openwrt/staging/svanheule.git
16 months agorealtek: clock driver: switch RTL930X DT syntax PR-930X-CLK
Markus Stockhausen [Tue, 6 Sep 2022 09:47:54 +0000 (11:47 +0200)]
realtek: clock driver: switch RTL930X DT syntax

Use new DT clockdriver syntax for RTL930X target. To make it work we
need to change some nodes:

- define SRAM
- add clock controller
- Add second CPU node for future SMP use
- map all devices to new clocks
- Remove dummy LXB clock
- add CPU OPP table

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
16 months agorealtek: clock driver: add for RTL930X target
Markus Stockhausen [Tue, 6 Sep 2022 09:46:53 +0000 (11:46 +0200)]
realtek: clock driver: add for RTL930X target

Make use the new clock driver for RTL930X target devices. Add its primary
consumer (cpufreq-dt) too. To be careful just set the default governor to
userspace. As we rely on SRAM activate that module too.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
16 months agorealtek: clock driver: enable basic RTL930X config for cpufreq
Markus Stockhausen [Tue, 6 Sep 2022 09:46:06 +0000 (11:46 +0200)]
realtek: clock driver: enable basic RTL930X config for cpufreq

Before we enable the driver for the new target we must tell the config that
RTL930X allows CPU frequency changing.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
16 months agorealtek: clock driver: enhancement for RTL930X
Markus Stockhausen [Tue, 6 Sep 2022 09:44:49 +0000 (11:44 +0200)]
realtek: clock driver: enhancement for RTL930X

The previous version only covered the RTL83XX series. Add all required
features to support also the RTL930X devices.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
[remove trailing spaces in clk-rtl930x-sram.S]
Signed-off-by: Sander Vanheule <sander@svanheule.net>
16 months agorealtek: clock driver: adapt RTL839X DT
Markus Stockhausen [Tue, 6 Sep 2022 09:40:14 +0000 (11:40 +0200)]
realtek: clock driver: adapt RTL839X DT

The clock driver is self contained now. Do the same as for RTL838X.
Remove the oscillator and add some scratch SRAM area for future use.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
16 months agorealtek: clock driver: adapt RTL838X DT
Markus Stockhausen [Tue, 6 Sep 2022 09:38:38 +0000 (11:38 +0200)]
realtek: clock driver: adapt RTL838X DT

The clock driver is self contained now. Remove old oscillator definition
from the DT. We need a small SRAM scratchpad area for future use. Add it.
While we are here adapt the SRAM size to the real 128K as the manuals say.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
16 months agorealtek: clock driver: add switch clock
Markus Stockhausen [Tue, 6 Sep 2022 09:34:11 +0000 (11:34 +0200)]
realtek: clock driver: add switch clock

The internal switch clock will become relevant when we add RTL930X support.
Add it to the list.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
16 months agorealtek: clock driver: allow multiple clock parents
Markus Stockhausen [Tue, 6 Sep 2022 09:01:44 +0000 (11:01 +0200)]
realtek: clock driver: allow multiple clock parents

RTL930X LXB clock is not derived from the crystal. In a future commit
we will add that and need the possibility to derive a clock from something
else than the crystal. Make this configurable per SOC type.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
16 months agorealtek: clock: only allow CPU reclocking
Sander Vanheule [Wed, 9 Nov 2022 21:37:19 +0000 (22:37 +0100)]
realtek: clock: only allow CPU reclocking

Since memory reclocking support is incomplete, and LXB reclocking would cause
issues with UART and timer rates, disable changing the rates of these clocks.

16 months agorealtek: clock: remove cached clock rate
Sander Vanheule [Wed, 9 Nov 2022 21:35:45 +0000 (22:35 +0100)]
realtek: clock: remove cached clock rate

The cached clocked rate isn't used anywhere, so there's no need to store it.

16 months agorealtek: clock: restyle RTCL_ROUND_SET()
Sander Vanheule [Wed, 9 Nov 2022 21:34:58 +0000 (22:34 +0100)]
realtek: clock: restyle RTCL_ROUND_SET()

16 months agorealtek: clock driver: provide crystal/oscillator clock
Markus Stockhausen [Tue, 6 Sep 2022 08:53:54 +0000 (10:53 +0200)]
realtek: clock driver: provide crystal/oscillator clock

Until now the driver depends on an external DT oscillator definition.
For now we can not see any benefit of setting that in the device tree.
Lots of other Linux kernel drivers hardcode these values. Usually the
name xtal (crystal) is used there. Align that and register a fixed
frequency clock (named xtal_clk) in the driver.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
[move cosmetic changes to separate patches]
Signed-off-by: Sander Vanheule <sander@svanheule.net>
16 months agorealtek: clock driver: cosmetic fixes
Markus Stockhausen [Tue, 6 Sep 2022 08:59:46 +0000 (10:59 +0200)]
realtek: clock driver: cosmetic fixes

Especially do the following:

- Fix wrong DRAM function definitions (although never used).
- Print speed ranges during startup only if changing is really supported.
- Add some namespace prefixes.
- Use kernel common type "unsigned long" for clock rate

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
[SPDX already has right comment type, keep it]
Signed-off-by: Sander Vanheule <sander@svanheule.net>
16 months agorealtek: clock driver: cut comments to 80 chars
Markus Stockhausen [Tue, 6 Sep 2022 08:57:49 +0000 (10:57 +0200)]
realtek: clock driver: cut comments to 80 chars

Try to keep line size with kernel recommendation.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
[merge other reflowed comment blocks]
Signed-off-by: Sander Vanheule <sander@svanheule.net>
16 months agorealtek: clock driver: reorganize documentation
Markus Stockhausen [Tue, 6 Sep 2022 08:56:11 +0000 (10:56 +0200)]
realtek: clock driver: reorganize documentation

Move the documentation into the device tree.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
[remove trailing spaces]
Signed-off-by: Sander Vanheule <sander@svanheule.net>
16 months agorealtek: clock driver: reorganize frequency reading
Markus Stockhausen [Tue, 6 Sep 2022 08:51:27 +0000 (10:51 +0200)]
realtek: clock driver: reorganize frequency reading

To better adapt to future devices reorganize the register reading. A small
macro helps us to build a switch statement for SOC and clock.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
[add space between macro arguments]
Signed-off-by: Sander Vanheule <sander@svanheule.net>
16 months agorealtek: clock driver: care about locked register
Markus Stockhausen [Tue, 6 Sep 2022 08:48:45 +0000 (10:48 +0200)]
realtek: clock driver: care about locked register

some RTL838X devices are shipped with locked registers. Remove the locks
so we can read/write the required PLL registers.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
16 months agorealtek: clock driver: get away with register macros
Markus Stockhausen [Tue, 6 Sep 2022 07:24:33 +0000 (09:24 +0200)]
realtek: clock driver: get away with register macros

