openwrt/staging/dangole.git
9 months agouboot-mediatek: fix patch order
Daniel Golle [Sun, 10 Mar 2024 13:29:14 +0000 (13:29 +0000)]
uboot-mediatek: fix patch order

Make sure patch sequence number is unique by moving patch
440-add-jdcloud_re-cp-03.patch -> 441-add-jdcloud_re-cp-03.patch

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
9 months agolldpd: Implement location parameter
Paul Donald [Thu, 8 Feb 2024 18:15:55 +0000 (19:15 +0100)]
lldpd: Implement location parameter

Previously only partially implemented. After commit
5007f488bb5e1b556052ef246bd3822fed8d09d2 lldp_location was never removed

Now, add the value of lldp_location to the generated config.

The location param has a few syntaxes, so the config acquires the first
usage from the man page: 'address country EU'

Supplementary fix for PR #14193 (this param was included in the original
PR #13018 but the lldp_location fixes were absent from PR #14193).

Tested on 22.03.5, 22.03.6

Signed-off-by: Paul Donald <newtwen@gmail.com>
9 months agolldpd: fix error "sh: XXXms: bad number"
Paul Donald [Thu, 8 Feb 2024 22:37:13 +0000 (23:37 +0100)]
lldpd: fix error "sh: XXXms: bad number"

from commit 3ce909914a12647bec52bcee0a162dd6d158a4f6

The lldpd man page says that "configure lldp tx-interval" can
specify an interval value in milliseconds by appending a "ms" suffix to
the figure. Thus mandating string handling, and not integer comparison.

Tested on 22.03.5

Signed-off-by: Paul Donald <newtwen@gmail.com>
9 months agolldpd: refactor out ifaces derivation; reuse function
Paul Donald [Thu, 8 Feb 2024 21:05:11 +0000 (22:05 +0100)]
lldpd: refactor out ifaces derivation; reuse function

from commit 909f0630663d7ae1471c76154516e3299992a150

Now pass two params to get_config_cid_ifaces() for:

cid_interface
interface

Each of which is a CSV of interfaces.

Tested on 22.03.5

Signed-off-by: Paul Donald <newtwen@gmail.com>
9 months agolldpd: remove unneeded quotes
Paul Donald [Thu, 8 Feb 2024 20:54:35 +0000 (21:54 +0100)]
lldpd: remove unneeded quotes

from commit a5f715da713304972467612d6934130ce3aa2837

Tested on 22.03.5

Signed-off-by: Paul Donald <newtwen@gmail.com>
9 months agolldpd: remove unneeded quotes and variable quoting
Paul Donald [Thu, 8 Feb 2024 20:44:56 +0000 (21:44 +0100)]
lldpd: remove unneeded quotes and variable quoting

from commit ac771313ebedd2c4bfda8adef47650d45d77c32d

portidsubtype takes 1 of 2 possible keywords which do not need quoting:

         configure lldp portidsubtype ifname | macaddress

The third keyword 'local' is used in the syntax when individual ports
are being defined:

         configure [ports ethX [,…]] lldp portidsubtype local value

When this syntax is used, quoting is useful (see test cases for lldpd).
In the init file, the 'local' syntax is unused.

Tested on 22.03.5

Signed-off-by: Paul Donald <newtwen@gmail.com>
9 months agolldpd: remove unneeded quotes and variable quoting
Paul Donald [Thu, 8 Feb 2024 20:42:11 +0000 (21:42 +0100)]
lldpd: remove unneeded quotes and variable quoting

from commit c98ee4dbb3db0f064d990941cdd82e872da76946

agent-type takes 1 of 3 possible keywords which do not require quoting:

         configure lldp agent-type nearest-bridge | nearest-non-tpmr-bridge
         | nearest-customer-bridge

Tested on 22.03.5

Signed-off-by: Paul Donald <newtwen@gmail.com>
9 months agolldpd: remove unneeded quotes and variable quoting
Paul Donald [Thu, 8 Feb 2024 20:29:37 +0000 (21:29 +0100)]
lldpd: remove unneeded quotes and variable quoting

from commit 3ce909914a12647bec52bcee0a162dd6d158a4f6

'capabilities enabled x' where x is a string of CSV

Tested on 22.03.5

Signed-off-by: Paul Donald <newtwen@gmail.com>
9 months agolldpd: remove unneeded quotes
Paul Donald [Thu, 8 Feb 2024 20:31:29 +0000 (21:31 +0100)]
lldpd: remove unneeded quotes

from commit 3ce909914a12647bec52bcee0a162dd6d158a4f6

Tested on 22.03.5

Signed-off-by: Paul Donald <newtwen@gmail.com>
9 months agolldpd: remove unneeded quotes and variable quoting
Paul Donald [Thu, 8 Feb 2024 20:31:57 +0000 (21:31 +0100)]
lldpd: remove unneeded quotes and variable quoting

from commit 24176a6bdd8f26040a97960868fd0d9ee968d695

Tested on 22.03.5

Signed-off-by: Paul Donald <newtwen@gmail.com>
9 months agolldpd: fix a paste error
Paul Donald [Thu, 8 Feb 2024 19:26:13 +0000 (20:26 +0100)]
lldpd: fix a paste error
from commit 1be2088a5247b2cfabe8be991c1e52ddaf780a16

The original PR #13018 did not exhibit this.

Tested on 22.03.5

Signed-off-by: Paul Donald <newtwen@gmail.com>
9 months agolldpd: spell fixes
Paul Donald [Thu, 8 Feb 2024 14:59:27 +0000 (15:59 +0100)]
lldpd: spell fixes

Supplementary fix for PR #14193

Tested on 22.03.5

Signed-off-by: Paul Donald <newtwen@gmail.com>
9 months agolldpd: fix -k 'lldp_no_version' row
Paul Donald [Thu, 8 Feb 2024 14:57:28 +0000 (15:57 +0100)]
lldpd: fix -k 'lldp_no_version' row

Supplementary fix for PR #14193 and commit
b67182008fd124706be0ec3ce67347447554ffd5

Tested on 22.03.5

Signed-off-by: Paul Donald <newtwen@gmail.com>
9 months agodnsmasq: version 2.90
Nathaniel Wesley Filardo [Sun, 18 Feb 2024 13:12:10 +0000 (13:12 +0000)]
dnsmasq: version 2.90

Bump to 2.90 to get upstream's fix for DNSSEC KeyTrap (CVE-2023-50387,
CVE-2023-50868) among many other goodies and fixes (notably, upstream
568fb024... fixes a UAF in cache_remove_uid that was routinely crashing
dnsmasq in my deployment).

Catch up our 200-ubus_dns.patch, too.

Signed-off-by: Nathaniel Wesley Filardo <nwfilardo@gmail.com>
9 months agoscripts: Kernel bumper script
Olliver Schinagl [Fri, 23 Feb 2024 21:54:48 +0000 (22:54 +0100)]
scripts: Kernel bumper script

For years, we have struggled and been frustrated at loosing history of
files in git, due to the 'copy + add' strategy. This could have been
prevented with a double-commit 'mv + add' trick.

On the mailing list [0] the discussion was started to put the
instructions in a wiki. Instead, it is much better to just script it and
put it in the repo.

Instead of doing mv + copy, which leads to two commits, but no history
on the copied files, it uses move, + copy and merge, which results in
three (merge) commits, but keeps the history of all files. As always
with renames, `--follow` will be needed.

The tool is trivial and works either in the OpenWrt git root directory,
or in the actual target directory.

Tested on the `realtek` and generic targets.

Note, that the tool does not do any of the labor needed after the move,
such as updating configs, dropping patches etc.

To make sure this script is easily found by any developer, who just
wants to do a kernel bump, the script is added here and not to
maintainer-tools repo as those scripts are a little bit more specialized.
Bumping a kernel is a trivial task that often regular developers do,
where most do not even know the existence of maintainer tools, are not
part of the main repo they'd clone, not part of the docker container
they'd use and so discoverability is probably much more important.

[0]: https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html

Signed-off-by: Olliver Schinagl <oliver@schinagl.nl>
Tested-by: Robert Marko <robimarko@gmail.com>
Tested-by: Weijie Gao <hackpascal@gmail.com>
9 months agobmips: add support for SmartRG SR505n
Kyle Hendry [Sun, 18 Feb 2024 20:26:38 +0000 (12:26 -0800)]
bmips: add support for SmartRG SR505n

Specifications:
- SoC: Broadcom BCM63168 dual 400MHz MIPS
- Flash: 16MB SPI NOR W25Q128WFG
- RAM: 128MB DDR3 W631GG6KB-15
- Ethernet: 1x 1000M, 3x 100M
- Wifi: BCM435F
- 1x USB 2.0 port
- 3x Button
- 12x LED

Flashing via serial
- Connect to the 3.3V TTL UART on the board
  (J6 pinout Vcc Rx Tx Gnd) at 115200-8-N-1
- Press any key in the serial console when powering up the board to enter
  the CFE prompt
- Configure an interface on your workstation to static IP 192.168.1.100
  and connect it to the board
- Start a TFTP server with the firmware image
- On the CFE prompt, enter the command
  "f 192.168.1.100:openwrt-bmips-bcm63268-smartrg_sr505n-squashfs-cfe.bin"

Signed-off-by: Kyle Hendry <kylehendrydev@gmail.com>
[Remove unneeded LED labels]
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
9 months agomac80211: select BRCMFMAC_SDIO on starfive
Zoltan HERPAI [Tue, 22 Aug 2023 21:39:14 +0000 (23:39 +0200)]
mac80211: select BRCMFMAC_SDIO on starfive

As the Visionfive V1 board has an Ampak module connected via SDIO, enable
support for SDIO in the brcmfmac module.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
9 months agostarfive: add new target for StarFive JH7100/7110 SoC
Zoltan HERPAI [Sun, 28 May 2023 13:14:02 +0000 (15:14 +0200)]
starfive: add new target for StarFive JH7100/7110 SoC

This target adds support for the StarFive JH7100 and JH7110 SoCs, based on
6.1, as well as a couple boards equipped with these.

