite in image/Makefile is
removed.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Tested-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
This commit adds support for the MikroTik RouterBOARD wAP G-5HacT2HnD
(wAP AC), a small weatherproof dual band, dual-radio 802.11ac
wireless AP with integrated omnidirectional anntennae and one
10/100/1000 Mbps Ethernet port.
See https://mikrotik.com/product/RBwAPG-5HacT2HnD for more info.
Specifications:
- SoC: Qualcomm Atheros QCA9556
- RAM: 64 MB
- Storage: 16 MB NOR
- Wireless:
· Atheros AR9550 (SoC) 802.11b/g/n 2x2:2, 2 dBi antennae
· Qualcomm QCA9880 802.11a/n/ac 3x3:3, 2 dBi antennae
- Ethernet: Atheros AG71xx (SoC, AR8033), 1x 1000/100/10 port,
passive PoE in
Working:
- Board/system detection
- Sysupgrade
- Serial console
- Ethernet
- 2.4 GHz radio
- 5 GHz radio and LED
- Reset button
Not working/Unsupported:
- 2.4 GHz LED
- AP/CAP LED
- ZT2046Q SPI temperature and voltage sensor
This adds the basic features for supporting MikroTik devices:
- a common recipe for mikrotik images in common-mikrotik.mk
- support for minor (MikroTik NOR) split firmware (only for
generic subtarget so far)
Acknowledgments: Robert Marko <robimarko@gmail.com>
Andrew Cameron <apcameron@softhome.net>
Koen Vandeputte <koen.vandeputte@ncentric.com>
Chuanhong Guo <gch981213@gmail.com>
Signed-off-by: Roger Pueyo Centelles <roger.pueyo@guifi.net>
Co-developed-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Tested-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
In order to make the QCA955x SGMII workaround work, the unsuccessful
SGMII autonegotiation on the AR8033 should not block the PHY
state-machine.
Otherwise, the ag71xx driver never becomes aware of the copper-side
link-establishment and the workaround is never executed.
Signed-off-by: David Bauer <mail@david-bauer.net>
[remove one trailing whitespace per file]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This commit adds a workaround for the loss of the SGMII link observed on
the QCA955x generation of SoCs. The workaround originates part from the
U-Boot source code, part from the implementation from AVM found in the
GPL tarball for the AVM FRITZ!WLAN Repeater 450E.
The bug results in a stuck SGMII link between the PHY device and the SoC
side. This has only been observed with the Atheros AR8033 PHY and most
likely all devices using such combination are affected.
It is worked around by reading a hidden SGMII status register and
issuing a SGMII PHY reset until the link becomes useable again.
Signed-off-by: David Bauer <mail@david-bauer.net>
Older ath79-based MikroTik devices have the ERD calibration data
compressed and stored different to newer IPQ40xx ones. This commit
adds support for these former ones.
Signed-off-by: Roger Pueyo Centelles <roger.pueyo@guifi.net>
Acked-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
This utility extracts the radio calibration data, as well as other
board-related information (model, serial number, etc.), from MikroTik
Routerboard devices' flash.
Signed-off-by: Roger Pueyo Centelles <roger.pueyo@guifi.net>
Acked-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
hostapd sets minimum values for CWmin/CWmax/AIFS and maximum for TXOP.
The code for applying those values had a few bugs leading to bogus values,
which caused significant latency and packet loss.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Due to the history of the target, all devices added before a certain
point have the same device string in BOARD_NAME, DEVICE_DTS and added
to SUPPORTED_DEVICES.
Thus, we can set this one automatically for all devices where
BOARD_NAME is specified, removing the explicit DEVICE_DTS and
SUPPORTED_DEVICES addition there.
For new devices, nothing has changed, and just DEVICE_DTS has to
be set manually.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This tidies up the image Makefiles for the sunxi target by:
- Move the if-condition for the subtarget to the parent Makefile
- Remove lots of unnecessary empty lines
- Sort device definitions alphabetically
- Harmonize line wrapping for DEVICE_PACKAGES
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Add ubootenv uci config for gl-ar150, gl-domino and gl-mifi
Signed-off-by: Kimmo Vuorinen <kimmo.vuorinen@gmail.com>
[commit message/title facelift]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Change u-boot-env partitions to be mounted as read-write for gl-ar150,
gl-domino and gl-mifi so uboot-envtools support is possible.
