Some patches were slightly cleaned up. One things worth mentioning is
that adding:
phy-mode = "rgmii"
broke SF2 driver. It made it access random register breaking switch
setup.
That's why this commit also adds a quick sf2 fix.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Add linux 5.10 as testing kernel, copy generic/subtarget configs and target
patches from 5.4.
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
x86: update target patches for kernel 5.10
Refresh all patches while also dropping:
800-hwmon-w83627ehf-dont-claim-nct677x.patch
which is now upstreamed as:
3207408ab4cb ("hwmon: (w83627ehf) remove nct6775 and nct6776 support")
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
x86: refresh 5.10 target config
Refresh config using "make kernel_oldconfig".
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
x86: refresh and test all subtarget configs
Refresh configs using "make kernel_oldconfig CONFIG_TARGET=subtarget".
Build and run-tested using QEMU: x86/64, x86/legacy, x86/generic.
Build-tested only: x86/geode.
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
[squashed commits]
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
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>
The ZyXEL GS1900-8 is a 8 port switch without any PoE functionality or
SFP ports, but otherwise similar to the other GS1900 switches.
Specifications
--------------
* Device: ZyXEL GS1900-8 v1.2
* SoC: Realtek RTL8380M 500 MHz MIPS 4KEc
* Flash: Macronix MX25L12835F 16 MiB
* RAM: Nanya NT5TU128M8GE-AC 128 MiB DDR2 SDRAM
* Ethernet: 8x 10/100/1000 Mbit
* LEDs: 1 PWR LED (green, not configurable)
1 SYS LED (green, configurable)
8 ethernet port status LEDs (green, SoC controlled)
* Buttons: 1 on-off glide switch at the back (not configurable)
1 reset button at the right side, behind the air-vent
(not configurable)
1 reset button on front panel (configurable)
* Power 12V 1A barrel connector
* UART: 1 serial header (JP2) with populated standard pin connector on
the left side of the PCB, towards the back. Pins are labelled:
+ VCC (3.3V)
+ TX (really RX)
+ RX (really TX)
+ GND
the labelling is done from the usb2serial connector's point of
view, so RX/ TX are mixed up.
Serial connection parameters for both devices: 115200 8N1.
Installation
------------
Instructions are identical to those for the GS1900-10HP and GS1900-8HP.
* Configure your client with a static 192.168.1.x IP (e.g. 192.168.1.10).
* Set up a TFTP server on your client and make it serve the initramfs
image.
* Connect serial, power up the switch, interrupt U-boot by hitting the
space bar, and enable the network:
> rtk network on
* Since the GS1900-10HP is a dual-partition device, you want to keep the
OEM firmware on the backup partition for the time being. OpenWrt can
only boot off the first partition anyway (hardcoded in the DTS). To
make sure we are manipulating the first partition, issue the following
commands:
> setsys bootpartition 0
> savesys
* Download the image onto the device and boot from it:
> tftpboot 0x84f00000 192.168.1.10:openwrt-realtek-generic-zyxel_gs1900-8-initramfs-kernel.bin
> bootm
* Once OpenWrt has booted, scp the sysupgrade image to /tmp and flash it:
> sysupgrade /tmp/openwrt-realtek-generic-zyxel_gs1900-8-squashfs-sysupgrade.bin
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This adds experimental ethernet support for BCM6318, BCM6328, BCM6362, BCM6368
and BCM63268.
BCM6358 needs a different driver, so there's no support for now.
Working devices:
- Comtrend AR-5315u
- Comtrend AR-5387un
- Comtrend VR-3025u
- Comtrend VR-3032u
Not working devices:
- Netgear DGND3700 v2 (no idea on how the external switch is connected)
- Huawei HG556a ver B (BCM6358 needs a separate driveer)
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Make some cosmetic changes in the Buffalo LinkStation LS421DE NAS:
- Delete pointless #xxx-cells
- bootargs: replace earlyprintk with earlycon and remove unneeded args.
- Separate pinmux nodes with empty lines.
Signed-off-by: Daniel González Cabanelas <dgcbueu@gmail.com>
dd on Mac OS X apparently fails when using 'M' unit for bs.
dd: bs: illegal numeric value
Use 'k' unit instead for 'pad-to' to fix that.
Reported-by: Georgi Valkov <gvalkov@abv.bg>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
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>
NVRAM access may be needed early in boot process. Reading it using mtd
happens quite late in the init process. Add NVRAM initialization to the
NVMEM driver which comes up early and depends on IO mapping only.
