openwrt/staging/jow.git
4 years agoath79: harmonize model names for Ubiquiti devices
Adrian Schmutzler [Wed, 15 Jul 2020 16:41:57 +0000 (18:41 +0200)]
ath79: harmonize model names for Ubiquiti devices

This harmonizes the model names for the ath79 Ubiquiti devices by
applying a few minor cosmetic adjustments:

- Removes hyphens where they are not found in the product names
  (Ubiquiti uses hyphens only for the abbreviated version names
  like UAP-AC-PRO which we don't use anyway.)

- Add (XM) suffix for DTS model strings to help with distinguishing
  them from their XW counterparts.

- Remove DEVICE_VARIANT for LAP-120 which actually was an alternate
  device name.

- Generally make DTS model names and those from generic-ubnt.mk
  more consistent.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agobase-files: replace backticks in lib/upgrade/nand.sh
Adrian Schmutzler [Sun, 10 May 2020 14:20:55 +0000 (16:20 +0200)]
base-files: replace backticks in lib/upgrade/nand.sh

This replaces deprecated backticks by more versatile $(...) syntax.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agopackage: drop PKG_VERSION for purely local packages
Adrian Schmutzler [Mon, 24 Feb 2020 15:57:38 +0000 (16:57 +0100)]
package: drop PKG_VERSION for purely local packages

In the package guidelines, PKG_VERSION is supposed to be used as
"The upstream version number that we're downloading", while
PKG_RELEASE is referred to as "The version of this package Makefile".
Thus, the variables in a strict interpretation provide a clear
distinction between "their" (upstream) version in PKG_VERSION and
"our" (local OpenWrt trunk) version in PKG_RELEASE.

For local (OpenWrt-only) packages, this implies that those will only
need PKG_RELEASE defined, while PKG_VERSION does not apply following
a strict interpretation. While the majority of "our" packages actually
follow that scheme, there are also some that mix both variables or
have one of them defined but keep them at "1".

This is misleading and confusing, which can be observed by the fact
that there typically either one of the variables is never bumped or
the choice of the variable to increase depends on the person doing the
change.

Consequently, this patch aims at clarifying the situation by
consistently using only PKG_RELEASE for "our" packages. To achieve
that, PKG_VERSION is removed there, bumping PKG_RELEASE where
necessary to ensure the resulting package version string is bigger
than before.

During adjustment, one has to make sure that the new resulting composite
package version will not be considered "older" than the previous one.

A useful tool for evaluating that is 'opkg compare-versions'. In
principle, there are the following cases:

1. Sole PKG_VERSION replaced by sole PKG_RELEASE:
   In this case, the resulting version string does not change, it's
   just the value of the variable put in the file. Consequently, we
   do not bump the number in these cases so nobody is tempted to
   install the same package again.

2. PKG_VERSION and PKG_RELEASE replaced by sole PKG_RELEASE:
   In this case, the resulting version string has been "version-release",
   e.g. 1-3 or 1.0-3. For this case, the new PKG_RELEASE will just
   need to be higher than the previous PKG_VERSION.
   For the cases where PKG_VERSION has always sticked to "1", and
   PKG_RELEASE has been incremented, we take the most recent value of
   PKG_RELEASE.

Apart from that, a few packages appear to have developed their own
complex versioning scheme, e.g. using x.y.z number for PKG_VERSION
_and_ a PKG_RELEASE (qos-scripts) or using dates for PKG_VERSION
(adb-enablemodem, wwan). I didn't touch these few in this patch.

Cc: Hans Dedecker <dedeckeh@gmail.com>
Cc: Felix Fietkau <nbd@nbd.name>
Cc: Andre Valentin <avalentin@marcant.net>
Cc: Matthias Schiffer <mschiffer@universe-factory.net>
Cc: Jo-Philipp Wich <jo@mein.io>
Cc: Steven Barth <steven@midlink.org>
Cc: Daniel Golle <dgolle@allnet.de>
Cc: John Crispin <john@phrozen.org>
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agovxlan: bump and change to PKG_RELEASE
Adrian Schmutzler [Wed, 15 Jul 2020 14:49:36 +0000 (16:49 +0200)]
vxlan: bump and change to PKG_RELEASE

Bumping package version has been overlooked in a previous commit.

While at it, use PKG_RELEASE instead of PKG_VERSION, as the latter
is meant for upstream version number only.
(The effective version string for the package would be "3" in both
cases, so there is no harm done for version comparison.)

Fixes: 0453c3866feb ("vxlan: fix udp checksum control")
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agoath79: add support for Ubiquiti PowerBridge M
Vieno Hakkerinen [Fri, 10 Jul 2020 21:23:01 +0000 (23:23 +0200)]
ath79: add support for Ubiquiti PowerBridge M

This adds support for the Ubiquiti PowerBridge M, which has the same
board/LEDs as the Bullet M XM, but different case and antennas.

Specifications:
- AR7241 SoC @ 400 MHz
- 64 MB RAM
- 8 MB SPI flash
- 1x 10/100 Mbps Ethernet, 24 Vdc PoE-in
- Internal antenna: 25 dBi
- POWER/LAN green LEDs
- 4x RSSI LEDs (red, orange, green, green)
- UART (115200 8N1) on PCB

Flashing via WebUI:
Upload the factory image via the stock firmware web UI.

Attention: airOS firmware versions >= 5.6 have a new bootloader with
an incompatible partition table!

Please downgrade to <= 5.5 _before_ flashing OpenWrt!
Refer to the device's Wiki page for further information.

Flashing via TFTP:
Same procedure as other Bullet M (XM) boards.

- Use a pointy tool (e.g., pen cap, paper clip) and keep the reset
button on the device or on the PoE supply pressed
- Power on the device via PoE (keep reset button pressed)
- Keep pressing until LEDs flash alternatively LED1+LED3 =>
LED2+LED4 => LED1+LED3, etc.
- Release reset button
- The device starts a TFTP server at 192.168.1.20
- Set a static IP on the computer (e.g., 192.168.1.21/24)
- Upload via tftp the factory image:
$ tftp 192.168.1.20
  tftp> bin
  tftp> trace
  tftp> put openwrt-ath79-generic-xxxxx-ubnt_powerbridge-m-squashfs-factory.bin

Signed-off-by: Vieno Hakkerinen <vieno@hakkerinen.eu>
4 years agoravpower-mcu: bump PKG_RELEASE
David Bauer [Wed, 15 Jul 2020 07:56:05 +0000 (09:56 +0200)]
ravpower-mcu: bump PKG_RELEASE

Fixes commit b94d1a2d2786 ("ravpower-mcu: remove target dependency")

Signed-off-by: David Bauer <mail@david-bauer.net>
4 years agovxlan: fix udp checksum control
Johannes Kimmel [Mon, 8 Jun 2020 14:14:43 +0000 (16:14 +0200)]
vxlan: fix udp checksum control

So far, passing "rxcsum" and "txcsum" had no effect.

Fixes: 95ab18e0124e ("vxlan: add options to enable and disable UDP
checksums")

Signed-off-by: Johannes Kimmel <fff@bareminimum.eu>
[add Fixes:]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agoravpower-mcu: remove target dependency
David Bauer [Tue, 14 Jul 2020 22:01:36 +0000 (00:01 +0200)]
ravpower-mcu: remove target dependency

Remove the target dependency as the tool is also usable with other
battery-powered routers.

Signed-off-by: David Bauer <mail@david-bauer.net>
4 years agoiproute2: tc: fix missing em_ipset module
Tony Ambardar [Tue, 30 Jun 2020 05:07:19 +0000 (22:07 -0700)]
iproute2: tc: fix missing em_ipset module

Feature detection doesn't recognize ipset v7 use on kernel v5.x systems
and thus disables the tc ematch function em_ipset.