Specifications:

SoCs:

JH7100:
 - StarFive JH7100 dual-core RISC-V (U74, RC64GC)
 - additional monitoring (S7) and control (E24) cores
 - 2Mb L2 cache

JH7110:
 - StarFive JH7110 quad-core RISC-V (U74, RV64GC)
 - additional monitoring (S7) and control (E24) cores
 - 2Mb L2 cache

Boards:

VisionFive1:
 - JH7100 @ 1GHz
 - Memory: 8Gb LPDDR4
 - 4x USB3.0
 - 1x GBit ethernet
 - AMPak 6236 wifi / bluetooth
 - audio
 - powered via USB-C

VisionFive2:
 - JH7110 @ 1.5GHz
 - Memory: 2/4/8Gb DDR4
 - 2x Gbit ethernet
 - 2x USB3.0 / 2x USB2.0
 - eMMC / SDIO
 - various multimedia input/outputs (MIPI CSI, HDMI, audio)
 - M.2 key M slot
 - PoE support
 - powered via USB-C

Installation:
Standard SD-card installation via dd-ing the generated image to
an SD-card of at least 256Mb.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
9 months agousb: add cdns3 support
Zoltan HERPAI [Mon, 1 Jan 2024 12:50:46 +0000 (13:50 +0100)]
usb: add cdns3 support

CDNS3 is a SuperSpeed (SS) USB 3.0 Dual-Role-Device (DRD) controller from
Cadence. Add support for this device, and add the required symbols into
the generic configs.

Compile-tested: apm821xx, bcm4908, imx, mpc85xx, pistachio, starfive

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
9 months agoramips: improve support for STORYLiNK SAP-G3200U3
Sungbo Eo [Sun, 10 Mar 2024 04:55:24 +0000 (13:55 +0900)]
ramips: improve support for STORYLiNK SAP-G3200U3

* Increase flash SPI frequency to 50MHz

  The maximum SPI frequency of MX25L6406EM2I is 86 MHz,
  but in this patch 50 MHz was chosen as a safe value.

* Update Ethernet MAC addresses

  Till now LAN/WAN MAC addresses were flipped
  compared to stock firmware.

* Fix Wi-Fi LEDs by adding mt76 led nodes

* Fix LAN port order

  LAN ports are in reverse order of switch ports.

* Fix the well-known "LZMA ERROR 1" error by using lzma-loader

* Set uImage name, which enables installation via stock web interface:

  1.  Upload **initramfs** image file to the web page.
  2.  Boot into OpenWrt and perform sysupgrade with sysupgrade image.

Signed-off-by: Sungbo Eo <mans0n@gorani.run>
9 months agoramips: add factory image for ipTIME AX2004M
Sungbo Eo [Sun, 4 Feb 2024 18:57:27 +0000 (03:57 +0900)]
ramips: add factory image for ipTIME AX2004M

Unlike the recovery image, this initramfs-factory image can be flashed
using the stock firmware web interface (from any active boot partition),
as well as the bootloader recovery web page. Drop the recovery image in
favor of the factory image.

Installation via stock/recovery web interface:
1.  Flash **initramfs-factory** image through the web page.
2.  Boot into OpenWrt and perform sysupgrade with sysupgrade image.

Signed-off-by: Sungbo Eo <mans0n@gorani.run>
9 months agoramips: mt7621: use variable for relocate-kernel argument
Sungbo Eo [Sun, 4 Feb 2024 13:10:29 +0000 (22:10 +0900)]
ramips: mt7621: use variable for relocate-kernel argument

The "0x80001000" address logically comes from "loadaddr-y" variable for
mt7621 subtarget. Let's replace the hardcoded value with the predefined
variable. This change is purely cosmetic.

Signed-off-by: Sungbo Eo <mans0n@gorani.run>
9 months agoramips: rename mtd partition of ipTIME NAND devices
Sungbo Eo [Sun, 4 Feb 2024 12:46:14 +0000 (21:46 +0900)]
ramips: rename mtd partition of ipTIME NAND devices

Contrary to common ipTIME NOR devices, the "Config" partition of T5004
and AX2004M contain normal U-Boot environment variables. Renaming the
partition into "u-boot-env" serves for better description, and it also
conforms to common naming practice in OpenWrt.

This patch might also be extended to A3004T, but its u-boot-env
partition layout has not been confirmed yet.

Signed-off-by: Sungbo Eo <mans0n@gorani.run>
9 months agoramips: add NMBM support for ipTIME AX2004M
Sungbo Eo [Sun, 17 Dec 2023 16:06:21 +0000 (01:06 +0900)]
ramips: add NMBM support for ipTIME AX2004M

AX2004M uses NMBM on its NAND flash, but it was not enabled in DTS as the
device support [1] had been added before NMBM feature in mtk_bmt driver [2].
Let's enable it now.

With this change, there is a low possibility of boot failure after
sysupgrade from older versions. As AX2004M already has gone through
two stable releases in the meantime, it would be safe to warn users
by bumping DEVICE_COMPAT_VERSION.

[1] 37753f34ac68 ("ramips: add support for ipTIME AX2004M")
[2] 06382d1af7b2 ("kernel: add support for mediatek NMBM flash mapping support")

Signed-off-by: Sungbo Eo <mans0n@gorani.run>
9 months agomac80211: rtl8xxxu: sync with linux-next 20240229
Shiji Yang [Thu, 22 Feb 2024 21:36:50 +0000 (05:36 +0800)]
mac80211: rtl8xxxu: sync with linux-next 20240229

Backporting upstream patches to improve RTL8188F support.

Signed-off-by: Shiji Yang <yangshiji66@qq.com>
9 months agofirmware: add firmware package for Realtek RTL8188FU
Shiji Yang [Thu, 22 Feb 2024 01:56:36 +0000 (09:56 +0800)]
firmware: add firmware package for Realtek RTL8188FU

Realtek RTL8188F is an 802.11n 1x1 USB Wi-Fi adapter. It has been
supported by the upstream rtl8xxxu driver since Linux 6.2 kernel.

Signed-off-by: Shiji Yang <yangshiji66@qq.com>
9 months agoath79: add support for D-Link COVR-C1200 A1
Sebastian Schaper [Sun, 11 Feb 2024 17:55:43 +0000 (18:55 +0100)]
ath79: add support for D-Link COVR-C1200 A1

The COVR-C1200 devices are sold as "Whole Home Mesh Wi-Fi"
sets in packs of two (COVR-C1202) and three (COVR-C1203).

Specifications:
 * QCA9563, 16 MiB flash, 128 MiB RAM, 2x3:2 802.11n
 * QCA9886 2x2:2 801.11ac Wave 2
 * AR8337, 2 Gigabit ports (1: WAN; 2: LAN)
 * USB Type-C power connector (5V, 3A)

Installation COVR Point A:
 * In factory reset state: OEM Web UI is at 192.168.0.50
   no DHCP, skip wizard by directly accessing:
     http://192.168.0.50/UpdateFirmware_Simple.html
 * After completing setup wizard: Web UI is at 192.168.0.1
     DHCP enabled, login with empty password
 * Flash factory.bin
 * Perform a factory reset to restore OpenWrt UCI defaults

Installation COVR Points B:
 * OEM Web UI is at 192.168.0.50, no DHCP, empty password
 * Flash factory.bin
 * Perform a factory reset to restore OpenWrt UCI defaults

Recovery:
 * Keep reset button pressed during power on
 * Recovery Web UI is at 192.168.0.50, no DHCP
 * Flash factory.bin
   used to work best with Chromium-based browsers or curl:
     curl -F firmware=@factory.bin \
       http://192.168.0.50/upgrade.cgi
   since this fails to work on modern Linux systems,
   there is also a script dlink_recovery_upload.py

Signed-off-by: Sebastian Schaper <openwrt@sebastianschaper.net>
9 months agoath79: split dtsi for D-Link COVR-P2500
Sebastian Schaper [Sun, 11 Feb 2024 17:53:07 +0000 (18:53 +0100)]
ath79: split dtsi for D-Link COVR-P2500

in preparation of adding COVR-C1200

Signed-off-by: Sebastian Schaper <openwrt@sebastianschaper.net>
9 months agofirmware: intel-microcode: update to 20231114
Christian Lamparter [Fri, 8 Mar 2024 20:16:44 +0000 (21:16 +0100)]
firmware: intel-microcode: update to 20231114

