Oversight from 8d302f5ebc , which requires
this to work. I couldn't test it as I ended up wiping my UBI volume by
accident.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16580
Signed-off-by: Robert Marko <robimarko@gmail.com>
This is needed in order to get a MAC address from UBI.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16580
Signed-off-by: Robert Marko <robimarko@gmail.com>
linux,ubi expects the volname variable, not volume.
This will fix NVMEM-on-UBI.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16580
Signed-off-by: Robert Marko <robimarko@gmail.com>
Since kernel_oldconfig doesn't work properly with it, I ran that first
and then moved all the config symbols to config-6.6 and found the
differences.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16580
Signed-off-by: Robert Marko <robimarko@gmail.com>
make kernel_oldconfig doesn't seem to work with this. Handle manually.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16580
Signed-off-by: Robert Marko <robimarko@gmail.com>
Adjusts the default config to modern kernels.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16580
Signed-off-by: Robert Marko <robimarko@gmail.com>
The ImageBuilder when using the profile for the OpenWrt One has been
failing because the initramfs image included in the ubinized image could
not be found.
Fix that by using the staged initramfs instead when using the
ImageBuilder.
Fixes: 797904b3cb ("mediatek/filogic: add OpenWrt One support")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
This allows to obtain detailed diagnostic information about the DSL
connection from the modem. The script is copied from the corresponding
packages for VDSL modems.
Signed-off-by: Jan Hoffmann <jan@3e8.eu>
Link: https://patchwork.ozlabs.org/project/openwrt/patch/20241019174041.1281093-4-jan@3e8.eu/
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This is a preparation for adding support for dsl_cpe_pipe.sh with a
similar set of commands compared to the VDSL variant.
The configuration is simplified by using the "--enable-model" option.
Other options are chosen to match the VDSL variant, while also making
sure that previously enabled options stay like that. However, ReTx
options stay disabled, because of incompatibility with the ubus code.
Signed-off-by: Jan Hoffmann <jan@3e8.eu>
Link: https://patchwork.ozlabs.org/project/openwrt/patch/20241019174041.1281093-3-jan@3e8.eu/
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This is a preparation for adding support for dsl_cpe_pipe.sh with a
similar set of commands compared to the VDSL variant.
The configuration is simplified by using the "--enable-model" option.
Other options are chosen to match the VDSL variant, while also making
sure that previously enabled options stay like that. However, ReTx
options stay disabled, because of incompatibility with the ubus code.
Signed-off-by: Jan Hoffmann <jan@3e8.eu>
Link: https://patchwork.ozlabs.org/project/openwrt/patch/20241019174041.1281093-2-jan@3e8.eu/
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
65bb027 CMakeLists.txt: bump minimum cmake version
252a9b0 libubus: Make UBUS_* macros work cleanly in C++
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This commit adds the build date during compilation to the os-release file.
This information can then be used later to output this via the the ubus,
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
Link: https://github.com/openwrt/openwrt/pull/15987
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Introduce EN7581 SoC support with currently rfb board supported.
This is a new 64bit SoC from Airoha that is currently almost fully
supported upstream with only the DTS missing. Setting source-only
waiting for the full upstream support to be completed.
Link: https://github.com/openwrt/openwrt/pull/16730
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Backport alloc_netdev_dummy patch from upstream needed for new Airoha
EN7581 SoC.
Refresh all affected patch with make target/linux/refresh.
Link: https://github.com/openwrt/openwrt/pull/16730
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
In preparation for EN7581 SoC support, move en7523 in dedicated
subtarget.
This is needed as EN7581 is now 64bit but en7523 is 32bit hence have
very different kernel config.
Also rename patch to a more friendly number sequence.
Link: https://github.com/openwrt/openwrt/pull/16730
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
COMFAST CF-E355AC v2 is a ceiling mount AP with PoE support,
based on Qualcomm/Atheros QCA9531 + QCA9886.
Short specification:
- 1x 10/100 Mbps Ethernet, with PoE support (wan/eth1)
- 1x 10/100/1000 Mbps Ethernet, with PoE support (lan/eth0)
- 128MB of RAM (DDR2)
- 16 MB of FLASH
- 2T2R 2.4 GHz, 802.11b/g/n (wlan2g)
- 2T2R 5 GHz, 802.11ac/n/a, WAVE 2 (wlan5g)
- built-in 4x 3 dBi antennas
- output power (max): 500 mW (27 dBm)
- 1x RGB LED, 1x button
- separate watchdog chip via GPIO (bottom of PCB?)
- UART header on PCB with proper labelling
Markings on PCB:
* R121QH_VER2.1 (silkscreen, bottom)
* CF-WA800 (sticker, top)
Initial flashing instructions:
Original firmware is based on OpenWrt.
a) Use sysupgrade image directly in vendor GUI.
b) Or via tftp:
ipaddr=192.168.1.1
serverip=192.168.1.10
bootfile="firmware.bin"
c) Or possibly via u-boot's `httpd` command.