- backport patch:
  * 002-configure-support-ipset-v7.patch:
    650591a7a70c configure: support ipset version 7 with kernel version 5

Fixes: 4e0c54bc5bc8 ("kernel: add support for kernel 5.4")
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
4 years agoiproute2: tc: fix dynamic symbol table size optimization
Tony Ambardar [Tue, 30 Jun 2020 05:04:54 +0000 (22:04 -0700)]
iproute2: tc: fix dynamic symbol table size optimization

Recent iproute2 5.x versions modified the symbols resolved for plugins,
causing "tc .. action xt .." to fail. Update the list of symbols to fix.

Fixes: b61495409bb3 ("iproute2: tc: reduce size of dynamic symbol table")
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
4 years agoipq40xx: fix ethernet vlan double tagging
John Crispin [Tue, 14 Jul 2020 08:07:32 +0000 (10:07 +0200)]
ipq40xx: fix ethernet vlan double tagging

As the the SoC uses implicit vlan tagging for dual MAC support, the
offload feature breaks when using double tagging.

Signed-off-by: Sven Eckelmann <sven@narfation.org>
Signed-off-by: John Crispin <john@phrozen.org>
4 years agoipq40xx: merge all ar40xx patches into one single patch
John Crispin [Tue, 14 Jul 2020 07:50:31 +0000 (09:50 +0200)]
ipq40xx: merge all ar40xx patches into one single patch

Signed-off-by: John Crispin <john@phrozen.org>
4 years agoipq806x: really fix aliases node for Qualcomm IPQ8064/DB149
Adrian Schmutzler [Tue, 14 Jul 2020 14:59:44 +0000 (16:59 +0200)]
ipq806x: really fix aliases node for Qualcomm IPQ8064/DB149

Having looked at this again, it appears that only gsbi2_serial
is actually enabled for this device, so the entry in the broken
aliases node was correct.

Therefore, this needs to set its own serial0 instead of inheriting
"serial0 = &gsbi4_serial;" from DTSI. Do this with the correctly
named aliases node now.

Fixes: c83f7b6d217f ("ipq806x: fix aliases node name for Qualcomm
IPQ8064/DB149")

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agoipq806x: use consistent DT labels for gsbiX_serial
Adrian Schmutzler [Tue, 14 Jul 2020 10:23:53 +0000 (12:23 +0200)]
ipq806x: use consistent DT labels for gsbiX_serial

Kernel uses the label gsbiX_serial, so let's adjust our labels to
this naming scheme.

This is cosmetic, and actually only already existing gsbi4_serial
has been used at all.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agoipq806x: fix aliases node name for Qualcomm IPQ8064/DB149
Adrian Schmutzler [Fri, 3 Jul 2020 13:50:42 +0000 (15:50 +0200)]
ipq806x: fix aliases node name for Qualcomm IPQ8064/DB149

This device uses a node incorrectly named "alias" instead of
"aliases" since it was introduced.

Remove it without replacement, as the definitions in it don't
seem to be required anyway:

The serial0 definition has never been effective anyway and this
would be the only device deviating from the common setting
"serial0 = &gsbi4_serial;" for ipq8064. (So, maybe the wrong
node prevented us from finding out about the wrong serial
definition?)

The mdio-gpio0 alias was supposed to be removed in d2a2eb7e48f6
anyway, the redundant definition in the alias node was just
overlooked back then.

Fixes: 0fd202f3e563 ("ipq806x: add db149 dts files")
Fixes: d2a2eb7e48f6 ("ipq806x: replace caf nss-gmac driver by upstream stmmac")
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agoprocd: allow optional watchdog instance parameter
Daniel Bailey [Mon, 13 Jul 2020 22:21:50 +0000 (15:21 -0700)]
procd: allow optional watchdog instance parameter

Optional instance watchdog timeout and watchdog mode can be set by
adding: procd_set_param $mode $timeout

$mode is an integer [0-1] representing instance watchdog mode of
operation:
0 = disabled
1 = passive mode, client must periodically poke watchdog via ubus

$timeout is an integer representing how often, in seconds, the watchdog must be poked.

Signed-off-by: Daniel Bailey <danielb@meshplusplus.com>
4 years agoprocd: update to git HEAD
Daniel Golle [Mon, 13 Jul 2020 23:23:11 +0000 (00:23 +0100)]
procd: update to git HEAD

 639df57 uxc: fix build with uClibc-ng
 b2230e4 procd: add service instance watchdog

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
4 years agoath79: use rgmii-id instead of rgmii in ubnt,lap-120
Roger Pueyo Centelles [Tue, 10 Mar 2020 18:31:20 +0000 (19:31 +0100)]
ath79: use rgmii-id instead of rgmii in ubnt,lap-120

Since commit 6f2e1b7485 (ath79: disable delays on AT803X config init)
the incoming incoming traffic on the ubnt,lap-120 devices Ethernet
port was not making it through. Using rgmii-id instead of rgmii (same
configuration as ubnt,litebeam-ac-gen2) fixes it.

Fixes FS#2893.

Signed-off-by: Roger Pueyo Centelles <roger.pueyo@guifi.net>
4 years agobuild,json: fix compatibility with Python 3.5
Paul Spooren [Sun, 12 Jul 2020 04:44:55 +0000 (18:44 -1000)]
build,json: fix compatibility with Python 3.5

The f-string feature was introduced in Python 3.6. As Buildbots may run
on Debian 9, which comes per default with Python 3.5, this would cause
an issue. Instead of f-strings use the *legacy* `.format()` function.

Signed-off-by: Paul Spooren <mail@aparcar.org>
4 years agoramips: add support for D-Link DIR-878 A1
Mathieu Martin-Borret [Fri, 10 Jul 2020 03:49:20 +0000 (13:49 +1000)]
ramips: add support for D-Link DIR-878 A1

Specifications:

    SoC: MT7621AT
    RAM: 128MB
    Flash: 16MB NOR SPI flash
    WiFi: MT7615N (2.4GHz) and MT7615N (5Ghz)
    LAN: 5x1000M
    Firmware layout is Uboot with extra 96 bytes in header
    Base PCB is AP-MTKH7-0002
    LEDs Power Green,Power Orange,Internet Green,Internet Orange
    LEDs "2.4G" Green & "5G" Green connected directly to wifi module
    Buttons Reset,WPS,WIFI

Flashing instructions:

Upload image via emergency recovery mode
Push and hold reset button (on the back of the device) until power led
starts flashing (about 10 secs or so) while powering the device on.
Give it ~30 seconds, to boot the recovery mode GUI
Connect your client computer to LAN1 of the device
Set your client IP address manually to 192.168.0.2 / 255.255.255.0.
Call the recovery page for the device at http://192.168.0.1
Use the provided emergency web GUI to upload and flash a new firmware to
the device. Some browsers/OS combinations are known not to work, so if
you don't see the percentage complete displayed and moving within a few
seconds, restart the procedure from scratch and try anoher one,
or try the command line way.

Alternative method using command line on Linux:
curl -v -i -F "firmware=@openwrt-xxxx-squashfs-factory.bin" 192.168.0.1

Signed-off-by: Mathieu Martin-Borret <mathieu.mb@protonmail.com>
[use of generic uimage-padhdr in image generation code]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agokernel: mtdsplit_uimage: add SGE parser
Mathieu Martin-Borret [Fri, 10 Jul 2020 03:40:58 +0000 (13:40 +1000)]
kernel: mtdsplit_uimage: add SGE parser

Adding SGE parser for SGE (T&W) Shenzhen Gongjin Electronics
to add 96 bytes padding the the firmware image

Signed-off-by: Mathieu Martin-Borret <mathieu.mb@protonmail.com>
4 years agoglibc: fix avx2 strncmp offset compare condition check [BZ #25933]
Hans Dedecker [Mon, 13 Jul 2020 20:36:47 +0000 (22:36 +0200)]
glibc: fix avx2 strncmp offset compare condition check [BZ #25933]

4e8a33a959 NEWS: Mention BZ 25933 fix
fd15ba932d Fix avx2 strncmp offset compare condition check [BZ #25933

Signed-off-by: Hans Dedecker <dedeckeh@gmail.com>
4 years agorockchip: don't disable timer LED trigger
David Bauer [Mon, 13 Jul 2020 15:44:22 +0000 (17:44 +0200)]
rockchip: don't disable timer LED trigger

The timer LED trigger is enabled in all targets (except for lantiq
xway-legacy). It's necessary for the OpenWrt preinit LED pattern to
work.