Debian changelog:

intel-microcode (3.20231114.1) unstable; urgency=medium

  * New upstream microcode datafile 20231114 (closes: #1055962)
    Mitigations for "reptar", INTEL-SA-00950 (CVE-2023-23583)
    Sequence of processor instructions leads to unexpected behavior for some
    Intel(R) Processors, may allow an authenticated user to potentially enable
    escalation of privilege and/or information disclosure and/or denial of
    service via local access.
    Note: "retvar" on 4th gen Xeon Scalable (sig 0x806f8 pfm 0x87), 12th gen
    Core mobile (sig 0x906a4 pfm 0x80), 13th gen Core desktop (sig 0xb0671 pfm
    0x01) were already mitigated by a previous microcode update.
  * Fixes for unspecified functional issues
  * Updated microcodes:
    sig 0x000606a6, pf_mask 0x87, 2023-09-01, rev 0xd0003b9, size 299008
    sig 0x000606c1, pf_mask 0x10, 2023-09-08, rev 0x1000268, size 290816
    sig 0x000706e5, pf_mask 0x80, 2023-09-03, rev 0x00c2, size 113664
    sig 0x000806c1, pf_mask 0x80, 2023-09-07, rev 0x00b4, size 111616
    sig 0x000806c2, pf_mask 0xc2, 2023-09-07, rev 0x0034, size 98304
    sig 0x000806d1, pf_mask 0xc2, 2023-09-07, rev 0x004e, size 104448
    sig 0x000806f8, pf_mask 0x87, 2023-06-16, rev 0x2b0004d0, size 572416
    sig 0x000806f8, pf_mask 0x87, 2023-06-16, rev 0x2b0004d0
    sig 0x000806f7, pf_mask 0x87, 2023-06-16, rev 0x2b0004d0
    sig 0x000806f6, pf_mask 0x87, 2023-06-16, rev 0x2b0004d0
    sig 0x000806f5, pf_mask 0x87, 2023-06-16, rev 0x2b0004d0
    sig 0x000806f4, pf_mask 0x87, 2023-06-16, rev 0x2b0004d0
    sig 0x000806f8, pf_mask 0x10, 2023-06-26, rev 0x2c000290, size 605184
    sig 0x000806f8, pf_mask 0x10, 2023-06-26, rev 0x2c000290
    sig 0x000806f6, pf_mask 0x10, 2023-06-26, rev 0x2c000290
    sig 0x000806f5, pf_mask 0x10, 2023-06-26, rev 0x2c000290
    sig 0x000806f4, pf_mask 0x10, 2023-06-26, rev 0x2c000290
    sig 0x00090672, pf_mask 0x07, 2023-06-07, rev 0x0032, size 222208
    sig 0x00090672, pf_mask 0x07, 2023-06-07, rev 0x0032
    sig 0x00090675, pf_mask 0x07, 2023-06-07, rev 0x0032
    sig 0x000b06f2, pf_mask 0x07, 2023-06-07, rev 0x0032
    sig 0x000b06f5, pf_mask 0x07, 2023-06-07, rev 0x0032
    sig 0x000906a3, pf_mask 0x80, 2023-06-07, rev 0x0430, size 220160
    sig 0x000906a3, pf_mask 0x80, 2023-06-07, rev 0x0430
    sig 0x000906a4, pf_mask 0x80, 2023-06-07, rev 0x0430
    sig 0x000906a4, pf_mask 0x40, 2023-05-05, rev 0x0005, size 117760
    sig 0x000a0671, pf_mask 0x02, 2023-09-03, rev 0x005d, size 104448
    sig 0x000b0671, pf_mask 0x32, 2023-08-29, rev 0x011d, size 210944
    sig 0x000b06a2, pf_mask 0xe0, 2023-08-30, rev 0x411c, size 216064
    sig 0x000b06a2, pf_mask 0xe0, 2023-08-30, rev 0x411c
    sig 0x000b06a3, pf_mask 0xe0, 2023-08-30, rev 0x411c
    sig 0x000b06e0, pf_mask 0x11, 2023-06-26, rev 0x0012, size 136192
  * Updated 2023-08-08 changelog entry:
    Mitigations for "retvar" on a few processors, refer to the 2023-11-14
    entry for details.  This information was disclosed in 2023-11-14.
  * source: update symlinks to reflect id of the latest release, 20231114

 -- Henrique de Moraes Holschuh <hmh@debian.org>  Thu, 16 Nov 2023 08:09:43 -0300

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
9 months agoRevert "ipq-wifi: fix upstream board-2.bin ZTE M289F snafu"
Christian Lamparter [Fri, 8 Mar 2024 20:12:45 +0000 (21:12 +0100)]
Revert "ipq-wifi: fix upstream board-2.bin ZTE M289F snafu"

This reverts commit 75505c5ec724b9b961dcb411bac1d4b9aede3e1d.
The issue has been fixed upstream.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
9 months agomediatek: linksys-e8450: improve compat warning
Daniel Golle [Sat, 9 Mar 2024 13:50:40 +0000 (13:50 +0000)]
mediatek: linksys-e8450: improve compat warning

Include a statement about having to run the installer in the
sysupgrade compat warning for the Linksys E8450 (UBI).

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
9 months agomediatek: wait for fitblk rootfs
Daniel Golle [Sat, 9 Mar 2024 13:48:03 +0000 (13:48 +0000)]
mediatek: wait for fitblk rootfs

Probing of the fitblk driver in some situations happens after the kernel
attempts to mount rootfs, which then fails.
Always use 'rootwait' when using fitblk for rootfs.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
9 months agouboot-mediatek: add 'rootwait' to bootargs where needed
Daniel Golle [Sat, 9 Mar 2024 13:58:29 +0000 (13:58 +0000)]
uboot-mediatek: add 'rootwait' to bootargs where needed

Probing of the fitblk driver in some situations happens after Linux
attempts to mount rootfs, which then fails.
Always use 'rootwait' kernel parameter when using fitblk for rootfs.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
9 months agomac80211: fix a regression in the broadcast AQL patch
Felix Fietkau [Fri, 8 Mar 2024 21:42:50 +0000 (22:42 +0100)]
mac80211: fix a regression in the broadcast AQL patch

The AQL limit for buffered broadcast packets is higher than the maximum
total pending airtime limit. This can get unicast data stuck whenever there
is too much pending broadcast data. Fix this by excluding broadcast AQL from
the total limit.

Signed-off-by: Felix Fietkau <nbd@nbd.name>
9 months agombedtls: enable threading support
Felix Fietkau [Tue, 5 Mar 2024 18:13:52 +0000 (19:13 +0100)]
mbedtls: enable threading support

Fixes libssh, which requires it. Bump ABI_VERSION, since enabling this
option affects data structures in mbedtls include files.

Signed-off-by: Felix Fietkau <nbd@nbd.name>
9 months agoath79: cleanup and enable uart1 on ELECOM WAB-I1750-PS
INAGAKI Hiroshi [Mon, 4 Mar 2024 10:12:07 +0000 (19:12 +0900)]
ath79: cleanup and enable uart1 on ELECOM WAB-I1750-PS

Cleanup uart1-related node on ELECOM WAB-I1750-PS and enable it for
"SERIAL" port on the case.

"SERIAL" port can be used for OpenWrt console by adding the following
line to /etc/inittab and rebooting:

ttyATH1::askfirst:/usr/libexec/login.sh

Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
9 months agoath79: drop unnecessary &uart node from dtsi of QCA955x devices
INAGAKI Hiroshi [Mon, 4 Mar 2024 10:11:36 +0000 (19:11 +0900)]
ath79: drop unnecessary &uart node from dtsi of QCA955x devices

&uart (uart0) is already enabled in qca955x.dtsi by default.

Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
9 months agoath79: add aliases to qca955x.dtsi for uart0/1
INAGAKI Hiroshi [Mon, 4 Mar 2024 11:28:23 +0000 (20:28 +0900)]
ath79: add aliases to qca955x.dtsi for uart0/1

Add aliases with "serialN = &uartN;" of uart0/1 on QCA955x SoCs to
qca955x.dtsi, to enable uart1 on Linux Kernel.

without this:

[    0.342915] ar933x-uart 18500000.uart: unable to get alias id, err=-19

Additionally, remove "serial0 = &uart;" alias from QCA955x device
dts/dtsi files.

Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
9 months agoath79: add HighSpeed UART (uart1) support for QCA955x
INAGAKI Hiroshi [Mon, 4 Mar 2024 10:08:07 +0000 (19:08 +0900)]
ath79: add HighSpeed UART (uart1) support for QCA955x

Add HighSpeed UART support to QCA955x series SoCs as a secondary UART
(uart1). This UART is compatible with qca,ar9330-uart.

Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
9 months agoath79: rename label of primary UART on QCA955x to "uart0"
INAGAKI Hiroshi [Mon, 4 Mar 2024 10:07:05 +0000 (19:07 +0900)]
ath79: rename label of primary UART on QCA955x to "uart0"

Rename the DT label of the primary UART on Qualcomm Atheros QCA955x
series SoCs to "uart0" from "uart" for the preparation to add HighSpeed
UART (uart1) support.

Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
9 months agompc85xx: p1020: convert Aerohive AP330/AP350 to simpleImage
Christian Lamparter [Thu, 7 Mar 2024 21:43:41 +0000 (22:43 +0100)]
mpc85xx: p1020: convert Aerohive AP330/AP350 to simpleImage

with 6.1, the kernel no longer fitted into the 16 MiB and
kicking down the can and increasing KERNEL_SIZE to 20 MiB
didn't help as the device failed to boot.