Just use plain io functions. In a future version we will simplify the
(void *) modifications but for now keep the commit simple.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
16 months agorealtek: clock driver: simplify magic register values
Markus Stockhausen [Tue, 6 Sep 2022 07:22:11 +0000 (09:22 +0200)]
realtek: clock driver: simplify magic register values

For simplicity we use register magic values. We simply read/write full registers
but only some of the bits are relevant. So we might overwrite data that is
different depending on the vendor. For simplification we will handle constants
as follows:

- constants are right shifted and contain only relevant bits
- the code will left shift the values as needed
- the code will run a read-modify-write cycle
- As syncs are not needed for registers we can drop that part

Btw. SRAM can be access cached/uncached. When using parts of it make clear that
we access the cached part.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
[svanheule: remove reflowed comments]

16 months agorealtek: clock driver: rename variables
Markus Stockhausen [Tue, 6 Sep 2022 07:15:46 +0000 (09:15 +0200)]
realtek: clock driver: rename variables

Use our common prefix for several variables

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
16 months agoiproute2: Fix build with GCC 12 and glibc 2.36
Hauke Mehrtens [Wed, 21 Dec 2022 22:32:23 +0000 (23:32 +0100)]
iproute2: Fix build with GCC 12 and glibc 2.36

This fixes the detection of name_to_handle_at() when GCC 12 and glibc
2.36 are used.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agouhttpd: update to latest Git HEAD
Hauke Mehrtens [Mon, 26 Dec 2022 12:58:47 +0000 (13:58 +0100)]
uhttpd: update to latest Git HEAD

2397755 client: fix incorrectly emitting HTTP 413 for certain content lengths

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agorules: drop -Wno-error additional flags from default TARGET_CFLAGS
Christian Marangi [Sun, 17 Jul 2022 11:26:14 +0000 (13:26 +0200)]
rules: drop -Wno-error additional flags from default TARGET_CFLAGS

We currently enable -Wno-error=unused-but-set-variable and
-Wno-error=unused-result by default on every compile package.

While this is (relatively) unharmful, we should follow other project
direction and starts enforcing good code quality. For example the linux
kernel recently started to enforce Wall by default and clean code is
mandatory for inclusion.

Drop for good these flags and and make it mandatory to correctly handle
return values at least with a warning log if they are not strictly error
condition.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agolibbsd: fix libpath to not use host path
Xuefer H [Sat, 24 Dec 2022 12:16:11 +0000 (20:16 +0800)]
libbsd: fix libpath to not use host path

libpath.so uses host path in ld script causing other packages fail to
cross compile, e.g. perl:
"ld: cannot find /usr/lib/libbsd.so.0.11.6: No such file or directory"

Fixes: openwrt/packages#19390
Signed-off-by: Xuefer H <xuefer@gmail.com>
16 months agodnsmasq: bump to v2.88
Kevin Darbyshire-Bryant [Mon, 17 Oct 2022 18:28:07 +0000 (19:28 +0100)]
dnsmasq: bump to v2.88

Most relevant feature for openwrt in this release, supports dynamically
removing hosts from 'hostsdir' supplied host files.

Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
16 months agouboot-omap: build with omap generic subtarget
Kuan-Yi Li [Sat, 24 Dec 2022 20:06:38 +0000 (04:06 +0800)]
uboot-omap: build with omap generic subtarget

Fix image build fail by using the generic subtarget.

Fixes: b2bfea48 ("omap: add generic subtarget")
Ref: 40e3f660 ("uboot-fritz4040: build with ipq40xx generic subtarget")
Signed-off-by: Kuan-Yi Li <kyli@abysm.org>
[ add commit description ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agouboot-tegra: build with tegra generic subtarget
Kuan-Yi Li [Sat, 24 Dec 2022 20:06:26 +0000 (04:06 +0800)]
uboot-tegra: build with tegra generic subtarget

Fix image build fail by using the generic subtarget.

Fixes: c028e1b1 ("tegra: add generic subtarget")
Ref: 40e3f660 ("uboot-fritz4040: build with ipq40xx generic subtarget")
Signed-off-by: Kuan-Yi Li <kyli@abysm.org>
[ add commit description ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agouboot-zynq: build with zynq generic subtarget
Kuan-Yi Li [Sat, 24 Dec 2022 20:06:14 +0000 (04:06 +0800)]
uboot-zynq: build with zynq generic subtarget

Fix image build fail by using the generic subtarget.

Fixes: 6d7129ef ("zynq: add generic subtarget")
Ref: 40e3f660 ("uboot-fritz4040: build with ipq40xx generic subtarget")
Signed-off-by: Kuan-Yi Li <kyli@abysm.org>
[ add commit description ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agouboot-mxs: build with mxs generic subtarget
Kuan-Yi Li [Sat, 24 Dec 2022 20:06:03 +0000 (04:06 +0800)]
uboot-mxs: build with mxs generic subtarget

Fix image build fail by using the generic subtarget.

Fixes: 64ef920b ("mxs: add generic subtarget")
Ref: 40e3f660 ("uboot-fritz4040: build with ipq40xx generic subtarget")
Signed-off-by: Kuan-Yi Li <kyli@abysm.org>
[ add commit description ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agouboot-kirkwood: build with kirkwood generic subtarget
Kuan-Yi Li [Sat, 24 Dec 2022 20:05:03 +0000 (04:05 +0800)]
uboot-kirkwood: build with kirkwood generic subtarget

Fix image build fail by using the generic subtarget.

Fixes: cada395a ("kirkwood: add generic subtarget")
Ref: 40e3f660 ("uboot-fritz4040: build with ipq40xx generic subtarget")
Signed-off-by: Kuan-Yi Li <kyli@abysm.org>
[ add commit description ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agosunxi: ensure NanoPi R1 has unique MAC address
Jan-Niklas Burfeind [Fri, 23 Dec 2022 21:03:14 +0000 (22:03 +0100)]
sunxi: ensure NanoPi R1 has unique MAC address

Ensure the MAC address for all NanoPi R1 boards is assigned uniquely for
each board.

The vendor ships the device in two variants; one with and one without
eMMC; but both without static mac-addresses.
In order to assign both board types unique MAC addresses, fall back on
the same method used for the NanoPi R2S and R4S in case the EEPROM
chip is not present by generating the board MAC from the SD card CID.

[0] https://wiki.friendlyelec.com/wiki/index.php/NanoPi_R1#Hardware_Spec

Similar too and based on:

commit b5675f500daf ("rockchip: ensure NanoPi R4S has unique MAC address")

Co-authored-by: David Bauer <mail@david-bauer.net>
Signed-off-by: Jan-Niklas Burfeind <git@aiyionpri.me>
16 months agokernel: bump 5.10 to 5.10.161
John Audia [Wed, 21 Dec 2022 19:19:12 +0000 (14:19 -0500)]
kernel: bump 5.10 to 5.10.161

All patches automatically rebased.

