The HG MXPD-483II 1310nm SFP module is meant to operate with 2500Base-X,
however, in their EEPROM they incorrectly specify:
Transceiver type : Ethernet: 1000BASE-LX
...
BR, Nominal : 2600MBd
Use sfp_quirk_2500basex for this module to allow 2500Base-X mode anyway.
https://forum.banana-pi.org/t/bpi-r3-sfp-module-compatibility/14573/60
X-Patchwork-Id: 13197378
X-Mailing-List: netdev@vger.kernel.org
X-Patchwork-Delegate: kuba@kernel.org
Reported-by: chowtom <chowtom@gmail.com>
Tested-by: chowtom <chowtom@gmail.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
The XMC XM25QH64C is a 8MB SPI NOR chip. The patch is verified on TL-WPA8631P v3.
Datasheet available at https://www.xmcwh.com/uploads/442/XM25QH64C.pdf
Signed-off-by: Joe Mullally <jwmullally@gmail.com>
Pahole version is being autodetected during runtime since kernel 5.15.96
via in-kernel scripts/pahole-version.sh so add CONFIG_PAHOLE_VERSION to
kernel filter in order to prevent it from being added to target configs.
Signed-off-by: Robert Marko <robimarko@gmail.com>
Use scratch buffer for DMA operetations. Passing a pointer to a stack
variable won't work and results in bogus bit flips being reported.
Patch was submitted upstream and is part of Linux 6.3.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
The MT7531 switch IC comes with SerDes ports with PCS identical to
what is also used in MediaTek's SoCs. Make use of the shared driver
to ease maintainance and reduce code duplication.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Using 2500Base-T SFP modules e.g. on the BananaPi R3 requires manually
disabling auto-negotiation, e.g. using ethtool. While a proper fix
using SFP quirks is being discussed upstream, bring a work-around to
restore user experience to what it was before the switch to the
dedicated SGMII PCS driver.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Backport patch allowing to set the MDIO bus clock frequency.
By default the MDIO bus clock runs on 2.5 MHz, allow increasing it
up to 25 MHz.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Replace patches for MediaTek Ethernet driver SGMII/SerDes unit with
their corresponding upstream patches. Not all of the patches in our
tree went upstream as-is, some are slightly different implementations,
and they require the phylink_pcs helpers now made available.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
It isn't feasible to literally backport all upstream phylink_pcs changes
down to Linux 5.15: It's just too many patches, and many downstream
drivers and hacks are likely to break. We are too close to branching off
to risk this, and it's also just too much work.
Instead just add helper functions used by modern PCS drivers while keeping
the original functions instact as well. While this may add a kilobyte or
two of extra kernel size, it has the advantage that we get the best of both
worlds: None of the existing codepaths are touched, but yet we have the
option to backport singular improvements to Ethernet drivers where needed.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
The current patches are old, update them from mainline.
Backports taken from https://github.com/yuzhaogoogle/linux/commits/mglru-5.15
Tested-by: Kazuki H <kazukih0205@gmail.com> #mt7622/Linksys E8450 UBI
Signed-off-by: Kazuki H <kazukih0205@gmail.com>
Add a patch to allow modification of the PHY LED configuration. This is
required for boards, where the reset configuration of LED functions is
incompatibe with the usage of the device LEDs.
This is the case for the ASUS TUF-AX4200 Wireless router. It requires
modification of the LED configuration because as the WAN LED on the
front of the device is driven by the PHY. Without patching, it would
only illuminate in case the Link speed is 100 Mbit/s.
Signed-off-by: David Bauer <mail@david-bauer.net>
This flash-chip is used on the Asus TUF-AX4200 and TUF-AX6000 routers.
As the filogic target only uses kernel 5.15, skip the 5.10 backport.
Signed-off-by: David Bauer <mail@david-bauer.net>
This can improve load balancing by pushing backlog (and RPS) processing
to separate threads, allowing the scheduler to distribute the load.
It can be enabled with: echo 1 > /proc/sys/net/core/backlog_threaded
Signed-off-by: Felix Fietkau <nbd@nbd.name>
These patches have now received a positive review upstream, so let's add them
to pending patches.