Signed-off-by: David Bauer <mail@david-bauer.net>
4 years agorockchip: use downstream GPIO button implementation
David Bauer [Mon, 13 Jul 2020 15:44:07 +0000 (17:44 +0200)]
rockchip: use downstream GPIO button implementation

Use the OpenWrt netlink GPIO button implementation to forward button
presses to procd. This is necessary to make failsafe-mode access
using a button possible.

Signed-off-by: David Bauer <mail@david-bauer.net>
4 years agoprocd: update to git HEAD
Daniel Golle [Mon, 13 Jul 2020 11:19:08 +0000 (12:19 +0100)]
procd: update to git HEAD

 aed7fb3 procd: fix compilation with uClibc-ng
 9d0f831 jail: fix segfault with len(uidmap/gidmap) > 1
 42a6217 jail: consider PATH for argv in OCI container
 83f4b72 jail: actually chdir into OCI defined CWD
 fc9f614 jail: parse and run OCI hooks
 02eec92 jail: memory allocation fixes
 71e75f4 jail: refactor mount support to cover OCI spec
 b586e7d jail: don't make mount source read-only
 dacab12 uxc: fix 'stop' command

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
4 years agobcm47xx: disable image for Linksys WRT160n v3
Adrian Schmutzler [Wed, 8 Jul 2020 12:32:31 +0000 (14:32 +0200)]
bcm47xx: disable image for Linksys WRT160n v3

The device can only hold 3.6 MB, but newer images (since 18.06)
are bigger, so flashing a new version fails.

This disables default build for this device based on the bug report
referenced below.

Fixes: FS#1963
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agoramips: consolidate Netgear devices for mt7628
Adrian Schmutzler [Fri, 10 Jul 2020 10:00:08 +0000 (12:00 +0200)]
ramips: consolidate Netgear devices for mt7628

This creates a common DTSI and shared image definition for the
relatively similar Netgear devices for mt7628 platform.

As a side effect, this raises SPI flash frequency for the R6120,
as it's expected to work there as well if it works for R6080 and
R6020.

Based on the data from the other devices, it also seems probable
the 5g MAC address for R6120 could be extracted from the caldata,
and the mtd-mac-address there could be dropped.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agowireguard: bump to 1.0.20200712
Jason A. Donenfeld [Mon, 13 Jul 2020 03:37:11 +0000 (21:37 -0600)]
wireguard: bump to 1.0.20200712

This release brings parity with the commits Linus released a few hours
ago into 5.8-rc5.

* receive: account for napi_gro_receive never returning GRO_DROP

The napi_gro_receive function no longer returns GRO_DROP ever, making
handling GRO_DROP dead code. This commit removes that dead code.
Further, it's not even clear that device drivers have any business in
taking action after passing off received packets; that's arguably out of
their hands.

* device: implement header_ops->parse_protocol for AF_PACKET

WireGuard uses skb->protocol to determine packet type, and bails out if
it's not set or set to something it's not expecting. For AF_PACKET
injection, we need to support its call chain of:

    packet_sendmsg -> packet_snd -> packet_parse_headers ->
      dev_parse_header_protocol -> parse_protocol

Without a valid parse_protocol, this returns zero, and wireguard then
rejects the skb. So, this wires up the ip_tunnel handler for layer 3
packets for that case.

* queueing: make use of ip_tunnel_parse_protocol

Now that wg_examine_packet_protocol has been added for general
consumption as ip_tunnel_parse_protocol, it's possible to remove
wg_examine_packet_protocol and simply use the new
ip_tunnel_parse_protocol function directly.

* compat: backport ip_tunnel_parse_protocol and ip_tunnel_header_ops

These are required for moving wg_examine_packet_protocol out of
wireguard and into upstream.

Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
4 years agoscripts/env: Fix 56f813674a scripts/env: use command -v instead of which
Kevin Darbyshire-Bryant [Sun, 12 Jul 2020 12:25:39 +0000 (13:25 +0100)]
scripts/env: Fix 56f813674a scripts/env: use command -v instead of which

We don't need to see how git will be executed and it produces non silent
output on 'scripts/env diff' commands when there are no differences
unlike before.

Re-introduce original silent behaviour.

Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
4 years agoath79: fix dsa binding for TP-Link TL-WR941ND v2
Chuanhong Guo [Sun, 12 Jul 2020 08:14:45 +0000 (16:14 +0800)]
ath79: fix dsa binding for TP-Link TL-WR941ND v2

upstream changed dt-bindings for marvell 88e6060 to use mdio-device
and dropped support for legacy bindings.
fix it in our local dts.

Fixes: FS#2524
Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
4 years agoath79: qihoo,c301: use phy-supply for usb vbus
Chuanhong Guo [Sun, 12 Jul 2020 07:39:18 +0000 (15:39 +0800)]
ath79: qihoo,c301: use phy-supply for usb vbus

Linux phy subsystem provides support for a phy regulator defined via
phy-supply property. Use it to turn on usb power only when usb is
probed.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
4 years agomvebu: add support for MACCHIATObin Single Shot
Tomasz Maciej Nowak [Fri, 10 Jul 2020 17:35:22 +0000 (19:35 +0200)]
mvebu: add support for MACCHIATObin Single Shot

Add support for Marvell MACCHIATObin Single Shot, cortex-a72 based
Marvell ARMADA 8040 Community board. Single Shot was broken as the
device tree is different on the Double Shot Board.

Specifications:
- Quad core Cortex-A72 (up to 2GHz)
- DDR4 DIMM slot with optional ECC and single/dual chip select support
- Dual 10GbE (1/2.5/10GbE) SFP+
  2.5GbE (1/2.5GbE) via SFP
  1GbE via copper
- SPI Flash
- 3 X SATA 3.0 connectors
- MicroSD connector
- eMMC
- PCI x4 3.0 slot
- USB 2.0 Headers (Internal)
- USB 3.0 connector
- Console port (UART) over microUSB connector
- 20-pin Connector for CPU JTAG debugger
- 2 X UART Headers
- 12V input via DC Jack
- ATX type power connector
- Form Factor: Mini-ITX (170 mm x 170 mm)

More details at http://macchiatobin.net

Installation:

Write the Image to your Micro SD Card and insert it in the
MACCHIATObin Single Shot SD Card Slot.

In the U-Boot Environment:
   1. reset U-Boot environment:
      env default -a
      saveenv

   2. prepare U-Boot with boot script:
      setenv bootcmd "load mmc 1:1 0x4d00000 boot.scr; source 0x4d00000"
      saveenv

   or manually (hanging lines indicate wrapped one-line command):
      setenv fdt_name armada-8040-mcbin-singleshot.dtb
      setenv image_name Image
      setenv bootcmd 'mmc dev 1; ext4load mmc 1:1 $kernel_addr
         $image_name;ext4load mmc 1:1 $fdt_addr $fdt_name;setenv
         bootargs $console root=/dev/mmcblk1p2 rw rootwait; booti
         $kernel_addr - $fdt_addr'
      saveenv

   On newer Bootloaders (18.12) the Variables have been changed, use:
      setenv fdt_name armada-8040-mcbin-singleshot.dtb
      setenv image_name Image
      setenv bootcmd 'mmc dev 1; ext4load mmc 1:1 $kernel_addr_r
         $image_name;ext4load mmc 1:1 $fdt_addr_r $fdt_name;setenv
         bootargs $console root=/dev/mmcblk1p2 rw rootwait; booti
         $kernel_addr_r - $fdt_addr_r'

Reported-by: Alexandra Alth <alexandra@alth.de>
Signed-off-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
Tested-by: Alexandra Alth <alexandra@alth.de>
[add specs and installation as provided by Alexandra Alth]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agoscripts/mkits.sh: fix use of printf
Adrian Schmutzler [Sat, 11 Jul 2020 16:46:53 +0000 (18:46 +0200)]
scripts/mkits.sh: fix use of printf

Due to a line break, printf was accidentally called with three
arguments instead of two, causing a different output than before.