Build system: x86_64
Build-tested: ramips/tplink_archer-a6-v3
Run-tested: ramips/tplink_archer-a6-v3

Signed-off-by: John Audia <therealgraysky@proton.me>
16 months agokernel: bump 5.15 to 5.15.85
John Audia [Wed, 21 Dec 2022 19:13:53 +0000 (14:13 -0500)]
kernel: bump 5.15 to 5.15.85

All patches automatically rebased.

Build system: x86_64
Build-tested: bcm2711/RPi4B
Run-tested: bcm2711/RPi4B

Signed-off-by: John Audia <therealgraysky@proton.me>
16 months agomac80211: Do not build brcmsmac on bcm47xx_legacy
Hauke Mehrtens [Sat, 24 Dec 2022 13:39:17 +0000 (14:39 +0100)]
mac80211: Do not build brcmsmac on bcm47xx_legacy

brcmsmac needs bcma. bcma is build into the kernel for the other bcm47xx
subtargets, but not for the legacy target because it only uses ssb. We
could build bcma as a module for bcm47xx_legacy, but none of these old
devices uses a wifi card supported by brcsmac.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agompc85xx: drop empty patch as it backported to LTS kernel 5.15.70
Josef Schlehofer [Thu, 22 Dec 2022 18:37:36 +0000 (19:37 +0100)]
mpc85xx: drop empty patch as it backported to LTS kernel 5.15.70

This patch is part of Linux kernel 5.15.70, the content was removed, but
the empty file is still being carried over the repository, so remove it
once for all

Fixes: e1b009c ("kernel: bump 5.15 to 5.15.70")
Signed-off-by: Josef Schlehofer <pepe.schlehofer@gmail.com>
16 months agoCI: tools: reogranize and split workflow
Christian Marangi [Fri, 9 Dec 2022 17:09:32 +0000 (18:09 +0100)]
CI: tools: reogranize and split workflow

Generilize tools workflow for future usage in shared workflow for tools
build.

Split tools workflow to tools and push-containers:
- tools just execute build test
- push-containers build and push prebuilt containers

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agorealtek: 5.15: Fix incorrect switch patches
Olliver Schinagl [Mon, 3 Oct 2022 12:45:21 +0000 (14:45 +0200)]
realtek: 5.15: Fix incorrect switch patches

Add correct header to patche(s) to be correctly used
by git am and have better tracking of it.

See commit f1f97db627b2 ("realtek: Convert incorrect v5.10 patches").

Signed-off-by: Olliver Schinagl <oliver@schinagl.nl>
16 months agorealtek: 5.10: Fix incorrect switch patches
Olliver Schinagl [Mon, 3 Oct 2022 12:45:21 +0000 (14:45 +0200)]
realtek: 5.10: Fix incorrect switch patches

Add correct header to patche(s) to be correctly used
by git am and have better tracking of it.

See commit f1f97db627b2 ("realtek: Convert incorrect v5.10 patches").

Signed-off-by: Olliver Schinagl <oliver@schinagl.nl>
16 months agotools: Improve diffability/maintainability
Olliver Schinagl [Wed, 14 Dec 2022 10:33:01 +0000 (11:33 +0100)]
tools: Improve diffability/maintainability

There's no purpose to squish multiple tools into a single line (and
spread those out over multiple lines). It might look 'nice' in certain
conditions, but it's annoying to maintain.

For example, but not limited to:
* adding/removing tools, causes hard to read diffs
* Duplicates are harder to spot
* Sorting can not be (easily?) automated

With this proposed change, the above annoyances go away. Inserting a new
tool can be done with a single line-change-diff, sorting can be done by
any editor (in vi, select, :sort for example) and dupes are much easier
to spot.

Signed-off-by: Olliver Schinagl <oliver@schinagl.nl>
16 months agoomap: add generic subtarget
Christian Marangi [Mon, 19 Dec 2022 15:47:35 +0000 (16:47 +0100)]
omap: add generic subtarget

Same game as for 853e4dd3062df7cb5704b15d6af6730e3194b571. Add generic
to the filenames.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agopistachio: add generic subtarget
Christian Marangi [Mon, 19 Dec 2022 15:47:16 +0000 (16:47 +0100)]
pistachio: add generic subtarget

Same game as for 853e4dd3062df7cb5704b15d6af6730e3194b571. Add generic
to the filenames.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agoocteontx: add generic subtarget
Christian Marangi [Mon, 19 Dec 2022 15:46:56 +0000 (16:46 +0100)]
octeontx: add generic subtarget

Same game as for 853e4dd3062df7cb5704b15d6af6730e3194b571. Add generic
to the filenames.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agotegra: add generic subtarget
Christian Marangi [Mon, 19 Dec 2022 15:46:44 +0000 (16:46 +0100)]
tegra: add generic subtarget

Same game as for 853e4dd3062df7cb5704b15d6af6730e3194b571. Add generic
to the filenames.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agozynq: add generic subtarget
Christian Marangi [Mon, 19 Dec 2022 15:46:24 +0000 (16:46 +0100)]
zynq: add generic subtarget

Same game as for 853e4dd3062df7cb5704b15d6af6730e3194b571. Add generic
to the filenames.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agomxs: add generic subtarget
Christian Marangi [Mon, 19 Dec 2022 15:46:10 +0000 (16:46 +0100)]
mxs: add generic subtarget

Same game as for 853e4dd3062df7cb5704b15d6af6730e3194b571. Add generic
to the filenames.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agokirkwood: add generic subtarget
Christian Marangi [Mon, 19 Dec 2022 15:45:39 +0000 (16:45 +0100)]
kirkwood: add generic subtarget

Same game as for 853e4dd3062df7cb5704b15d6af6730e3194b571. Add generic
to the filenames.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agogemini: add generic subtarget
Christian Marangi [Mon, 19 Dec 2022 15:45:11 +0000 (16:45 +0100)]
gemini: add generic subtarget

Same game as for 853e4dd3062df7cb5704b15d6af6730e3194b571. Add generic
to the filenames.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agoath25: add generic subtarget
Christian Marangi [Mon, 19 Dec 2022 15:44:29 +0000 (16:44 +0100)]
ath25: add generic subtarget

Same game as for 853e4dd3062df7cb5704b15d6af6730e3194b571. Add generic
to the filenames.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agoramips: Fix drivers/pwm/pwm-mediatek-ramips.c compile
Hauke Mehrtens [Thu, 22 Dec 2022 23:33:44 +0000 (00:33 +0100)]
ramips: Fix drivers/pwm/pwm-mediatek-ramips.c compile