MAC-address mapping follows original firmware:
* eth1 (wan) is the lowest mac address (art @ 0x0)
* eth0 (lan) uses eth1 + 1 (art @ 0x1002)
* wlan2g (phy1) uses eth1 + 2 (art @ 0x06)
* wlan5g (phy0) uses eth1 + 10 (not present in art)
* unused MAC (eth1 + 3) (art @ 0x5006)
Art dump (`hexdump /dev/mtd1 |grep ZZZZ`):
0000000 ZZZZ XXXX XXX0 ZZZZ XXXX XXX2 ffff ffff
0001000 0202 ZZZZ XXXX XXX1 0000 0000 0000 0000
0005000 202f bd21 0101 ZZZZ XXXX XXX3 0000 2000
Root access to original firmware (only via UART) can be achieved by
making a backup of configuration from web interface. Backup contains
whole `/etc` directory...
Signed-off-by: Priit Laes <plaes@plaes.org>
Link: https://github.com/openwrt/openwrt/pull/16556
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
don't use configuration directories which are relative
Signed-off-by: Paul Donald <newtwen+github@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/14975
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Add support for handling of DNS RR (Resource Records) requests, which
are needed for the HTTPS Type 65 records, introduced to support the
DNS-based Service Discovery (DNS-SD) mechanism for HTTPS services and
defined in the RFC 9460 (9.1. Query Names for HTTPS RRs).
Ref: https://forum.openwrt.org/t/resolving-query-type-65-to-local-address-for-ios-clients-in-dnsmasq/179504/11
uci config usage:
config dnsrr
option rrname 'foo.example.com'
option rrnumber '65'
option hexdata '00'
hexdata is optional.
Available since dnsmasq 2.62 (for around 12 years at this point).
Note: dnsmasq dns-rr are not affected by filter-rr
Tested on 22.03.5
Signed-off-by: Paul Donald <newtwen+github@gmail.com>
Tested-by: Vladimir Kochkovski <ask@getvladimir.com>
Link: https://github.com/openwrt/openwrt/pull/14975
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Prevents problems when variables contain spaces.
Tested on: 23.05.3
Signed-off-by: Paul Donald <newtwen+github@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/14975
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Users can now freely add new dnsmasq parameters (i.e. a whole config)
via extraconf. This means users can add their own parameters without
changes to init or GUI.
Co-opted the default of confdir also to include the instance name.
This way each instance gets its own .d directory (and separate instances
do not all inherit the same 'extraconftext').
Usage:
config dnsmasq 'config'
...
option extraconftext 'cache-size=2048\nlog-async=20'
config dnsmasq 'blah'
...
option extraconftext 'cache-size=128\nlog-async=5'
or even (which would produce staggered output but still valid)
config dnsmasq 'blah'
...
option extraconftext 'cache-size=128
log-async=5'
See https://forum.openwrt.org/t/add-dnsmasq-custom-options-field-in-luci-gui/193184
Tested on: 23.05.3, 22.03.6
Signed-off-by: Paul Donald <newtwen+github@gmail.com>
Tested-by: Vladimir Kochkovski <ask@getvladimir.com>
Link: https://github.com/openwrt/openwrt/pull/14975
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
dnsmasq v2.90 introduced `--cache-rr=<rrtype>[,<rrtype>...]`.
uci config usage:
config dnsmasq
...
option cache_rr 'AAAA,CNAME,NXDOMAIN,SRV,...'
The dnsmasq instance internally builds a linked list of RR to cache
from the individually supplied parameters, so it's allowed to provide
multiples:
... --cache-rr=AAAA --cache-rr=A ...
See https://forum.openwrt.org/t/resolving-query-type-65-to-local-address-for-ios-clients-in-dnsmasq/179504
Tested on: 23.05.2
Signed-off-by: Paul Donald <newtwen+github@gmail.com>
Tested-by: Vladimir Kochkovski <ask@getvladimir.com>
Link: https://github.com/openwrt/openwrt/pull/14975
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
dnsmasq v2.90 introduced `--filter-rr=<rrtype>[,<rrtype>...]`.
uci config usage:
config dnsmasq
...
option filter_rr 'AAAA,CNAME,NXDOMAIN,SRV,...'
The dnsmasq instance internally builds a linked list of RR to filter
from the individually supplied parameters, so it's harmless to provide
synonyms:
... --filter-A --filter-rr=A ...
See https://forum.openwrt.org/t/resolving-query-type-65-to-local-address-for-ios-clients-in-dnsmasq/179504/23
Tested on: 23.05.2
Signed-off-by: Paul Donald <newtwen+github@gmail.com>
Tested-by: Vladimir Kochkovski <ask@getvladimir.com>
Link: https://github.com/openwrt/openwrt/pull/14975
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This patch allows the use of SAE when using PPSK after
https://w1.fi/cgit/hostap/commit/?id=fcbdaae8a52e542705a651ee78b39b02935fda20
added support for it.
It also implements a fix so that this option works with SAE. The reason this
doesn't work out of the box is because OpenWRT deviates from hostapd defaults
by setting `sae_pwe` option to 2 which makes this mode not function properly
(results in every auth attempt being denied).