Fix it by splitting the printf command into two lines.

Fixes: 907053193a1c ("scripts/mkits.sh: replace echo -e with printf")
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agoramips: consolidate recipes with uimage_padhdr
Adrian Schmutzler [Fri, 10 Jul 2020 12:16:23 +0000 (14:16 +0200)]
ramips: consolidate recipes with uimage_padhdr

There are already two very similar recipes using uimage_padhdr
in ramips target, and a third one is about to be added.

Make the recipe more generic, so redefinitions are not necessary
anymore.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Tested-by: Bjørn Mork <bjorn@mork.no> [Zyxel WAP6805]
4 years agomvebu: fix support for Marvell 8040 MACCHIATOBin
Adrian Schmutzler [Fri, 10 Jul 2020 08:28:13 +0000 (10:28 +0200)]
mvebu: fix support for Marvell 8040 MACCHIATOBin

Between kernels 4.20 and 5.0, a new variant of this board has been
introduced ("Single Shot"), and the existing one has been renamed
with the appendix "Double Shot". [1]
This also adjusted the first compatible in the list:

marvell,armada8040-mcbin -> marvell,armada8040-mcbin-doubleshot

This patch updates the OpenWrt implementation of this device by
adjusting the relevant references to that compatible (i.e., our
board name).

To still provide support for 4.19 with our setup, this adds a
small patch to change the compatible there as well.

[1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=b1f0bbe2700051886b954192b6c1751233fe0f52

Cc: Tomasz Maciej Nowak <tomek_n@o2.pl>
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
4 years agobcm63xx: R5010UNv2: fix flash partitions for 16MB flash
Daniel González Cabanelas [Mon, 22 Jun 2020 19:44:44 +0000 (21:44 +0200)]
bcm63xx: R5010UNv2: fix flash partitions for 16MB flash

The router Nucom R5010UN v2 has the partitions defined for a 8MB flash,
but the flash chip is 16MB size. We are wasting half of the flash.

Fix it and use generic names for partitions.

Fixes: 474cde61234c ("brcm63xx: probe SPI flash through DT")
Signed-off-by: Daniel González Cabanelas <dgcbueu@gmail.com>
4 years agobuild: improve ccache support
Roman Yeryomin [Fri, 12 Jun 2020 18:43:46 +0000 (21:43 +0300)]
build: improve ccache support

Set CCACHE_DIR to $(TOPDIR)/.ccache and CCACHE_BASEDIR to $(TOPDIR).
This allows to do clean and dirclean. Cache hit rate for test build
after dirclean is ~65%.
If CCACHE is enabled stats are printed out at the end of building process.
CCACHE_DIR config variable allows to override default, which could be useful
when sharing cache with many builds.
cacheclean make target allows to clean the cache.

Changes from v1:
- remove ccache directory using CCACHE_DIR variable
- remove ccache leftovers from sdk and toolchain make files
- introduce CONFIG_CCACHE_DIR variable
- introduce cacheclean make target

Signed-off-by: Roman Yeryomin <roman@advem.lv>
4 years agoramips: mt7621: refactor set affinity script
DENG Qingfang [Thu, 4 Jun 2020 04:00:44 +0000 (12:00 +0800)]
ramips: mt7621: refactor set affinity script

The current one only looks for mt76x2e and mt7603e, and
does not work for 2 or more same Wi-Fi chips.
Refactor the script to cover those cases.

Signed-off-by: DENG Qingfang <dengqf6@mail2.sysu.edu.cn>
4 years agoltq-vmmc: update permission handling
Sebastian Kemper [Tue, 9 Jun 2020 17:18:55 +0000 (19:18 +0200)]
ltq-vmmc: update permission handling

The firmware is currently just copied. It can end up with o= on the
device (this is the case for voice_ar9_firmware.bin for instance).
Instead of copying it the Makefile is changed to use the macro
"$(INSTALL_DATA)" in order for the file to be world-readable.

While at it refactor the device node creation in the init script with
loop.

Signed-off-by: Sebastian Kemper <sebastian_ml@gmx.net>
[removed 2nd part with custom group handling for device nodes]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agolayerscape: define IMAGE_SIZE and LS_SYSUPGRADE_IMAGE_SIZE
Yangbo Lu [Mon, 29 Jun 2020 08:18:57 +0000 (16:18 +0800)]
layerscape: define IMAGE_SIZE and LS_SYSUPGRADE_IMAGE_SIZE

Define 64m IMAGE_SIZE for flash firmware.bin since the flash size
is 64MB. Define 48m LS_SYSUPGRADE_IMAGE_SIZE for flash sysupgrade.bin
which contains maximum 16MB kernel and 32MB rootfs according to
memory map.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
4 years agolayerscape: support sysupgrade for SD card ext4 rootfs
Yangbo Lu [Thu, 28 May 2020 08:02:28 +0000 (16:02 +0800)]
layerscape: support sysupgrade for SD card ext4 rootfs

Support sysupgrade for SD card ext4 rootfs.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
4 years agolayerscape: support sysupgrade for squashfs rootfs
Yangbo Lu [Thu, 28 May 2020 04:48:29 +0000 (12:48 +0800)]
layerscape: support sysupgrade for squashfs rootfs

Support sysupgrade for all Layerscape boards with squashfs
rootfs.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
4 years agolayerscape: fix board name for SD card boot
Yangbo Lu [Thu, 28 May 2020 05:50:42 +0000 (13:50 +0800)]
layerscape: fix board name for SD card boot

Many Layerscape boards support both flash boot and SD card boot.
And different firmware and sysupgrade.bin are built for the two
boot methods. To identify them, a fix could be done on board name
by adding a postfix "-sdboot" for SD card boot.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
4 years agolayerscape: add compatible for board ls1043ardb
Yangbo Lu [Thu, 28 May 2020 04:06:20 +0000 (12:06 +0800)]
layerscape: add compatible for board ls1043ardb

Backport a dts patch which added compatible for ls1043ardb.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
4 years agolayerscape: convert to squashfs rootfs for QSPI NOR boot
Yangbo Lu [Wed, 27 May 2020 03:10:19 +0000 (11:10 +0800)]
layerscape: convert to squashfs rootfs for QSPI NOR boot

There had been an issue in Layerscape QSPI driver for very long
time, which made squashfs,jffs2 rootfs not work on QSPI NOR.
And the ubifs had been used as a workaround.
Now the issue has been fixed. So convert to use squashfs,jffs2
rootfs on QSPI NOR for Layerscape boards (LS1012ARDB/LS1046ARDB/
LS1088ARDB), and update u-boot bootargs for booting.

Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
4 years agobusybox: store applet usage messages uncompressed
Rui Salvaterra [Tue, 7 Jul 2020 13:16:47 +0000 (14:16 +0100)]
busybox: store applet usage messages uncompressed

The rootfs squashfs is already highly (XZ) compressed. Storing the applet
messages in compressed form will increase the entropy and reduce the overall
image compression ratio.