The pwmchip_remove() function returns void now. Fix a compile problem in
the drivers/pwm/pwm-mediatek-ramips.c driver.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agokernel: Move CONFIG_PWM_IMG and CONFIG_PWM_MEDIATEK to generic configuration
Hauke Mehrtens [Thu, 22 Dec 2022 23:28:51 +0000 (00:28 +0100)]
kernel: Move CONFIG_PWM_IMG and CONFIG_PWM_MEDIATEK to generic configuration

In the build of the ramips/mt76x8 target the user gets asked about these
two configuration options, add them to the generic kernel configuration.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agokernel: Move CONFIG_DRM_XEN_FRONTEND to generic configuration
Hauke Mehrtens [Thu, 22 Dec 2022 23:14:03 +0000 (00:14 +0100)]
kernel: Move CONFIG_DRM_XEN_FRONTEND to generic configuration

The CONFIG_DRM_XEN_FRONTEND configuration symbol is also used by the
layerscape target, move it to the generic kernel configuration.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agomt76: update to the latest version
Felix Fietkau [Thu, 22 Dec 2022 15:19:28 +0000 (16:19 +0100)]
mt76: update to the latest version

a03ef0aab93e wifi: mt76: mt7921: fix deadlock in mt7921_abort_roc
5b509e80384a wifi: mt76: dma: fix a regression in adding rx buffers

Signed-off-by: Felix Fietkau <nbd@nbd.name>
16 months agouboot-sunxi: use UUID of bootdev and bootpart
Jan-Niklas Burfeind [Tue, 20 Dec 2022 15:43:23 +0000 (16:43 +0100)]
uboot-sunxi: use UUID of bootdev and bootpart

Several sunxi devices come with multiple mmc devices. On such devices,
the mmc device order is unpredictable, so using /dev/mmcblk0p2 as root
device doesn't always work, which results in unbootable devices.

For the Banana Pi BPI-R3 in the mediatek target, this has been solved by
defining aliases for the mmc devices in the DTS. Ideally we would do the
same here, but for sunxi-a64 we already use UUID probing, so let's start
with that (5f2ff607e286 ("uboot-sunxi: a64: allow booting directly from
eMMC")).

Since we're building and including u-boot in each supported device
image, and this method has been proven to work fine for a64, let's just
change the default u-boot env file to do the same.

Fixes: #10080
Fixes: e6d9f6fdff ("sunxi: add support for FriendlyARM NanoPi R1")
Co-authored-by: Karl Palsson <karlp@etactica.com>
Signed-off-by: Jan-Niklas Burfeind <git@aiyionpri.me>
[use UUID in default u-boot env, rewrite commit message]
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
16 months agosunxi: remove frequency for NanoPi R1
Jan-Niklas Burfeind [Sun, 11 Dec 2022 13:04:46 +0000 (14:04 +0100)]
sunxi: remove frequency for NanoPi R1

The frequency appears as unlisted initial frequency.
Removed it as Hauke suggested.

Signed-off-by: Jan-Niklas Burfeind <git@aiyionpri.me>
16 months agogithub: point to forum feature request category
Stijn Tintel [Wed, 21 Dec 2022 15:23:23 +0000 (17:23 +0200)]
github: point to forum feature request category

This should have been amended in the previous commit.

Fixes: 65e72f8e380e ("github: add issue template")
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
16 months agogithub: add issue template
Stijn Tintel [Mon, 19 Dec 2022 17:56:55 +0000 (19:56 +0200)]
github: add issue template

Add an issue template with required fields, instructions how to easily
get some of that required data from the device. Aside from that, also
add some links to the OpenWrt contact page, and for feature requests,
link to the forum.

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
16 months agokernel: Add missing kernel configuration options
Hauke Mehrtens [Wed, 21 Dec 2022 13:22:46 +0000 (13:22 +0000)]
kernel: Add missing kernel configuration options

This fixes compile of the bmips target.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agokernel: Reorder kernel configuration options
Hauke Mehrtens [Wed, 21 Dec 2022 13:22:26 +0000 (13:22 +0000)]
kernel: Reorder kernel configuration options

./scripts/kconfig.pl '+' target/linux/generic/config-5.10 /dev/null > target/linux/generic/config-5.10-new
mv target/linux/generic/config-5.10-new target/linux/generic/config-5.10

./scripts/kconfig.pl '+' target/linux/generic/config-5.15 /dev/null > target/linux/generic/config-5.15-new
mv target/linux/generic/config-5.15-new target/linux/generic/config-5.15

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agotoolchain: glibc: Update to glibc 2.36
Hauke Mehrtens [Sat, 6 Aug 2022 22:00:36 +0000 (00:00 +0200)]
toolchain: glibc: Update to glibc 2.36

This updates to glibc to version 2.36.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agotoolchain: gcc: backport patches to fix build with glibc 2.36
Hauke Mehrtens [Sat, 6 Aug 2022 22:36:17 +0000 (00:36 +0200)]
toolchain: gcc: backport patches to fix build with glibc 2.36

glibc 2.36 changed the definition of enum fsconfig_command, it now
collides with the same definition from sys/mount.h. Remove the include
of linux/fs.h This still compiled with musl too.

This backports a patch which is already in the stable branch of GCC 11
and GCC 12.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agoarm-trusted-firmware-sunxi: drop CPE ID
Stijn Tintel [Tue, 20 Dec 2022 18:04:54 +0000 (20:04 +0200)]
arm-trusted-firmware-sunxi: drop CPE ID

The CPE ID is already set in trusted-firmware-a.mk.

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
16 months agotrusted-firmware-a.mk: use correct CPE ID
Stijn Tintel [Tue, 20 Dec 2022 18:04:53 +0000 (20:04 +0200)]
trusted-firmware-a.mk: use correct CPE ID

There are 2 different CPE IDs on the NVD website:
cpe:/a:arm:trusted_firmware-a
cpe:/o:arm:arm_trusted_firmware

The ID as currently used in trusted-firmware-a.mk does not exist. The
CPE ID using the arm_trusted_firmware product name only lists a few
records for versions 2.2 and 2.3 on the NVD site. The CPE ID using the
trusted_firmware-a product name lists many more records, and actually
has a CVE linked to it. Therefore, use the CPE ID using the
trusted_firmware-a product name.