This is required by Linksys devices which use NVRAM content for proper
partitioning (detecting current firmware partition).
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Demote a number of debugging printk's to pr_debug to avoid log
nosie. Several of these functions are called as a result of
userspace activity. This can cause a lot of log noise when
userspace does periodic polling.
Most of this could probably be removed completely, but let's
keep it for now since these drivers are still in development.
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Tested-by: Stijn Tintel <stijn@linux-ipv6.be>
It supports NVRAM access described using DT binding. Right now NVRAM
data is exposed using /sys/bus/nvmem/ only.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Introduce new patch for automatically detecting RAM size.
Some boards have a different amount of RAM depending on the HW revision.
Therefore, automatically detecting the RAM size instead of hard-coding it will
reduce the number of device definitions.
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
The Sercomm AD1018 has a NAND flash. We recently added support for NANDs
in this target.
Use the internal NAND as additional storage.
Signed-off-by: Daniel González Cabanelas <dgcbueu@gmail.com>
This was added recently and thus overlooked in 85b1f4d8ca
("treewide: remove execute bit and shebang from board.d files").
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The bootloader of many ipq806x boards seems to require the config node
of the FIT image to be 'config@1' (or a secific different value).
This requirement used to be implicitely satisfied because OpenWrt used
to also call the configuration node inside a FIT image 'config@1'.
However, as recent U-Boot now prohibits the use of the '@' symbol as
part of node names, this was changed by
commit 5ec60cbe9d ("scripts: mkits.sh: replace @ with - in nodes")
Explicitely restore the default name of the configuration node to
'config@1' on ipq806x.
(ipq807x is unaffected as DEVICE_DTS_CONFIG default is set
"config@hk01" in target/linux/ipq807x/image/Makefile)
Reported-by: Chen Minqiang <ptpt52@gmail.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
The bootloader of many ipq40xx boards seems to require the config node
of the FIT image to be 'config@1' (or a secific different value).
This requirement used to be implicitely satisfied because OpenWrt used
to also call the configuration node inside a FIT image 'config@1'.
However, as recent U-Boot now prohibits the use of the '@' symbol as
part of node names, this was changed by
commit 5ec60cbe9d ("scripts: mkits.sh: replace @ with - in nodes")
Explicitely restore the default name of the configuration node to
'config@1' on ipq40xx.
Reported-by: Chen Minqiang <ptpt52@gmail.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
The in-SoC RTC of the Bananapi R64 is more disruptive than useful
without a battery connected. Disable it to not have Linux use the
RTC provided time 2000-01-01 00:00:00 after power-loss.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Make packages depending on usb-serial selective, so we do not have
to add kmod-usb-serial manually for every device.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
So far, board.d files were having execute bit set and contained a
shebang. However, they are just sourced in board_detect, with an
apparantly unnecessary check for execute permission beforehand.
Replace this check by one for existance and make the board.d files
"normal" files, as would be expected in /etc anyway.
Note:
This removes an apparantly unused '#!/bin/sh /etc/rc.common' in
target/linux/bcm47xx/base-files/etc/board.d/01_network
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Enable testing kernel.
Fix compile errors by using new kernel APIs.
Fix fuzz by manually editing patches to ensure the code goes in the
right place.
For 721-NET-no-auto-carrier-off-support.patch, revert upstream commit
a307593a6 to keep the OpenWrt ralink driver operational.
Add mt7621-pci-phy patch to select REGMAP_MMIO as discussed in PR #3693
and #3952.
Run automatic quilt refresh on the rest.
Signed-off-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
0098-disable_cm.patch is not needed because upstream fixed CM handling.
The rest are straightforward removals of upstreamed patches.
Signed-off-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
This reverts commit b4aad29a1d.
This was accidentally folded into a single commit. Remove it and
apply it properly again.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
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>
Enable testing kernel.
Delete upstreamed patches:
0098-disable_cm.patch can be dropped, upstream fixed CM handling.
Fix compile errors by using new kernel APIs.
Fix fuzz by manually editing patches to ensure the code goes in the
right place.
For 721-NET-no-auto-carrier-off-support.patch, revert upstream commit
a307593a6 to keep the OpenWrt ralink driver operational.
Add mt7621-pci-phy patch to select REGMAP_MMIO as discussed in PR #3693
and #3952.
