Disable ASPM support for this NIC, fixing strange behavior problems, such as
increased latency, strange uneven throughput, etc.
With this option disabled the NIC achieves stable performance.
Upsteam r8169 driver disables ASPM by default for this NIC.
Signed-off-by: Pavel Kubelun <be.dissent@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18509
[bump release, fix commit description, split 2nd tx queue part]
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
This allows the network interface naming to be stable, free from any
possible interaction from external USB network devices that might
claim usb* interface names.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
This allows the network interface naming to be stable, free from any
possible interaction from external USB network devices that might
claim usb* interface names.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
This allows the network interface naming to be stable, free from any
possible interaction from external USB network devices that might
claim usb* interface names.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
This allows the network interface naming to be stable, free from any
possible interaction from external USB network devices that might
claim usb* interface names.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Currently, bcm63xx-cfe is being installed into kernel build dir, however
that does not work for Image Builder as only certain artifacts from kernel
build dir are included in Image Builder.
So, simply install bcm63xx-cfe into image staging dir so its artifacts can
be used in Image Builder as well.
Fixes: #18408Fixes: #18409
Link: https://github.com/openwrt/openwrt/pull/18463
Signed-off-by: Robert Marko <robimarko@gmail.com>
Partially revert removal of 003 patch to fix compilation error with
kernel 6.6.
In 6.6 .remove still require int and .remove_new needs to be used.
Fixes: bfb106e8ae ("mac80211: backport API change for sta rate control update")
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
The ath10k dt-binding property "qcom,coexist-support" was
explicitly defined as type uint8 since upstream commit
ed09c61eb19d ("dt-bindings: net: Convert ath10k to YAML").
Therefore, this hack patch no longer makes sense.
Link: ed09c61eb1
Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18393
Signed-off-by: Robert Marko <robimarko@gmail.com>
Backport two commits to resolve issues with ath1kk causing it to fail
driver registration on iommuless systems with DRAM outside of 32bit
addressing such as a 4GiB imx8mm:
commit 1bcd20981834 ("wifi: ath11k: Fix DMA buffer allocation
to resolve SWIOTLB issues")
commit eeadc6baf8b3 ("wifi: ath11k: Use dma_alloc_noncoherent
for rx_tid buffer allocation")
Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Link: https://github.com/openwrt/openwrt/pull/17751
Signed-off-by: Robert Marko <robimarko@gmail.com>
This new version fixes some memory leak and NULL pointer issues.
Upstreamed patches:
[1] 010-api_update.patch
[2] 201-wifi-ath10k-add-LED-and-GPIO-controlling-support-for.patch
The new 003 patch was introduced to fix the build error on old LTS
kernel. The patch 004 was used to fix the issue of IPQ4019 radio
not being able to start.
Tested on QCA9882 and IPQ4019.
[1] bca8bc0399
[2] 8e1debd824
Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18368
Signed-off-by: Robert Marko <robimarko@gmail.com>
Fix compilation error for old stable version caused by
genlmsg_multicast_allns backport fix pushed middle version.
Version 5.15 version 0-169, 6.1 version 0-115, 6.6 version 0-58 have the
old genlmsg_multicast_allns version with flags variable.
Compiling backport project with these version result in a compilation
error. To handle this, introduce a backport function for the affected
kernel version.
Link: https://github.com/openwrt/openwrt/pull/18373
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
MR7500 features IPQ6018 which is CPPE, however, it also utilizes AQR114CB0 as WAN.
Previously, without this patch, aquantia_phy_api_ops was never called.
Therefore, enable `IN_AQUATIA_PHY` so that AQR PHY gets initialized properly on boot.
[ 3.973942] Aquantia AQR114C 90000.mdio-1:08: loading firmware version 'v5.6.5 Cybertan Divo 090221 14:43:44' from 'NVMEM'
[ 16.059953] aquantia_phy_api_ops_init[1485]:INFO:qca probe aquantia phy driver succeeded!
[ 17.930149] Aquantia AQR114C 90000.mdio-1:08: attached PHY driver (mii_bus:phy_addr=90000.mdio-1:08, irq=POLL)
Signed-off-by: Weikai Kong <priv@pppig236.com>
Link: https://github.com/openwrt/openwrt/pull/18185
Signed-off-by: Robert Marko <robimarko@gmail.com>
Sets the boot flag for the i40e network device driver
to load it at a more early stage of the boot process.
With commit 0a47d518df,
I added a boot priority for the mlx4 and mlx5 drivers.
Also, increase those priorities because I think they
are too low since there is currently no "room" for
built-in network device drivers.
That can cause interface order, i.e., name inconsistencies,
when Mellanox ConnectX cards are inserted or removed.
Signed-off-by: Til Kaiser <mail@tk154.de>
Link: https://github.com/openwrt/openwrt/pull/17990
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The CONFIG_NET_SWITCHDEV option is needed by CONFIG_DSA and some other
options. It is boolean, we have to compile it into the kernel it self.
Activate it for all targets in the generic configuration, it is already
activated for most of them. This allows to install DSA drivers as a
module.
On the ramips/mt7620 target the kernel would grown by 4.5kB.
For some small targets which do not support a DSA switch by default the
option is deactivated.
Link: https://github.com/openwrt/openwrt/pull/17668
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This allows the network interface naming to be stable, free from any
possible interaction from external USB network devices that might
claim usb* interface names.
(This was a real problem I encountered with a nanopi R6S device and
an external rtl8152 usb3 network controller - the USB controller would
claim the eth1 name, causing much confusion).
Signed-off-by: Michel Lespinasse <michel@lespinasse.org>
Link: https://github.com/openwrt/openwrt/pull/17638
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
It doesn't depend on either usb-net or usb-net-cdc-ncm. It does, however, depend
on mii. Fix thusly, and make it depend explicitly on usb, not usb-net.
While at it, add a conditional dependency on libphy, for future kernel versions.
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
Now that CONFIG_USB_NET_DRIVERS is unconditionally enabled, remove it from the
usb-net kconfig symbol list.
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
The kmod-mlxsw-spectrum driver activated CONFIG_DCB indirectly already
on all targets which are building this driver. All other DCB capable
driver did not activate their DCB support.
CONFIG_DCB increases the uncompressed kernel size by about 7.8KB.
CONFIG_DCB is only needed some data center Ethernet cards and not used
on normal routers. Activate it only on the x86_64 and the armsr_arm64
target which are used on normal servers or in VMs.
Link: https://github.com/openwrt/openwrt/pull/17672
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