Using 'kernel-bin | gzip | uimage gzip' didn't work since the
uboot does not have enough heap to decompress these big kernels.

And finally playing around with uboot was more a hassle than
converting this device to take the simpleImage-boot-route in
the future.

Note: The device now takes even longer on the first boot-up after
the flash due to JFFS2 initializing all the remaining flash.
Be prepared to wait up to 10 minutes before the green status LED
stops blinking and will shine a solid green!

(On the plus site: the device now has ~10 MiB of additional
space for rootfs+rootfs_data).

Note2: This patch includes a kernel patch refresh.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
9 months agompc85xx: add 6.1 testing kernel
Pawel Dembicki [Tue, 21 Nov 2023 09:50:23 +0000 (10:50 +0100)]
mpc85xx: add 6.1 testing kernel

Configure 6.1 testing kernel for mpc85xx target.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
9 months agompc85xx: p1010: change wrapper address of simple image devices
Pawel Dembicki [Mon, 19 Feb 2024 12:39:10 +0000 (13:39 +0100)]
mpc85xx: p1010: change wrapper address of simple image devices

Kernel and initramfs size grows. Now uncompressed initramfs image and
regular kernel image overlaps configured area:

Initramfs:
  WRAP    arch/powerpc/boot/simpleImage.br200-wp
INFO: Uncompressed kernel (size 0x1428688) overlaps the address of the
wrapper(0x1000000)
INFO: Fixing the link_address of wrapper to (0x1500000)
  WRAP    arch/powerpc/boot/simpleImage.tl-wdr4900-v1
INFO: Uncompressed kernel (size 0x1428688) overlaps the address of the
wrapper(0x1000000)
INFO: Fixing the link_address of wrapper to (0x1500000)
  WRAP    arch/powerpc/boot/simpleImage.ws-ap3715i
INFO: Uncompressed kernel (size 0x1428688) overlaps the address of the
wrapper(0x1000000)
INFO: Fixing the link_address of wrapper to (0x1500000)

Regular image:
  WRAP    arch/powerpc/boot/simpleImage.br200-wp
INFO: Uncompressed kernel (size 0x10e0688) overlaps the address of the
wrapper(0x1000000)
INFO: Fixing the link_address of wrapper to (0x1100000)
  WRAP    arch/powerpc/boot/simpleImage.tl-wdr4900-v1
INFO: Uncompressed kernel (size 0x10e0688) overlaps the address of the
wrapper(0x1000000)
INFO: Fixing the link_address of wrapper to (0x1100000)
  WRAP    arch/powerpc/boot/simpleImage.ws-ap3715i
INFO: Uncompressed kernel (size 0x10e0688) overlaps the address of the
wrapper(0x1000000)
INFO: Fixing the link_address of wrapper to (0x1100000)

Let's change wrapper address to safe value.

Tested on: TL-WDR4900, BR200-WP

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
9 months agompc85xx: kernel: refresh 6.1 config
Pawel Dembicki [Tue, 21 Nov 2023 09:45:22 +0000 (10:45 +0100)]
mpc85xx: kernel: refresh 6.1 config

Kernel config needs to be refreshed after copy from 5.15.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
9 months agompc85xx: refresh 6.1 patches
Pawel Dembicki [Tue, 21 Nov 2023 09:43:46 +0000 (10:43 +0100)]
mpc85xx: refresh 6.1 patches

Most changes done by "make target/linux/refresh"

Manually refreshed:
102-powerpc-add-cmdline-override.patch

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
9 months agompc85xx: copy patches 5.15 to 6.1
Pawel Dembicki [Tue, 21 Nov 2023 08:38:37 +0000 (09:38 +0100)]
mpc85xx: copy patches 5.15 to 6.1

Just copy of config.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
9 months agompc85xx: copy config 5.15 to 6.1
Pawel Dembicki [Tue, 21 Nov 2023 08:38:20 +0000 (09:38 +0100)]
mpc85xx: copy config 5.15 to 6.1

Simple copy of patches.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
9 months agompc85xx: Get rid of 'of_flat_dt_is_compatible'
Pawel Dembicki [Tue, 21 Nov 2023 09:46:13 +0000 (10:46 +0100)]
mpc85xx: Get rid of 'of_flat_dt_is_compatible'

This patch replaces 'of_flat_dt_is_compatible' with 'of_machine_is_compatible'.
The TL-WDR4900 platform file won't compile in the 6.1 kernel. The platform
files for the rest of the routers have been reworked or based on newer
solutions.

Let's make the TL-WDR4900 consistent with them.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
9 months agokernel: crypto: use ARM64 SHA256 CE optimized module for more targets
Robert Marko [Wed, 6 Mar 2024 21:21:47 +0000 (22:21 +0100)]
kernel: crypto: use ARM64 SHA256 CE optimized module for more targets

At start I only set qualcommax to use the Crypto Extensions optimized
version of SHA256 as I knew it supports the optional Crypto Extensions.

However, after looking into the tree there are more targets/subtargets
that I could find at least a specification sheet that says the support
Cryptographic Extensions, so lets add them.

Signed-off-by: Robert Marko <robimarko@gmail.com>
9 months agokernel: crypto: use ARM64 SHA1 CE optimized module for more targets
Robert Marko [Wed, 6 Mar 2024 21:19:54 +0000 (22:19 +0100)]
kernel: crypto: use ARM64 SHA1 CE optimized module for more targets

At start I only set qualcommax to use the Crypto Extensions optimized
version of SHA1 as I knew it supports the optional Crypto Extensions.

However, after looking into the tree there are more targets/subtargets
that I could find at least a specification sheet that says the support
Cryptographic Extensions, so lets add them.

Signed-off-by: Robert Marko <robimarko@gmail.com>
9 months agolantiq: Fix build after kernel 5.15.150
Hauke Mehrtens [Fri, 8 Mar 2024 13:43:08 +0000 (14:43 +0100)]
lantiq: Fix build after kernel 5.15.150

This fixes the following compile problem:
````
  CC      arch/mips/kernel/vpe-mt.o
arch/mips/kernel/vpe-mt.c: In function 'vpe_run':
arch/mips/kernel/vpe-mt.c:32:23: error: unused variable 'physical_memsize' [-Werror=unused-variable]
   32 |         unsigned long physical_memsize = 0L;
      |                       ^~~~~~~~~~~~~~~~
cc1: all warnings being treated as errors
make[6]: *** [scripts/Makefile.build:289: arch/mips/kernel/vpe-mt.o] Error 1
````

physical_memsize was removed from upstream kernel, see:
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-5.15.y&id=5b4f6c5ff65c8551018ccea40c569afd759734c0

Fixes: 387fde0da0e8 ("kernel: bump 5.15 to 5.15.150")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
9 months agolibiwinfo: update to Git HEAD (2024-03-08)
Christian Marangi [Fri, 8 Mar 2024 13:25:54 +0000 (14:25 +0100)]
libiwinfo: update to Git HEAD (2024-03-08)

163a640fef30 devices: add device id for Qualcomm Atheros QCA6174
8ffb8bfd1115 devices: add add Qualcomm Atheros IPQ6018 WiSoC compatible

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
9 months agomediatek: clean duplicated #include in Xiaomi Redmi AX6000's dts
Furong Xu [Fri, 8 Mar 2024 08:43:00 +0000 (16:43 +0800)]
mediatek: clean duplicated #include in Xiaomi Redmi AX6000's dts

Clean duplication of #include <dt-bindings/leds/common.h>.
Thanks musashino205

Fixes: 1493e8f8cbe2 ("mediatek: convert LED color/function format for Xiaomi Redmi AX6000")
Signed-off-by: Furong Xu <xfr@outlook.com>
9 months agokernel: enable KASAN option for more architectures
Qingfang Deng [Fri, 8 Mar 2024 03:07:49 +0000 (11:07 +0800)]
kernel: enable KASAN option for more architectures

KASAN has supported more architectures, such as ARM, PPC32 and RISC-V 64.
Enable KASAN option for those architectures.

Signed-off-by: Qingfang Deng <dqfext@gmail.com>
9 months agoubox: make logread as an alternative
Sergey Ponomarev [Sat, 2 Mar 2024 21:50:58 +0000 (23:50 +0200)]
ubox: make logread as an alternative

The logread can be replaced with syslog-ng.
To support this it should be an alternative itself.

Signed-off-by: Sergey Ponomarev <stokito@gmail.com>
9 months agomediatek: filogic: switch to fitblk for Xiaomi Redmi AX6000
Furong Xu [Thu, 7 Mar 2024 02:56:18 +0000 (10:56 +0800)]
mediatek: filogic: switch to fitblk for Xiaomi Redmi AX6000

Use the new fitblk driver.

Run-tested: filogic/mt7986a-xiaomi-redmi-router-ax6000-ubootmod

Signed-off-by: Furong Xu <xfr@outlook.com>
9 months agomediatek: convert LED color/function format for Xiaomi Redmi AX6000
Furong Xu [Mon, 4 Mar 2024 12:54:44 +0000 (20:54 +0800)]
mediatek: convert LED color/function format for Xiaomi Redmi AX6000

Commit 2d63d42f5e2f ("mediatek: convert to new LED color/function
format where possible") leaves Xiaomi Redmi AX6000 un-converted,
the two LEDs become dead.
Now, LEDs are alive again.

