Add support for qca8327 internal phy needed for correct init of the
switch port. It does use the same qca8337 function and reg just with a
different id.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Signed-off-by: Matthew Hagan <mnhagan88@gmail.com>
Backport fixes including:
net: dsa: qca8k: fix missing unlock on error in qca8k_vlan_(add|del)
net: dsa: qca8k: check return value of read functions correctly
net: dsa: qca8k: add missing check return value in qca8k_phylink_mac_config()
net: dsa: qca8k: fix an endian bug in qca8k_get_ethtool_stats()
net: dsa: qca8k: check the correct variable in qca8k_set_mac_eee()
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Signed-off-by: Matthew Hagan <mnhagan88@gmail.com>
This is a backport of Ansuel Smith's "Multiple improvement to qca8k stability"
series. The QCA8337 switch is available on multiple platforms including
ipq806x, ath79 and bcm53xx.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Signed-off-by: Matthew Hagan <mnhagan88@gmail.com>
Rosen reported strange dtc warnings that had their origin in
an upstream patch to 5.8-rc1. Upon further digging this
revealed an ongoing thread [0] discussing the topic:
> [...]I don't think we need a bunch of warning fix patches to add
> these everywhere. Also, the need for #address-cells pretty much makes
> no sense on any modern system. It is a relic from days when the bus
> (address) topology and interrupt topology were related.
and later on:
> So really, we only need to be checking for #address-cells in nodes
> with interrupt-map.
This patch backports just the patch which removed the warning message
(this is from the upstream dtc project [1] - but not the kernel).
the patch does not add the checking of the #address-cells in nodes
with interrupt-map.
[0] <https://lore.kernel.org/linux-devicetree/91e3405245c89f134676449cf3822285798d2ed2.1612189652.git.michal.simek@xilinx.com/>
[1] <https://git.kernel.org/pub/scm/utils/dtc/dtc.git/commit/?id=d8d1a9a77863a8c7031ae82a1d461aa78eb72a7b>
Link: <https://github.com/openwrt/openwrt/pull/4685>
Reported-by: Rosen Penev <rosenp@gmail.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
This allows to specify and control switch LEDs on devices using mt7530
(typically mediatek and ramips targets).
Normally these LED GPIOs are 0, 3, 6, 9, and 12. wan/lan assignment is
per device. GPIO 9 is normally inverted. so GPIO_ACTIVE_HIGH instead of
GPIO_ACTIVE_LOW.
Tested on Linksys E7350.
Refreshed all patches.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Backport upstream SFP support for the Marvell 88E1510/2 PHY-s.
Globalscale MOCHAbin uses this PHY for the hybrid
WAN port that has 1G SFP and 1G RJ45 with PoE PD
connected to it.
This allows the SFP port to be used on it as well as
parsing the SFP module details with ethtool.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Backport upstream support for 100Base-FX, 100Base-LX, 100Base-PX and
100Base-BX10 SFP modules.
This is a prerequisite for the Globalscale MOCHAbin hybrid 1G
SFP/Copper support backporting.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Removed upstreamed:
backport-5.4/070-v5.5-MIPS-BPF-Restore-MIPS32-cBPF-JIT.patch
All other patches automatically rebased.
Signed-off-by: John Audia <graysky@archlinux.us>
Now that we have fully switched to nvmem interface we can drop
the use of mtd-mac-address patches as it's not used anymore and
the new nvmem implementation should be used for any new device.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
This fixes:
ERROR: "switch_generic_set_link" [drivers/net/phy/b53/b53_common.ko] undefined!
At some point all packages for swconfig drivers were dropped and targets
were meant to have them built into kernels. It seems b53 (re-)gained its
kmod-switch-bcm53xx however and b53 needs to be built as module.
Fixes: b2cfed48f6 ("Revert "swconfig: fix Broadcom b53 support"")
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Add Vladimir Oltean's "net: dsa: don't set skb->offload_fwd_mark when
not offloading the bridge"
This covers cases where packets received by an upstream switch must be
forwarded back on the same port, which skb->offload_fwd_mark normally
prevents.
Signed-off-by: Matthew Hagan <mnhagan88@gmail.com>
This reverts commit 8f9cd1af0f.
That commit was meant to add a single EXPORT_SYMBOL_GPL() but it
actually also added few .of_match_table-s. One commit should handle one
thing and should not introduce unrelated changes.
Regarding actual changes:
1. EXPORT_SYMBOL_GPL is not required as we don't build swconfig drivers
as modules.
2. PHY drivers must not have .of_match_table. That is allowed for MDIO
drivers. This could work for some time (although is didn't for me on
bcm53xx) but does not with kernel 5.10. It causes a soft lockup and
upstream developers confirmed it's an unsupported design.
Link: https://lore.kernel.org/netdev/2b1dc053-8c9a-e3e4-b450-eecdfca3fe16@gmail.com/t/#mf80e472f35ee23f7a75cbf5b1e101a17ab3a64a3
Cc: Tobias Schramm <tobleminer@gmail.com>
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Three missing symbols were found during mpc85xx/p2020 compilation.
While at it, CONFIG_FSL_ENETC_MDIO is moved to generic config
for consistency.
Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
[move CONFIG_FSL_ENETC_MDIO, remove redundant definitions, adjust
commit message/title]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
SATA_PMP option is no longer exposed when no SATA host driver is enabled
since upstream linux commit bd322af15ce9 ("ata: make SATA_PMP option
selectable only if any SATA host driver is enabled").
Commit 1bb3f593ee ("kirkwood: update config for kernel 5.10") manually
added CONFIG_SATA_PMP=y to config file, but the config will disappear for
every kernel_oldconfig refresh.
To prevent this, a small hack is added, which selects SATA_HOST
automatically when SATA_PMP is enabled. This patch can be dropped if
SATA_MV is ever re-added into kernel config file.
Tested-by: Pawel Dembicki <paweldembicki@gmail.com>
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
[Move patch to generic/hack-5.10]
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This backports the upstream commit:
leds: lp55xx: Initialize enable GPIO direction to output
Without it under kernel 5.10 on Asus MAP-AC2200
the LED driver will fail probing:
[ 1.947521] lp5523x: probe of 0-0032 failed with error -22
After the backported fix:
[ 1.873236] lp5523x 0-0032: lp5523 Programmable led chip found
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Tested-by: Szabolcs Hubai <szab.hu@gmail.com> [ipq4029/gl-b1300]
Tested-by: Nick Hainke <vincent@systemli.org> [ipq4019/fritzbox-7530
ipq4019/fritzbox-4040
ipq4019/sxtsq-5ac]
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de> [ipq4019/map-ac2200]
This fixes:
Distributed Switch Architecture (NET_DSA) [Y/n/m/?] y
Tag driver for Atheros AR9331 SoC with built-in switch (NET_DSA_TAG_AR9331) [N/m/y/?] n
Tag driver for Broadcom switches using in-frame headers (NET_DSA_TAG_BRCM) [N/m/y/?] n
Tag driver for Broadcom legacy switches using in-frame headers (NET_DSA_TAG_BRCM_LEGACY) [N/m/y/?] (NEW)
Error in reading or end of file.
