Marvell mv88e6xxx switch series cannot perform MAC learning from
CPU-injected (FROM_CPU) DSA frames, which results in 2 issues.
- excessive flooding, due to the fact that DSA treats those addresses
as unknown
- the risk of stale routes, which can lead to temporary packet loss
Backport those patch series from netdev mailing list, which solve these
issues by adding and clearing static entries to the switch's FDB.
Add a hack patch to set default VID to 1 in port_fdb_{add,del}. Otherwise
the static entries will be added to the switch's private FDB if VLAN
filtering disabled, which will not work.
The switch may generate an "ATU violation" warning when a client moves
from the CPU port to a switch port because the static ATU entry added by
DSA core still points to the CPU port. DSA core will then clear the static
entry so it is not fatal. Disable the warning so it will not confuse users.
Link: https://lore.kernel.org/netdev/20210106095136.224739-1-olteanv@gmail.com/
Link: https://lore.kernel.org/netdev/20210116012515.3152-1-tobias@waldekranz.com/
Ref: https://gitlab.nic.cz/turris/turris-build/-/issues/165
Signed-off-by: DENG Qingfang <dqfext@gmail.com>
The update to gettext 0.21 broke packages that use autotools and
gettext because the sed line was failing with the new version. Fix with
a better sed expression.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Add symbol to generic config (this was added between 5.4 and 5.10),
and remove it from the targets where it was added by kernel_oldconfig
in the meantime.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
There are efforts underway to bring wireguard in-tree for Linux 5.4 and
to have a common build infrastructure for both 5.4 and 5.10 for
kmod-wireguard[0]. Until then, restrict kmod-wireguard to build only on
Linux 5.4, because the wireguard-compat package will not build on Linux
5.10.
[0]: https://github.com/openwrt/openwrt/pull/3885
Signed-off-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
Modify existing modules to reflect their new location in Linux 5.10. Add
missing dependenices.
Signed-off-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
[enable CRYPTO_USER_API_ENABLE_OBSOLETE; add kmod-crypto-hash dependency
to usb-net-rtl8152]
Signed-off-by: David Bauer <mail@david-bauer.net>
Now that mirrors have picked it up, switch to using the @OPENWRT
mirror instead of hosting those files on Github.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Description:
1. From key and led config setting, we can find only "uartf" and "i2c" are used
as gpio by check mt7620 datasheet. It's time to remove unused pin group.
2. PSG1218 only have three led, so we can remove ethernet led pinctrl. refer to
Phicomm K2G.
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Keep 5.4 as stable until further validation.
Tested on Turris Omnia (Rui Salvaterra) and ESPRESSObin v5 (Tomasz
Maciej Nowak). Cortex-A{53,72} subtargets are only build-tested.
Reviewed-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Tested-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
[added comment about tests]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Enable and fix hardware buffer management. Also fix the IRQ storm caused by a
misconfiguration of the PCA9538 interrupt pin.
Reviewed-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Tested-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
Make each logical patch group numbering contiguous.
Reviewed-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Tested-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
[add kernel version to backport]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Also delete already upstreamed patches/changes.
Reviewed-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Tested-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
Just a simple copy, no refresh yet.
Reviewed-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Tested-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
[do not duplicate files]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Remove the implicit/inherited symbols. While not strictly necessary, this will
make reviewing the diff between 5.4 and 5.10 easier.
Reviewed-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Tested-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
This will make the specific kconfig smaller.
Reviewed-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Tested-by: Tomasz Maciej Nowak <tmn505@gmail.com>
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
[improved commit title]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
SMP isn't supported on BCM6358 since it has a shared TLB. Some boards boot
with CPU #1 instead of CPU #0, and this is currently not supported do to a
smp-bmips bug.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
2be57ed cosmetics: provide compatible system info on Aarch64
37eed13 system: expose if system was booted from initramfs
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
The current driver has some troubles:
- Some groupings are wrong.
- The pinctrl group0 owns pins never used (at least in Openwrt) for any
pinmux. The driver hijacks all the pins on the group avoiding any other
use, spite they're free. I.e. for buttons, causing this kernel error:
[ 4.735928] gpio-keys-polled keys: unable to claim gpio 479, err=-22
[ 4.742642] gpio-keys-polled: probe of keys failed with error -22
- Minor errors about groupings on the documentation
- Missing "diag" grouping in dtsi
- Wrong groupings in dtsi
Fix it by setting the correct groups.