Size diffs (compressed vs uncompressed):

busybox (the executable): 364596 vs 384804 bytes.

OpenWrt target images (the kernel image is unchanged, obviously):

 omnia-medkit-openwrt-mvebu-cortexa9-cznic_turris-omnia-initramfs.tar.gz:

  9163597 vs 9162531 bytes (1066 bytes difference).

openwrt-mvebu-cortexa9-cznic_turris-omnia-initramfs-kernel.bin:

  9161688 vs 9160600 bytes (1088 bytes difference).

openwrt-mvebu-cortexa9-cznic_turris-omnia-sysupgrade.img.gz:

 9729550 vs 9729230 bytes (320 bytes difference).

All in all, we save just a little bit over 1 kiB. As an added bonus, we
also don't have to decompress the messages twice, (first from squashfs,
then from the bzip2 message storage).

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
[added additional size comparision diff detaisl]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agobase-files/functions.sh: use command -v instead of which
Rosen Penev [Fri, 17 Jan 2020 22:38:19 +0000 (14:38 -0800)]
base-files/functions.sh: use command -v instead of which

which must be executed. command -v is a shell builtin.

https://github.com/koalaman/shellcheck/wiki/SC2230

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agoscripts/mkits.sh: fix remaining shellcheck warning
Petr Štetiar [Sat, 11 Jul 2020 11:31:35 +0000 (13:31 +0200)]
scripts/mkits.sh: fix remaining shellcheck warning

Fixes following shellcheck warning:

 In scripts/mkits.sh line 19:
   "-k kernel [-D name -d dtb] -o its_file" "$(basename $0)"
                                                                       ^-- SC2086: Double quote to prevent globbing and word splitting.

Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agoscripts/mkits.sh: switch from bash to sh
Rosen Penev [Mon, 6 Jan 2020 02:31:20 +0000 (18:31 -0800)]
scripts/mkits.sh: switch from bash to sh

This no longer needs bash.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agoscripts/mkits.sh: fix improper string and array concatenation
Rosen Penev [Mon, 6 Jan 2020 02:31:19 +0000 (18:31 -0800)]
scripts/mkits.sh: fix improper string and array concatenation

Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agoscripts/mkits.sh: add missing quotes
Rosen Penev [Mon, 6 Jan 2020 02:31:18 +0000 (18:31 -0800)]
scripts/mkits.sh: add missing quotes

Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agoscripts/mkits.sh: replace echo -e with printf
Rosen Penev [Mon, 6 Jan 2020 02:31:17 +0000 (18:31 -0800)]
scripts/mkits.sh: replace echo -e with printf

echo flags are not POSIX. printf does the same with added \n.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agoscripts/gen_image_generic.sh: fix more shellcheck warnings
Petr Štetiar [Sat, 11 Jul 2020 11:24:17 +0000 (13:24 +0200)]
scripts/gen_image_generic.sh: fix more shellcheck warnings

Fixes following shellcheck warnings:

 In scripts/gen_image_generic.sh line 20:
 cyl=$(( (KERNELSIZE + ROOTFSSIZE) * 1024 * 1024 / (head * sect * 512)))
 ^-^ SC2034: cyl appears unused. Verify use (or export if used externally).
 --
 In scripts/gen_image_generic.sh line 34:
     [ -n "$PADDING" ] && dd if=/dev/zero of="$OUTPUT" bs=512 seek="$(($ROOTFSOFFSET + $ROOTFSSIZE))" conv=notrunc count="$sect"
                                                                       ^-----------^ SC2004: $/${} is unnecessary on arithmetic variables.
 --
 In scripts/gen_image_generic.sh line 35:
     mkfs.fat -n kernel -C "$OUTPUT.kernel" -S 512 "$(($KERNELSIZE / 1024))"
                                                       ^---------^ SC2004: $/${} is unnecessary on arithmetic variables.

Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agoscripts/gen_image_generic.sh: use /bin/sh
Rosen Penev [Mon, 6 Jan 2020 02:22:55 +0000 (18:22 -0800)]
scripts/gen_image_generic.sh: use /bin/sh

This has nothing bash specific.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agoscripts/gen_image_generic.sh: replace -o with if/&&
Rosen Penev [Mon, 6 Jan 2020 02:22:54 +0000 (18:22 -0800)]
scripts/gen_image_generic.sh: replace -o with if/&&

-o is not well defined.

Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agoscripts/env: fix remaining shellcheck warning
Petr Štetiar [Sat, 11 Jul 2020 11:20:34 +0000 (13:20 +0200)]
scripts/env: fix remaining shellcheck warning

Fixes following shellcheck warning:

 In scripts/env line 25:
  exit ${1:-1}
             ^-----^ SC2086: Double quote to prevent globbing and word splitting.

Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agoscripts/env: use read -r instead of read
Rosen Penev [Mon, 6 Jan 2020 02:32:38 +0000 (18:32 -0800)]
scripts/env: use read -r instead of read

read mangles backslashes.

Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agoscripts/env: exit in case of failure to cd
Rosen Penev [Mon, 6 Jan 2020 02:32:37 +0000 (18:32 -0800)]
scripts/env: exit in case of failure to cd

Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agoscripts/env: replace \! with !
Rosen Penev [Mon, 6 Jan 2020 02:32:36 +0000 (18:32 -0800)]
scripts/env: replace \! with !

The latter is more standard. The former throws an error under
shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agoscripts/env: use command -v instead of which
Rosen Penev [Mon, 6 Jan 2020 02:32:35 +0000 (18:32 -0800)]
scripts/env: use command -v instead of which

Simpler and built in to the shell.

Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agobuild: reduce number of files passed to ipk-remove
Eneas U de Queiroz [Thu, 20 Feb 2020 21:29:04 +0000 (18:29 -0300)]
build: reduce number of files passed to ipk-remove

Instead of using xargs to pass a huge number of files to
script/ipkg-remove, which will usually pick only one, use a more
restrictive wildcard so that, currently, at the most 325 files are
examined, instead of up to over 2,300.  The 325-file package is python,
which is picking up python3* ipks.  It is about to be removed.
Runner-up is ddns-scripts with 7 files.

This makes a second run of make package/luci/compile go from
real 16.40s; user 17.42s; sys 2.73s
to
real 10.71s; user 9.51s; sys 1.27s

There is a caveat though: if one were to remove the ABI_VERSION of a
package that ends in a digit [0-9], then the old package ipk will not be
removed from the bin directory by make package/abc2/clean.

Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
4 years agobuild: call ipkg-remove using xargs if #args>=512
Eneas U de Queiroz [Thu, 20 Feb 2020 21:29:03 +0000 (18:29 -0300)]
build: call ipkg-remove using xargs if #args>=512

The wildcard call to clean up luci package (luci*) can pick up over
2,300 files when the full tree is built. Running make package/luci/clean
or a second run of make package/luci/compile would fail with an
'Argument list too long' error.

To avoid that, a maybe_use_xargs function was created that runs the
command straight as usual if the number of arguments is < 512, or saves
the list in a temporary file and feeds it to xargs otherwise.

Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
4 years agobuild: package-ipkg: avoid calling wildcard twice
Eneas U de Queiroz [Thu, 20 Feb 2020 21:29:02 +0000 (18:29 -0300)]
build: package-ipkg: avoid calling wildcard twice

Instead of calling $(wildcard) to check if the removal list is empty,
then calling it again to actually remove the files, define a function so
that the arguments are expanded only once when it gets called.

Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
4 years agotoolchain/wrapper.sh: fix remaining shellcheck warnings
Petr Štetiar [Sat, 11 Jul 2020 11:06:36 +0000 (13:06 +0200)]
toolchain/wrapper.sh: fix remaining shellcheck warnings

Fixes following warnings:

 In target/toolchain/files/wrapper.sh line 18:
 REALNAME=$(readlink -f $0)
                        ^-- SC2086: Double quote to prevent globbing and word splitting.
 --
 In target/toolchain/files/wrapper.sh line 20:
 REALNAME_BASE=$(basename $REALNAME)
                          ^-------^ SC2086: Double quote to prevent globbing and word splitting.
 --
 In target/toolchain/files/wrapper.sh line 21:
 REALNAME_DIR=$(dirname $REALNAME)
                        ^-------^ SC2086: Double quote to prevent globbing and word splitting.
 --
 In target/toolchain/files/wrapper.sh line 74:
  exec $TARGET_TOOLCHAIN_TRIPLET-$BINARY.bin $GCC_SYSROOT_FLAGS $TARGET_FUNDAMENTAL_CFLAGS $TARGET_ROOTFS_CFLAGS "$@"
                     ^-----------------------^ SC2086: Double quote to prevent globbing and word splitting.
 --
 In target/toolchain/files/wrapper.sh line 77:
  exec $TARGET_TOOLCHAIN_TRIPLET-$BINARY.bin $LD_SYSROOT_FLAGS $TARGET_FUNDAMENTAL_LDFLAGS "$@"
                      ^-----------------------^ SC2086: Double quote to prevent globbing and word splitting.
 --
 In target/toolchain/files/wrapper.sh line 80:
  exec $TARGET_TOOLCHAIN_TRIPLET-$BINARY.bin $TARGET_FUNDAMENTAL_ASFLAGS "$@"
                      ^-----------------------^ SC2086: Double quote to prevent globbing and word splitting.
 --
 In target/toolchain/files/wrapper.sh line 83:
  exec $TARGET_TOOLCHAIN_TRIPLET-$BINARY.bin "$@"
                      ^-----------------------^ SC2086: Double quote to prevent globbing and word splitting.

Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agotoolchain/wrapper.sh: Replace read with read -r
Rosen Penev [Mon, 30 Dec 2019 03:41:17 +0000 (19:41 -0800)]
toolchain/wrapper.sh: Replace read with read -r

Without -r, backslashes would get mangled.

Found with shellcheck.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agotoolchain/gdb: Don't use gdb-arc
Rosen Penev [Fri, 22 Nov 2019 23:09:01 +0000 (15:09 -0800)]
toolchain/gdb: Don't use gdb-arc

GDB got support for ARC with version 8.2. No need for this fork.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agokernel: kmod-zram: break the strict dependency from lz4
Rui Salvaterra [Thu, 25 Jun 2020 14:20:23 +0000 (15:20 +0100)]
kernel: kmod-zram: break the strict dependency from lz4

Zram is only strictly dependent on lzo, not lz4. Break this dependency and
make the lz4 module visible in the configuration, in order for the user to
have the choice of enabling/disabling it, if (s)he sees fit.

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
4 years agozram-swap: correctly express the required dependencies
Rui Salvaterra [Wed, 8 Jul 2020 16:20:28 +0000 (17:20 +0100)]
zram-swap: correctly express the required dependencies

The block-mount swapon implementation doesn't support discard, so make zram-swap
depend only on the default BusyBox implementation or, when unavailable, on the
one present in the swap-utils package.

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
4 years agozram-swap: enable swap discard
Rui Salvaterra [Tue, 23 Jun 2020 11:29:54 +0000 (12:29 +0100)]
zram-swap: enable swap discard

Zram block devices have supported trim/discard for over six years, let's
enable it. This allows the zram device to actually free up allocated memory
when it's marked as unused in the filesystem metadata, as explained in more
detail in the original commit message [1].

[1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/drivers/block/zram/zram_drv.c?h=linux-4.14.y&id=f4659d8e620d08bd1a84a8aec5d2f5294a242764

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
4 years agobusybox: use CLOCK_MONOTONIC instead of gettimeofday
Rui Salvaterra [Wed, 8 Jul 2020 08:18:14 +0000 (09:18 +0100)]
busybox: use CLOCK_MONOTONIC instead of gettimeofday

The clock_gettime(CLOCK_MONOTONIC) syscall exists for so long that the first
kernel version to support it is not even specified in the man page [1]. Let's
enable it on BusyBox by default. Otherwise, gettimeofday will be used instead,
which will give wrong results if the date/time is reset (time moving backwards).

[1] https://linux.die.net/man/2/clock_gettime

Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
4 years agolzo: switch to building with CMake
Rosen Penev [Fri, 10 Jul 2020 23:13:10 +0000 (16:13 -0700)]
lzo: switch to building with CMake

CMake is less error prone that autotools and also compiles faster.

Fixed license information.

Added pkgconfig file to InstallDev so that packages that use it can
find lzo.

Before:

time make package/lzo/compile -j 12
________________________________________________________
Executed in   20.87 secs   fish           external
   usr time   26.95 secs    0.00 micros   26.95 secs
   sys time    5.49 secs  305.00 micros    5.49 secs

After:

time make package/lzo/compile -j 12
________________________________________________________
Executed in   13.22 secs   fish           external
   usr time   19.59 secs  328.00 micros   19.59 secs
   sys time    4.03 secs   10.00 micros    4.03 secs

Time output is with fish shell. make clean was ran before both attempts.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
4 years agolibubox: update to version 2020-07-11
Petr Štetiar [Mon, 20 Jan 2020 15:22:07 +0000 (16:22 +0100)]
libubox: update to version 2020-07-11

 f4e9bf73ac5c examples/lua: attempt to highlight some traps
 53b9a2123fc6 lua/uloop: fd_add: use absolute indices for arguments
 c0941d3289fc lua/uloop: make get_sock_fd capable of absolute addresses
 161c25960ba2 lua/uloop: fd_add() better args checking

Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agoramips: kernel: fix awake-rt305x-dwc2 patch
Alexey Dobrovolsky [Wed, 6 May 2020 20:48:15 +0000 (23:48 +0300)]
ramips: kernel: fix awake-rt305x-dwc2 patch

At this point in v5.4 kernel we cannot use dwc2_readl() and
dwc2_writel() since they rely on the value hsotg->needs_byte_swap
which cannot be obtained before the controller wakes up.
We should use readl() and writel() to wake up the controller before
calling dwc2_check_core_endianness().

Fixes: 6be0da90a165 ("ramips: refresh patches")
Signed-off-by: Alexey Dobrovolsky <dobrovolskiy.alexey@gmail.com>
[fixed Fixes: tag]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agolinux-firmware: package EIP197 mini firmware
Tomasz Maciej Nowak [Thu, 9 Jul 2020 19:16:15 +0000 (21:16 +0200)]
linux-firmware: package EIP197 mini firmware

Quoting part of original message from eefb5f741015 commit in
linux-firmware repository:

This adds the "minifw" version of the EIP197 firmware, which the inside-
secure driver will use as a fallback if the original full-featured
firmware cannot be found. This allows for using the inside-secure driver
and hardware without access to "official" firmware only available under
NDA.

Signed-off-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
4 years agofirmware-utils: mkfwimage: fix memcpy and strncpy usage
Petr Štetiar [Fri, 26 Jul 2019 12:45:32 +0000 (14:45 +0200)]
firmware-utils: mkfwimage: fix memcpy and strncpy usage

Firmware is binary blob, so there are barely any NULL terminated strings
expected, so we should probably convert all chars into u8 types, and
after that it's clear, that using strcpy doesn't make sense anymore.

This is rather theoretical stuff, but `uint8_t name[PART_NAME_LENGTH]`
means, that you can supply PART_NAME_LENGTH sized name, not
PART_NAME_LENGTH-1 name when NULL terminated.

