ath10k and ath11k support reading calibration variants from the device
tree to locate the correct Board Description File (BDF). The ath12k-wsi
binding already describes using qcom,calibration-variant but it is not
implemented in the driver.
Many ath12k designs expose all the radios under a single phy, each of
which typically require a separate BDF. Without this, the radios may not
come up or will not be calibrated correctly.
Fix this by parsing the device tree for the generation-agnostic
qcom,calibration-variant. This allows the driver to properly select,
read and apply the correct BDF.
Additionally, the ath12k-wsi binding documentation describes using the
generation-specific qcom,ath12k-calibration-variant binding as well as
the generation-agnostic qcom,calibration-variant binding to load
board-specific calibration data from the device tree. However, the
driver never implemented either of these.
Given that no devices currently supported use
qcom,ath12k-calibration-variant and the previous patch implements
qcom,calibration-variant, drop the generation-specific version from the
binding to prevent future confusion.
Tested on Askey SBE1V1K.
Signed-off-by: Andrew LaMarche <andrewjlamarche@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21767
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
The in-tree realtek 5G/10G driver picks the firmware patch table from the
PHY ID (RTL8264 -> rtl8264b.bin). The vendor driver instead picks it from
VEND1 0x104[2:0] (0 = RTL8264B, else RTL8261N-C), independent of PHY ID.
On the S1300WP the RTL8264 dies read 0x104=0xf802 ([2:0]=2), so they need
the RTL8261N-C table; the RTL8264B table leaves the PCS unlocked (0xA600
stuck 0x1000) and config_init fails with -ETIME. Read 0x104[2:0] and load
rtl8261n.bin when non-zero. Boards with such a die must ship rtl8261n.bin.
Fixes: 6369c9e5c7 ("generic: net: phy: realtek: add 5G and 10G PHY support")
Co-authored-by: Sven Eckelmann <sven@narfation.org>
Signed-off-by: Carlo Szelinsky <github@szelinsky.de>
Link: https://github.com/openwrt/openwrt/pull/23947
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
The RTL8367S-VB (Family D) uses a software-emulated VLAN MC table
(emu_vlanmc) since it lacks hardware support for the VLAN MC registers
used by earlier chips.
- Add missing NULL check after kzalloc() in rtl8367b_init_regs() to
prevent NULL pointer dereference on allocation failure.
- Add missing kfree(smi->emu_vlanmc) in rtl8367b_remove() to prevent
memory leak on driver unload.
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/23970
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Remove boilerplate.
Also added deregister for bcm_wifi for consistency. Not needed as it's
builtin but still good to have.
There's a slight change from subsys_initcall to module_init. Not really
an issue for mtd drivers. No driver in linux upstream's drivers/mtd uses
subsys_initcall.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/23893
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
strlcat takes the total buffer size, not the remaining space.
Passing strlen(cmdline) + l caused it to truncate by one byte
when the appended string fit within the buffer.
This is also more consistent with the rest of the file.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/23892
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
BUG() and BUG_ON() are deprecated - they destabilize the system
and make debugging impossible. Replace with:
- WARN_ON() + return error for assertion failures
- WARN_ON_ONCE() for bounds checks
- WARN_ON() + return -EPERM for in_interrupt() guards
Assisted-by: Opencode:Big-Pickle
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/23892
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Backport upstream net-next commit 8ac44d24c3a1. An SFP cage whose MOD_DEF0
signal is not wired to a readable GPIO currently uses sff_gpio_get_state(),
which always reports the module present: an empty cage gets stuck in
MOD_ERROR, hot-insertion is never detected and empty cages spam -EIO at
boot. Derive presence from a throttled single-byte I2C read of the module
EEPROM instead, so hot-plug works and the boot spam stops. A soldered-down
"sff,sff" module stays always-present.
This helps RTL93xx SFP boards that route no cage presence signal to a GPIO.
Remaining sfp.c patches refreshed.