776-net-dsa-b53-mmap-add-phy-ops.patch:
This is mostly bmips/bcm63xx-specific to get external switches working
without hanging the device when accessing certain registers.
777-net-dsa-b53-mdio-add-support-for-BCM53134:
This adds support for BCM53134 switch on DSA B53, so any target using DSA B53
can benefit from it.
Also fix sercomm-h500-s external switch IMP port phy-mode.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Refreshing the patches for fff07085fb moved the b53_adjust_63xx_rgmii() call
from b53_phylink_mac_link_up() to b53_phylink_mac_link_down().
In order to properly configure the RGMII ports we need to restore it to its
correct place.
Fixes: fff07085fb ("kernel: add pending bmips patches")
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
bcm63268-timer-clocks have been sent upstream with a positive review, so let's
add them to pending v5.15.
Also add devm_clk_hw_register_gate() patch from v5.17 to backports since it's
needed for upstream bcm63268-timer-clocks patches.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Manually rebased:
ramips/patches-5.10/810-uvc-add-iPassion-iP2970-support.patch
All other patches automatically rebased.
Signed-off-by: John Audia <therealgraysky@proton.me>
Move it to pending, since it wasn't actually accepted upstream yet.
Fixes potential issues when doing offload between multiple MACs.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Make it possible to change the kernel configuration option
CONFIG_HARDLOCKUP_DETECTOR from OpenWrt.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
This sets the CONFIG_KCOV_IRQ_AREA_SIZE kernel configuration option to its default value.
This is shown when I set CONFIG_KERNEL_KCOV=y in the OpenWrt configuration on x86/64.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
This deactivates some kernel configuration options I see when
CONFIG_KERNEL_KASAN=y is set in the OpenWrt configuration on x86/64.
Set CONFIG_STACK_HASH_ORDER to its default value.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
This sets some kernel configuration options to their default values. I saw
these as warnings when I set CONFIG_KERNEL_UBSAN=y is set in the OpenWrt
configuration on x86/64.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
This deactivates some kernel configuration options I see when
CONFIG_KERNEL_DYNAMIC_FTRACE=y is set in the OpenWrt configuration on x86/64.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
This deactivates some kernel configuration options I see when
CONFIG_KERNEL_HIST_TRIGGERS=y is set in the OpenWrt configuration on x86/64.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
This deactivates some kernel configuration options I see when
CONFIG_KERNEL_DEBUG_VIRTUAL=y is set in the OpenWrt configuration on x86/64.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
This deactivates some kernel configuratoion options I see when
CONFIG_KERNEL_DEBUG_VM=y is set in the OpenWrt configuration on x86/64.
Signed-off-by: Hauke Mehrtens <hmehrtens@maxlinear.com>
Backport patches from kernel 6.0 which are fixing building of perf with
binutils 2.40.
perf with kernel 5.10 is also not building but the backporting is more
complicated and only a few targets are still using kernel 5.10.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This fixes spurious boot-errors with some ath79 MIPS 74Kc boards such
as the AC Lite as well as Archer C7 v2.
The missing barrier leads to the icache flush being executed before the
dcache writeback, which results in the CPU executing the dummy infinite
loop in tlbmiss_handler_setup_pgd.
Applying this patch from upstream ensures the dcache is written back
before flushing the icache.
Signed-off-by: David Bauer <mail@david-bauer.net>
This backports the third version [1], which is awaiting upstream merge. It
adds support for watchdog max6370, which is connected via GPIO. It
is useful primarily for P2020 RDB and Turris 1.x routers, which are
not yet supported.
[1] https://www.spinics.net/lists/linux-watchdog/msg23299.html
Signed-off-by: Josef Schlehofer <pepe.schlehofer@gmail.com>
The padding intended to avoid corrupted non-zero padding payload was
accidentally adding too many padding bytes, tripping up some setups.
Fix this by using eth_skb_pad instead.
Fixes#11942.
