IPQ60xx and IPQ807x use threaded NAPI for EDMA. The NAPI threads
inherit CPU0 affinity, and running both there can eventually starve the
RX fill ring under sustained traffic.
Add an IPQ60xx smp_affinity service and extend the existing IPQ807x
service to discover the EDMA IRQs and NAPI IDs at runtime. Place RX on
CPU1 and TX on CPU2. Retry EDMA setup asynchronously because its NAPI
threads may appear after the init service runs.
Provide the same UCI enable and logging controls on both subtargets.
IPQ50xx uses a different Ethernet driver and is left unchanged.
Link: https://github.com/openwrt/openwrt/pull/24203
Signed-off-by: Robert Marko <robimarko@gmail.com>
Run the EDMA RX and TX NAPI poll loops in dedicated kernel threads by
default. This allows the threads and their interrupts to be placed on
separate CPUs and avoids overloading a single softirq context.
If thread creation fails, dev_set_threaded() leaves NAPI in softirq mode
and the driver reports the failure.
Link: https://github.com/openwrt/openwrt/pull/24203
Signed-off-by: Robert Marko <robimarko@gmail.com>
A malformed RX preheader can occur repeatedly after the descriptor ring
overruns. Rate-limit the warning to prevent excessive console output
from making the receive stall worse.
Decode the little-endian preheader metadata before printing it and use
consistent hexadecimal formatting.
Link: https://github.com/openwrt/openwrt/pull/24203
Signed-off-by: Robert Marko <robimarko@gmail.com>
RX descriptor fields use little-endian byte order. Decode the DMA
address and status with the matching helpers before consuming them.
This makes the byte order explicit and keeps the code portable. The
conversions remain no-ops on little-endian IPQ60xx.
Link: https://github.com/openwrt/openwrt/pull/24203
Signed-off-by: Robert Marko <robimarko@gmail.com>
c3bf07a66f7e make: honor SOURCE_DATE_EPOCH in BUILD_DATE
bb4f3f0b947e Revert "MP: fix uniphy reset in phy to phy link scenario"
Signed-off-by: Robert Marko <robimarko@gmail.com>
7df188543e26 libfstools: enable f2fs overlay compression formatting
16718b6e3c0f libfstools: mount f2fs overlay with zstd compression
Signed-off-by: Robert Marko <robimarko@gmail.com>
Backport the Microchip EMC2305 driver changes up to current linux-next.
These are required in order to actually be able to use the driver with
DTS support to drive fans as cooling devices.
Replace the bcm27xx RPi patch with the one from RPi kernel 7.0 that was
already adapted to all of these backports.
Link: https://github.com/openwrt/openwrt/pull/22942
Signed-off-by: Robert Marko <robimarko@gmail.com>
CPR driver was converted to use of_device_get_match_data and it locally
constified the CPR match data structures but they are also stored per
regulator in their struct, these were not constified and would cause a
compile time error.
So, constify those as well.
Fixes: c1fa85f659 ("treewide: use of_device_get_match_data")
Signed-off-by: Robert Marko <robimarko@gmail.com>
No need to add another aliases node just for the label-mac-device,
luckily DTC was smart enough to combine them together in the final DTB.
Signed-off-by: Robert Marko <robimarko@gmail.com>
CONFIG_TOOLS_SUPPORT_RELR is set during runtime by the kernel so it should
not end up in our static configs.
So, filter it out by default.
Signed-off-by: Robert Marko <robimarko@gmail.com>
If DSA support is enabled, then these will pop up.
So instead of ending up in target configs, add them to the generic configs.
Signed-off-by: Robert Marko <robimarko@gmail.com>
There is no point in printing the missing M3 memory dump adress message
on each boot under the warning level, as not all boards need it at all.
So, degrate it to a debug print with QMI mask.
Signed-off-by: Robert Marko <robimarko@gmail.com>
22fb70661799 fix flex array not at end of struct
6a5c4716ca25 convert memcpy + ETH_GSTRING_LEN to ethtool_puts
Signed-off-by: Robert Marko <robimarko@gmail.com>
46f373b47f69 tplink-safeloader: add support for TP-Link Archer AX21 v4.6
7324b0ba8e05 tplink-safeloader: fix segfault when partition name is NULL
7593018845d8 asusuimage: Cleanup code and fix typo
caac8b133aca tplink-safeloader: fix soft_ver for Archer AX21
c0d7de851c9a ptgen: fix bug caused by not completely correct reverts
5b6ef84eaa74 ptgen: allow to specify index of gpt entries to be used
467685270cf0 ptgen: add an option to disable stub partition creation
a2c06c39b41b ptgen: add long option support
6a87eaf434cb ptgen: add support for marking multiple partitions as bootable
Fixes: https://github.com/openwrt/firmware-utils/issues/59
Signed-off-by: Robert Marko <robimarko@gmail.com>
This reverts commit c053b22573.