Link: https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?id=8ac44d24c3a1
Signed-off-by: Carlo Szelinsky <github@szelinsky.de>
Link: https://github.com/openwrt/openwrt/pull/23836
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
The firmware-version gate in mxl862xx_phylink_get_caps() leaves
config->supported_interfaces empty for the CPU/SerDes ports (9, 13) when
the switch runs firmware older than 1.0.84. phylink rejects an empty
supported_interfaces bitmap, so the switch fails to probe at all:
mxl862xx mdio-bus:10: switch ready after 2480ms, firmware 1.0.70 (build 70)
mxl862xx mdio-bus:10: phylink: error: empty supported_interfaces
error creating PHYLINK: -22
mxl862xx mdio-bus:10: probe with driver mxl862xx failed with error -22
This regresses boards still shipping the older vendor firmware (e.g. the
BananaPi BPi-R4 Pro 8X, which ships 1.0.70) where all LAN ports disappear.
Ungate only the CPU/SerDes ports (9, 13) so the switch probes and the user
can update the firmware; ports 10..12 and 14..16 stay gated as they are
genuinely unsupported by the old firmware.
Tested on a BPi-R4 Pro 8X with switch firmware 1.0.70 (kernel 6.18); the
6.12 patch receives the identical change for parity.
Fixes: 028dc3f57a ("generic: 6.18: update MxL862xx DSA switch driver")
Fixes: 5b69e6a4a6 ("generic: 6.12: update MxL862xx DSA switch driver")
Signed-off-by: Aliaksandr Babrykovich <abobrikovich@gmail.com>
[refreshed patches]
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
LED trigger for network interfaces.
- Aggregated per-family (lan/wan/wlan).
- Family and flags are taken from device tree properties:
- "family" : simple family string "lan" | "wan" | "wlan"
- "mode" : any combination of "link", "tx", "rx" flags
Priority/combination:
* If "mode" present: flags come from "mode" and take precedence.
Family is taken from "family" if present, otherwise from the LED name.
* If only "family" present: use its family and default flags = link+tx+rx.
* If neither present: fall back to LED device name parsing.
- Suffix "-online" is valid ONLY in the LED name (label),
e.g. "green:wlan-online". It indicates the online variant but is applied
only when DT "mode" is absent.
Behaviour:
- wlan (normal): blink/solid driven by throughput table
- lan/wan (normal): one-shot blink on TX/RX packet change
- *-online variants: steady ON while any interface of the family has carrier
Interfaces are auto-tracked by name match (lan0, wan1, wlan2, phy0, wl1,
ath0, ra0...). Up to MAX_IFACES (16) interfaces per family
This trigger is intended for board/device authors and drivers to provide simple
network-activity LED behaviour without per-interface wiring in userspace.
Refresh bcm27xx patches in the same commit to account for line shifts in
drivers/leds/trigger/Kconfig and Makefile, ensuring clean applies for
bisectability.
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/19903
Signed-off-by: Robert Marko <robimarko@gmail.com>
Add a pending patch to fix the WED flow timeout.
OpenWrt has recently migrated many platforms to kernel 6.18. On the
MediaTek platform, which supports hardware network offloading, WiFi
connections accelerated via the WED path were observed to drop after
roughly 300 seconds.
After several debugging sessions, assisted by the Claude LLM, the
problem was narrowed down as follows:
nf_flow_table_extend_ct_timeout() extends ct->timeout for offloaded
flows using:
cmpxchg(&ct->timeout, expires, new_timeout);
'expires' comes from nf_ct_expires(ct) and is a relative value, while
ct->timeout holds an absolute timestamp. The two are never equal, so
the cmpxchg always fails and the timeout is never extended.
This goes unnoticed for most flows, but a long-lived hardware (WED)
offloaded flow on MediaTek MT7986 eventually has ct->timeout decay to
zero, the conntrack entry is reaped and the connection breaks.
Open-code the relative value from a single READ_ONCE(ct->timeout)
snapshot and compare against that same absolute snapshot in the
cmpxchg, so the timeout extension actually takes effect while the
datapath remains authoritative if it updates ct->timeout concurrently.