Fixes: 8fa1b576bb ("linux: update b53 upstream driver")
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
This patch fixes the forwarding behavior of bridge in bridge
configurations with DSA.
Without it, the configuration of the upper bridge might overwrite
settings of the lower bridge. For example, a vlan-aware bridge
with DSA interfaces in it might be offloaded to the DSA hardware. If the
bridge interface itself gets slave of a different bridge without vlan
filtering, the vlan filtering setting of the lower bridge is overwritten
by the upper bridge, which results in an incorrect hardware
configuration.
This was backported from kernel 5.7.
Ref: https://lore.kernel.org/netdev/20191222192235.GK25745@shell.armlinux.org.uk/
Fixes: FS#3996
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Designated initializers are required when using the randstruct GCC
plugin, otherwise an error like the following is seen:
./include/linux/lzma.h:60:31: error: positional initialization of field in 'struct' declared with 'designated_init' attribute [-Werror=designated-init]
Signed-off-by: Matt Merhar <mattmerhar@protonmail.com>
These options are selectable when some of the kernel debug options like
KERNEL_SOFTLOCKUP_DETECTOR are selected.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The default value for CONFIG_RCU_CPU_STALL_TIMEOUT was changed from 60
seconds to 21 seconds in 2012 in the upstream kernel. Some targets
already use 21 seconds.
This patch changes the default value in the generic configuration to 21
seconds and removes the target specific configuration options.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Acked-by: Rui Salvaterra <rsalvaterra@gmail.com>
CONFIG_RCU_{NEED_SEGCBLIST,STALL_COMMON} are set basically everywhere. Move them
to the generic kconfigs. And resort the generic kconfigs while at it.
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
Based on the existing documentation [1][2], I dare anyone to demonstrate that
we need to fine-tune these RCU parameters. The (performance) breakage potential
for doing so is immense, so let's just please put down this loaded footgun.
Disable CONFIG_RCU_EXPERT and its dependent symbols. Additionally, remove the
CONFIG_RCU_EXPERT symbol from the target kconfigs which contain it.
[1] https://www.kernel.org/doc/Documentation/RCU/Design/Data-Structures/Data-Structures.html
[2] https://lwn.net/Articles/777214/
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
There is a longlasting warn about unused defined try_to_force_load.
Fix this by including this function only when actually used.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
The uImage.FIT partition uses page mapping without properly handling
paging errors. This can lead to Kernel Oops in case of read errors
while trying to parse uImage.FIT partitions.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
The current patch produces the following error when CONFIG_DMABUF_HEAPS is
enabled:
drivers/built-in.a: member drivers/dma-buf/heaps in archive is not an object
Fixes: b10d604459 ("kernel: add linux 5.10 support")
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
When a target configuration has unser Kconfig symbols, the build will
fail when OpenWrt is compiled with V=s and stdin is connected to a tty.
In case OpenWrt is compiled without either of these preconditions, the
build will uscceed with the symbols in question being unset.
Modify the kernel configuration in a way it fails on unset symbols
regardless of the aformentioned preconditions.
Signed-off-by: David Bauer <mail@david-bauer.net>
Calling free for the OF property can result in a kernel panic, as the
buffer in question might be referenced elsewhere. Also, it is not
removed from the tree.
Always allocate a new property and updating the tree with it fixes both
issues.
Fixes commit 91a52f22a1 ("treewide: backport support for nvmem on non platform devices")
Signed-off-by: David Bauer <mail@david-bauer.net>
The code from ks8851.c was moved to ks8851_common.c, so it was not
backported. This broke the compile of the omap target which uses this
driver.
Fixes commit 91a52f22a1 ("treewide: backport support for nvmem on non platform devices")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Add the missing CONFIG_KCSAN (disabled). Found while making kernel_oldconfig on
an x86-64 subtarget.
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
kirkwood build broke due to missing include needed for ETH_ALEN.
Add patch (sent upstream as well) to address that.
Refresh patches for 5.4 and 5.10.
Fixes: 91a52f22a1 ("treewide: backport support for nvmem on non platform devices")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Add the missing ARM_SCMI_PROTOCOL symbol. Apparently it was exposed
for 5.10.53 with a kernel dependency change.
Missing symbol observed with mediatek/7622 E8450/RT3200 router.
Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
The virtual cable tester depends on the netlink interface for ethtool.
Thus, enable it in the generic kernel configuration.
Signed-off-by: David Bauer <mail@david-bauer.net>
In the current state, nvmem cells are only detected on platform device.
To quickly fix the problem, we register the affected problematic driver
with the of_platform but that is more an hack than a real solution.
Backport from net-next the required patch so that nvmem can work also
with non-platform devices and rework our current patch.
Drop the mediatek and dsa workaround and rework the ath10k patches.
Rework every driver that use the of_get_mac_address api.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
bcm27xx-bcm2710 builds are stalling when compiled with V=s.
Explitily disable these unset symbols to avoid stalling
builds.
Signed-off-by: David Bauer <mail@david-bauer.net>
Fixes a unused variable warning:
drivers/of/of_net.c: In function 'of_get_mac_address_mtd':
drivers/of/of_net.c:92:19: warning: unused variable 'prop' [-Wunused-variable]
Signed-off-by: David Bauer <mail@david-bauer.net>
The 1st generation MediaTek PCIe host bridge cannot handle Message
Signaled Interrupts (MSIs). The core PCI code is not aware that MSI is
not available. This results in warnings of the form:
WARNING: CPU: 2 PID: 112 at include/linux/msi.h:219
pci_msi_setup_msi_irqs.constprop.8+0x64/0x6c
Modules linked in: ahci(+) libahci libata sd_mod scsi_mod
gpio_button_hotplug
CPU: 2 PID: 112 Comm: kmodloader Not tainted 5.10.52 #0
Hardware name: Mediatek Cortex-A7 (Device Tree)
Import patches that introduce the 'no_msi' attribute to signal missing
MSI support to the core PCI.
Refresh patches:
- 000-spi-fix-fifo.patch
- 330-mtk-bmt-support.patch
- 510-net-mediatek-add-flow-offload-for-mt7623.patch
- 601-PCI-mediatek-Use-regmap-to-get-shared-pcie-cfg-base.patch
- 610-pcie-mediatek-fix-clearing-interrupt-status.patch
- 700-net-ethernet-mtk_eth_soc-add-support-for-coherent-DM.patch
- 710-pci-pcie-mediatek-add-support-for-coherent-DMA.patch
Signed-off-by: Nick Hainke <vincent@systemli.org>
The label-mac logic relies on the mac-address property of a netdev
devices of-node. However, the mac address can also be stored as a
different property or read from e.g. an mtd device.