KSZ DSA driver is the only thing in the kernel selecting DCB support
instead of depending on it if required.
So, it will enable DCB support without asking and we do not want the
kernel size increase, as well as current Layerscape ARMv8 build failure.
So, revert this until its fixed upstream or worked around.
Signed-off-by: Robert Marko <robimarko@gmail.com>
Not all targets need regulator support, so they dont enable it as its
disabled in the generic config by default.
So, in order to allow kernel modules to depend on regulator support lets
add a new feature flag "regulator" and set it automatically if target
kernel config enables CONFIG_REGULATOR.
Link: https://github.com/openwrt/openwrt/pull/22172
Signed-off-by: Robert Marko <robimarko@gmail.com>
CONFIG_REGULATOR_QCOM_LABIBB is being disabled by targets manually, so
instead lets disable it in generic config.
Signed-off-by: Robert Marko <robimarko@gmail.com>
Instead of disabling CONFIG_REGULATOR_QCOM_SPMI manually per target,
lets simply disable it in generic config.
Signed-off-by: Robert Marko <robimarko@gmail.com>
CONFIG_REGULATOR_QCOM_USB_VBUS showed up on Layerscape recently, and it
looks like multiple targets disable it manually, so simply disable it in
generic configuration instead.
Signed-off-by: Robert Marko <robimarko@gmail.com>
Instead of redifining the DEVICE_PACKAGES, make sure to expand the list
so yafut is included by default as well.
Fixes: 42cd486787 ("ath79: remove kmod-ath9k and wpad from Mikrotik RB750r2")
Signed-off-by: Robert Marko <robimarko@gmail.com>
Follow the recent change where all DTS files were moved to a dedicated dir.
Fixes: 3a39f682df ("qualcommax: ipq50xx: add support for CMCC MR3000D-CI")
Signed-off-by: Robert Marko <robimarko@gmail.com>
I botched the rebase of ipq-wifi, so fix it.
Fixes: 67e6baf05e ("qualcommax: ipq60xx: add Netgear RBx350 support")
Signed-off-by: Robert Marko <robimarko@gmail.com>
Currently, kmod-crypto-kpp is only implied for curve25519 and thus it
can get unselected and then building will fail due to missing kpp.ko.
Same issue happens with kmod-crypto-hash for poly1305.
So, lets drop the whole implies logic and make sure to always select
them as dependencies.
Fixes: #21386
Link: https://github.com/openwrt/openwrt/pull/21483
Signed-off-by: Robert Marko <robimarko@gmail.com>
BDF for EAP623-Outdoor HD v1 was renamed in the repo, but ipq-wifi
was not updated to reflect that, so do it now to fix broken wifi.
Fixes: #21432
Signed-off-by: Robert Marko <robimarko@gmail.com>
This reverts commit 423cda97f9.
It appears that this broke remoteproc on some devices, so revert for now.
Signed-off-by: Robert Marko <robimarko@gmail.com>
c6202981872e ipq5018: use correct board for GL-iNet B3000 IPQ5018 radio
ec72376cadf1 qca4019: Add Meraki Underdog
059b7114c77b ipq6018: add tp-link eap620hd v2
Signed-off-by: Robert Marko <robimarko@gmail.com>
cdedfd891bdf nss-dp: fix hardware STP state sync with the kernel
19c51af0c5be nss-dp: switchdev: improve FDB cleanup for roaming
Signed-off-by: Robert Marko <robimarko@gmail.com>
a284e7751fa7 file: bump sz_size to 64 bits
ffb9961c1f8b file: provide user name and group name lookups for stat listings
Signed-off-by: Robert Marko <robimarko@gmail.com>
CPR conversion to of_property_present is broken in multiple places, either
by leaving the comma after now the last argument or by attempting to replace
an instance that actually uses the returned length.
So, lets fix these in order for qualcommax to compile again.