Signed-off-by: Qingfang Deng <dqfext@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/23611
Signed-off-by: Robert Marko <robimarko@gmail.com>
Kernel ae18b94099b0 (2025-08-20, in 6.18) introduced a dependency on the
Linux-specific O_LARGEFILE, and kernel 97169cd6d95b (2025-08-20, in
6.18) introduced a dependency on the Linux-specific copy_file_range.
Both of these commits were a part of the
https://lore.kernel.org/all/20250819032607.28727-1-ddiss@suse.de/
series. These new dependencies may not be available on non-Linux
systems, although it is possible to cross-build Linux on non-Linux build
hosts, and it is appropriate to run tools like gen_init_cpio on such
build hosts. It is straightforward to avoid these Linux-specific
features when not building on Linux.
This fixes:
HOSTCC usr/gen_init_cpio
usr/gen_init_cpio.c:460:16: error: call to undeclared
function 'copy_file_range'; ISO C99 and later do not support implicit
function declarations [-Wimplicit-function-declaration]
460 | this_read = copy_file_range(file, NULL, outfd, N...
| ^
usr/gen_init_cpio.c:677:31: error: use of undeclared
identifier 'O_LARGEFILE'
677 | O_WRONLY | O_CREAT | O_LARGEFILE | O_TRUNC,
| ^~~~~~~~~~~
2 errors generated.
gmake[7]: *** [scripts/Makefile.host:114: usr/gen_init_cpio] Error 1
Signed-off-by: Mark Mentovai <mark@mentovai.com>
Link: https://github.com/openwrt/openwrt/pull/23612
Signed-off-by: Robert Marko <robimarko@gmail.com>
Since kernel 5a406031d0719 (2025-03-17, in 6.15), objtool calls
sendfile, but sendfile is Linux-specific and a sendfile with the
requisite capabilities is not available on other operating systems.
Provide a more portable fallback implementation to be used when
cross-building on non-Linux hosts.
This fixes:
CC …/linux-6.18.33/tools/objtool/builtin-check.o
builtin-check.c:13:10: fatal error: 'sys/sendfile.h' file
not found
13 | #include <sys/sendfile.h>
| ^~~~~~~~~~~~~~~~
1 error generated.
gmake[8]: *** […/linux-6.18.33/tools/build/Makefile.build:86: …/linux-6.18.33/tools/objtool/builtin-check.o] Error 1
Fixes: https://github.com/openwrt/openwrt/issues/23356
Signed-off-by: Mark Mentovai <mark@mentovai.com>
Link: https://github.com/openwrt/openwrt/pull/23612
Signed-off-by: Robert Marko <robimarko@gmail.com>
Between the upgrade from 6.12 to 6.18, the airoha-cpufreq driver
now has a `-Wincompatible-pointer-types` error while building due
to the `802-OPP-Provide-old-opp-to-config_clks-on-_set_opp` patch.
This patch revises the driver with the proper signature.
Signed-off-by: Kenneth Kasilag <kenneth@kasilag.me>
Link: https://github.com/openwrt/openwrt/pull/21019
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
A recently added patch needs to be refreshed. Do that with make
target/linux/refresh.
Fixes: 89ef8aaf3e ("kernel: backport lm63 enhancements")
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Add CONFIG_NET_SCH_BPF disabled.
NET_SCH_BPF depends on DEBUG_INFO_BTF.
When CONFIG_KERNEL_DEBUG_INFO_BTF is enabled, the symbol
becomes visible and Kconfig prompts for it because it is
missing from the generic 6.18 kernel config.
Signed-off-by: Til Kaiser <mail@tk154.de>
Link: https://github.com/openwrt/openwrt/pull/23498
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Rosen Penev
Joshua Covington
Andrew LaMarche
John Audia
Carlo SzelinskyandSven Eckelmann
Mieczyslaw Nalewaj
Mikhail Zhilkin
Qingfang Deng
Shiji Yang
Christian Marangi
Jonas Jelonek
Aliaksandr Babrykovich
Luiz Angelo Daros de Luca
Andre Heider
Mark Mentovai
Kenneth Kasilag
Daniel Golle
Jan-Henrik Bruhn
Til Kaiser