Fixes: 104d60fe94ce ("trusted-firmware-a.mk: add PKG_CPE_ID")
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
16 months agoxdp-tools: update to version 1.2.9
Daniel Golle [Mon, 19 Dec 2022 18:58:42 +0000 (18:58 +0000)]
xdp-tools: update to version 1.2.9

Changes since v1.2.8:
 32aaf32 libxdp: Fix incorrect rx_ring_setup_done
 6049671 headers: add bpf_endian.h for parsing_helpers.h
 2682c1c export-man: Ignore errors when executing git shell command
 8afda7a xdp-loader/README: Mention lack of support for HW mode in most cards
 dc69919 libxdp: fix prog_fd checks for fd >= 0
 3d7c22a libxdp: Allow falling back to single-program attachment for loaded programs
 af00429 libxdp: Fix check in xdp_program__attach_single()
 41703d2 libxdp: Make sure to set the the program autoload when loading a program
 b1fd2e5 test-xdpdump: Only run tshark attribute test on newer versions of tshark
 5dfe342 libxdp: Convert xdp-dispatcher to use strict section names
 929a22e configure: Try to auto-detect versioned clang binaries
 074fcfb libxdp: Check program name when determining if a program is a dispatcher
 e13a191 Bump TOOLS_VERSION to 1.2.9

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
16 months agoRevert "Revert "mvebu: switch default kernel to 5.15""
Petr Štetiar [Mon, 19 Dec 2022 10:14:41 +0000 (11:14 +0100)]
Revert "Revert "mvebu: switch default kernel to 5.15""

This reverts commit 97c77fff28cf001399f33c7bc1ec6687ba18450b as commit
8be6350f6646 ("generic: 5.15: allow MV88E6xxx built-in when PTP support
disabled") contains the fix, so lets enable kernel 5.15 on mvebu again.

Signed-off-by: Petr Štetiar <ynezz@true.cz>
16 months agokernel: bump 5.10 to 5.10.160
John Audia [Mon, 19 Dec 2022 14:18:02 +0000 (09:18 -0500)]
kernel: bump 5.10 to 5.10.160

No patches affected by this update.

Build system: x86_64
Build-tested: ramips/tplink_archer-a6-v3
Run-tested: ramips/tplink_archer-a6-v3

Signed-off-by: John Audia <therealgraysky@proton.me>
16 months agokernel: bump 5.15 to 5.15.84
John Audia [Mon, 19 Dec 2022 14:17:10 +0000 (09:17 -0500)]
kernel: bump 5.15 to 5.15.84

All patches automatically rebased

Build system: x86_64
Build-tested: bcm2711/RPi4B
Run-tested: bcm2711/RPi4B

Signed-off-by: John Audia <therealgraysky@proton.me>
16 months agotools/xz: update to 5.2.10
Nick Hainke [Wed, 14 Dec 2022 10:00:07 +0000 (11:00 +0100)]
tools/xz: update to 5.2.10

Update to latest version.

Signed-off-by: Nick Hainke <vincent@systemli.org>
16 months agohostapd: add mbedtls variant
Rosen Penev [Thu, 15 Sep 2022 07:48:45 +0000 (00:48 -0700)]
hostapd: add mbedtls variant

This adds the current WIP mbedtls patches for hostapd. The motivation
here is to reduce size.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
16 months agoath79: image: don't depend on other COMPILE targets
Alexander Couzens [Sat, 3 Dec 2022 14:25:12 +0000 (15:25 +0100)]
ath79: image: don't depend on other COMPILE targets

A device COMPILE target should not depend on another COMPILE.
Otherwise race condition may happen.
The loader is very small. Compiling it twice shouldn't
have a huge impact.

Signed-off-by: Alexander Couzens <lynxis@fe80.eu>
16 months agoimage: fix device profile specific COMPILE targets
Michael Pratt [Tue, 22 Nov 2022 00:37:39 +0000 (00:37 +0000)]
image: fix device profile specific COMPILE targets

Commit a01d23e75 ("image: always rebuild kernel loaders")
is a step in the right direction, but exposed some issues
and regressions in the makefile.

Some of the files made by device specific COMPILE targets
start with an "append" command (i.e. >> instead of > redirection)
and if the file already exists, the target file is the
input to itself before the first recipe-specified input.

Fixes: a01d23e75 ("image: always rebuild kernel loaders")
Fixes: a7fb589e8 ("image: always rebuild kernel loaders")
Signed-off-by: Michael Pratt <mcpratt@pm.me>
16 months agomac80211: add ath11k PCI support
Robert Marko [Sun, 6 Feb 2022 14:50:24 +0000 (15:50 +0100)]
mac80211: add ath11k PCI support

ath11k is the upstream driver for Qualcomm 802.11ax radios, both for the
internal AHB and PCI based cards.
This commit does however only provide PCI support while AHB will follow
but its SoC specific so it will require an OpenWrt target first.

It differs a bit from ath10k as it requires stuff like QRTR, MHI and QMI
helpers.

PCI variant requires qrtr-mhi and mhi-bus which backports do provide,
however we are dropping those in a patch as they will conflict with
support for the AHB variant as that one requires qrtr-smd which in turn
requires RPMSG and GLINK and its not feasable to provide those in
backports as they are really SoC specific.

QRTR and MHI in kernel 5.10 are not usable and backporting the changes
is not easy as they have changed drastically from 5.10 to 5.15 ath11k will
only be available on targets that use kernel 5.15.

Signed-off-by: Robert Marko <robimarko@gmail.com>
16 months agolinux-firmware: package ath11k consumer cards firmware
Robert Marko [Tue, 13 Dec 2022 22:01:20 +0000 (23:01 +0100)]
linux-firmware: package ath11k consumer cards firmware

Package firmware for ath11k supported QCA consumer cards from linux-firmware.

Signed-off-by: Robert Marko <robimarko@gmail.com>
16 months agokernel: package QRTR support
Robert Marko [Tue, 21 Dec 2021 12:47:03 +0000 (13:47 +0100)]
kernel: package QRTR support

QRTR is Qualcomm IPC router protocol and ath11k requires it for both
AHB and PCI support, so package it as a kernel module so it can be
added as a dependency to the ath11k module.

Only kernel 5.15 is currently supported due to various things missing in
5.10 whose backporting is out of scope for this patch.

SMD, TUN and MHI variants are packaged.

SMD variant depends on the ipq807x
target as it has dependency on the RPMSG drivers which are Qualcomm
and SoC specific anyway.

Signed-off-by: Robert Marko <robimarko@gmail.com>
16 months agotools/patchelf: update to 0.17.0
Linhui Liu [Fri, 16 Dec 2022 06:40:49 +0000 (14:40 +0800)]
tools/patchelf: update to 0.17.0

Update to the latest released version.

Signed-off-by: Linhui Liu <liulinhui36@gmail.com>
16 months agotools/sed: Fix handling of symlinks of 128 chars
Hauke Mehrtens [Sun, 18 Dec 2022 02:09:22 +0000 (03:09 +0100)]
tools/sed: Fix handling of symlinks of 128 chars

If the absolute path a symlink is pointing to is 128 bytes long sed
failed with an error message like this: "<path>/sedstbU8O: Not a directory"

This fixes a problem building python seen in the build bot.