Signed-off-by: Felix Fietkau <nbd@nbd.name>
The CONFIG_PPC_QUEUED_SPINLOCKS configuration option is not defined for
kernel 5.15, it is defined for kernel 5.10.
This fixes the compilation of mpc85xx/p2020 with kernel 5.15.
Reviewed-by: Robert Marko <robimarko@gmail.com>
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
They were backported to stable kernels but we backport more stuff on our
own so we have to pick up few remaining.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
This commit fixes the following commit
f584fb2f7e kernel: import accepted MediaTek Ethernet patches
Unrefreshed patches caused the CI to fail.
Signed-off-by: Paul Spooren <mail@aparcar.org>
Import some accepted and pending upstream patches for mtk_eth_soc,
replacing some semantically equivalent local patches and fixing issues
when operating the PCS in 1G SGMII mode.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Kconfig docs say:
> The default value deliberately defaults to 'n' in order to avoid
> bloating the build.
Apply this rule everywhere, to avoid more cloning of bad examples
Signed-off-by: Tony Butler <spudz76@gmail.com>
"842" is a compression scheme and this is the software implementation
which is too slow to really use beyond a proof of concept. It can be
selected in ZRAM, ZSWAP, or `fs/pstore`, and is here for completeness.
In general you need a Power8 or better with 842-in-hardware for it to
be fast, but other 842-accelerators are emerging.
Signed-off-by: Tony Butler <spudz76@gmail.com>
Recent backport of NVMEM layout support as well as acommpanying OF changes
introduced a false #nvmem-cell-cells warning as #nvmem-cell-cells are
fully optional.
So, backport an upstream fix for this.
Fixes: 11759a5bf3 ("kernel: backport of changes & helpers")
Signed-off-by: Robert Marko <robimarko@gmail.com>
_oob_read returns number of bitflips on success while
bbt_nand_read should return 0.
Fixes: 2d49e49b18 ("mediatek: bmt: use generic mtd api")
Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
The patch that adds support for hw flow-offloading counters on newer
MediaTek SoCs tries to prints acct->packets and acct->bytes in debugfs,
without checking that acct isn't null. This causes a kernel panic when
trying to read /sys/kernel/debug/ppe0/entries on older MediaTek SoCs.
Fix this by adding a check for acct.
Fixes: openwrt#11756
Fixes: 9721a42a27 ("kernel: support hw flow-offloading counters on newer MediaTek SoCs")
Reported-by: Stijn Tintel <stijn@linux-ipv6.be>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
This adds the MHI SBL callback that ath11k will utilize in order to
support multiple PCI cards or AHB+PCI combo which currently does not
work due to QRTR ID-s conflicting.
This is a prerequisite for the mac80211 patch targeting ath11k as it
uses MHI from kernel.
Signed-off-by: Robert Marko <robimarko@gmail.com>
1. Check for -EPROBE_DEFER
If it occurs we have to return immediately. Trying other properties
could result in another error and ignoring -EPROBE_DEFER which has a
special meaning.
2. Check for read result
Assuming property->read() success can result in NULL pointer
dereference. It happens e.g. for "mac-address" with NVMEM cell
containing invalid MAC.
3. Simplify code
Don't move cell reading & nvmem_cell_put() into a loop. Simplify loop
code.
Fixes: ecd81de7a5 ("ath79: add nvmem cell mac-address-ascii support")
Cc: Yousong Zhou <yszhou4tech@gmail.com>
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
The EN25QH256A variant of the EN25QH256 doesn't initialize correctly from SFDP
alone and only accesses memory below 8m (addr_width is 4 but read_opcode takes
only 3 bytes).
Set SNOR_F_4B_OPCODES if the flash chip variant was detected using hwcaps.
The fix submitted upstream uses the PARSE_SFDP initializer that is not
available in the kernel used with Openwrt.
Signed-off-by: Leon M. Busch-George <leon@georgemail.eu>
Fix Silicon Labs bindings in the spidev driver
Some bindings for Silicon Labs chips already exists upstream.
These bindings can be found in trivial-devices.yaml.