Rename patches to follow the 3-digit classification from the OpenWrt
Developer Guide.
Run automatic quilt refresh.
Signed-off-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
This patch copies over refreshed patches from 5.4.
- dropped crypto patches (they got upstreamed)
- dropped renesas USB 3 firmware loader (they got upstreamed)
- NAND now needs extra device-properties for ECC settings.
Signed-off-by: Christian Lamparter <chunkeey@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>
It's a BCM4906 based device (2 CPU cores). It has 512 MiB of RAM, 4 LAN
ports, 1 WAN port, 2 USB ports, NAND flash. WiFi unknown at this point.
Flashing is possible using CFE only, proper image will be worked on
later.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
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>
Patch to fix kernel panic was recently accepted upstream so rename patch
and add acked lines to reflect that.
Signed-off-by: Sieng Piaw Liew <liew.s.piaw@gmail.com>
(add the same patch for v5.10)
Signed-off-by: Álvaro Fernández Rojas <noltari@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>
While rebasing into setting bits instead of magic values,
I accidentally forgot to actually set the force bit.
Without it using the pins as GPIO-s did not actually work.
Fixes: b5c93ed ("ipq40xx: add Qualcomm QCA807x driver")
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Debugging the SPI CS issue with kernel 5.10 resulted in a better
understanding for the root cause and a proper patch with a better
explanation.
Exchange the old hack patch with a more efficient (and upstreamable)
solution.
Signed-off-by: David Bauer <mail@david-bauer.net>
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>
All mt7622 devices except for the UBI-variant of the mt7622-rfb1 carry
metadata appended to the sysupgrade image.
Add it for the mt7622-rfb1-ubi as well and check it on sysupgrade to
avoid accidentally flashing firmware for the wrong device (or variant
or future DEVICE_COMPAT_VERSION).
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
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>
Just by running `make kernel_oldconfig` and unsetting following options
manually as those cores are cortex-a7 based and thus irrelevant for the
currently default cortex-a9 used cores.
CONFIG_CLK_IMX6SL is not set
CONFIG_CLK_IMX6SX is not set
CONFIG_CLK_IMX6UL is not set
Signed-off-by: Petr Štetiar <ynezz@true.cz>
mt7622 uses MBR partition for booting from SD card.
Add hybrid MBR entry with boot flag after PMBR entry.
Signed-off-by: Oskari Lemmela <oskari@lemmela.net>
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 is required for devices that use NVRAM data for detecting currently
used firmware partition (e.g. Linksys devices).
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
The previous approach of referencing artifacts in follow-up artifacts
can't work with parallel builds in the current way image.mk is built.
Refactor things so this is not needed.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Write everything needed for eMMC install into the gaps between
partitions on SD card. In that way, installation to eMMC only needs
the SD card, no additional files need to be loaded via TFTP any more.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
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 profile is meant to be used on MT7622 rfb1 AP, indicate that in
the name to make things less confusing.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
At this moment driver start fail with error:
[ 3.771991] fsl,elbc-fcm-nand: probe of ffa00000.nand failed with error -22
elbc-fcm-nand driver use legacy method of ecc mode detection. It detect hw/sw
ecc mode when system configure it to "none". [1]
This patch adds 'nand-ecc-mode = "none"' propoerty to use generic driver
ecc mode detection.
[1] https://elixir.bootlin.com/linux/v5.10.18/source/drivers/mtd/nand/raw/fsl_elbc_nand.c#L730
Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
At this moment p2020rdb has broken images, because NOR memory connected
to eLBC bus isn't detected.
In 642b1e8dbed7 linux tree commit, config dependencies of MTD_PHYSMAP_OF
was changed and now MTD_PHYSMAP is required.
This patch adds MTD_PHYSMAP option to kernel config in p2020 subtarget
and fix booting of p2020rdb.
Fixes: 13b1db795f ("mpc85xx: add support for kernel 5.4")
Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
Amazon AWS T3 cloud instances require kernel support
for the Elastic Fabric Adapter to access storage
and for Elastic Network Adapter to use network
interfaces.
Since the Fabric Adapter is needed to access
root filesystem, enable in x86_64 kernel.
Elastic Network Adapter goes in a module,
and add this module to default list in x86_64.
The module is set to AutoLoad because AutoProbe does
not seem to load it.