This patch is on its way into upstream sed.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agoramips: add support for Linksys E5400 and clones
Julien Dusser [Sun, 20 Mar 2022 09:11:15 +0000 (10:11 +0100)]
ramips: add support for Linksys E5400 and clones

Linksys E5400 is a 2.4/5GHz dual band AC router, based on MediaTek
MT7628AN. This device is also sold as Linksys E2500v4, E5300 and E5350
with the same hardware, but software speed limitations in vendor
firmwares.

Specification:
 * SoC: MT7628AN (580 MHz)
 * RAM: DDR2 64 MiB
 * Flash: 16 MiB NOR (W25Q128BV)
 * Wi-Fi:
   * 2.4GHz: SoC Built-in
   * 5GHz: MT7612EN
 * Ethernet: 5x 100Mbps
   * Switch: SoC built-in
 * UART:
   * 115200, 8N1, 3.3V (real u-boot speed: 119050)
   * Pinout JB4: 1:[3V3] (TXD) (RXD) [NC] (GND)

Flash Layout:
  0x0000000-0x0030000 : "Bootloader"
  0x0030000-0x0040000 : "Config"
  0x0040000-0x0050000 : "Factory"
  0x0050000-0x0ff0000 : "Kernel"
  0x0ff0000-0x1000000 : "CBTinfo"

MAC address:
  LAN: factory 0x28
  WAN: LAN + 1
  2.4G: LAN + 2
  5G: LAN + 3

Installation:
1. Connect to UART, use LF as EOL (not CRLF)
2. Set IP address
  - Press 4 during boot
  - setenv serverip <TFTPSERVER_IP>
  - setenv ipaddr <DEVICE_IP>
  - setenv bootfile openwrt-ramips-mt76x8-linksys_e5400-initramfs-kernel.bin
  - saveenv
  - reset
3. Load Openwrt Kernel image from TFTP:
  - Press 1 during boot
  - IP and filename should be already set
4. Boot into OpenWrt and perform sysupgrade with sysupgrade image.

Signed-off-by: Julien Dusser <julien.dusser@free.fr>
16 months agogeneric: 5.15: allow MV88E6xxx built-in when PTP support disabled
Johnny S. Lee [Thu, 15 Dec 2022 10:57:31 +0000 (18:57 +0800)]
generic: 5.15: allow MV88E6xxx built-in when PTP support disabled

As described in commit 97c77fff28cf001399f33c7bc1ec6687ba18450b
MV88E6xxx driver (NET_DSA_MV88E6XXX) cannot be built-in when PTP
(PTP_1588_CLOCK) is a module in Linux 5.15. But actually it should be
allowed to be built-in when its PTP support (NET_DSA_MV88E6XXX_PTP) is
disabled.

This adds a patch to fix that.

Signed-off-by: Johnny S. Lee <foss@jsl.io>
16 months agotools/libressl: update to 3.7.0
Linhui Liu [Thu, 15 Dec 2022 07:29:38 +0000 (15:29 +0800)]
tools/libressl: update to 3.7.0

Release notes:
https://ftp.openbsd.org/pub/OpenBSD/LibreSSL/libressl-3.6.0-relnotes.txt
https://ftp.openbsd.org/pub/OpenBSD/LibreSSL/libressl-3.6.1-relnotes.txt
https://ftp.openbsd.org/pub/OpenBSD/LibreSSL/libressl-3.7.0-relnotes.txt

Signed-off-by: Linhui Liu <liulinhui36@gmail.com>
16 months agoiproute2: update to 6.1.0
Nick Hainke [Tue, 13 Dec 2022 19:17:58 +0000 (20:17 +0100)]
iproute2: update to 6.1.0

Announcement:
https://lore.kernel.org/netdev/20221214094130.7b11ec2e@hermes.local/T/#t

Refresh patch:
- 170-ip_tiny.patch

Signed-off-by: Nick Hainke <vincent@systemli.org>
16 months agorules: fix broken commitcount on alpine system
Christian Marangi [Sat, 17 Dec 2022 15:39:00 +0000 (16:39 +0100)]
rules: fix broken commitcount on alpine system

To generate commitcount we use grep --max-count. This is not present on
alpine grep and cause wrong generation. Use -m as it's just the short
version of --max-count and more portable.

Fixes: #11200
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agoimage-commands.mk: Be consistent in command invocation
Olliver Schinagl [Wed, 14 Dec 2022 10:39:17 +0000 (11:39 +0100)]
image-commands.mk: Be consistent in command invocation

Most/all other tools use the staging dir prefix, gzip should as well.

Signed-off-by: Olliver Schinagl <oliver@schinagl.nl>
Acked-by: Christian Marangi <ansuelsmth@gmail.com>
16 months agoramips: add support for SNR-CPE-W4N-MT router
Alexey Bartenev [Mon, 24 Oct 2022 15:50:33 +0000 (22:50 +0700)]
ramips: add support for SNR-CPE-W4N-MT router

General specification:
- SoC Type: MediaTek MT7620N (580MHz)
- ROM: 8 MB SPI-NOR (W25Q64FV)
- RAM: 64 MB DDR (M13S5121632A)
- Switch: MediaTek MT7530
- Ethernet: 5 ports - 5×100MbE (WAN, LAN1-4)
- Wireless 2.4 GHz: b/g/n
- Buttons: 1 button (RESET)
- Bootloader: U-Boot 1.1.3, MediaTek U-Boot: 5.0.0.5
- Power: 12 VDC, 1.0 A

Flash by the native uploader in 2 stages:
1. Use the native uploader to flash an initramfs image. Choose
 openwrt-ramips-mt7620-snr_cpe-w4n-mt-initramfs-kernel.bin file by
 "Administration/Management/Firmware update/Choose File" in vendor's
 web interface (ip: 192.168.1.10, login: Admin, password: Admin).
 Wait ~160 seconds.
2. Flash a sysupgrade image via the initramfs image. Choose
 openwrt-ramips-mt7620-snr_cpe-w4n-mt-squashfs-sysupgrade.bin
 file by "System/Backup/Flash Firmware/Flash image..." in
 LuCI web interface (ip: 192.168.1.1, login: root, no password).
 Wait ~240 seconds.

Flash by U-Boot TFTP method:
1. Configure your PC with IP 192.168.1.131
2. Set up TFTP server and put the
 openwrt-ramips-mt7620-snr_cpe-w4n-mt-squashfs-sysupgrade.bin
 image on your PC
3. Connect serial port (57600 8N1) and turn on the router.
 Then interrupt "U-Boot Boot Menu" by hitting 2 key (select "2:
 Load system code then write to Flash via TFTP.").
Press Y key when show "Warning!! Erase Linux in Flash then burn
 new one. Are you sure? (Y/N)"
Input device IP (192.168.1.1) ==:192.168.1.1
Input server IP (192.168.1.131) ==:192.168.1.131
Input Linux Kernel filename () ==:
openwrt-ramips-mt7620-snr_cpe-w4n-mt-squashfs-sysupgrade.bin
3. Wait ~120 seconds to complete flashing

Signed-off-by: Alexey Bartenev <41exey@proton.me>
16 months agogdb: Do not link against xxhash
Hauke Mehrtens [Sat, 17 Dec 2022 21:10:39 +0000 (22:10 +0100)]
gdb: Do not link against xxhash

libxxhash is now available in the OpenWrt package feed and gdb will link
against it if gdb finds this library. Explicitly deactivate the usage
of xxhash.