The existing bindings are using "silabs" instead of "siliconlabs" to
identify the manufacturer.
This commit add two submitted patches for silabs chips and rename the
manufacturer in the different DTS for more coherence.
Signed-off-by: Vincent Tremblay <vincent@vtremblay.dev>
Backport qca8k fixup for mgmt and mdio read/write for kernel 5.15 fixup
port dropping and configuration issue.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Robert Marko <robimarko@gmail.com>
Backport patch from kernel 5.14.
Treat only the highest, not the lowest, IPv4 address within a local
subnet as a broadcast address, as subnets do not need two different
broadcast addresses and networking documentation consistently prefers
the highest address as broadcast.
This patch was merged in upstream net-next tree in May 2021 at
https://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next.git/commit/?id=94c821c74bf5
This eventually frees up one address per subnet. It matches behavior
suggested in our Internet-Draft, and also the default behavior of OpenBSD
and FreeBSD.
Signed-off-by: Seth David Schoen <schoen@loyalty.org>
Removed upstreamed:
pending-5.15/101-Use-stddefs.h-instead-of-compiler.h.patch[1]
ipq806x/patches-5.15/122-01-clk-qcom-clk-krait-fix-wrong-div2-functions.patch[2]
bcm27xx/patches-5.15/950-0198-drm-fourcc-Add-packed-10bit-YUV-4-2-0-format.patch[3]
Manually rebased:
ramips/patches-5.15/100-PCI-mt7621-Add-MediaTek-MT7621-PCIe-host-controller-.patch[4]
Added patch/backported:
ramips/patches-5.15/107-PCI-mt7621-Add-sentinel-to-quirks-table.patch[5]
All other patches automatically rebased.
1. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v5.15.86&id=c160505c9b574b346031fdf2c649d19e7939ca11
2. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v5.15.86&id=a051e10bfc6906d29dae7a31f0773f2702edfe1b
3. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v5.15.86&id=ec1727f89ecd6f2252c0c75e200058819f7ce47a
4. Quilt gave this output when I applied the patch to rebase it:
% quilt push -f
Applying patch platform/100-PCI-mt7621-Add-MediaTek-MT7621-PCIe-host-controller-.patch
patching file arch/mips/ralink/Kconfig
patching file drivers/pci/controller/Kconfig
patching file drivers/pci/controller/Makefile
patching file drivers/staging/Kconfig
patching file drivers/staging/Makefile
patching file drivers/staging/mt7621-pci/Kconfig
patching file drivers/staging/mt7621-pci/Makefile
patching file drivers/staging/mt7621-pci/TODO
patching file drivers/staging/mt7621-pci/mediatek,mt7621-pci.txt
patching file drivers/staging/mt7621-pci/pci-mt7621.c
Hunk #1 FAILED at 1.
Not deleting file drivers/staging/mt7621-pci/pci-mt7621.c as content differs from patch
1 out of 1 hunk FAILED -- saving rejects to file drivers/staging/mt7621-pci/pci-mt7621.c.rej
patching file drivers/pci/controller/pcie-mt7621.c
Applied patch platform/100-PCI-mt7621-Add-MediaTek-MT7621-PCIe-host-controller-.patch (forced; needs refresh)
Upon inspecting drivers/staging/mt7621-pci/pci-mt7621.c.rej, it seems that
the original patch wants to delete drivers/staging/mt7621-pci/pci-mt7621.c
but upstream's version was not an exact match. I opted to delete that
file.
5. Suggestion by hauke: 19098934f9
"This patch is in upstream kernel, but it was backported to the old
staging driver in kernel 5.15."
Build system: x86_64
Build-tested: bcm2711/RPi4B, filogic/xiaomi_redmi-router-ax6000-ubootmod
Run-tested: bcm2711/RPi4B, filogic/xiaomi_redmi-router-ax6000-ubootmod
Signed-off-by: John Audia <therealgraysky@proton.me>
In the build of the ramips/mt76x8 target the user gets asked about these
two configuration options, add them to the generic kernel configuration.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The CONFIG_DRM_XEN_FRONTEND configuration symbol is also used by the
layerscape target, move it to the generic kernel configuration.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
As described in commit 97c77fff28
MV88E6xxx driver (NET_DSA_MV88E6XXX) cannot be built-in when PTP
(PTP_1588_CLOCK) is a module in Linux 5.15. But actually it should be
allowed to be built-in when its PTP support (NET_DSA_MV88E6XXX_PTP) is
disabled.
