However, mainline driver doesn't work very well on RTL8111
devices, so let's move to vendor driver for better performance.
Signed-off-by: Tianling Shen <cnsztl@gmail.com>
It's stable enough to overclock cpu frequency to 2.2/1.8 GHz,
and for better performance.
Signed-off-by: Tianling Shen <cnsztl@gmail.com>
Co-authored-by: gzelvis <gzelvis@gmail.com>
Some boards have SD card connectors where the power rail cannot be switched
off by the driver. If the card has not been power cycled, it may still be
using 1.8V signaling after a warm re-boot. Bootroms expecting 3.3V signaling
will fail to boot from a UHS card that continue to use 1.8V signaling.
Set initial signal voltage in mmc_power_off() to allow re-boot to function.
This fixes re-boot with UHS cards on Asus Tinker Board (Rockchip RK3288),
same issue have been seen on some Rockchip RK3399 boards.
Backport from https://lore.kernel.org/linux-rockchip/AM3PR03MB09664161A7FA2BD68B2800A7AC620@AM3PR03MB0966.eurprd03.prod.outlook.com/
Signed-off-by: Tianling Shen <cnsztl@gmail.com>
sysupgrade.bin has been added to IMAGES twice, resulting in
warnings like:
Makefile:86: warning: overriding recipe for target
'[...]/tmp/openwrt-ath79-generic-dlink_dap-2660-a1-squashfs-sysupgrade.bin'
Makefile:86: warning: ignoring old recipe for target
'[...]/tmp/openwrt-ath79-generic-dlink_dap-2660-a1-squashfs-sysupgrade.bin'
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The current support for MikroTik NAND-based devices relies on a
gross hack that packs the kernel into a static YAFFS stub, as the
stock bootloader only supports booting a YAFFS-encapsulated kernel.
The problem with this approach is that since the kernel partition is
blindly overwritten without any kind of wear or badblock management
(due to lack of proper support for YAFFS in OpenWRT), the NAND flash
is not worn uniformly and eventually badblocks appear, leading to
unbootable devices.
Until a proper fix is found (or the stock bootloader supports other
filesystems), we disable building these images to prevent unknowing
users from risking their devices.
Thanks to Thibaut Varène for summarizing the details above.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
A few devices in ath79 and ramips use mtd-concat to concatenate
individual partitions into a bigger "firmware" or "ubi" partition.
However, the original partitions are still present and visible,
and one can write to them directly although this might break the
actual virtual, concatenated partition.
As we cannot do much about the former, let's at least choose more
descriptive names than just "firmwareX" in order to indicate the
concatenation to the user. He might be less tempted into overwriting
a "fwconcat1" than a "firmware1", which might be perceived as an
alternate firmware for dual boot etc.
This applies the new naming consistently for all relevant devices,
i.e. fwconcatX for virtual "firmware" members and ubiconcatX for
"ubi" members.
While at it, use DT labels and label property consistently, and
also use consistent zero-based indexing.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Otherwise the missing symbol is added to target config for every kernel
config refresh.
While at it, remove the disabled symbol from target configs.
Fixes: 4943bc5cff ("kernel: only strip proc for small flash devices")
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
- Important improvement
Rewrite startup code to speed up startup
Same as global server, no need to start multiple processes
Use unified path and configuration
- Changes
Add Reset default settings
Optimize the use of shadowsocks plugin
Remove the sh script used in update and modify the UI usage
Remove Easylist China from adblock and add NEO DEV HOST
Adjust the parameters used by wget to ensure compatibility
- Fixes
Fix the issue of starting twice after apply in the server node list
Fix the problem that dnsmasq fails to start when the access control list contains wrong entries
This reverts the usage of the S-Tag for separating LAN and WAN port on
the embedded switch. Many users complained about not being able to
manage C-Tag addition / removal on the switch as well as degraded
performance.
Fixes: commit 9da2b56760 ("ipq40xx: fix ethernet vlan double tagging")
Signed-off-by: David Bauer <mail@david-bauer.net>
Since generic has the option set to y and other targets now inherit that
choice, there is no behaviour change
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
The option was introduced in upstream linux commit a6484045 ("[TCP]: Do
not present confusing congestion control options by default.").
The option is set to y in generic config and to the moment does not
incur additional size increment. Make it y for all so that packages
such as kmod-tcp-bbr do not have to set it on every occasion
Signed-off-by: Yousong Zhou <yszhou4tech@gmail.com>
original idea by chunkeey
The essedma driver has its own unique take on VLAN management
and its configuration. In the original SDK, each VLAN is
assigned one virtual ethernet netdev.
However, this is non-standard. So, this patch does away
with the default_vlan_tag property the driver is using
and therefore forces the user to use the kernel's vlan
feature.
Unfortunately, this change will cause the essedma driver
to leak LAN<->WAN during LEDE bootup.
Signed-off-by: Chen Minqiang <ptpt52@gmail.com>
This commit adds support for Qualcomm IPQ8062 SoC.
IPQ8062 is a lower clock variant of IPQ8064.
CPU and NSS clocks:
- CPU: 384 MHz - 1 GHz
- NSS: 110 MHz - 550 MHz
opp and l2 clock values are taken from WG2600HP3 GPL source code [1].
Due to a lack of devices, I didn't test the following features.
- SATA
- NAND flash memory controller
- SD
- USB
- GSBI2, GSBI7
- PCIE2
- GMAC0, GMAC3
Works properly:
- GSBI4 UART
- GSBI5 SPI
- GMAC1, GMAC2
- PCIE0, PCIE1
- MDIO0
Does not work properly:
- CPU SPC
- This can cause a system hang. Same as IPQ8065.
See 2336c2dbb1
[1] https://www.aterm.jp/function/wg2600hp3/appendix/opensource.html
Signed-off-by: Yanase Yuki <dev@zpc.sakura.ne.jp>