This should fix the build of gdb in build bots.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agoath79: add support for KuWFi C910
Davide Fioravanti [Tue, 13 Dec 2022 22:28:08 +0000 (23:28 +0100)]
ath79: add support for KuWFi C910

KuWFi C910 is an 802.11n (300N) indoor router with LTE support.

I can't find anywhere the OEM firmware. So if you want to restore the
original firmware you must do a dump before the OpenWrt flash.

According to the U-Boot, the board name is Iyunlink MINI_V2.

Hardware
--------
SoC:   Qualcomm QCA9533 650/400/200/25/25 MHz (CPU/RAM/AHB/SPI/REF)
RAM:   128 MB DDR2 16-bit CL3-4-4-10 (Nanya NT5TU64M16HG-AC)
FLASH: 16 MB Winbond W25Q128
ETH:
  - 2x 100M LAN (QCA9533 internal AR8229 switch, eth0)
  - 1x 100M WAN (QCA9533 internal PHY, eth1)
WIFI:
  - 2.4GHz: 1x QCA9533 2T2R (b/g/n)
  - 2 external non detachable antennas (near the power barrel side)
LTE:
  - Quectel EC200T-EU (or -CN or -AU depending on markets)
  - 2 external non detachable antennas (near the sim slot side)
BTN:
  - 1x Reset button
LEDS:
  - 5x White leds (Power, Wifi, Wan, Lan1, Lan2)
  - 1x RGB led (Internet)
UART: 115200-8-N-1 (Starting from lan ports in order: GND, RX, TX, VCC)

Everything works correctly.

MAC Addresses
-------------
LAN XX:XX:XX:XX:XX:48 (art@0x1002)
WAN XX:XX:XX:XX:XX:49 (art@0x1002 + 1)
WIFI XX:XX:XX:XX:XX:48

LABEL XX:XX:XX:XX:XX:48

Installation
------------
Turn the router on while pressing the reset button for 4 seconds.
You can simply count the flashes of the first lan led. (See notes)
If done correctly you should see the first lan led glowing slowly and
you should be able to enter the U-Boot web interface.

Click on the second tab ("固件") and select the -factory.bin firmware
then click "Update firmware".

A screen "Update in progress" should appear.

After few minutes the flash should be completed.

This procedure can be used also to recover the router in case of soft
brick.

Backup the original firmware
----------------------------
The following steps are intended for a linux pc. However using the
right software this guide should also work for Windows and MacOS.

1) Install a tftp server on your pc. For example tftpd-hpa.

2) Create two empty files in your tftp folder called:
kuwfi_c910_all_nor.bin
kuwfi_c910_firmware_only.bin

3) Give global write permissions to these files:
chmod 666 kuwfi_c910_all_nor.bin
chmod 666 kuwfi_c910_firmware_only.bin

4) Start a netcat session on your pc with this command:
nc -u -p 6666 192.168.1.1 6666

5) Set the static address on your pc: 192.168.1.2. Connect the router
to your pc.

6) Turn the router on while pressing the reset button for 8-9 seconds.
You can simply count the flashes of the first lan led. If you
press the reset button for too many seconds it will continue
the normal boot, so you have to restart the router. (See notes)

7) If done correctly you should see the U-Boot network console and you
should see the following lines on the netcat session:
Version and build date:
  U-Boot 1.1.4-55f1bca8-dirty, 2020-05-07

Modification by:
  Piotr Dymacz <piotr@dymacz.pl>
  https://github.com/pepe2k/u-boot_mod

u-boot>

8) Start the transfer of the whole NOR:
tftpput 0x9f000000 0x1000000 kuwfi_c910_all_nor.bin

9) The router should start the transfer and it should end with a
message like this (pay attention to the bytes transferred):
TFTP transfer complete!

Bytes transferred: 16777216 (0x1000000)

10) Repeat the same transfer for the firmware:
tftpput 0x9f050000 0xfa0000 kuwfi_c910_firmware_only.bin

11) The router should start the transfer and it should end with a
message like this (pay attention to the bytes transferred):
TFTP transfer complete!

Bytes transferred: 16384000 (0xfa0000)

12) Now you have the backup for the whole nor and for the firmware
partition. If you want to restore the OEM firmware from OpenWrt
you have to flash the kuwfi_c910_firmware_only.bin from the
U-Boot web interface.

WARNING: Don't use the kuwfi_c910_all_nor.bin file. This file
is only useful if you manage to hard brick the router or you
damage the art partition (ask on the forum)

Notes
-----
This router (or at least my unit) has the pepe2k's U-Boot. It's a
modded U-Boot version with a lot of cool features. You can read more
here: https://github.com/pepe2k/u-boot_mod

With this version of U-Boot, pushing the reset button while turning on
the router starts different tools:
 - 3-5 seconds: U-Boot web interface that can be used to replace the
  firmware, the art or the U-Boot itself
 - 5-7 seconds: U-Boot uart console
 - 7-10 seconds: U-Boot network console
 - 11+ seconds: Normal boot

The LTE modem can be used in cdc_ether (ECM) or RNDIS mode.
The default mode is ECM and in this commit only the ECM software is
included. In order to set RNDIS mode you must use this AT command:
AT+QCFG="usbnet",3
In order to use again the ECM mode you must use this AT command:
AT+QCFG="usbnet",1

Look for "Quectel_EC200T_Linux_USB_Driver_User_Guide_V1.0.pdf" for
other AT commands

Signed-off-by: Davide Fioravanti <pantanastyle@gmail.com>
16 months agoramips: improve support for SIM SIMAX1800T and Haier HAR-20S2U1
Shiji Yang [Mon, 12 Dec 2022 11:25:45 +0000 (11:25 +0000)]
ramips: improve support for SIM SIMAX1800T and Haier HAR-20S2U1

1. Use ARTIFACTS to build factory image. This change allows users to
   generate initramfs factory image using OpenWrt ImageBuilder.

2. Override the default bootargs property defined in "mt7621.dtsi".
   Although we use the "bootargs-override" property to set bootargs,
   the default "bootargs" property will still be written into the
   device tree, so it is better to override it.