Create this node when reading a mac-address from OF if it does not
already exist and copy the mac-address used for the device to this
property. This way, the MAC address can be accessed using procfs.
Signed-off-by: David Bauer <mail@david-bauer.net>
This reverts commit b309248730.
This commit could create a property without allocated memory, breaking
subsequent reads over a nodes property. Also, the mac-address-increment
was not applied when reading from nvmem.
Revert this commit for now, which breaks the label-mac-address logic.
Possibly, traversing the device-tree from the netdev side is easier
anyways.
Signed-off-by: David Bauer <mail@david-bauer.net>
Add CONFIG_HAVE_ARM_ARCH_TIMER (disabled). A make kernel_oldconfig on cortexa9
will otherwise prompt for its selection. The 5.4 configuration already contains
the same symbol.
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
Add the new CONFIG_BATTERY_RT5033 to the generic configuration, as reported by
Paul Blazejowski. Resort the kconfig while at it.
No deleted or manually refreshed patches.
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
Adds support for GPON SFP modules based on the Realtek RTL8672 and
RTL9601C chips, including but not limited to:
* V-SOL V2801F
* C-Data FD511GX-RM0
* OPTON GP801R
* BAUDCOM BD-1234-SFM
* CPGOS03-0490 v2.0
* Ubiquiti U-Fiber Instant
* EXOT EGS1
Signed-off-by: Vladimir Markovets <abam_a@yahoo.com>
Latest binutils (2.37) exposed a long-standing bug. The kernel linking stage
would break at the SORTTAB step, due to the exception table having been
previously purged from vmlinux, as its section wasn't marked as unconditionally
kept. Fix thusly.
Additionally, the "#define ARM_MMU_DISCARD(x) KEEP(x)" change is bogus. It
would only apply to !CONFIG_MMU devices (which we don't support in OpenWrt), and
it would even break the build if referenced. Drop it.
While at it, rename the patch in order to make it obvious that it's
arm-specific.
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
[Add same changes for kernel 5.4 too]
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Nvmem require the device node to be registered with the of_platform.
Register the device node so that nvmem can correctly find the dev and
correctly load the mac-addr stored in the nvmem cell declared in the dts.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
Traversing the device-tree by referencing a network device to determine
a devices labe-mac does not work with the generic nvmem implementation,
as the userspace expects the MAC-address to be available as a
device-tree property.
The legacy mtd-mac-address implementation did create such a node. Do the
same when using the nvmem implementation to allow reading the MAC
address.
Fixes commit d284e6ef0f ("treewide: convert mtd-mac-address-increment*
to generic implementation")
Signed-off-by: David Bauer <mail@david-bauer.net>
Introduce new partition type 0x2e representing uImage.FIT and trigger
FIT partition parser on partitions having that type.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Rework patch 681-NET-add-mtd-mac-address-support to implement
only the function to read the mac-address from mtd.
Generalize mtd-mac-address-increment function so it can be applied
to any source of of_get_mac_address.
Rename any mtd-mac-address-increment to mac-address-increment.
Rename any mtd-mac-address-increment-byte to mac-address-increment-byte.
This should make simplify the conversion of target to nvmem implementation.
Signed-off-by: Ansuel Smith <ansuelsmth@gmail.com>
When compiling with all modules enabled, Kconfig complains about
CONFIG_I2C_DESIGNWARE_SLAVE being unset. Disable this symbol by default.
Fixes commit e9c9b5ec72 ("kernel: package Synopsys Designware PCI to I2C controller")
Signed-off-by: David Bauer <mail@david-bauer.net>
AR8031/AR8033 have different status registers for copper
and fiber operation. However, the extended status register
is the same for both operation modes.
As a result of that, ESTATUS_1000_XFULL is set to 1 even when
operating in copper TP mode.
Remove this mode from the supported link modes, as this driver
currently only supports copper operation.
Signed-off-by: David Bauer <mail@david-bauer.net>
Manually rebased:
bcm27xx/patches-5.4/950-0089-cgroup-Disable-cgroup-memory-by-default.patch
All other patches automatically rebased.
Signed-off-by: John Audia <graysky@archlinux.us>
Without this patch we have to manually bring up the CPU interface in
failsafe mode.
This was backported from kernel 5.12.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Tested-by: Rafał Miłecki <rafal@milecki.pl>
These are the latest patches that just landed upstream for 5.13, will be
backported by Greg into 5.10 (because of stable@), and are now in the
5.4 backport branch of wireguard: https://git.zx2c4.com/wireguard-linux/log/?h=backport-5.4.y
Cc: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Tested-by: Stijn Segers <foss@volatilesystems.org>
The option was added in 5.9 and for some reason, it is causing performance
issues at least on an APU2 board with the igb device.
Switch CONFIG_PCIE_BUS_DEFAULT to fix the performance issues and match the
older kernel's behavior
Signed-off-by: Felix Fietkau <nbd@nbd.name>
This reduces the flash space impact, since built-in code is much smaller
than a bunch of kernel modules on squashfs
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Atheros PHYs using the at803x driver apart from the AR8031/AR8033 fail
to probe with kernel 5.4, due to ret in at803x_probe being
uninitialized.
[ 1.403461] Atheros 8035 ethernet: probe of 4ef600c00.ethernet:01
failed with error -1066114012
Initialize ret in order to successfully prove the PHYs on kernel 5.4.
Kernel 5.10 is not affected, as the ret is always assigned prior to
returning.
Tested on OCEDO Koala.
Reported-by: Russell Senior <russell@personaltelco.net>
Signed-off-by: David Bauer <mail@david-bauer.net>
As patches for the AR8031/AR8033 copper page selection were merged
upstream, we can backport these patches.
This also fixes a PHY capabilities detection issue on the Ubiquiti
ER-X-SFP.
Signed-off-by: David Bauer <mail@david-bauer.net>
ar8xxx_soft_reset is effectively a NOP function. In the PHY state
machine, the availability and result of a soft_reset function
makes no difference for the code flow.
Thus, we can safely remove this method.
Signed-off-by: David Bauer <mail@david-bauer.net>
Modifying PHY capabilities in the probe function broke with upstream
commit 92ed2eb7f4b7 ("net: phy: probe the PHY before determining the
supported features").
AR8316 switches only support 10/100 Mbit/s link modes because of this
change.
Provide a get_features method for the PHY driver, so Gigabit link mode
will be advertised to link partners again.
Signed-off-by: David Bauer <mail@david-bauer.net>
Add missing symbol CONFIG_AHCI_XGENE.
Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
[also add for 5.10, remove from layerscape target config]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Rid of kernel error message:
[ 0.780828] orion-mdio d0072004.mdio: IRQ index 0 not found
on Marvell targets backporting the kernel commit fa2632f74e57
Signed-off-by: Daniel González Cabanelas <dgcbueu@gmail.com>
The higher 16-bit of EEE register was overwritten by mistake, fix that.
Fixes: 5b9ba4a93e ("generic: mt7530: support adjusting EEE")
Signed-off-by: DENG Qingfang <dqfext@gmail.com>
Removed upstreamed:
generic/pending-5.4/770-02-net-ethernet-mtk_eth_soc-fix-rx-vlan-offload.patch
All other patches automatically rebased.
Build system: x86_64
Build-tested: ipq806x/R7800
Note that since I rebased the previous commit, I removed my Run-tested line
although I confirm building the image successfully.
Signed-off-by: John Audia <graysky@archlinux.us>
Removed upstreamed:
generic/backport-5.4/050-gro-fix-napi_gro_frags-Fast-GRO-breakage-due-to-IP-a.patch
bcm63xx/patches-5.4/434-nand-brcmnand-fix-OOB-R-W-with-Hamming-ECC.patch*
Removed/code was included upstream and therefore redundant:
ramips/patches-5.4/999-fix-pci-init-mt7620.patch
All other patches automatically rebased.
* update_kernel.sh did not flag this yet it was included in 5.4.119[1], as a
result of the rebase, I removed my testing lines since I did not go back to
test built or to run test 5.4.119 with the removed patch present.
1. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v5.4.119&id=e5b3e69eb36ac1178a7a2392616fd29afd288c4e
Signed-off-by: John Audia <graysky@archlinux.us>
The code uses get_mtd_device_nm() which must be followed by a call to
put_mtd_device() once the handle is no longer used.
This fixes spurious shutdown console messages such as:
[ 2256.334562] Removing MTD device #7 (soft_config) with use count 1
Reported-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Tested-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
Signed-off-by: Thibaut VARÈNE <hacks@slashdirt.org>
311-MIPS-zboot-put-appended-dtb-into-a-section.patch
commit d2e850e96183 in kernel, part of v5.11
499-mtd-don-t-lock-when-recursively-deleting-partitions.patch
commit cb4543054c5c in kernel, part of v5.13
103-MIPS-select-CPU_MIPS64-for-remaining-MIPS64-CPUs.patch
commit 5a4fa44f5e1b in kernel, part of v5.13
Move them to backports folder to make maintainance easier.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
ath79, lantiq, ipq40xx, ramips all use the OpenWrt-specific gpio-export
functionality. Consolidate the patch that adds it under hack-5.10 since
this logic is obviously not target-specific. For those who want to
disable it, unsetting CONFIG_GPIO_SYSFS symbol will disable this code.
Signed-off-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
This partly reverts commit 20a924d2ae.
This commit broke the build of the Mediatek target with kernel 5.4, for
kernel 5.10 it is fine.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
These have long been obsolete. For reference, here's the Linux version where
each symbol has been dropped:
CONFIG_IP6_NF_QUEUE - 3.5
CONFIG_IP6_NF_TARGET_LOG - 3.4
CONFIG_IP_NF_MATCH_DSCP - 2.6.19
CONFIG_NF_CONNTRACK_IPV4 - 4.19
CONFIG_NF_CONNTRACK_IPV6 - 4.19
CONFIG_NF_CONNTRACK_RTCACHE - out-of-tree, superseded by flow offloading
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
This kmod is similar to macvlan with the difference being that the
endpoints have the same mac address.
It is useful on cloud where only one mac address allowed on port,
where macvlan not works but ipvlan would.
One use case is where multiple IPs and gateways assign on one net port
Signed-off-by: Chen Minqiang <ptpt52@gmail.com>
CONFIG_FORTIFY_SOURCE=y is already set in the generic kernel
configuration, but it is not working for MIPS on kernel 5.4, support for
MIPS was only added with kernel 5.5, other architectures like aarch64
support FORTIFY_SOURCE already since some time.
This patch adds support for FORTIFY_SOURCE to MIPS with kernel 5.4,
kernel 5.10 already supports this and needs no changes.
This backports one patch from kernel 5.5 and one fix from 5.8 to make
fortify source also work on our kernel 5.4.
The changes are not compatible with the
306-mips_mem_functions_performance.patch patch which was also removed
with kernel 5.10, probably because of the same problems. I think it is
not needed anyway as the compiler should automatically optimize the
calls to memset(), memcpy() and memmove() even when not explicitly
telling the compiler to use the build in variant.
This increases the size of an uncompressed kernel by less than 1 KB.
Acked-by: Rosen Penev <rosenp@gmail.com>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The new EEE patch is accepted upstream, so backport it and replace the
current one.
Cc: René van Dorst <opensource@vdorst.com>
Signed-off-by: DENG Qingfang <dqfext@gmail.com>
This CFI patch was accepted upstream for 5.13. Move it away from under
ath79 and place under backports to be removed in due time.
Signed-off-by: Mauri Sandberg <sandberg@mailfence.com>
Manually rebased*
generic/backport-5.4/700-v5.5-net-core-allow-fast-GRO-for-skbs-with-Ethernet-heade.patch
Added new backport*
generic/backport-5.4/050-gro-fix-napi_gro_frags-Fast-GRO-breakage-due-to-IP-a.patch
All others updated automatically.
The new backport was included based on this[1] upstream commit that will be
mainlined soon. This change is needed because Eric Dumazet's check for
NET_IP_ALIGN (landed in 5.4.114) causes huge slowdowns on drivers which use
napi_gro_frags().
Build system: x86_64
Build-tested: ipq806x/R7800
Run-tested: ipq806x/R7800
No dmesg regressions, everything functional
*Credit to Alexander Lobakin
1. https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net.git/commit/?id=7ad18ff6449cbd6beb26b53128ddf56d2685aa93
Signed-off-by: John Audia <graysky@archlinux.us>
It's in backports-5.4, but it wasn't ever merged. Upstream followed another
approach, with flow offloading, which has much better performance. Drop this
obsolete patch and refresh the kernel patches.
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
Acked-by: Stijn Tintel <stijn@linux-ipv6.be>
Acked-by: Felix Fietkau <nbd@nbd.name>
Some targets select HZ=100, others HZ=250. There's no reason to select a higher
timer frequency (and 100 Hz are available in every architecture), so change all
targets to 100 Hz.
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
Make sure all compression algorithms are enabled, for the time being, in order
not to break sysupgrade. In the future, we'd like to disable all except zstd, as
it's the best all-around performer.