Signed-off-by: Kimmo Vuorinen <kimmo.vuorinen@gmail.com>
[commit title/message facelift]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
TP-Link TL-WA850RE v2 is a wall-plug N300 Wi-Fi range extender,
based on Qualcomm/Atheros QCA9533 v2.
Short specification:
- 550/391/195 MHz (CPU/DDR/AHB)
- 1x 10/100 Mbps Ethernet
- 32 MB of RAM (DDR1)
- 4 MB of FLASH
- 2T2R 2.4 GHz
- 2x internal antennas (embedded on PCB)
- 9x LED (all can be turned off with GPIO15), 2x button
- UART (J3) header on PCB
Flash instruction: use "factory" image directly in vendor GUI.
Warning: this device does not include any kind of recovery mechanism
in the bootloader and disassembling process is not trivial.
You can access vendor firmware over serial line using:
- login: root
- password: sohoadmin
Stock firmware uses label MAC address for WiFi and same with local
bit set for ethernet. Since this is difficult to reproduce with
the toolset of OpenWrt, we just keep both ethernet and WiFi to
the same address here.
This is the first tiny device with tplink-safeloader in ath79.
Firmware partition is only 3648k and thus even smaller than for
the tplink-4m(lzma) devices.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This moves the TP-Link TL-WA850RE v2 to dynamic partitioning and
will allow to use this for ath79 as well.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
- Remove the "RGMII TX delay fixup" hack and the associated
DT-property. It was never used in a DT-based platform and
solved a problem which can be mitigated by using correct
delays on the MAC side.
- Remove the patch to enable platform-data support for the
at803x driver. It was only used by ar71xx which does not
(and never will) support kernel 4.19 or later.
- Remove the SmartEEE DT-configuration patch. As explained
previously, this patch never disabled the Atheros SmartEEE
implementation, but rather "standard" EEE. This can be done
on device-tree compatible platforms by adding the
"eee-broken-1000t" or "eee-broken-100tx" properties to the PHY
node. As all usages of the old properties are migrated, this
patch can be removed.
Signed-off-by: David Bauer <mail@david-bauer.net>
Acked-by: Christian Lamparter <chunkeey@gmail.com>
The hack-patch which introduced the "at803x-disable-smarteee" for
disabling SmartEEE did in fact not disable SmartEEE but rather disabled
802.3az "standard" EEE.
This can be done by using the upstream properties "eee-broken-100tx" and
"eee-broken-1000t". EEE is then disabled by the PHY subsystem.
Tested on devolo WiFi pro 1200e.
Signed-off-by: David Bauer <mail@david-bauer.net>
This ports support for the TL-WA850RE v1 range extender from ar71xx
to ath79.
Specifications:
Board: AP123 / AR9341 rev. 3
Flash/RAM: 4/32 MiB
CPU: 535 MHz
WiFi: 2.4 GHz b/g/n
Ethernet: 1 port (100M)
Flashing instructions:
Upload the factory image via the vendor firmware upgrade option.
Recovery:
Note that this device does not provide TFTP via ethernet like many
other TP-Link devices do. You will have to open the case if you
require recovery beyond failsafe.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
hostapd allows putting WDS (4addr mode) clients into a separate bridge
other than the bridge regular (3addr mode) clients end up in. This is
useful for example giving WDS clients access to several VLANs
(trunking) while regular clients will end up inside a specific VLAN.
Add 'wds_bridge' config parameter for wifi-iface which contains the
name of the bridge. hostapd-mini already supports this feature, so all
needed is to add the UCI wrapping in mac80211.sh.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Add patches retrying to probe the PHY to restore support for PHYs taking
longer to initialize without breaking modules without PHYs.
Patches taken from http://git.armlinux.org.uk/cgit/linux-arm.git/log/?h=phy
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
Update the uDPU kernel support, fixing a number of issues:
* make ethernet work again by adding comphy definitions
* slow the I2C bus to give it more chance of actually working
* the SFP cages are designed to support up to 3W modules, which
would be prevented from initialising without this patch.