That issue was addressed by not overriding hostapd's default for the `sae_pwe`
option when the PPSK option is in use. This should be fine because hostapd's
test cases specifically test this mode with the default SAE parameters. See:
https://w1.fi/cgit/hostap/commit/?id=c34b35b54e81dbacd9dee513b74604c87f93f6a3
Signed-off-by: Rany Hany <rany_hany@riseup.net>
Link: https://github.com/openwrt/openwrt/pull/16343
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Fix conditions for handling offloaded packets
Fixes: #13430
Signed-off-by: Leon M. Busch-George <leon@georgemail.eu>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Set the physical switch to KEY_RFKILL, since its previous value
(KEY_SETUP) is unsupported. This should also make the KEY_RESET button
functional, by allowing the gpio-button-hotplug kmod to load.
Signed-off-by: Chris Jones <cmsj@tenshu.net>
Link: https://github.com/openwrt/openwrt/pull/16564
Signed-off-by: Robert Marko <robimarko@gmail.com>
Allows removal of the remove function by letting devm handle everything.
Avoids gotos as well.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16507
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
AT803X_PHY was replaced with QCA83XX_PHY which handles external switches.
But for internal ones like in the Meraki MR24, AT803X_PHY is needed.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16737
Signed-off-by: Robert Marko <robimarko@gmail.com>
The qpic DMA controller is used by the parallel NAND Flash
interface. We don't need to enable it when nand-controller node
is marked as disabled.
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Link: https://github.com/openwrt/openwrt/pull/16654
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Mikrotik seems to prefer "hEX S" as general name for this model,
therefore include this in devicetree model name as well.
Signed-off-by: Priit Laes <plaes@plaes.org>
Link: https://github.com/openwrt/openwrt/pull/16658
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Update package to the latest stable version and drop upstreamed patches:
0001-arm-mvebu-turris_omnia-Enable-LTO-by-default-on-Turr.patch
100-mvebu-armada-8k-respect-CONFIG_DISTRO_DEFAULTS.patch
Other patches automatically refreshed (line numbers only)
Add custom config flags to disable building efimkcapsule by default.
This introduces a dependency to GnuTLS which is not present and we do
not need it here.
Signed-off-by: Stefan Kalscheuer <stefan@stklcode.de>
Link: https://github.com/openwrt/openwrt/pull/16676
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This allows the llvm toolchain to be executed on different host.
Also add it to strip list.
Fixes: 0ac0840088 ("sdk: ship llvm toolchain")
Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16674
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
From the upstream repo:
Instead of assuming only one register is used, track all 16 regs
individually.
This avoids need for the 'PREV_PAYLOAD' hack and also avoids the need to
clear out old flags:
When we see that register 'x' will be written to, that register state is
reset automatically.
Existing dissector decodes
ip saddr 1.2.3.4 meta l4proto tcp
... as
-s 6.0.0.0 -p tcp
iptables-nft -s 1.2.3.4 -p tcp is decoded correctly because the expressions
are ordered like:
meta l4proto tcp ip saddr 1.2.3.4
|
... and 'meta l4proto' did clear the PAYLOAD flag.
The simpler fix is:
ctx->flags &= ~NFT_XT_CTX_PAYLOAD;
in nft_parse_cmp(), but that breaks dissection of '1-42', because
the second compare ('cmp lte 42') will not find the
payload expression anymore.
This commit fixes#11169 and openwrt/packages#22727, and potentially anyone that uses iptables-nft legacy support.
Signed-off-by: Rodrigo B. de Sousa Martins <rodrigo.sousa.577@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16504
[Added patch header]
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Increase usage of devm to get rid of goto and _remove.
Get rid of hw_reset_count. It's not really used for anything.
Use dev_err_probe to handle potential EPROBE_DEFER.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16588
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
If LEDs are disabled, they should not be handled.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16651
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
No longer need normal _remove function. Replaced with _disable.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16651
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The already existing uci function ucidef_set_network_device_path
can be used to specify a unique PCI address to name a network interface.
However, I noticed that some NIC ports share the same PCI address
but are still distinguishable by the dev_port value of the network
interface's sysfs entry.
This commit adds a new uci function ucidef_set_network_device_path_port,
which is similar to ucidef_set_network_device_path but takes an
additional argument where the user can specify the dev_port value.
The internal function preinit_config_port loops through
all network interfaces at the given PCI address and chooses the one
where the dev_port value matches.
This was tested on an x86_64 device using a Mellanox ConnectX-3 card.
Signed-off-by: Til Kaiser <mail@tk154.de>
Link: https://github.com/openwrt/openwrt/pull/16560
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Set the boot flag for the igc, mlx4-core, and mlx5-core network device drivers
to load them at a more early stage of the boot process.
This is required for network drivers whose network interface PCI paths are set
via ucidef_set_network_device_path inside the 02_network script since it is
called after kernel modules are loaded from modules-boot.d but before they are
loaded from the modules.d directory.
Signed-off-by: Til Kaiser <mail@tk154.de>
Link: https://github.com/openwrt/openwrt/pull/16560
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>