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
This makes no sense on anything but the IXP4xx platform
that we do not even support anymore. If we bring it back,
it can be selectively enabled for that platform only.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
This config is for platforms that have:
A) CONFIG_THERMAL set (we disable it by default)
B) Has thermal zones that need to control cpufreq
This is not suitable for default config, it needs to be enabled
on a per-target basis in individual configs.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
We do not enable GPIO_CDEV so do not enable CDEV_V1
either. As we rebuild the whole userspace and kernel
in OpenWRT it doesn't make any sense to support the
old ABI anyway: the new one should be used by default.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
This CMA memory allocation option only applies to NUMA
(Non-Uniform Memory Access) systems which are seldom
the kind of systems that OpenWRT address.
It is safe to assume that any system that need this
option would turn it on locally.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
The non-MQ CFQ, deadline and noop IO schedulers are
deleted from the Linux kernel
since commit f382fb0bcef4c37dc049e9f6963e3baf204d815c
"block: remove legacy IO schedulers".
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Ran update_kernel.sh in a fresh clone without any existing toolchains.
Removed upstreamed:
mvebu/patches-5.4/319-ARM-dts-turris-omnia-configure-LED-2--INTn-pin-as-interrupt-pin.patch
Build system : x86_64
Build-tested : ipq806x/R7800
Run-tested : ipq806x/R7800
No dmesg regressions, everything functional
Signed-off-by: John Audia <graysky@archlinux.us>
This fixes incorrect supported link modes indicated when using ethtool
on the Ubiquiti UniFi AC lite / LR / Mesh.
eroot@OpenWrt:~# ethtool eth0
Settings for eth0:
Supported ports: [ TP MII ]
Supported link modes: 100baseT/Half 100baseT/Full
1000baseT/Full
1000baseX/Full
The reason for this is, that since kernel 5.4 the PHY capabilities are
read dynamically from the PHY. As Ubiquiti leaves the PHY in a wrong
state, the capabilities of the SGMII side of the PHY are read.
As a side effect, 10 Mbit/s link modes did not work on these boards.
Signed-off-by: David Bauer <mail@david-bauer.net>
From Frank Wunderlich <frank-w@public-files.de>:
Issue were traffic problems after a while with increased ping times if
flow offload is active.
It turns out that key_offset with cookie is needed in rhashtable_params
and head_offset was defined twice.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Following changes are made to the Lantiq kernel patches:
0001-MIPS-lantiq-add-pcie-driver.patch
The pci header isn't included by the of_pci header any longer
0024-MIPS-lantiq-revert-DSA-switch-driver-PMU-clock-chang.patch
Due to the merge of grx390 and ar10 clocks, extend support to grx390
0025-NET-MIPS-lantiq-adds-xrx200-legacy.patch
The do_carrier arguments was dropped from phy_link_change. The
phylib has always sets the third parameter to true so the flag is
always changed anyway.
of_get_phy_mode() returns an error, or 0 on success, and pass a
pointer, of type phy_interface_t, where the phy mode should be
stored now. So far an error wasn't considered. Print at least an
error message if something unexpected happens.
The stuck queue is now passed to xrx200_tx_timeout (the timeout
handler) but not used so far.
0028-NET-lantiq-various-etop-fixes.patch
ioremap has provided non-cached semantics by default since the Linux
2.6 days and was removed with kernel version 5.6.
of_get_phy_mode() returns an error, or 0 on success, and pass a
pointer, of type phy_interface_t, where the phy mode should be
stored now. So far an error wasn't considered. Print at least an
error message if something unexpected happens.
0042-arch-mips-increase-io_space_limit.patch
Move IO space extension to laniq specific file
Signed-off-by: Mathias Kresin <dev@kresin.me>
Check if gendisk is already up instead of trying to figure out parent
mtd device (which didn't work well).
Reported-by: DENG Qingfang <dqfext@gmail.com>
Reported-by: Koen Vandeputte <koen.vandeputte@citymesh.com>
Fixes: 2809d00007 ("kernel: support FIT partition parser on mtdblock devices")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Ran update_kernel.sh in a fresh clone without any existing toolchains.
Manually rebased due to movement of rx-offload.c in 5.4.110:
layerscape/patches-5.4/802-can-0002-can-rx-offload-fix-long-lines.patch
layerscape/patches-5.4/802-can-0003-can-rx-offload-can_rx_offload_compare-fix-typo.patch
layerscape/patches-5.4/802-can-0004-can-rx-offload-can_rx_offload_irq_offload_timestamp-.patch
layerscape/patches-5.4/802-can-0005-can-rx-offload-can_rx_offload_reset-remove-no-op-fun.patch
layerscape/patches-5.4/802-can-0006-can-rx-offload-Prepare-for-CAN-FD-support.patch
layerscape/patches-5.4/802-can-0018-can-flexcan-use-struct-canfd_frame-for-CAN-classic-f.patch
Build system: x86_64
Build-tested: ipq806x/R7800
Run-tested: ipq806x/R7800
No dmesg regressions, everything functional
Signed-off-by: John Audia <graysky@archlinux.us>
Move 480-mtd-set-rootfs-to-be-root-dev.patch to hack-5.10 and fix assumption
about the block device index.
Fixes: 2809d00007 ("kernel: support FIT partition parser on mtdblock devices")
Acked-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Allow using the FIT block-device partition partition parser to work on
top of mtdblock devices.
This is more tricky than it sounds as it requires to reorganize the
procedure of registering mtdblock devices in order to avoid locking
troubles caused by the block partition parsers then trying to open
the mtdblock device for reading while locks are still being held.
Fix that by moving the adding of the disks after the entire device
has been probed when locks no longer need to be held.
Also address issue with mtdsuper surfaced when using sub-partitions
which prevented mounting JFFS2 using the /dev/mtdblock* device.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
U-Boot has recently added support for having data blobs in uImage.FIT
images stored at offsets after the FDT structure rather than embedding
the data into the FDT structure itself. This is useful as it allows
parts of the image to be mapped by the FIT partition parser, and it
allows the FIT structure itself to be parsed more easily as it usually
fits into single page.
mtdsplit_fit assumed that the total length of an image is identical
to the length of the FDT structure. For uImage.FIT with external data
this assumption no longer holds true.
Add support for uImage.FIT with external data to mtdsplit_fit and in
in that case only split-off rootfs_data -- selecting and mapping rootfs
partition is left to the block partition parser just like on UBI and
block/GPT based platforms.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Use update_kernel to refresh all patches, required manual updates to:
610-netfilter_match_bypass_default_checks.patch
611-netfilter_match_bypass_default_table.patch
762-net-bridge-switchdev-Refactor-br_switchdev_fdb_notif.patch
764-net-bridge-switchdev-Send-FDB-notifications-for-host.patch
Run-tested: x86_64
Nothing screamed out but any funny business with linux bridging should
suspect this update first.
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
Ran update_kernel.sh in a fresh clone without any existing toolchains.
Manually rebased:
pending-5.4/611-netfilter_match_bypass_default_table.patch
The upstream change affecting this patch is the revert of an earlier
kernel commit. Therefore, we just revert our corresponding changes
in [1].