Fixes: 2d63d42f5e2f ("mediatek: convert to new LED color/function
format where possible")

Signed-off-by: Furong Xu <xfr@outlook.com>
9 months agokernel: 6.1: add missing config symbol for tracing
Tony Ambardar [Thu, 7 Mar 2024 09:21:40 +0000 (01:21 -0800)]
kernel: 6.1: add missing config symbol for tracing

Add missing option CONFIG_FPROBE to generic config, found building x86_64
arch with tracing enabled.

Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
9 months agolinux-firmware: Update to version 20240220
Hauke Mehrtens [Thu, 22 Feb 2024 20:33:45 +0000 (21:33 +0100)]
linux-firmware: Update to version 20240220

This updates the following firmware files:
airoha-en8811h-firmware/lib/firmware/airoha/EthMD32.DSP.bin
airoha-en8811h-firmware/lib/firmware/airoha/EthMD32.dm.bin
amdgpu-firmware (Many files)
ar3k-firmware/lib/firmware/qca/hpnv21.bin
ar3k-firmware/lib/firmware/qca/hpnv21g.bin
ath10k-board-qca4019/lib/firmware/ath10k/QCA4019/hw1.0/board-2.bin
ath10k-board-qca9888/lib/firmware/ath10k/QCA9888/hw2.0/board-2.bin
ath10k-firmware-qca6174/lib/firmware/ath10k/QCA6174/hw3.0/firmware-6.bin
ath11k-firmware-wcn6750/lib/firmware/ath11k/WCN6750/hw1.0/board-2.bin
ath11k-firmware-wcn6855/lib/firmware/ath11k/WCN6855/hw2.0/amss.bin
ath11k-firmware-wcn6855/lib/firmware/ath11k/WCN6855/hw2.0/board-2.bin
ath11k-firmware-wcn6855/lib/firmware/ath11k/WCN6855/hw2.1/amss.bin
ath11k-firmware-wcn6855/lib/firmware/ath11k/WCN6855/hw2.1/board-2.bin
ibt-firmware/lib/firmware/intel/ibt-0040-0041.ddc
ibt-firmware/lib/firmware/intel/ibt-0040-0041.sfi
ibt-firmware/lib/firmware/intel/ibt-0040-1050.sfi
ibt-firmware/lib/firmware/intel/ibt-0040-4150.sfi
ibt-firmware/lib/firmware/intel/ibt-0041-0041.sfi
ibt-firmware/lib/firmware/intel/ibt-0180-0041.sfi
ibt-firmware/lib/firmware/intel/ibt-0180-1050.sfi
ibt-firmware/lib/firmware/intel/ibt-0180-4150.sfi
ibt-firmware/lib/firmware/intel/ibt-1040-0041.sfi
ibt-firmware/lib/firmware/intel/ibt-1040-4150.sfi
iwlwifi-firmware-be200/lib/firmware/iwlwifi-gl-c0-fm-c0-83.ucode
iwlwifi-firmware-be200/lib/firmware/iwlwifi-gl-c0-fm-c0.pnvm

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
9 months agokernel: Remove unused schedulers
Hauke Mehrtens [Sun, 3 Mar 2024 11:34:19 +0000 (12:34 +0100)]
kernel: Remove unused schedulers

These schedulers were removed in kernel 5.15.150 and 6.1.180.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
9 months agokernel: bump 5.15 to 5.15.150
Hauke Mehrtens [Sun, 3 Mar 2024 11:25:27 +0000 (12:25 +0100)]
kernel: bump 5.15 to 5.15.150

Removed because it is upstream:
bcm53xx/patches-5.15/030-v5.16-0019-ARM-dts-BCM53573-Describe-on-SoC-BCM53125-rev-4-swit.patch
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=cb1003c07e746e4e82bdd3959c9ea37018ed41a3

Removed because it is upstream:
bcm53xx/patches-5.15/037-v6.6-0004-ARM-dts-BCM53573-Drop-nonexistent-default-off-LED-tr.patch
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=c65a23e98e38dc991f495d6bdb3cfa6163a88a0c

Removed because it is upstream:
bcm53xx/patches-5.15/037-v6.6-0005-ARM-dts-BCM53573-Drop-nonexistent-usb-cells.patch
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=71475bcee001cae3844644c2787eef93b26489d1

Adapted hack-5.15/650-netfilter-add-xt_FLOWOFFLOAD-target.patch to match
the changes from the upstream flow offload patch:
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=7c71b831220edeab7ce603d818dc1708d9ea4137

Manually Adapted the following patch:
bcm53xx/patches-5.15/035-v6.2-0004-ARM-dts-broadcom-align-LED-node-names-with-dtschema.patch

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
9 months agokernel: bump 5.15 to 5.15.149
Hauke Mehrtens [Sat, 2 Mar 2024 18:14:22 +0000 (19:14 +0100)]
kernel: bump 5.15 to 5.15.149

Removed because it is upstream:
generic/pending-5.15/540-ksmbd-only-v2-leases-handle-the-directory.patch
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=81ea755b8fa57ce186f56f14b709cf491c2a5505

Manually Adapted the following patch:
ath79/patches-5.15/900-unaligned_access_hacks.patch

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
9 months agoath25: drop target
Robert Marko [Tue, 5 Mar 2024 09:31:48 +0000 (10:31 +0100)]
ath25: drop target

ath25 has been on life support for the last couple of releases, eventually
leading to marking it as source-only in 2023.

It has been basically only touched to do a kernel bump so that we can make
the new OpenWrt release which was a challenge due to small RAM amount.

However, with the attempt of kernel 6.1 update it turns out that kernel
cannot even finish booting due to RAM constraints, so its time to let this
target go.

Signed-off-by: Robert Marko <robimarko@gmail.com>
9 months agoath11k-firmware: Move from kvalo to new upstream repository
Sven Eckelmann [Thu, 7 Mar 2024 08:24:20 +0000 (09:24 +0100)]
ath11k-firmware: Move from kvalo to new upstream repository

It was announced [1] that the original staging repositories are no longer
used for staging of new firmware binaries. And that the old repository will
be removed [2] in June 2024.

The ath11k-firmware package must therefore point to the new repository
before the old one is no longer accessible.

[1] https://lore.kernel.org/r/bac97f31-4a70-4c4c-8179-4ede0b32f869@quicinc.com
[2] https://github.com/kvalo/ath11k-firmware/commit/8d2cc160f390badd62970a66483214773c3fbea1

Signed-off-by: Sven Eckelmann <sven@narfation.org>
9 months agokernel: crypto: add SHA512 ARM64 ASM optimized module
Robert Marko [Wed, 6 Mar 2024 20:54:07 +0000 (21:54 +0100)]
kernel: crypto: add SHA512 ARM64 ASM optimized module

Kernel has an ASM optimized version of SHA512 that was ported from
OpenSSL, so lets package it as it provides significant perfomance
improvement compared to the generic implementation.

There is a Cryptographic Extension based version as well, but that relies
on ARMv8.2 ISA which I am not aware any of the OpenWrt supported SoC-s use.

Signed-off-by: Robert Marko <robimarko@gmail.com>
9 months agokernel: crypto: add SHA2(224 and 256) ARM64 CE optimized module
Robert Marko [Wed, 6 Mar 2024 17:38:28 +0000 (18:38 +0100)]
kernel: crypto: add SHA2(224 and 256) ARM64 CE optimized module

Kernel has optimized version of SHA2(224 and 256) using the ARMv8 Crypto
Extensions, so lets package it.

Use it by default for qualcommax as it uses Cortex-A53 core and has ARMv8
CE extensions present.

Signed-off-by: Robert Marko <robimarko@gmail.com>
9 months agokernel: crypto: add SHA256 ARM64 ASM optimized module
Robert Marko [Wed, 6 Mar 2024 17:30:21 +0000 (18:30 +0100)]
kernel: crypto: add SHA256 ARM64 ASM optimized module

Kernel has an ASM optimized version of SHA256 that was ported from
OpenSSL, so lets package it as it provides significant perfomance
improvement compared to the generic implementation.

Signed-off-by: Robert Marko <robimarko@gmail.com>
9 months agokernel: crypto: add SHA1 ARM64 CE optimized module
Robert Marko [Wed, 6 Mar 2024 10:47:45 +0000 (11:47 +0100)]
kernel: crypto: add SHA1 ARM64 CE optimized module

Kernel has optimized version of SHA1 using the ARMv8 Crypto Extensions,
so lets package it.

Use it by default for qualcommax as it uses Cortex-A53 core and has ARMv8
CE extensions present.

Signed-off-by: Robert Marko <robimarko@gmail.com>
9 months agocryptodev-linux: update to 1.13
Weijie Gao [Sun, 7 Jan 2024 05:42:00 +0000 (13:42 +0800)]
cryptodev-linux: update to 1.13

Update to 1.13 with upstream backports.

Signed-off-by: Weijie Gao <hackpascal@gmail.com>
9 months agokernel: qca-ssdk: refresh PCS patch
Robert Marko [Thu, 29 Feb 2024 21:18:20 +0000 (22:18 +0100)]
kernel: qca-ssdk: refresh PCS patch

Recently added PCS patch requires a refresh, so lets do it.