Tested on SIM SIMAX1800T

Signed-off-by: Shiji Yang <yangshiji66@qq.com>
16 months agoramips: improve support for H3C TX180x series devices
Shiji Yang [Mon, 12 Dec 2022 11:18:41 +0000 (11:18 +0000)]
ramips: improve support for H3C TX180x series devices

1. Explicitly declare gpio pin groups to ensure that gpio works properly.

2. Override bootargs in device tree to avoid modifying u-boot envs during
   initial installation.

Tested on H3C TX1801 Plus

Signed-off-by: Shiji Yang <yangshiji66@qq.com>
16 months agosunxi: fix board_name for MarsBoard A10
Chukun Pan [Mon, 10 Oct 2022 15:08:27 +0000 (23:08 +0800)]
sunxi: fix board_name for MarsBoard A10

The compatible in the device tree is "haoyu,a10-marsboard",
modify the board_name to keep it consistent.

Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
16 months agosunxi: fix typo in device packages for MarsBoard A10
Chukun Pan [Sun, 9 Oct 2022 15:15:28 +0000 (23:15 +0800)]
sunxi: fix typo in device packages for MarsBoard A10

The kmod prefix for sound-soc-sunxi is missing, fix it.
Also add kmod-sound-core as dependence.

Fixes: 6a35659 ("sunxi: Added profile for HAOYU Electronics Marsboard A10")
Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
16 months agoath79: convert WiFis based on ar7241_ubnt_unifi.dtsi to nvmem-cells
Nick Hainke [Thu, 15 Dec 2022 16:04:14 +0000 (17:04 +0100)]
ath79: convert WiFis based on ar7241_ubnt_unifi.dtsi to nvmem-cells

Pull the calibration data from the nvmem subsystem. This allows us to
move userspace caldata extraction into the device-tree definition.

While working on it remove stale uboot partition label and merge art
into partition node.

Signed-off-by: Nick Hainke <vincent@systemli.org>
16 months agoath79: convert Buffalo WZR-HP-G302H A1A0 WiFis to nvmem-cells
Nick Hainke [Thu, 15 Dec 2022 14:17:26 +0000 (15:17 +0100)]
ath79: convert Buffalo WZR-HP-G302H A1A0 WiFis to nvmem-cells

Pull the calibration data from the nvmem subsystem. This allows us to
move userspace caldata extraction into the device-tree definition.

Merge art into partition node.

Signed-off-by: Nick Hainke <vincent@systemli.org>
16 months agoath79: convert OpenMesh OM2P v1 WiFis to nvmem-cells
Nick Hainke [Thu, 15 Dec 2022 17:18:16 +0000 (18:18 +0100)]
ath79: convert OpenMesh OM2P v1 WiFis to nvmem-cells

Pull the calibration data from the nvmem subsystem. This allows us to
move userspace caldata extraction into the device-tree definition.

Merge art into partition node.

Signed-off-by: Nick Hainke <vincent@systemli.org>
16 months agoath79: convert OpenMesh OM5P-AN WiFis to nvmem-cells
Nick Hainke [Thu, 15 Dec 2022 17:13:55 +0000 (18:13 +0100)]
ath79: convert OpenMesh OM5P-AN WiFis to nvmem-cells

Pull the calibration data from the nvmem subsystem. This allows us to
move userspace caldata extraction into the device-tree definition.

Merge art into partition node.

Signed-off-by: Nick Hainke <vincent@systemli.org>
16 months agoath79: convert boards based on ar9344_openmesh_mr600.dtsi to nvmem-cells
Nick Hainke [Thu, 15 Dec 2022 17:09:21 +0000 (18:09 +0100)]
ath79: convert boards based on ar9344_openmesh_mr600.dtsi to nvmem-cells

Pull the calibration data from the nvmem subsystem. This allows us to
move userspace caldata extraction into the device-tree definition.

Merge art into partition node.

Signed-off-by: Nick Hainke <vincent@systemli.org>
16 months agostrace: update to 6.1
Nick Hainke [Tue, 13 Dec 2022 19:05:07 +0000 (20:05 +0100)]
strace: update to 6.1

Release Notes:
https://github.com/strace/strace/releases/tag/v6.1

Signed-off-by: Nick Hainke <vincent@systemli.org>
16 months agotools/bash: update to 5.2.15
Linhui Liu [Thu, 15 Dec 2022 12:11:16 +0000 (20:11 +0800)]
tools/bash: update to 5.2.15

Update to the latest released version.

Signed-off-by: Linhui Liu <liulinhui36@gmail.com>
16 months agotrace-cmd: update to v3.1.5
Nick Hainke [Thu, 15 Dec 2022 13:42:35 +0000 (14:42 +0100)]
trace-cmd: update to v3.1.5

Update to latest release.

Signed-off-by: Nick Hainke <vincent@systemli.org>
16 months agolibtracefs: update to 1.6.3
Nick Hainke [Thu, 15 Dec 2022 13:40:55 +0000 (14:40 +0100)]
libtracefs: update to 1.6.3

Update to latest release.

Signed-off-by: Nick Hainke <vincent@systemli.org>
16 months agolibtraceevent: update to 1.7.0
Nick Hainke [Thu, 15 Dec 2022 13:38:50 +0000 (14:38 +0100)]
libtraceevent: update to 1.7.0

Update to latest release.

Signed-off-by: Nick Hainke <vincent@systemli.org>
16 months agokernel: Make use of KERNEL_MAKE
Hauke Mehrtens [Wed, 14 Dec 2022 21:34:19 +0000 (22:34 +0100)]
kernel: Make use of KERNEL_MAKE

Make use of KERNEL_MAKE in kernel packages were easily possible.
This moves some more code to common places and reduces the number of
lines.

It is defined like this:
KERNEL_MAKE = $(MAKE) $(KERNEL_MAKEOPTS)
KERNEL_MAKEOPTS = -C $(LINUX_DIR) $(KERNEL_MAKE_FLAGS)

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
16 months agokernel: Make KERNEL_MAKEOPTS recursively expanded
Hauke Mehrtens [Wed, 14 Dec 2022 15:29:16 +0000 (16:29 +0100)]
kernel: Make KERNEL_MAKEOPTS recursively expanded

KERNEL_MAKEOPTS will get expanded when it is used and not when it is
defined in the kernel.mk file now. This fixes problems finding dependent
kernel modules when it is used by a kernel module package.

Without this change the build of packages which depend on other out of
tree modules failed when they used KERNEL_MAKE because some symbols could
not be found. This happened because KERNEL_MAKE_FLAGS which contains a
"if $(__package_mk)" was  evaluated where KERNEL_MAKEOPTS was defined
and not when the KERNEL_MAKE was used. For packages which included
kernel.mk before package.mk we saw this problem. One workaround
was to use the correct include order and the other one was to not
use KERNEL_MAKE_FLAGS, but copy its content.

Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>