Build system: x86_64
Build-tested: ipq806x/R7800
[1] 9b1b89229f ("kernel: bump 5.4 to 5.4.86")
Signed-off-by: John Audia <graysky@archlinux.us>
[adjust manually rebased patch, add explanation]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The uImage.FIT partition parser used to squeeze in FIT partitions in
the range where partition editor tools (fdisk and such) expect the
regular partition. This is confusing people and tools when adding
additional partitions on top of the partition used for OpenWrt's
uImage.FIT.
Instead of squeezing in the additional partitions, rather start with
all uImage.FIT partitions at offset 64.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
1. Use upstream accepted NVMEM patches
2. Minor fix for BCM4908 partitioning
3. Support for Linksys firmware partitions on Northstar
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
The CONFIG_USERIO option is unset in multiple target configurations. On
the sunxi target it is activated. Move the kernel configuration option
to the generic kernel configuration.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Instead of deactivating this in every target config, deactivate it once
in the generic kernel config. I was asked for this config option in a
x86 64 build in OpenWrt 21.02.
Fixes: 87046e87e2 ("kernel: add missing kernel config symbol")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Bump to 5.10.26 reversed dependencies on IOMMU for CONFIG_VFIO thus
malta (at least) prompted for this new symbol.
Set it to 'false' in the default config. rockchip & X86 enable it in
target specific configs.
Thanks to Tony Ambardar for falling over this one
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
It's helpful for accessing booting data (DTS, kernel, etc.). It has to
be used carefully as CFE's JFFS2 support is quite dumb. It doesn't
recognize deleted files and has problems handling 0 inode.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Straightforward refresh of patches using update_kernel.
Removed (reverse-applicable):
bmips/patches-5.10/010-v5.11-net-dsa-implement-a-central-TX-reallocation-procedur.patch
Run tested: x86_64 (apu2)
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
Straightforward refresh of patches using update_kernel.
Run tested: x86_64 (apu2)
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
Ran update_kernel.sh in a fresh clone without any existing toolchains.
Manually rebased:
bcm27xx/950-0993-xhci-quirks-add-link-TRB-quirk-for-VL805.patch
layerscape/701-net-0231-enetc-Use-DT-protocol-information-to-set-up-the-port.patch
Build system: x86_64
Build-tested: ipq806x/R7800
Run-tested: ipq806x/R7800
No dmesg regressions, everything functional
Signed-off-by: John Audia <graysky@archlinux.us>
[remove accidental whitespace edit]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
With kernel 5.10, exfat is out of staging and in tree.
Added small hack to make it work with kernel 5.4 as well.
Added removed config options for 5.4 to generic config.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
[Set CONFIG_EXFAT_ config options to default values]
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Update 5.10 generic config to add missing options prompting during builds,
when setting CONFIG_ALL_KMODS. Refresh the x86 subtargets to drop symbols
added to the generic config.
Also fix a typo in 5.4/5.10 generic configs: "CONFIG_CONFIG_KVM_AMD_SEV".
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
A SPI NOR flash should not be target-specific, so move it to generic in
case another platform needs to use it.
Signed-off-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
Backport upstream patch that fixes TRGMII mode now that mt7530 is
actually resetting the switch on ramips devices.
Patches apply to both Linux 5.4 and 5.10, since TRGMII is broken on both.
Fixes: 69551a2442 ("ramips: manage low reset lines")
Signed-off-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
Add kmod-mdio-devres package and kernel config symbol, and update related
kmod-r8169 and kmod-ixgbe package dependencies.
Build tested module dependencies on x86/64 with CONFIG_ALL_KMODS=y.
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
It's meant to provide upstream support for mtd & NVMEM. It's required
e.g. for reading MAC address from mtd partition content. It seems to be
in a final shape so it's worth testing.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
It appears to be an automatic Kconfig symbol that varies depending on
the host platform. There is no need to define it in target configs, so
filter it out.
Also sort config-filter entries alphabetically.
Cc: Adrian Schmutzler <mail@adrianschmutzler.de>
Cc: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
Also add a new kconfig symbol (CONFIG_KCMP) to the generic config,
disabling the SYS_kcmp syscall (it was split from
CONFIG_CHECKPOINT_RESTORE, which is disabled by default, so the
previous behaviour is kept).
Removed (upstreamed) patches:
070-net-icmp-pass-zeroed-opts-from-icmp-v6-_ndo_send-bef.patch
081-wireguard-device-do-not-generate-ICMP-for-non-IP-pac.patch
082-wireguard-queueing-get-rid-of-per-peer-ring-buffers.patch
083-wireguard-kconfig-use-arm-chacha-even-with-no-neon.patch
830-v5.12-0002-usb-serial-option-update-interface-mapping-for-ZTE-P685M.patch
Manually rebased patches:
313-helios4-dts-status-led-alias.patch
104-powerpc-mpc85xx-change-P2020RDB-dts-file-for-OpenWRT.patch
Run tested:
ath79 (TL-WDR3600)
mvebu (Turris Omnia)
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
From the original commit message:
"With GCC 10, building usbip triggers error for multiple definition
of 'udev_context', in:
- libsrc/vhci_driver.c:18 and
- libsrc/usbip_host_common.c:27.
Declare as extern the definition in libsrc/usbip_host_common.c."
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
REFCOUNT_FULL was removed for linux 5.5:
commit fb041bb7c0a9 (locking/refcount: Consolidate implementations of refcount_t)
COMMON_CLK_VERSATILE was removed on linux 5.8:
commit 5f55f1fb187d (clk: versatile: Fix kconfig dependency on COMMON_CLK_VERSATILE)
Signed-off-by: Luis Araneda <luaraneda@gmail.com>
The BCM5365 UID was updated in the driver, but we should also update it in the
fixup.
Fixes: cbcac4fde8 ("kernel: b53: update the BCM5365 UID")
Signed-off-by: Daniel González Cabanelas <dgcbueu@gmail.com>
(Ammend commit description, add Fixes tag)
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
5.4.102 backported a lot of stuff that our WireGuard backport already
did, in addition to other patches we had, so those patches were
removed from that part of the series. In the process other patches were
refreshed or reworked to account for upstream changes.
This commit involved `update_kernel.sh -v -u 5.4`.
Cc: John Audia <graysky@archlinux.us>
Cc: David Bauer <mail@david-bauer.net>
Cc: Petr Štetiar <ynezz@true.cz>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Compile testing i.mx6 with ALL_KMODS=y, PACKAGE_perf=y and bunch of
tracing/probing symbols has unveiled bunch of missing config options so
add them.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
These make a big difference when doing WireGuard with small armv7
routers, and the 5.4 backport already has it.