Signed-off-by: Alberto Bursi <bobafetthotmail@gmail.com>
Changes:
* Increase "oem" partition size from 0x10000 to 0x20000
* Correct partition lables, synchronize with official firmware
Evidence:
It should be the same as hiwifi hc5x61a and the fact indeed the
case. Here is part of dmesg boot log read from official firmware:
[ 1.470000] Creating 7 MTD partitions on "raspi":
[ 1.470000] 0x000000000000-0x000000030000 : "u-boot"
[ 1.480000] 0x000000030000-0x000000040000 : "hw_panic"
[ 1.490000] 0x000000040000-0x000000050000 : "Factory"
[ 1.490000] 0x000000fc0000-0x000000fe0000 : "oem"
[ 1.500000] 0x000000fe0000-0x000000ff0000 : "bdinfo"
[ 1.510000] 0x000000ff0000-0x000001000000 : "backup"
[ 1.510000] 0x000000050000-0x000000fc0000 : "firmware"
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Assign the usbdev trigger via devicetree and drop the userspace
handling of the usb leds.
Drop the now unused userspace helper code as well.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Acked-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
These boards have a fixed size kernel partition but do not limit the
kernel size during image building.
Disable image building for both boards as well, since the kernel of the
last release as well as master are to big to fit into the 2 MByte kernel
partition.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Assign the usbdev trigger via devicetree and drop the userspace
handling of the usb leds
Add the PCI attached usb controller as trigger sources for the usb led
as well.
Signed-off-by: Mathias Kresin <dev@kresin.me>
The symbol CONFIG_CAVIUM_CN63XXP1 was disabled during the bump to
4.19 (see Fixes:) with the following reason:
No supported hardware uses CN63XXP1 and it causes "slight decrease
in performance"
However, it later turned out that the edgerouter image needed it,
which led to having the device disabled in [1].
Still, dropping support of a device seems a harsh action for just
removing a "slight" decrease in performance from the other devices.
Thus, this enables CONFIG_CAVIUM_CN63XXP1 again, and essentially
restores the situation present until (including) kernel 4.14 on
this target.
For OpenWrt as a platform, it seems more desirable to support all
devices (and have them tested regularly via the snapshots) in this
case.
Users interested in maximum performance might still just remove
the symbol again in their local build.
[1] 3824fa26d2 ("octeon: disable edgerouter image")
Fixes: 6c22545225 ("target/octeon: Add Linux 4.19 support")
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
These patches are required for the Ubiquiti UniFi 6 LR to work. They
were already present for kernel 5.4 but got lost when adding 5.10
support.
Signed-off-by: David Bauer <mail@david-bauer.net>
**What's new**
* Bring support for the Bananapi BPi-R64 to the level desirable for
a nice hackable routerboard.
* Use ARM Trusted Firmware A from source. (goodbye binary preloader)
* Use Das U-Boot from source. (see previous commit)
* Assemble SD-card image using OpenWrt image-commands.
(no gen_sd_cruz_foo.sh added, this is not Raspbian)
* Updated kernel options to support root filesystem.
* Updated DTS to match OpenWrt LAN ports, known LEDs, buttons, ...
* Detect root device, handle sysupgrade, config restore, ...
* Wire up (known) LEDs and buttons in OpenWrt-fashion.
* Build one set of images from SD-card and eMMC.
* Hopefully provide a good example of how things can be done right
from scratch.
**Installation and images**
* Have an empty SD-card at hand
* Write stuff to the card, as root (card device is /dev/mmcblkX)
- write header, gpt, bl2, atf, u-boot and recovery kernel:
`cat *bpi-r64-boot-sdcard.img *bpi-r64-initramfs-recovery.fit > /dev/mmcblkX`
- rescan partitions:
`blockdev --rereadpt /dev/mmcblkX`
- write main system to production partition:
`cat *bpi-r64-squashfs-sysupgrade.fit > /dev/mmcblkXp5`
* Installation to eMMC works using SD-card bootloader via TFTP
When running OpenWrt of SD-card, issue this to trigger installation
to eMMC:
`fw_setenv bootcmd run emmc_init`
Be prepared to serve the content of bin/targets/mediatek/mt7622 on
TFTP server address 192.168.1.254.
**What's missing**
* The red LED is always on, probably a hardware bug.
* AHCI (probably needs DTS changes)
* Ship SD-card image ready with every needed for eMMC install.
* The eMMC has a second, currently unused boot partition. This would
be ideal to store the WiFi EEPROM and Ethernet MAC address(es).
@sinovoip ideas?