Signed-off-by: Robert Marko <robimarko@gmail.com>
9 months agokernel: crypto: package SHA3
Robert Marko [Tue, 5 Mar 2024 19:15:55 +0000 (20:15 +0100)]
kernel: crypto: package SHA3

SHA3 is now required by jitterentropy_rng in kernel 6.6, so lets start
preparing by packaging SHA3 support as its supported in 5.15 and 6.1
kernels as well.

AFAIK, only ARMv8.2 has a crypto extension for SHA3, however I am not aware
of any SoC we support that uses ARMv8.2 ISA so its not enabled currently.

Signed-off-by: Robert Marko <robimarko@gmail.com>
9 months agobase-files: sysupgrade: handle errors when generating backup
Rafał Miłecki [Mon, 4 Mar 2024 14:18:13 +0000 (15:18 +0100)]
base-files: sysupgrade: handle errors when generating backup

1. Return error if any step of generating tar file fails
2. Use pipefail to avoid calling "gzip" if tar failed

Fixes: e36cc530927c ("base-files: sysupgrade: use tar helper to include installed_packages.txt")
Reported-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
Cc: Luiz Angelo Daros de Luca <luizluca@gmail.com>
Cc: Jo-Philipp Wich <jo@mein.io>
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
9 months agoqualcommax: ipq60xx: add support for netgear wax214
Dirk Buchwalder [Sun, 3 Mar 2024 15:00:46 +0000 (16:00 +0100)]
qualcommax: ipq60xx: add support for netgear wax214

    Netgear WAX214 is a 802.11 ax dual-band AP
    with PoE. (similar to Engenius EWS357APV3)

    Specifications:

        •     CPU: Qualcomm IPQ6010 Quad core Cortex-A53
        •     RAM: 512MB of DDR3
        •     Storage: 128MB NAND (Macronix MX30UF1G18AC)
        •     Ethernet: 1x 1G RJ45 port (QCA8072) PoE
        •     WIFI:
              2.4GHz: Qualcomm QCN5022 2x2 802.11b/g/n/ax 574 Mbps PHY rate
              5GHz: Qualcomm QCN5052 2x2 802.11a/b/g/n/ac/ax 1201 PHY rate

        •     LEDs:
              4 x GPIO-controlled LEDs
                - 1 Power LED (orange)
                - 1 LAN LED (blue)
                - 1 WIFI 5g LED (blue)
                - 1 WIFI 2g LED (blue)
                black_small_square  Buttons: 1x soft reset
                black_small_square  Power: 12V DC jack or PoE (802.3af )

            An populated serial header is onboard, format is
             1.25mm 4p (DF13A-4P-1.25H)
            RX/TX is working, bootwait is active, secure boot is not
            enabled.

            The root password of the stock firmware is unknown,
            but failsafe mode can be entered to reset the password.

            Installation Instructions:

                - obtain serial access
                - stop auto boot (press "4", Entr boot command line
  interface)
                - setenv active_fw 0 (to boot from the primary rootfs,
                  or set to 1 to boot from the secondary rootfs
                  partition)
                - saveenv

                - tftpboot the initramfs image
                - bootm

                - copy
  openwrt-qualcommax-ipq60xx-netgear_wax214-squashfs-factory.ubi
                  to the device
                - write the image to the NAND:
                   - cat /proc/mtd and look for rootfs partition (should
     be mtd11,
                     or mtd12 if you choose active_fw 1)
                   - ubiformat /dev/mtd11 -f -y
     openwrt-qualcommax-ipq60xx-netgear_wax214-squashfs-factory.ubi
                - reboot

            Note: the firmware is senao-based. But I was unable to build
                  a valid senao-header into the image.
                  Maybe they changed the header format and senaoFW isn't
                  working any more.

Signed-off-by: Dirk Buchwalder <buchwalder@posteo.de>
9 months agoipq-wifi: update to version 2024-03-04
Dirk Buchwalder [Mon, 4 Mar 2024 20:33:32 +0000 (21:33 +0100)]
ipq-wifi: update to version 2024-03-04

The new version adds support for the Netgear WAX214.

Signed-off-by: Dirk Buchwalder <buchwalder@posteo.de>
9 months agokernel: bump 6.1 to 6.1.80
John Audia [Fri, 1 Mar 2024 21:08:40 +0000 (16:08 -0500)]
kernel: bump 6.1 to 6.1.80

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.1.80

Manually rebased:
generic/hack-6.1/650-netfilter-add-xt_FLOWOFFLOAD-target.patch[1]

All other patches automatically rebased.

1. Acknowledgement to @heheb and @DragonBluep. Upstream commit for ref: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.1.80&id=9c5662e95a8dcc232c3ef4deb21033badcd260f6

Build system: x86/64
Build-tested: x86/64/AMD Cezanne, ramips/tplink_archer-a6-v3
Run-tested: x86/64/AMD Cezanne, ramips/tplink_archer-a6-v3

Signed-off-by: John Audia <therealgraysky@proton.me>
9 months agokernel: Remove dsmark support
John Audia [Sun, 3 Mar 2024 14:49:03 +0000 (09:49 -0500)]
kernel: Remove dsmark support

dsmark support was removed in kernel 5.15.150 and 6.1.80. Remove it from
the kmod package as well

Signed-off-by: John Audia <therealgraysky@proton.me>
9 months agobuild: do not depend on $(STAGING_DIR)/.prepared when in SDK
Matthias Schiffer [Mon, 4 Mar 2024 22:45:15 +0000 (23:45 +0100)]
build: do not depend on $(STAGING_DIR)/.prepared when in SDK

The dependency can't be satisfied when building using the SDK, breaking
package builds. As the staging and bin dirs are distributed with the SDK
archive, ignoring the dependency is fine when SDK is set.

Fixes: fbb924abff8a ("build: add $(STAGING_DIR) and $(BIN_DIR) ...")
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
9 months agobcm27xx: use PARTUUID for root in cmdline
Álvaro Fernández Rojas [Tue, 30 Jan 2024 20:30:20 +0000 (21:30 +0100)]
bcm27xx: use PARTUUID for root in cmdline

This allows booting from USB/NVME while keeping SD compatibility.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
9 months agobase-files: sysupgrade: allow overriding config restore
Álvaro Fernández Rojas [Wed, 31 Jan 2024 10:43:55 +0000 (11:43 +0100)]
base-files: sysupgrade: allow overriding config restore

Some platforms like Raspberry Pi require patching some backup files like
cmdline.txt in order to set the correct root PARTUUID.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
9 months agobcm27xx: pass signature to partition generation
Álvaro Fernández Rojas [Tue, 30 Jan 2024 20:28:54 +0000 (21:28 +0100)]
bcm27xx: pass signature to partition generation

Passes IMG_PART_SIGNATURE to ptgen instead of using the default "OWRT"
signature.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
9 months agobcm27xx: base-files: fix platform_copy_config
Álvaro Fernández Rojas [Mon, 4 Mar 2024 15:20:55 +0000 (16:20 +0100)]
bcm27xx: base-files: fix platform_copy_config

/boot/kernel*.img wildcard only works without quotation marks.

Fixes: 1a5e51ab0011 ("bcm27xx: base-files: fix platform_copy_config")
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
9 months agobcm63xx: drop target
Álvaro Fernández Rojas [Mon, 5 Feb 2024 17:39:57 +0000 (18:39 +0100)]
bcm63xx: drop target

BCM63xx maintainance is a PITA since there's no full Device Tree support and
therefore every board needs it's own definition in board_bcm963xx.c
There's no DSA support for this board and there have been very few
contributions in the last years.
Moreover, BCM63xx SoCs >= BCM6358 are already supported in bmips target, so
any efforts to keep this target alive aren't worth the time.
Let's focus the community efforts on improving bmips instead.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
9 months agobase-files: sysupgrade: fix streaming backup archives to stdout
Jo-Philipp Wich [Mon, 4 Mar 2024 13:50:46 +0000 (14:50 +0100)]
base-files: sysupgrade: fix streaming backup archives to stdout

Due to previous refactoring in sysupgrade, writing backup archives to
stdout became impossible since the hardcoded gzip output redirection
did not account for the `-` special case filename.

Fix this issue by substituting `-` with `/proc/self/fd/1` in the tar
archive output path variable.

Also remove a redundant `rm -f` of the target file path that occurs
before the file could've possibly been written.

Fixes: #14773
Fixes: https://github.com/openwrt/luci/issues/6961
Fixes: e36cc53092 ("base-files: sysupgrade: use tar helper to include installed_packages.txt")
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
9 months agobuild: add $(STAGING_DIR) and $(BIN_DIR) preparation to target and package subdir...
Matthias Schiffer [Thu, 1 Feb 2024 23:46:58 +0000 (00:46 +0100)]
build: add $(STAGING_DIR) and $(BIN_DIR) preparation to target and package subdir compile dependencies

In a pristine build, these directories are created as dependencies of
the tools subdir compile, however this step never runs when the tools
compile stamp already exists. Since commit ed6ba2801c0a ("tools: keep
stamp file in $(STAGING_DIR_HOST)"), this will happen after `make clean`:
$(STAGING_DIR) has been deleted, but the tools stamp still exists, so
the next build will fail because $(STAGING_DIR) has not been set up
correctly.