Suggested-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
Cc: David Bauer <mail@david-bauer.net>
Cc: Petr Štetiar <ynezz@true.cz>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Without this patch, the chacha block counter is not incremented on neon
rounds, resulting in incorrect calculations and corrupt packets.
This also switches to using `--no-numbered --zero-commit` so that future
diffs are smaller.
Reported-by: Hans Geiblinger <cybrnook2002@yahoo.com>
Reviewed-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
Cc: David Bauer <mail@david-bauer.net>
Cc: Petr Štetiar <ynezz@true.cz>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
This adds the latest version of ofpart commit. It hopefully
1. Doesn't break compilation
2. Doesn't break partitioning
(this time).
It's required to implement fixed partitioning with some quirks. It's
required by bcm53xx, bcm4908, kirkwood, lantiq and mvebu.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
This is useful for dual-boot setups where the loader sets variables depending
on the flash boot partition.
For example the Linksys E8450 sets mtdparts=master for the first partition
and mtdparts=slave for the second one.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Introduce a magic GUID_PARTITION_LINUX_FIT_GUID to designate a GPT
partition to be interpreted by the FIT partition parser.
In that way, sub-partitions for (external-data) uImage.FIT stored
directly in a partition can be split, similar like we do for devices
with raw flash storage.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
The CPU_MIPS64 and CPU_MIPS32 variables are supposed to be able to
distinguish broadly between 64-bit and 32-bit MIPS CPUs. However, they
weren't selected by the specialty CPUs, Octeon and Loongson, which meant
it was possible to hit a weird state of:
MIPS=y, CONFIG_64BIT=y, CPU_MIPS64=n
This commit rectifies the issue by having CPU_MIPS64 be selected when
the missing Octeon or Loongson models are selected.
In particular, this affects our octeonplus target.
It has been posted to LKML here:
https://lore.kernel.org/linux-mips/20210227122605.2680138-1-Jason@zx2c4.com/
Cc: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
Cc: David Bauer <mail@david-bauer.net>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
BCM63XX internal PHYs and BCM5365 SoC internal switch are both using the
same phy_driver->phy_id, causing conflicts and unnecessary probes. E.g
the BCM63XX phy internal IRQ is lost on the first probe.
The full BCM5365 UID is 0x00406370.
Use an additional byte to mask the BCM5365 UID to avoid duplicate driver
phy_id's. This will fix the IRQ issue in internal BCM63XX PHYs and avoid
more conflicts in the future.
Signed-off-by: Daniel González Cabanelas <dgcbueu@gmail.com>
Rather than using the clunky, old, slower wireguard-linux-compat out of
tree module, this commit does a patch-by-patch backport of upstream's
wireguard to 5.4. This specific backport is in widespread use, being
part of SUSE's enterprise kernel, Oracle's enterprise kernel, Google's
Android kernel, Gentoo's distro kernel, and probably more I've forgotten
about. It's definately the "more proper" way of adding wireguard to a
kernel than the ugly compat.h hell of the wireguard-linux-compat repo.
And most importantly for OpenWRT, it allows using the same module
configuration code for 5.10 as for 5.4, with no need for bifurcation.
These patches are from the backport tree which is maintained in the
open here: https://git.zx2c4.com/wireguard-linux/log/?h=backport-5.4.y
I'll be sending PRs to update this as needed.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
ZTE MF283+ is a dual-antenna LTE category 4 router, based on Ralink
RT3352 SoC, and built-in ZTE P685M PCIe MiniCard LTE modem.
Hardware highlighs:
- CPU: MIPS24KEc at 400MHz,
- RAM: 64MB DDR2,
- Flash: 16MB SPI,
- Ethernet: 4 10/100M port switch with VLAN support,
- Wireless: Dual-stream 802.11n (RT2860), with two internal antennas,
- WWAN: Built-in ZTE P685M modem, with two internal antennas and two
switching SMA connectors for external antennas,
- FXS: Single ATA, with two connectors marked PHONE1 and PHONE2,
internally wired in parallel by 0-Ohm resistors, handled entirely by
internal WWAN modem.
- USB: internal miniPCIe slot for modem,
unpopulated USB A connector on PCB.
- SIM slot for the WWAN modem.
- UART connector for the console (unpopulated) at 3.3V,
pinout: 1: VCC, 2: TXD, 3: RXD, 4: GND,
settings: 57600-8-N-1.
- LEDs: Power (fixed), WLAN, WWAN (RGB),
phone (bicolor, controlled by modem), Signal,
4 link/act LEDs for LAN1-4.
- Buttons: WPS, reset.
Installation:
As the modem is, for most of the time, provided by carriers, there is no
possibility to flash through web interface, only built-in FOTA update
and TFTP recovery are supported.
There are two installation methods:
(1) Using serial console and initramfs-kernel - recommended, as it
allows you to back up original firmware, or
(2) Using TFTP recovery - does not require disassembly.
(1) Using serial console:
To install OpenWrt, one needs to disassemble the
router and flash it via TFTP by using serial console:
- Locate unpopulated 4-pin header on the top of the board, near buttons.
- Connect UART adapter to the connector. Use 3.3V voltage level only,
omit VCC connection. Pin 1 (VCC) is marked by square pad.
- Put your initramfs-kernel image in TFTP server directory.
- Power-up the device.
- Press "1" to load initramfs image to RAM.
- Enter IP address chosen for the device (defaults to 192.168.0.1).
- Enter TFTP server IP address (defaults to 192.168.0.22).
- Enter image filename as put inside TFTP server - something short,
like firmware.bin is recommended.
- Hit enter to load the image. U-boot will store above values in
persistent environment for next installation.
- If you ever might want to return to vendor firmware,
BACK UP CONTENTS OF YOUR FLASH NOW.
For this router, commonly used by mobile networks,
plain vendor images are not officially available.
To do so, copy contents of each /dev/mtd[0-3], "firmware" - mtd3 being the
most important, and copy them over network to your PC. But in case
anything goes wrong, PLEASE do back up ALL OF THEM.
- From under OpenWrt just booted, load the sysupgrade image to tmpfs,
and execute sysupgrade.
(2) Using TFTP recovery
- Set your host IP to 192.168.0.22 - for example using:
sudo ip addr add 192.168.0.22/24 dev <interface>
- Set up a TFTP server on your machine
- Put the sysupgrade image in TFTP server root named as 'root_uImage'
(no quotes), for example using tftpd:
cp openwrt-ramips-rt305x-zte_mf283plus-squashfs-sysupgrade.bin /srv/tftp/root_uImage
- Power on the router holding BOTH Reset and WPS buttons held for around
5 seconds, until after WWAN and Signal LEDs blink.
- Wait for OpenWrt to start booting up, this should take around a
minute.
Return to original firmware:
Here, again there are two possibilities are possible, just like for
installation:
(1) Using initramfs-kernel image and serial console
(2) Using TFTP recovery
(1) Using initramfs-kernel image and serial console
- Boot OpenWrt initramfs-kernel image via TFTP the same as for
installation.