Fixes: c9e7f32c4c ("treewide: use of_property_present")
Signed-off-by: Robert Marko <robimarko@gmail.com>
APK will generate the compressed scripts.tar by default, however we rely
on using TAR support for updating the tarball on the fly and this does not
work when tarball is compressed and will cause the following errors:
staging_dir/host/bin/tar: Cannot update compressed archives
Try /staging_dir/host/bin/tar --help' or
staging_dir/host/bin/tar --usage' for more information.
So, lets simply decompress the scripts.tar.gz before usage and then
compress it after we are done.
Fixes: 5d85657f6d ("apk-tools: implement compression of on-device scripts.tar")
Signed-off-by: Robert Marko <robimarko@gmail.com>
fa62f8a2885c ipq8074: add BDF for Zyxel NWA210AX
7c6b7545f69a qca4019: qca9888: add bdfs for Netgear RBK20
Signed-off-by: Robert Marko <robimarko@gmail.com>
It seems that adding Yuncore AX850 was missed during the board support,
and this breaks buildbots for ipq50xx with:
ERROR: unable to select packages:
ipq-wifi-yuncore_ax850 (no such package):
required by: world[ipq-wifi-yuncore_ax850]
make[4]: *** [/builder/shared-workdir/build/include/image.mk:396: target-dir-90b51871] Error 1
make[4]: *** Waiting for unfinished jobs....
So, add the required meta package.
Fixes: 5d2994a73e ("qualcommax: ipq50xx: Add support for Yuncore AX850")
Signed-off-by: Robert Marko <robimarko@gmail.com>
Invalid hash was added during the 5.3 update and CI will fail since none
of the available sources has the tarball with incorrect hash.
Fixes: 423ed9fc5a ("tools: bash: update to 5.3")
Signed-off-by: Robert Marko <robimarko@gmail.com>
util-linux needs to be refreshed, so do so.
Fixes: 4fd6117129 ("tools/util-linux: fix libuuid linking")
Signed-off-by: Robert Marko <robimarko@gmail.com>
Since upstream commit:
b809fc656e763296f227b9b31e8f225e5977a8af ("perf build: Shellcheck support
for OUTPUT directory")
perf will attempt to run shellcheck on the test shell scripts, however
there is no point in doing this in OpenWrt and while perf checks for
shellcheck presence on your host it can fail to build in some cases.
So, simply disable it for now.
Link: https://github.com/openwrt/openwrt/pull/19361
Signed-off-by: Robert Marko <robimarko@gmail.com>
BUILD_BPF_SKEL was set to 1 by default in upstream commit:
9925495d96efc14d885ba66c5696f664fe0e663c ("perf build: Default
BUILD_BPF_SKEL, warn/disable for missing deps")
Prior to that, it was disabled by default and you had to enable it to
build BPF skeleton support.
So in order to fix perf compilation with kernel 6.12, lets disable
BUILD_BPF_SKEL.
Fixes: #19310
Link: https://github.com/openwrt/openwrt/pull/19361
Signed-off-by: Robert Marko <robimarko@gmail.com>
5be4c60a9084 ipq5018: add BDF for Xiaomi AX6000
22588512b844 qcn9074: add BDF for Xiaomi AX6000
4810aacf3b1c qca9889: add BDF for Xiaomi AX6000
Signed-off-by: Robert Marko <robimarko@gmail.com>
61ae5732adea iprule: amend ipproto netlink nla_put_u32 to nla_put_u8
d610d68c71b8 device: add support for configuring vrf
a1b6386a20a6 device: fix bonding primary port selection
e8bbf246ce2e system-linux: fix sysfs name for all_ports_active flag
723c699e84f4 Restore disable_ipv6 sysctl after removing a device from bridge or bond
d476e18e8d43 iprule: resolve ipproto by name
7901e66c5f27 netifd: iprule add sport and dport
Signed-off-by: Robert Marko <robimarko@gmail.com>
archs38 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.
Link: https://github.com/openwrt/openwrt/pull/19001
Signed-off-by: Robert Marko <robimarko@gmail.com>
There is no practical value in keeping GCC11 around, as even OpenWrt 23.05
uses GCC12 as the default one, so drop it.
Signed-off-by: Robert Marko <robimarko@gmail.com>
Currently, MaxLinear/Exar USB serial devices are supported via out-of-tree
usb-serial-xr_usb_serial_common driver which is broken on 6.12.
So since upstream has support for these devices since 6.5 lets package
the in-tree driver to be able to drop the out-of-tree one.