Fix builds after `make clean` by adding the preparation as dependencies
for the target and package directories as well.

Fixes: ed6ba2801c0a ("tools: keep stamp file in $(STAGING_DIR_HOST)")
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
9 months agokernel: lantiq: ltq-vmmc: introduce user group for vmmc
Daniel Golle [Sun, 3 Mar 2024 15:22:54 +0000 (15:22 +0000)]
kernel: lantiq: ltq-vmmc: introduce user group for vmmc

asterisk-chan-lantiq is by now the only user of the VMMC interface.
And asterisk runs as user 'asterisk' which doesn't give it permission
to open the /dev/vmmc* devices.
Introduce a new user group 'vmmc' and give permission to access the
/dev/vmmc* devices to that group.
Another commit for asterisk-chan-lantiq will add the 'asterisk' user
to that group.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
9 months agomvebu: add support for SolidRun ClearFog Pro
Tobias Schramm [Mon, 26 Feb 2024 11:47:29 +0000 (12:47 +0100)]
mvebu: add support for SolidRun ClearFog Pro

The SolidRun ClearFog Pro is a router based on the SolidRun CN9130 SOM.

Specs:
 - SoC: Quad-Core Cortex-A72 CN9130 SoC
 - RAM: 4GiB DDR4
 - Serial: Micro-USB port on front (FT232R, 115200 8n1)
 - Storage: 8GiB eMMC, microSD card slot, 8MiB SPI NOR flash
 - Ethernet: 7x GbE (1 port dedicated on SoC, 6 port switch with single GbE CPU port)
 - SFP: 1x SFP+
 - USB: 1x USB-A 3.1 Gen 1
 - PCIe: 2x mini PCIe (one slot with USB and SIM card socket)
 - SATA: 1x M.2 Key-B

In addition to the usual connectivity options this device also features
an internal mikroBUS expansion connector.

SATA is currently untested due to lack of a suitable M.2 SSD.

Installation
============

1. Write sdcard sysupgrade image to microSD card using dd or similar
2. Insert microSD card into router and apply power
3. Device boots into OpenWRT
4. (optional) dd sysupgrade image to /dev/mmcblk0 to install to eMMC

Signed-off-by: Tobias Schramm <tobias@t-sys.eu>
9 months agokernel: 6.1: backport Marvell Armada 8k AP807 cpufreq support
Tobias Schramm [Mon, 26 Feb 2024 09:40:50 +0000 (10:40 +0100)]
kernel: 6.1: backport Marvell Armada 8k AP807 cpufreq support

Upstream a patch adding support for cpufreq on AP807-based SoC like the
CN913x was submitted. Include it in patches to ensure best performance
under load and lowest power consumption in idle.

Signed-off-by: Tobias Schramm <tobias@t-sys.eu>
9 months agou-boot.mk: override default PATH to avoid pick hostpkg python
Tianling Shen [Sun, 25 Feb 2024 12:33:08 +0000 (20:33 +0800)]
u-boot.mk: override default PATH to avoid pick hostpkg python

hostpkg python from packages feed can be picked when do a incremental
build because hostpkg has higher priority in PATH. It may lead build
faliure as it's heavily trimmed (e.g. lacks necessary modules).

For uboot which uses binman and intree dtc, this is forced as hostpkg
python will never provide those modules by default.

Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
9 months agomediatek: filogic: fix failsafe mode on devices with no lan1
Paul Geraedts [Sat, 24 Feb 2024 08:01:01 +0000 (09:01 +0100)]
mediatek: filogic: fix failsafe mode on devices with no lan1

Default to lan4 port, instead of (missing) lan1 port,
to fix failsafe mode on Xiaomi AX3000T, WR30U, AX6000

Build system: x86/64
Build-tested: mediatek/filogic/xiaomi_mi-router-ax3000t-ubootmod
Run-tested: mediatek/filogic/xiaomi_mi-router-ax3000t-ubootmod

Signed-off-by: Paul Geraedts <p.f.j.geraedts@gmail.com>
9 months agols-mc: bump to 10.38.0
Pawel Dembicki [Fri, 1 Mar 2024 09:47:36 +0000 (10:47 +0100)]
ls-mc: bump to 10.38.0

Update ls-mc to latest 10.38.0 version.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
9 months agols-dpl: bump to 10.38.0
Pawel Dembicki [Fri, 1 Mar 2024 09:44:48 +0000 (10:44 +0100)]
ls-dpl: bump to 10.38.0

Update ls-dpl to latest 10.38.0 version.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
9 months agoath79: add support for ELECOM WAB-I1750-PS
INAGAKI Hiroshi [Thu, 22 Feb 2024 10:00:57 +0000 (19:00 +0900)]
ath79: add support for ELECOM WAB-I1750-PS

ELECOM WAB-I1750-PS is a 2.4/5 GHz band 11ac (Wi-Fi 5) access point,
based on QCA9558.

Specification:

- SoC              : Qualcomm Atheros QCA9558
- RAM              : DDR2 128 MiB (2x Winbond W9751G6KB251)
- Flash            : SPI-NOR 16 MiB (Macronix MX25L12835FMI-10G)
- WLAN             : 2.4/5 GHz 3T3R
  - 2.4 GHz        : Qualcomm Atheros QCA9558 (SoC)
  - 5 GHz          : Qualcomm Atheros QCA9880
- Ethernet         : 2x 10/100/1000 Mbps
  - phy ("PD")     : Atheros AR8035
  - phy ("PSE")    : Atheros AR8033
- LEDs/keys (GPIO) : 3x/3x
- UART             : 2x RJ-45 port
  - "SERVICE"      : TTL (3.3V)
    - port         : ttyS0
    - assignment   : 1:3.3V, 2:GND, 3:TX, 4:RX
    - settings     : 115200n8
    - note         : no compatibility with "Cisco console cable"
  - "SERIAL"       : RS232C (+-12V)
    - port         : ?
    - assignment   : 1:NC , 2:NC , 3:TXD, 4:GND,
                     5:GND, 6:RXD, 7:NC , 8:NC
    - settings     : 115200n8
    - note         : compatible with "Cisco console cable"
- Buzzer           : 1x GPIO-controlled
- USB              : 1x USB 2.0 Type-A
- Power            : DC jack or PoE
  - DC jack        : 12 VDC, 1.04 A (device only, rating)
  - PoE            : 802.3af/at, 48 VDC, 0.26 A (device only, rating)
  - note           : supports 802.3af supply on PSE (downstream) port
                     when powered by DC adapter or 802.3at PoE

Flash instruction using factory.bin image:

1. Boot WAB-I1750-PS without no upstream connection (or PoE connection
   without DHCP)
2. Access to the WebUI ("http://192.168.3.1") on the device and open
   firmware update page
   ("ツールボックス" -> "ファームウェア更新")
3. Select the OpenWrt factory.bin image and click update
   ("アップデート") button
4. Wait ~120 seconds to complete flashing

Revert to OEM firmware:

1. Download the latest OEM firmware
2. Remove 128 bytes(0x80) header from firmware image
3. Decode by xor with a pattern "8844a2d168b45a2d" (hex val)
4. Upload the decoded firmware to the device
5. Flash to "firmware" partition by mtd command
6. Reboot

Notes:

- To use the "SERVICE" port, the connection of 3.3V line is also
  required to enable console output.
  The uart line of "SERVICE" is branched out from the internal pin
  header with 74HC126D and 3.3V line is connected to OE pin on it.

- "SERIAL" port is provided by HS UART on QCA9558 SoC that has
  compatibility with qca,ar9330-uart, but QCA955x SoC's is not supported
  on Linux Kernel and OpenWrt.

- To supply 802.3af PoE on "PSE" port when powered by DC adapter, 12 VDC
  3.5 A adapter is recommended. (official: WAB-EX-ADP1)

MAC addresses:

Ethernet (PD, PSE): 00:90:FE:xx:xx:0A (Config, ethaddr (text))
2.4GHz            : 00:90:FE:xx:xx:0A (Config, ethaddr (text))
5GHz              : 00:90:FE:xx:xx:0B

[original work]
Signed-off-by: Yanase Yuki <dev@zpc.st>
[update for NVMEM and others]
Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
9 months agoath79: add support for ELECOM WAB-S1167-PS
INAGAKI Hiroshi [Thu, 22 Feb 2024 10:00:57 +0000 (19:00 +0900)]
ath79: add support for ELECOM WAB-S1167-PS

ELECOM WAB-S1167-PS is a 2.4/5 GHz band 11ac (Wi-Fi 5) access point,
based on QCA9557.