Thanks to Thomas Hühn @thuehn for providing the hardware!
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
The vendor flash layout of the Linksys E8450 is problematic as it uses
the SPI-NAND chip without any wear-leveling while at the same time
wasting a lot of space for padding.
Use an all-UBI layout instead, storing the kernel+dtb+squashfs in
uImage.FIT standard format in UBI volume 'fit', the read-write
overlay in UBI volume 'rootfs_data' as well as reduntant U-Boot
environments 'ubootenv' and 'ubootenv2', and a 'recovery'
kernel+dtb+initramfs uImage.FIT for dual-boot.
** WARNING **
THIS PROCEDURE CAN EASILY BRICK YOUR DEVICE PERMANENTLY IF NOT CARRIED
OUT VERY CAREFULLY AND EXACTLY AS DESCRIBED!
Step 0
* Configure your PC to have the static IPv4 address 192.168.1.254/24
* Provide bin/targets/mediatek/mt7622 via TFTP
Now continue EITHER with step 1A or 1B, depending on your preference
(and on having serial console wired up or not).
Step 1A (Using the vendor web interface (or non-UBI OpenWrt install))
In order to update to the new bootloader and UBI-based firmware,
use the web browser of your choice to open the routers web-interface
accessible on http://192.168.1.1
* Navigate to
'Configuration' -> 'Administration' -> 'Firmware Upgrade'
* Upload the file
openwrt-mediatek-mt7622-linksys_e8450-ubi-initramfs-recovery.itb
and proceed with the upgrade.
* Once OpenWrt comes up, use SCP to upload the new bootloader files to
/tmp on the router:
*-mt7622-linksys_e8450-ubi-preloader.bin
*-mt7622-linksys_e8450-ubi-bl31-uboot.fip
* Connect via SSH as you will now need to replace the bootloader in
the Flash.
ssh root@192.168.1.1
(the usual warnings)
* First of all, backup all the flash now:
for mtd in /dev/mtdblock*; do
dd if=$mtd of=/tmp/$(basename $mtd);
done
* Then use SCP to copy /tmp/mtdblock* from the router and keep them
safe. You will need them should you ever want to return to the
factory firmware!
* Now flow the uploaded files:
mtd -e /dev/mtd0 write /tmp/*linksys_e8450-ubi-preloader.bin /dev/mtd0
mtd -e /dev/mtd1 write /tmp/*linksys_e8450-ubi-bl31-uboot.fip /dev/mtd1
If and only if both writes look like the completed successfully
reboot the router. Now continue with step 2.
Step 1B (Using the vendor bootloader serial console)
* Use the serial to backup all /dev/mtd* devices before using the
stock firmware (you got root shell when connected to serial).
* Then reboot and select 'U-Boot Console' in the boot menu.
* Copy the following lines, one by one:
tftpboot 0x40080000 openwrt-mediatek-mt7622-linksys_e8450-ubi-preloader.bin
tftpboot 0x40100000 openwrt-mediatek-mt7622-linksys_e8450-ubi-bl31-uboot.fip
nand erase 0x0 0x180000
nand write 0x40080000 0x0 0x180000
reset
Now continue with step 2
Step 2
Once the new bootchain comes up, the loader will initialize UBI and the
ubootenv volumes. It will then of course fail to find any bootable
volume and hence resort to load kernel via TFTP from server
192.168.1.254 while giving itself the address 192.168.1.1
The requested file is called
openwrt-mediatek-mt7622-linksys_e8450-ubi-initramfs-recovery.itb
and your TFTP server should provide exactly that :)
It will be written to UBI as recovery image and booted.
You can then continue and flash the production OS image, either
by using sysupgrade in the booted initramfs recovery OS, or by using
the bootloader menu and TFTP.
That's it. Go ahead and mess around with a bootchain built almost
completely from source (only DRAM calibration blobs are fitted in bl2,
and the irreplacable on-chip ROM loader remains, of course).
And enjoy U-Boot built with many great features out-of-the-box.
You can access the bootloader environment from within OpenWrt using the
'fw_printenv' and 'fw_setenv' commands. Don't be afraid, once you got
the new bootchain installed the device should be fairly unbrickable
(holding reset button before and during power-on resets things and
allows reflashing recovery image via TFTP)
Special thanks to @dvn0 (Devan Carpenter) for providing amazingly fast
infra for test-builds, allowing for `make clean ; make -j$(nproc)` in
less than two minutes :)
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