Link: https://github.com/openwrt/openwrt/pull/18926
Signed-off-by: Robert Marko <robimarko@gmail.com>
Currently, we are filtering out images if DEFAULT:=n or BROKEN:=y are set,
so if you are building from scratch and want to build custom images that
are stripped down to fit, you must edit the image recipe or its just
filtered out.
So, to allow this behaviour when building from scratch as we can assume
that person doing that knows what they are attempting to do lets just limit
the filtering to ImageBuilder.
Fixes: f060615a78 ("image: respect DEFAULT and BROKEN when Default profile is selected")
Signed-off-by: Robert Marko <robimarko@gmail.com>
Default qualcommax target to 6.12 kernel.
It makes no sense for us to double the effort required for backporting
upstreamed patches or for downstream modifications.
Link: https://github.com/openwrt/openwrt/pull/18840
Signed-off-by: Robert Marko <robimarko@gmail.com>
New m4 release was made after a long time, for us most importantly
it ships newer gnulib so it compiles against GCC15 without forcing GNU17.
* Noteworthy changes in release 1.4.20 (2025-05-10) [stable]
** Fix a bug in the `eval' builtin where it does not suppress warnings
about division by zero that occurs within a more complex expression on
the right hand side of || or && (present since short-circuiting was
introduced in 1.4.8b).
** The `syscmd' and `esyscmd' builtins no longer mishandle a command line
starting with `-' or `+' (present since "the beginning").
** Fix regression introduced in 1.4.19 where trace output (such as with
`debugmode(t)') could read invalid memory when tracing a series of
pushed macros that are popped during argument collection.
** Fix regression introduced in 1.4.19 where the `format' builtin
inadvertently took on locale-dependent parsing and output of floating
point numbers as a side-effect of introducing message translations.
While it would be nice for m4 to be fully locale-aware, such a behavior
change belongs in a major version release such as 1.6, and not a minor
release.
** Fix regression introduced in 1.4.11 where the experimental `changeword'
builtin could cause a crash if given a regex that does not match all
one-byte prefixes of valid longer matches. As a reminder, `changeword'
is not recommended for production use, and will likely not be present
in the next major version release.
** On non-Unix platforms where binary files differ from text, loading a
frozen file (which should be cross-platform compatible) now correctly
uses binary mode.
** Several documentation improvements to the manual.
** Update to comply with newer C standards, and inherit portability
improvements from gnulib.
Link: https://github.com/openwrt/openwrt/pull/18821
Signed-off-by: Robert Marko <robimarko@gmail.com>
CONFIG_DRM_ACCEL_QAIC was not added to x86 generic config, but it needs to
be added as CONFIG_DRM_ACCEL_QAIC does not depend on x86_64 like others.
Fixes: 51b78aa ("kernel: package intel_vpu driver")
Signed-off-by: Robert Marko <robimarko@gmail.com>
Currently, when you select the Default profile it does not honor DEFAULT:=n
nor BROKEN:=y in device profiles but rather just tries to build all of them.
This may work when building directly, but when using Image Builder it will
always fail since no kernel or anything else is present for devices that
have DEFAULT:=n or BROKEN:=Y set since those are skipped during build.
So, lets look for DEFAULT being set to "n" or BROKEN being set to "y" and
then remove clear _PROFILE_SET so they dont end up being marked for
installation.
Fixes: #18410
Link: https://github.com/openwrt/openwrt/pull/18814
Signed-off-by: Robert Marko <robimarko@gmail.com>
There are NAND IC-s that define 1-bit ECC as the minimal strength,
however that is unsupported by QPIC-SNAND as it only supports 4 or 8 bit
ECC.
Since most of these chips also support 4-bit ECC just fine, instead of
erroring out if 1-bit ECC is requested lets instead default to 4-bit ECC.
Fixes: 01b72ce61e ("qualcommax: ipq50xx: remove ECC user config from board files")
Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18795
Signed-off-by: Robert Marko <robimarko@gmail.com>
Update to current SSDK head for 6.12 compatibility.
We must disable warnings for missing declarations and prototypes since
whole SSDK is litered with them and actually fixing them is like doing
whack a mole.
Link: https://github.com/openwrt/openwrt/pull/18795
Signed-off-by: Robert Marko <robimarko@gmail.com>
6.12 PWM core introduced a bunch of incompatible changes, namely removal
of manual module owner assignment, complete PWM struct allocation and usage
refactor, etc.
So, update the driver to follow other drivers in 6.12 so it compiles.
Link: https://github.com/openwrt/openwrt/pull/18795
Signed-off-by: Robert Marko <robimarko@gmail.com>
Robert Marko