And relax the pin capturing, letting the gpios belonging to any group to
be used for other purposes like buttons. This was the behavior with stock
firmwares and old OpenWrt versions which never caused any trouble.
Signed-off-by: Daniel González Cabanelas <dgcbueu@gmail.com>
IMAGE_NAME is set twice for ventana, and the second value
actually matches the default in image.mk. Remove both.
Fixes: ded905ce43 ("imx6: extend cubox support to hummingboard,
add support for building full images")
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The variables KERNEL_INITRAMFS_PREFIX and KERNEL_PREFIX are already
defined in include/image.mk and don't have to be redefined in the
target Makefiles.
Signed-off-by: Paul Spooren <mail@aparcar.org>
[also cover imx6]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Like done for several targets already, splitting base-files into
subtarget will provide smaller images due to more specific
distribution of files per subtarget and allow to use default cases
more effectively.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
IPSEC needs a reset before using its HW RNG.
Otherwise, the numbers generated won't be random at all.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
The lan port sequence was reversed compared to the labels.
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
[improve commit title/message]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The aliases node is expected as one of the first entries, and
having it there matches alphabetic sorting as well.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
These were redefines of the same value already set in the SoC dtsi
files.
Reported-by: Shiji Yang <yangshiji66@qq.com>
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This target has full device tree support, thus reducing the number of
patches needed for bcm63xx, in which there's a patch for every board.
The intention is to start with a minimal amount of downstream patches and
start upstreaming all of them.
Current status:
- Enabling EHCI/OHCI on BCM6358 causes a kernel panic.
- BCM63268 lacks Timer Clocks/Reset support.
- No PCI/PCIe drivers.
- No ethernet drivers.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Acked-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Since there are only 16 characters available, on most cases the vendor name
will fit in the metadata, but the model name won't fit.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Broadcom CFE bootloader relies on a tag for identifying the current firmware,
such as version, image start address, kernel address and size, rootfs size,
board id, signatures, etc.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Phicomm K2G:
add missing label_mac
Phicomm PSG1218A & PSG1218B:
The previous wan mac was set as factory@0x28 +1 (originally based
on the default case for the ramips target), but the correct wan mac
is factory@0x28 -1, being equal to factory@0x2e.
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
[minor commit title/message adjustments]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Regulator support was enabled on all subtargets except for ath79-nand.
With Kernel 5.10, AT803x requires Regulator support, thus enabling on
the complete target, as ath79-nand requires AT803x.
While this is only required on Kernel 5.10, enable it also on 5.4. We
have no major size-constraint, so enabling it on 5.4 allows us to clean
up the occurences in the subtarget configuration.
Signed-off-by: David Bauer <mail@david-bauer.net>
As mangix pointed out on IRC, ioremap and ioremap_nocache are
functionally equivalent on kenrel 5.4 and 5.10.
Therefore we can use ioremap regardless of the kernel the driver
gets compiled for.
Signed-off-by: David Bauer <mail@david-bauer.net>
This was done by executing this script:
find . -name "config-*" > ../configs.txt
for config in $(cat ../configs.txt); do
./scripts/kconfig.pl '+' $config /dev/null > $config-new
mv $config-new $config
done
rm ../configs.txt
Signed-off-by: Aleksander Jan Bajkowski <A.Bajkowski@stud.elka.pw.edu.pl>
Prerequisite patch:
Correct a typo in the Changelog and clean up a stray file
Fix changes in libusb which introduced a regression:
Commit e2be556bd2 ("linux_usbfs: Parse config descriptors during device
initialization") introduced a regression for devices with multiple
configurations. The logic that verifies the reported length of the
configuration descriptors failed to count the length of the
configuration descriptor itself and would truncate the actual length by
9 bytes, leading to a parsing error for subsequent descriptors.
Signed-off-by: Georgi Valkov <gvalkov@abv.bg>
The initramfs images are missing from the profiles.json files.
Signed-off-by: Moritz Warning <moritzwarning@web.de>
[fix code by exporting device variables]
Signed-off-by: Paul Spooren <mail@aparcar.org>