- Copy over the backed up "firmware.bin" image of "mtd3" to /tmp/
- Use "mtd write /tmp/firmware.bin /dev/mtd3", where firmware.bin is
your backup taken before OpenWrt installation, and /dev/mtd3 is the
"firmware" partition.
(2) Using TFTP recovery
- Follow the same steps as for installation, but replacing 'root_uImage'
with firmware backup you took during installation, or by vendor
firmware obtained elsewhere.
A few quirks of the device, noted from my instance:
- Wired and wireless MAC addresses written in flash are the same,
despite being in separate locations.
- Power LED is hardwired to 3.3V, so there is no status LED per se, and
WLAN LED is controlled by WLAN driver, so I had to hijack 3G/4G LED
for status - original firmware also does this in bootup.
- FXS subsystem and its LED is controlled by the
modem, so it work independently of OpenWrt.
Tested to work even before OpenWrt booted.
I managed to open up modem's shell via ADB,
and found from its kernel logs, that FXS and its LED is indeed controlled
by modem.
- While finding LEDs, I had no GPL source drop from ZTE, so I had to probe for
each and every one of them manually, so this might not be complete -
it looks like bicolor LED is used for FXS, possibly to support
dual-ported variant in other device sharing the PCB.
- Flash performance is very low, despite enabling 50MHz clock and fast
read command, due to using 4k sectors throughout the target. I decided
to keep it at the moment, to avoid breaking existing devices - I
identified one potentially affected, should this be limited to under
4MB of Flash. The difference between sysupgrade durations is whopping
3min vs 8min, so this is worth pursuing.
In vendor firmware, WWAN LED behaviour is as follows, citing the manual:
- red - no registration,
- green - 3G,
- blue - 4G.
Blinking indicates activity, so netdev trigger mapped from wwan0 to blue:wwan
looks reasonable at the moment, for full replacement, a script similar to
"rssileds" would need to be developed.
Behaviour of "Signal LED" in vendor firmware is as follows:
- Off - no signal,
- Blinking - poor coverage
- Solid - good coverage.
A few more details on the built-in LTE modem:
Modem is not fully supported upstream in Linux - only two CDC ports
(DIAG and one for QMI) probe. I sent patches upstream to add required device
IDs for full support.
The mapping of USB functions is as follows:
- CDC (QCDM) - dedicated to comunicating with proprietary Qualcomm tools.
- CDC (PCUI) - not supported by upstream 'option' driver yet. Patch
submitted upstream.
- CDC (Modem) - Exactly the same as above
- QMI - A patch is sent upstream to add device ID, with that in place,
uqmi did connect successfully, once I selected correct PDP context
type for my SIM (IPv4-only, not default IPv4v6).
- ADB - self-explanatory, one can access the ADB shell with a device ID
added to 51-android.rules like so:
SUBSYSTEM!="usb", GOTO="android_usb_rules_end"
LABEL="android_usb_rules_begin"
SUBSYSTEM=="usb", ATTR{idVendor}=="19d2", ATTR{idProduct}=="1275", ENV{adb_user}="yes"
ENV{adb_user}=="yes", MODE="0660", GROUP="plugdev", TAG+="uaccess"
LABEL="android_usb_rules_end"
While not really needed in OpenWrt, it might come useful if one decides to
move the modem to their PC to hack it further, insides seem to be pretty
interesting. ADB also works well from within OpenWrt without that. O
course it isn't needed for normal operation, so I left it out of
DEVICE_PACKAGES.
Signed-off-by: Lech Perczak <lech.perczak@gmail.com>
[remove kmod-usb-ledtrig-usbport, take merged upstream patches]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Ran update_kernel.sh in a fresh clone without any existing toolchains.
Build system: x86_64
Build-tested: ipq806x/R7800
Run-tested: ipq806x/R7800
No dmesg regressions, everything functional
Signed-off-by: John Audia <graysky@archlinux.us>
[refresh again]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
CONFIG_NET_DSA_MV88E6XXX_PTP was simply missing in generic config.
CONFIG_I2C_PXA_SLAVE was previously enabled via i2c-pxa package,
but got removed there without moving the symbol to generic config.
Fixes: dd13add3ce ("kernel: i2c-pxa: remove slave")
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Allow for single (external-data) FIT image to hold kernel, dtb and
squashfs. In that way, the bootloader verifies the system integrity
including the rootfs, because what's the point of checking that the
hash of the kernel is correct if it won't boot in case of squashfs
being corrupted? Better allow bootloader to check everything needed
to make it at least up to failsafe mode. As a positive side effect
this change also makes the sysupgrade process on nand potentially
much easier as it is now.
In short: mkimage has a parameter '-E' which allows generating FIT
images with 'external' data rather than embedding the data into the
device-tree blob itself. In this way, the FIT structure itself remains
small and can be parsed easily (rather than having to page around
megabytes of image content). This patch makes use of that and adds
support for adding sub-images of type 'filesystem' which are used to
store the squashfs. Now U-Boot can verify the whole OS and the new
partition parsers added in the Linux kernel can detect the filesystem
sub-images, create partitions for them, and select the active rootfs
volume based on the configuration in FIT (passing configuration via
device tree could be implemented easily at a later stage).
This new FIT partition parser works for NOR flash (on top of mtdblock),
NAND flash (on top of ubiblock) as well as classic block devices
(ie. eMMC, SDcard, SATA, NVME, ...).
It could even be used to mount such FIT images via `losetup -P` on a
user PC if this patch gets included in Linux upstream one day ;)
Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
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>
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>
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>
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>
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>
This patch allows devices without a high resolution timer to boot up faster.
It should speed up boots for bcm2708 and bcm63xx.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Ran update_kernel.sh in a fresh clone without any existing toolchains.
No manual changes needed.
Build system: x86_64
Build-tested: bcm27xx/bcm2711
Signed-off-by: John Audia <graysky@archlinux.us>
When recursively deleting partitions, don't acquire the masters
partition lock twice. Otherwise the process endy up in a deadlocked
state.
Signed-off-by: David Bauer <mail@david-bauer.net>
This patch has been added to 5.4, but not been copied to 5.10:
7495acb555 ("kernel: backport mtd commit converting partitions doc syntax")
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This applies another patch from 5.4 to 5.10 as well:
de09355f74 ("kernel/hack-5.4: make UDP tunneling user-selectable")
UDP tunneling support isn't user-selectable, but it's required by WireGuard
which is, for the time being, an out-of-tree module. We currently work around
this issue by selecting an unrelated module which depends on UDP tunnelling
(VXLAN). This is inconvenient, as it implies this unrelated module needs to be
built-in when doing a monolithic build.
Fix this inconvenience by making UDP tunneling user-selectable in the kernel
configuration.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