Signed-off-by: Russell King <linux@armlinux.org.uk>
[jonas.gorski: rename to mvebu, refresh patches]
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
Add support for Methode DM7052 NBASE-T module to OpenWRT. These
patches are taken from my "phy" branch, and will be sent for the
next kernel merge window.
Signed-off-by: Russell King <linux@armlinux.org.uk>
[jonas.gorski: move patches to pending, refresh patches]
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
Backport the phy/phylink/sfp patches currently queued in netdev or in
mainline necessary to support GPON popular modules, specifically to
support Huawei and Nokia GPON modules.
Signed-off-by: Russell King <linux@armlinux.org.uk>
[jonas.gorski: include kernel version in file names, refresh patches]
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
Move two phylink patches from mvebu to generic, so that everyone can
benefit from them.
Signed-off-by: Russell King <linux@armlinux.org.uk>
[jonas.gorski: add kernel version to file names]
Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>
Remove the old phylink/SFP patches from the OpenWRT build; these will
be updated with a new set in subsequent.
450-reprobe_sfp_phy is also removed for several reasons:
1) it is not in mainline.
2) it breaks copper modules that do not have a PHY.
3) it makes backporting the current patch set harder.
Discussion is ongoing with the patch author for a mainline Linux kernel
patch for this.
Signed-off-by: Russell King <linux@armlinux.org.uk>
Update bison to 3.5
Release notes:
https://lists.gnu.org/archive/html/info-gnu/2019-12/msg00002.html
Note for future: release notes mention that YYPRINT macro is
declared deprecated, but apparently still works for now. I found
one possible use of that in scripts/config/zconf.tab.c_shipped
That might be modernized at some point, but as the file is synced
with the one from upstream Linux, it might get fixed there.
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
TPLINK_BOARD_NAME has been renamed to TPLINK_BOARD_ID a long time
ago (7d6c63d87542: "build: rename TPLINK_BOARD_NAME to
TPLINK_BOARD_ID" for ar71xx), and before introducing ath79 target
at all.
TPLINK_BOARD_NAME seems to have been introduced into ath79 target
only by mistake. It has never been used. Remove it.
Fixes: 53c474abbd ("ath79: add new OF only target for QCA MIPS silicon")
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This reverts commit de80424f70 which needs
more work and testing as it broke at least jffs2 overlays at least on
ath79 platform, marking them as read-only, thus unusable:
jffs2_build_filesystem(): erasing all blocks after the end marker...
jffs2: Erase at 0x009e0000 failed immediately: -EROFS. Is the sector locked?
Ref: http://lists.infradead.org/pipermail/openwrt-devel/2020-January/021344.html
Reported-by: Steve Brown <sbrown@ewol.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
In zynq target, the DEVICE_DTS variable is always set consistent
with the model part of the device definition name.
This patch replaces the redundant definitions for the individual
devices with a common recipe.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
For the mvebu target in particular, there is a lot of files in
base-files that are only relevant for one subtarget. Improve
overview and reduce size per subtarget by moving/splitting
base-files depending on the subtarget they belong to.
While at it, consolidate 01_leds by using the model part of
the board name as variable.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Acked-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
This introduces the SOC variable to mvebu target to derive some of
the DEVICE_DTS variables based on the SOC prefix and the device
definition name.
Since DTS names and compatible are inconsistent also in the kernel
for this target, the scheme cannot be applied to all devices, though.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Acked-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
This moves the if conditions for choosing which image Makefiles
are used to the parent image/Makefile. It seems more convenient
to have "codeflow" in the parent while the subtarget-specific
files only contain the definitions.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Acked-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
Adrian Schmutzler
Kevin Darbyshire-Bryant
Koen Vandeputte
Jason A. Donenfeld
DENG Qingfang
Petr Štetiar
Felix Fietkau
Roger Pueyo Centelles
David Bauer
Kimmo Vuorinen
Daniel Golle
Jonas Gorski
Russell King
Jan Pavlinec
Hannu Nyman