This adds a patch to fix that.
Signed-off-by: Johnny S. Lee <foss@jsl.io>
Add support for EON EN25QX128A spi nor flash with no flags as it does
support SFDP parsing.
Fixes: #9442
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Tested-by: Szabolcs Hubai <szab.hu@gmail.com> [ramips/mt7621: xiaomi_mi-router-4a-gigabit]
Kernel 5.10.158 added a prompt for the FUNCTION_ERROR_INJECTION symbol.
This is exposed in builds with CONFIG_KERNEL_KPROBES enabled, causing
those builds to fail due to a missing symbol. Add the symbol to fix
this.
Fixes: 6801c460b6 ("kernel: bump 5.10 to 5.10.158")
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
Kernel 5.15.82 added a prompt for the FUNCTION_ERROR_INJECTION symbol.
This is exposed in builds with CONFIG_KERNEL_KPROBES enabled, causing
those builds to fail due to a missing symbol. Add the symbol to fix
this.
Fixes: 68426e54ed ("kernel: bump 5.15 to 5.15.82")
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
nvmem_cell_read return a pointer error when an error occurs. Currently
we convert the pointer error to an int while the rest of the function
return a void* and expcet an error pointer. Fix this PTR_ERR msuse
fixing compilation warning.
Fixes the following compilation warning:
net/ethernet/eth.c: In function 'nvmem_cell_get_mac_address':
net/ethernet/eth.c:547:24: warning: returning 'long int' from a function with return type 'void *' makes pointer from integer without a cast [-Wint-conversion]
547 | return PTR_ERR(mac);
| ^~~~~~~~~~~~
net/ethernet/eth.c: In function 'nvmem_cell_get_mac_address_ascii':
net/ethernet/eth.c:564:24: warning: returning 'long int' from a function with return type 'void *' makes pointer from integer without a cast [-Wint-conversion]
564 | return PTR_ERR(mac_ascii);
| ^~~~~~~~~~~~~~~~~~
Fixes: 7b863af180 ("kernel: move mac-address-ascii patches to generic")
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Removed a20 in description, since it
only works for a10, a13 and a31.
Also refreshed kernel config.
Fixes: #10466
Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
(moved fixes, refreshed, added CONFIG_DVB_USB=n)
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
enable option `CONFIG_CRYPTO_LZ4HC` to match default kernel config
this only adds the `lz4hc_compress` module, and has no effect on the
`lz4_decompress` module which already supports any flavor
Signed-off-by: Tony Butler <spudz76@gmail.com>
Make the patches apply cleanly again.
Fixes: 8327e0fb72 ("kernel: backport the dev_set_threaded export to 5.10")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Manually rebased: ath79/patches-5.10/910-unaligned_access_hacks.patch
All other patches automatically rebased.
Signed-off-by: John Audia <therealgraysky@proton.me>
The CDMQ ingress special tag flag needs to be set for 7986 even without DSA
untag offload, otherwise tx checksum offload seems to break
Fixes: 9b482ee22f ("kernel: add more fixes for mtk_eth_soc")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
The CDMQ ingress special tag flag needs to be set for 7986 even without DSA
untag offload, otherwise tx checksum offload seems to break
Fixes: 9b482ee22f ("kernel: add more fixes for mtk_eth_soc")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Protect the flow block cb list readers against concurrent updates
Reported-by: Chad Monroe <chad.monroe@smartrg.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Performance comparison (iperf3, mtu 1500):
Before: 53.9 Mbps
After: 87.9 Mbps
The tests were performed on a BT Home Hub 5A router.
The iperf3 server was running on the router, the client
on the host.
Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