Ref: https://github.com/openwrt/openwrt/pull/2274
Fixes: 04cb651376f9 ("firmware-utils: mkfwimage: fix more errors reported by gcc-6/7/9")
Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agotoolchain/gcc: Add GCC 10.1.0 patches
Syrone Wong [Tue, 7 Jul 2020 22:46:22 +0000 (06:46 +0800)]
toolchain/gcc: Add GCC 10.1.0 patches

Compared to GCC 9:

870-ppc_no_crtsavres.patch changes moved to another file following upstream
881-no_tm_section.patch keep the tm section disabled

patches refreshed to apply cleanly

See https://gcc.gnu.org/gcc-10/porting_to.html for more info

Compiled and run tested on x86_64

Signed-off-by: Syrone Wong <wong.syrone@gmail.com>
4 years agotoolchain/gcc: Add GCC 10.1.0 config
Syrone Wong [Tue, 7 Jul 2020 22:41:21 +0000 (06:41 +0800)]
toolchain/gcc: Add GCC 10.1.0 config

Add needed config changes and tarball hash for new GCC version.

Signed-off-by: Syrone Wong <wong.syrone@gmail.com>
[added missing commit description]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agotoolchain/gcc: Copy patches from 9.3.0 to 10.1.0
Syrone Wong [Tue, 7 Jul 2020 22:40:24 +0000 (06:40 +0800)]
toolchain/gcc: Copy patches from 9.3.0 to 10.1.0

No content changes in this commit

Signed-off-by: Syrone Wong <wong.syrone@gmail.com>
4 years agokirkwood: add support for Check Point L-50
Pawel Dembicki [Sun, 26 Jan 2020 16:22:28 +0000 (17:22 +0100)]
kirkwood: add support for Check Point L-50

This patch adds support for the Check Point L-50 from 600/1100 series
routers.

Specification:
-CPU: Marvell Kirkwood 88F6281 1200MHz
-RAM: 512MB
-Flash: NAND 512MB
-WiFi: mPCIe card based on Atheros AR9287 b/g/n
-WAN: 1 Gigabit Port (Marvell 88E1116R PHY)
-LAN: 9 Gigabit Ports (2x Marvell 88E6171(5+4))
-USB: 2x USB2.0
-Express card slot
-SD card slot
-Serial console: RJ-45 115200 8n1
-Unsupported DSL

Known issues:
-DSL not supported
-Expresscard not tested

Installation:

Step one -> backup:
make backup u-boot and env for revert stock posibility
make backup dsl_mac_addr, dmz_mac_addr, eth1addr, ethaddr and all lanX_mac_addr

Step two -> Use kwboot tool to upload openwrt u-boot to RAM:

run kwboot: "kwboot -B 115200 /dev/ttyUSB0 -b u-boot.kwb -p -t"
end start u-boot

Step three -> Restore macs (e.g. below):
setenv eth1addr 00:1C:XX:XX:XX:6C
setenv ethaddr 00:1C:XX:XX:XX:6B
setenv lan1_mac_addr 00:1C:XX:XX:XX:6C
setenv lan2_mac_addr 00:1C:XX:XX:XX:6D
setenv lan3_mac_addr 00:1C:XX:XX:XX:6E
setenv lan4_mac_addr 00:1C:XX:XX:XX:6F
setenv lan5_mac_addr 00:1C:XX:XX:XX:70
setenv lan6_mac_addr 00:1C:XX:XX:XX:71
setenv lan7_mac_addr 00:1C:XX:XX:XX:72
setenv lan8_mac_addr 00:1C:XX:XX:XX:73
setenv dmz_mac_addr 00:1C:XX:XX:XX:74
setenv dsl_mac_addr 00:1C:XX:XX:XX:75

Step four -> flash u-boot:
mw 0x0800000 0xffff 0x100000
nand erase 0x0 100000
tftp 0x0800000 openwrt-kirkwood-l50-u-boot.kwb
nand write 0x0800000 0x0 0x100000
saveenv

Step five -> run initramfs image:

tftpboot 0x02000000 openwrt.bin; bootm 0x02000000;

Step six -> install sysupgrade OpenWrt image:

copy to /tmp/ sysupgrade image
run sysupgrade

Back to stock:
Restore original u-boot end env.
Install factory image via stock u-boot.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
4 years agouboot-kirkwood: add uboot for CheckPoint L-50
Pawel Dembicki [Fri, 24 Jan 2020 06:13:24 +0000 (07:13 +0100)]
uboot-kirkwood: add uboot for CheckPoint L-50

This patch add u-boot for CheckPoint L-50 routers.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
4 years agokernel: add package for Seiko S-35390A I2C RTC
Pawel Dembicki [Fri, 24 Jan 2020 06:25:54 +0000 (07:25 +0100)]
kernel: add package for Seiko S-35390A I2C RTC

This patch adds kernel package for Seiko Instruments S-35390A.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
4 years agouboot-envtools: ath79: update ubootenv partion index for gl-ar300m
Huangbin Zhan [Wed, 8 Jul 2020 21:01:54 +0000 (05:01 +0800)]
uboot-envtools: ath79: update ubootenv partion index for gl-ar300m

The block index of u-boot-env changed from mtd1 to mtd3 after upgrading kernel to 5.4.
This patch search the mtd block by label name, work as expect when perform a clean flash.

Signed-off-by: Huangbin Zhan <zhanhb88@gmail.com>
4 years agokernel: bump 5.4 to 5.4.51
John Audia [Thu, 9 Jul 2020 21:22:35 +0000 (17:22 -0400)]
kernel: bump 5.4 to 5.4.51
update via update_kernel.sh -v -u 5.4

Removed upstreamed patches:
 350-MIPS-Add-missing-EHB-in-mtc0-mfc0-sequence-for-DSPen.patch

Script refreshed patches:
 902-debloat_proc.patch
 904-debloat_dma_buf.patch

Attempted merge conflict in following patches:
 0024-MIPS-lantiq-revert-DSA-switch-driver-PMU-clock-chang.patch

Build system: x86_64
Build tested: ipq806x (Netgear R7800)

Signed-off-by: John Audia <graysky@archlinux.us>
[fixed sha256sum of the tarball]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agoipq806x: enable Linksys EA8500 eth1 interface
Pawel Dembicki [Tue, 30 Jun 2020 09:29:46 +0000 (11:29 +0200)]
ipq806x: enable Linksys EA8500 eth1 interface

At this moment Linksys EA8500 uses only eth0.

This patch change switch registers, which allow to use eth1 as lan
and eth0 as wan. The method work with similar Linksys EA7500V1
and it work with EA8500.

Suggested-by: Sungbo Eo <mans0n@gorani.run>
Tested-by: Brian Onn <brian.a.onn@gmail.com>
Tested-by: Adrian Panella <ianchi74@outlook.com>
Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
4 years agobase-files: remove useless cat
Adrian Schmutzler [Sun, 10 May 2020 14:47:54 +0000 (16:47 +0200)]
base-files: remove useless cat

Check file contents directly instead of using cat.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agowireguard-tools: replace backticks by $(...)
Adrian Schmutzler [Sun, 10 May 2020 14:47:51 +0000 (16:47 +0200)]
wireguard-tools: replace backticks by $(...)

This replaces deprecated backticks by more versatile $(...) syntax.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agobase-files: replace backticks by $(...)
Adrian Schmutzler [Sun, 10 May 2020 14:47:46 +0000 (16:47 +0200)]
base-files: replace backticks by $(...)

This replaces deprecated backticks by more versatile $(...) syntax.

This does not touch lib/upgrade/nand.sh, as there replacement is
not trivial.

Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agoprocd: fix yet another build issue, this time with capabilities
Daniel Golle [Sat, 11 Jul 2020 10:27:33 +0000 (11:27 +0100)]
procd: fix yet another build issue, this time with capabilities

 3034eaf jail: use linux/capability.h instead of sys/capability.h

Fixes: b6e440a0f5 ("procd: update to git HEAD")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
4 years agoprocd: fix another seccomp-related build issue
Daniel Golle [Sat, 11 Jul 2020 09:52:28 +0000 (10:52 +0100)]
procd: fix another seccomp-related build issue

 3473671 ujail: add dependency on syscall-names-h

Fixes: b6e440a0f5 ("procd: update to git HEAD")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
4 years agoath79: disable TP-Link TL-WR941ND v2
Adrian Schmutzler [Sat, 11 Jul 2020 09:43:22 +0000 (11:43 +0200)]
ath79: disable TP-Link TL-WR941ND v2

The support for this device's Marvell MV88E6060 switch has been
reported to be broken with kernels 4.19/5.4 (see bug report).

Since this a 4/32 device and it has been confirmed to be working
with stable 19.07 release (kernel 4.14), and since fixing it does
not seem trivial, let's just disable it in master.

Fixes: FS#2524
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agoprocd: jail: fix build on platforms without seccomp support
Daniel Golle [Fri, 10 Jul 2020 23:08:04 +0000 (00:08 +0100)]
procd: jail: fix build on platforms without seccomp support

Fixes: b6e440a0f5 ("procd: update to git HEAD")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
4 years agoprocd: update to git HEAD
Daniel Golle [Fri, 10 Jul 2020 17:52:28 +0000 (18:52 +0100)]
procd: update to git HEAD

 ea7a790 jail: add support for running OCI bundle
 bb4a446 uxc: add container management CLI tool

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
4 years agomac80211: rt2x00: define RF5592 in init_eeprom routine
Daniel Golle [Tue, 7 Jul 2020 20:51:58 +0000 (21:51 +0100)]
mac80211: rt2x00: define RF5592 in init_eeprom routine

Make sure RF5592 is set for RT5592 chip which apparently sometimes
doesn't have RF defined (but always comes with RF5592).
This patch was originally submitted on linux-wireless by
Tom Psyborg <pozega.tomislav@gmail.com> but got rejected.
Turns out the patch is actually needed.

Reported-by: Sebastian Gottschall <s.gottschall@dd-wrt.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
4 years agoramips: add support for Netgear R6020
Tim Thorpe [Thu, 9 Jul 2020 03:07:35 +0000 (22:07 -0500)]
ramips: add support for Netgear R6020

This adds support for the Netgear R6020, aka Netgear AC750.

The R6020 appears to be the same hardware as the Netgear R6080,
aka Netgear AC1000, but it has a slightly different flash layout,
and no USB ports.

Specification:

SoC: MediaTek MT7628 (580 MHz)
Flash: 8 MiB
RAM: 64 MiB
Wireless: 2.4Ghz (builtin) and 5Ghz (MT7612E)
LAN speed: 10/100
LAN ports: 4
WAN speed: 10/100
WAN ports: 1
UART (57600 8N1) on PCB

MAC addresses based on vendor firmware:
LAN      *:88   0x4
WAN      *:89
WLAN2    *:88   0x4
WLAN5    *:8a   0x8004

The factory partition might have been corrupted beforehand. However,
the comparison of vendor firmware and OpenWrt still allowed to retrieve
a meaningful assignment that also matches the other similar devices.

Installation:

Flashing OpenWRT from stock firmware requires nmrpflash. Use an ethernet
cable to connect to LAN port 1 of the R6020, and power the R6020 off.
From the connected workstation, run
`nmrpflash -i eth0 -f openwrt-ramips-mt76x8-netgear_r6020-squashfs-factory.img`,
replacing eth0 with the appropriate interface (can be identified by
running `nmrpflash -L`). Then power on the R6020. After flashing has finished,
power cycle the R6020, and it will boot into OpenWRT. Once OpenWRT has been
installed, subsequent flashes can use the web interface and sysupgrade files.

Signed-off-by: Tim Thorpe <timfthorpe@gmail.com>
[slightly extend commit message, fix whitespaces in DTS, align From:
with Signed-off-by]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agolantiq: set correct gphy pins for Zyxel P-2812
Willem van den Akker [Thu, 9 Jul 2020 12:51:19 +0000 (14:51 +0200)]
lantiq: set correct gphy pins for Zyxel P-2812

This fixes the gphy LED pins for Zyxel P-2812 devices. The have
been accidentally altered during a tidy-up operation (see Fixes:
below).

No ports were available, dmesg reported:

[ 0.658577] pinctrl-xway 1e100b10.pinmux: pin io5 already requested by
            1e100bb0.stp; cannot claim for 1e108000.eth
[ 0.667566] pinctrl-xway 1e100b10.pinmux: pin-5 (1e108000.eth) status -22
[ 0.685238] lantiq,xrx200-net 1e108000.eth: Error applying setting,
            reverse things back
[ 0.693270] lantiq,xrx200-net: probe of 1e108000.eth failed with error -22

Fixes: FS#3188
Fixes: 660200e53d62 ("lantiq: dts: assign the GPHY LED pins to the
Ethernet controller node")

Suggested-by: Mathias Kresin <dev@kresin.me>
Signed-off-by: Willem van den Akker <wvdakker@wilsoft.nl>
[commit message facelift]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
4 years agomediatek: fix bashism in gen_banana_pi_img.sh
David Woodhouse [Thu, 9 Jul 2020 08:56:18 +0000 (09:56 +0100)]
mediatek: fix bashism in gen_banana_pi_img.sh

There was a bashism in the script. This fixes the script so that it
doesn't actually require bash, and can be run with any POSIX shell as
its shebang suggests.

Signed-off-by: David Woodhouse <dwmw2@infradead.org>
4 years agokernel: iscsi-initator: fix kernel config symbols
Lucian Cristian [Wed, 8 Jul 2020 19:49:57 +0000 (22:49 +0300)]
kernel: iscsi-initator: fix kernel config symbols

Enable SCSI low-level drivers on targets that don't have it already in
order to fix following build failures on few platforms:

  .config:4739:warning: symbol value 'm' invalid for SCSI_LOWLEVEL
 * Restart config...
 * SCSI low-level drivers
 SCSI low-level drivers (SCSI_LOWLEVEL) [Y/n] (NEW) aborted!

Fixes: b88f8202c4ce ("kernel: add iscsi-initator support")
Signed-off-by: Lucian Cristian <lucian.cristian@gmail.com>
[commit subject and description facelift]
Signed-off-by: Petr Štetiar <ynezz@true.cz>
4 years agomediatek: re-add u3phy2 phy-switch patch
Sungbo Eo [Fri, 29 May 2020 13:06:35 +0000 (22:06 +0900)]
mediatek: re-add u3phy2 phy-switch patch

This patch is required for pcie@2,0 in mt7623 to work. The patch was originally
added in kernel 4.14, but it has not been ported to 4.19 and later.

Fixes: FS#3217
Tested-by: David Woodhouse <dwmw2@infradead.org>
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
4 years agomediatek: do not create device nodes in uci-defaults script
Sungbo Eo [Fri, 29 May 2020 11:51:01 +0000 (20:51 +0900)]
mediatek: do not create device nodes in uci-defaults script

Since commit 298814e6be76 ("base-files: config_generate: split macaddr with
multiple ifaces") uci MAC address setup will create a device node for each
member iface. But this script might override the device nodes and interfere
with the MAC address setup.

Signed-off-by: Sungbo Eo <mans0n@gorani.run>
4 years agotreewide: kernel: drop redundant USB_EHCI_HCD=n config symbol
Petr Štetiar [Thu, 9 Jul 2020 07:04:54 +0000 (09:04 +0200)]
treewide: kernel: drop redundant USB_EHCI_HCD=n config symbol

Commit e53ec043bae1 ("kirkwood: move usb support to modules") has moved
this config symbol into generic configs, so it could be removed from
other configs.

Suggested-by: Aleksander Jan Bajkowski <A.Bajkowski@stud.elka.pw.edu.pl>
Signed-off-by: Petr Štetiar <ynezz@true.cz>