Specification:

- SoC              : Qualcomm Atheros QCA9557
- RAM              : DDR2 128 MiB (2x Winbond W9751G6KB251)
- Flash            : SPI-NOR 16 MiB (Macronix MX25L12835FMI-10G)
- WLAN             : 2.4/5 GHz 2T2R
  - 2.4 GHz        : Qualcomm Atheros QCA9557 (SoC)
  - 5 GHz          : Qualcomm Atheros QCA9882
- Ethernet         : 2x 10/100/1000 Mbps
  - phy ("PD")     : Atheros AR8035
  - phy ("PSE")    : Atheros AR8033
- LEDs/keys (GPIO) : 3x/3x
- UART             : 1x RJ-45 port
  - "SERVICE"      : TTL (3.3V)
    - port         : ttyS0
    - assignment   : 1:3.3V, 2:GND, 3:TX, 4:RX
    - settings     : 115200n8
    - note         : no compatibility with "Cisco console cable"
- Buzzer           : 1x GPIO-controlled
- USB              : 1x USB 2.0 Type-A
- Power            : DC jack or PoE
  - DC jack        : 12 VDC, 1 A (device only, rating)
  - PoE            : 802.3af/at, 48 VDC, 0.25 A (device only, rating)
  - note           : supports 802.3af supply on PSE (downstream) port
                     when powered by DC adapter or 802.3at PoE

Flash instruction using factory.bin image:

1. Boot WAB-S1167-PS without no upstream connection (or PoE connection
   without DHCP)
2. Access to the WebUI ("http://192.168.3.1") on the device and open
   firmware update page
   ("ツールボックス" -> "ファームウェア更新")
3. Select the OpenWrt factory.bin image and click update
   ("アップデート") button
4. Wait ~120 seconds to complete flashing

Revert to OEM firmware:

1. Download the latest OEM firmware
2. Remove 128 bytes(0x80) header from firmware image
3. Decode by xor with a pattern "8844a2d168b45a2d" (hex val)
4. Upload the decoded firmware to the device
5. Flash to "firmware" partition by mtd command
6. Reboot

Notes:

- To use the "SERVICE" port, the connection of 3.3V line is also
  required to enable console output.
  The uart line of "SERVICE" is branched out from the internal pin
  header with 74HC126D and 3.3V line is connected to OE pin on it.

- The same PCB is used with WAB-S600-PS.

- To supply 802.3af PoE on "PSE" port when powered by DC adapter, 12 VDC
  3.5 A adapter is recommended. (official: WAB-EX-ADP1)

MAC addresses:

Ethernet (PD, PSE): 00:90:FE:xx:xx:04 (Config, ethaddr (text))
2.4GHz            : 00:90:FE:xx:xx:04 (Config, ethaddr (text))
5GHz              : 00:90:FE:xx:xx:05

Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
9 months agoath79: add support for ELECOM WAB-S600-PS
INAGAKI Hiroshi [Thu, 22 Feb 2024 10:00:57 +0000 (19:00 +0900)]
ath79: add support for ELECOM WAB-S600-PS

ELECOM WAB-S600-PS is a 2.4/5 GHz band 11n (Wi-Fi 4) access point, based
on QCA9557.
This device also supports 11ac (Wi-Fi 5) with the another official
firmware.

Specification:

- SoC              : Qualcomm Atheros QCA9557
- RAM              : DDR2 128 MiB (2x Winbond W9751G6KB251)
- Flash            : SPI-NOR 16 MiB (Macronix MX25L12835FMI-10G)
- WLAN             : 2.4/5 GHz 2T2R
  - 2.4 GHz        : Qualcomm Atheros QCA9557 (SoC)
  - 5 GHz          : Qualcomm Atheros QCA9882
- Ethernet         : 2x 10/100/1000 Mbps
  - phy ("PD")     : Atheros AR8035
  - phy ("PSE")    : Atheros AR8033
- LEDs/keys (GPIO) : 3x/3x
- UART             : 1x RJ-45 port
  - "SERVICE"      : TTL (3.3V)
    - port         : ttyS0
    - assignment   : 1:3.3V, 2:GND, 3:TX, 4:RX
    - settings     : 115200n8
    - note         : no compatibility with "Cisco console cable"
- Buzzer           : 1x GPIO-controlled
- USB              : 1x USB 2.0 Type-A
- Power            : DC jack or PoE
  - DC jack        : 12 VDC, 1 A (device only, rating)
  - PoE            : 802.3af/at, 48 VDC, 0.25 A (device only, rating)
  - note           : supports 802.3af supply on PSE (downstream) port
                     when powered by DC adapter or 802.3at PoE

Flash instruction using factory.bin image:

1. Boot WAB-S600-PS without no upstream connection (or PoE connection
   without DHCP)
2. Access to the WebUI ("http://192.168.3.1") on the device and open
   firmware update page
   ("ツールボックス" -> "ファームウェア更新")
3. Select the OpenWrt factory.bin image and click update
   ("アップデート") button
4. Wait ~120 seconds to complete flashing

Revert to OEM firmware:

1. Download the latest OEM firmware
2. Remove 128 bytes(0x80) header from firmware image
3. Decode by xor with a pattern "8844a2d168b45a2d" (hex val)
4. Upload the decoded firmware to the device
5. Flash to "firmware" partition by mtd command
6. Reboot

Notes:

- To use the "SERVICE" port, the connection of 3.3V line is also
  required to enable console output.
  The uart line of "SERVICE" is branched out from the internal pin
  header with 74HC126D and 3.3V line is connected to OE pin on it.

- The same PCB is used with WAB-S1167-PS.

- To supply 802.3af PoE on "PSE" port when powered by DC adapter, 12 VDC
  3.5 A adapter is recommended. (official: WAB-EX-ADP1)

MAC addresses:

Ethernet (PD, PSE): BC:5C:4C:xx:xx:7C (Config, ethaddr (text))
2.4GHz            : BC:5C:4C:xx:xx:7C (Config, ethaddr (text))
5GHz              : BC:5C:4C:xx:xx:7D

[original work of common dtsi part for WAB-I1750-PS]
Signed-off-by: Yanase Yuki <dev@zpc.st>
[adding support for WAB-S600-PS]
Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
9 months agoramips: Unielec u7621-01: remove invalid led status
David Bentham [Sun, 24 Dec 2023 13:49:50 +0000 (13:49 +0000)]
ramips: Unielec u7621-01: remove invalid led status

The original commit had an invalid setting of the
led status for this device.

There is no gpio pin connected to that led so lets
remove these from the u7621-01 dtsi file.

Signed-off-by: David Bentham <db260179@gmail.com>
9 months agoocteontx: update default network config
Tim Harvey [Thu, 2 Mar 2023 22:29:39 +0000 (14:29 -0800)]
octeontx: update default network config

Update the default network configuration for Gateworks Newport boards
such that the left-most front-panel NIC is WAN and any additional are in
LAN bridge:
 - gw610x/gw6903; single NIC: eth0
 - gw620x/gw630x; two NIC's from left to right are: eth1 eth0
 - gw640x; multiple NIC's from left to right: eth4 eth0 eth1 eth2 eth3

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
9 months agoapm821xx: MBL DUO-only: enable USB-Storage for boot from USB support
Christian Lamparter [Sat, 17 Feb 2024 20:16:12 +0000 (21:16 +0100)]
apm821xx: MBL DUO-only: enable USB-Storage for boot from USB support

it is possible to boot and run OpenWrt from an image on an USB-Stick
on the MyBook Live DUO. (No, the MyBook Live Single does NOT have an
USB-Port and attempts at enabling it sadly all failed... so far).

To do that:
First, prepare a USB-Stick by writing the raw and uncompressed OpenWRT
factory image for the device onto the stick (i.e.
 # gunzip -c openwrt-*wd_mybooklive*-factory.img.gz > /dev/sd$XX).

Then enter the u-boot via an attached TTL/CMOS 3.3V cable adapter and
give the following commands a try in the:

 usb start; sata init
 ext2load usb 0:1 ${fdt_addr_r} /boot/apollo3g.dtb
 ext2load usb 0:1 ${kernel_addr_r} /boot/uImage
 setenv bootargs 'root=/dev/sdc2 rw rootfstype=squashfs,ext4 rootdelay=5'
 run addtty; bootm ${kernel_addr_r} - ${fdt_addr_r}

Notes:
 - booting from USB-Sticks takes a long time! Be prepared to wait a few
   minutes. (~3 minutes for 4 MiB /boot/uImage file on a USB 2.0 Stick)

 - the bootargs part 'root=/dev/sdX2' depends on how many HDDs/SSDs are
   slotted in. (if none: then use sda, if one: sdb. if two: sdc)

 - rootdelay is important as the storage on the USB-Sticks do not show
   up fast enough. 5 seconds might be excessive though.

 - it's possible to concat these commands together in one line and
   write it into u-boot's "bootcmd" environment variable and save the
   environment to make the device to always boot from USB from then on.

   if you have accidentally overridden the 'bootcmd' and want to return
   to 'spec' enter the following commands :

   setenv bootcmd 'run boot_sata_script_ap2nc'
   saveenv

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
9 months agoapm821xx: add u-boot,env compatible
Christian Lamparter [Fri, 9 Jun 2023 18:34:02 +0000 (20:34 +0200)]
apm821xx: add u-boot,env compatible

Note about the MyBook Live.
The MyBook Live didn't have the encompassing fixed-partitions.
This is needed for the u-boot,env compatible to be read as otherwise
the kernel assumes this is a legacy NOR/NAND node and this binding
is unsupported.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
9 months agoapm821xx: remove 5.15 remnants
Christian Lamparter [Sat, 17 Feb 2024 19:52:41 +0000 (20:52 +0100)]
apm821xx: remove 5.15 remnants

deletes config-5.15 and patches-5.15/.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
9 months agoapm821xx: make 6.1 kernel default
Christian Lamparter [Sat, 17 Feb 2024 19:51:13 +0000 (20:51 +0100)]
apm821xx: make 6.1 kernel default

6.1 has been available for some time for testing.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>