This patch backports several patches that went upstream into
Herbert Xu's cryptodev-2.6 tree:
crypto: Use zeroing memory allocator instead of allocator/memset
crypto: crypto4xx - performance optimizations
crypto: crypto4xx - convert to skcipher
crypto: crypto4xx - avoid VLA use
crypto: crypto4xx - add aes-ctr support
crypto: crypto4xx - properly set IV after de- and encrypt
crypto: crypto4xx - extend aead fallback checks
crypto: crypto4xx - put temporary dst sg into request ctx
The older, outstanding patches from 120-wxyz series have been
upstreamed as well and therefore they have been reassigned to
fit into the series.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
The swconfig package is part of the DEFAULT_PACKAGES list in
the apm821xx's nand subtarget. It's enabled by default because
the MX60(W) and WNDR4700 need it for their initramfs. However
the package is not necessary for the MR24's sysupgrade image, as
the MR24 does not have a switch.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Refreshed all patches
Added new ARM64 symbol: ARM64_ERRATUM_1024718
Compile-tested on: ar71xx
Runtime-tested on: ar71xx
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
- DGND3700v1: fix port numbers, they're in reverse order
- FaST2704n: there is no WAN port identified as such, only 4 LAN
- AD1018: configure the "FIBRE" port as WAN
Signed-off-by: Daniel Gonzalez Cabanelas <dgcbueu@gmail.com>
Mostly whitespace cleanups. Some unneeded code was removed.
MMC init was also moved to the probe function as in
6069bdd
The cleanup commits are over 100, making it hard to do them individually.
Tested on GnuBee PC1 with an SD card being used as swap.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
v2: Fixed compile issue with mt7620
Both initramfs and flashed images are built and boot. No Ethernet, no WLAN,
probably further issues, so the image is not added to TARGET_DEVICES for
now.
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Add the syscon compatible, otherwise used functions like
syscon_regmap_lookup_by_phandle() will return an error and setting the
ethernet pll data wont work at all.
Fix the pll register width. Writing to registers out of the range via
syscon isn't possible and returns an error. On ar7242 the last pll
register - Current Audio Modulation Logic Output - is at 0x1805003c.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Remove recipes for not yet migrated boards. Drop variables which aren't
required due to the switch to devicetree like boardname.
Drop the legacy image build code. Boards added to this target shouldn't
use the legacy image build code anyway.
Signed-off-by: Mathias Kresin <dev@kresin.me>
None of these checks are used at the moment and most of them aren't
required if the image metadata validation is used.
Signed-off-by: Mathias Kresin <dev@kresin.me>
The watchdog kill command was meant for busybox watchdog. Busybox watchdog
was replaced by the procd watchdog mid 2013 with commit df7ce9301a
("busybox: disable the watchdog utility by default"), which makes the kill
command obsolete since quite some time.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Add the SoC compatible to the individual dts files. Rename the dts files
to match the common pattern.
Remove dts files wich aren't used and no image in ar71xx exists.
Signed-off-by: Mathias Kresin <dev@kresin.me>
It isn't a phy supply, as the phy is still up if the GPIO is low. It
rather is the supply for the vbus. A correct setting/definition will be
relevant as soon as USB peripheral mode is supported.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Backport patch adding support for the vbus-supply devicetree property,
which allows to specific GPIOs (via fixed regulators) to enable
vbus/usb power.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Rename the image and use a compatible string which indicates that only
hardware revision 2 and higher is supported.
It allows to use the wireless LED, as HWRev 1 uses GPIO#39 for the
wireless LED and starting with HWRev 2 GPIO#35 is used for the wireless
LED and GPIO#39 for IFX_GPIO_MODULE_EXTPHY_MDIO.
The HWREV can be checked by connecting to the fritzbox right after power
on via ftp:
ftp> quote GETENV HWSubRevision
Within the same HW revision 5 of the Fritz!Box 3370 different NAND flash
chips are used. Usually it isn't a big deal but depending on the used
NAND flash chip, the ECC calculation is done different (and incompatible
of course).
Boards with a Micron MT29F1G08ABADA NAND flash chip are using the NAND
chip to calculate the ECC (on-die). Boards with a Hynix HY27UF081G2M NAND
flash chip are doing the ECC calculation in software.
Supporting both with a single DTS isn't possible. It might be possible
to add a patch selecting the ECC mode dynamicaly based on the found NAND
flash chip. But such a patch has no chance to get accepted upstream and
most likely need to be touched with every kernel update.
Instead two images are created. One for Micron NAND flash chip and one
for Hynix NAND flash chip. So far no pattern is known to identify the
used flash chip without opening the box.
Add the power off GPIO. At least EVA version 2186 sets/keeps the GPIO as
input, which will cause a reboot 30sec after power on. For boards with
EVA version 2186 the installation is tricky as it has to be finished
within the 30sec time frame.
The EVA version can be checked by connecting to the fritzbox right after
power on via ftp:
ftp> quote GETENV urlader-version
The ath9k eeprom/caldata is at a different and offset and stored in
reverse order (from the last byte to the beginning) on the flash.
Reverse the bits to bring the data into the format expected by the
ath9k driver.
Since the ath9k eeprom is stored in reverse order on flash, we can not
use the mac address from the on flash eeprom. Get the MAC address from
the tffs instead.
Within the same HW revision 5 of the Fritz!Box 3370 both version of the
vr9 SoC are used. During preparation of kernel 4.14 support, all
devicetree source files were changed to load the vr9 v1.1 and vr9 v1.2
gphy firmware, which fixed the embedded phys for boards using the
version 1.2 of the vr9 SoC.
While at it, add a trigger to make use of the LAN LED. Setup the
build-in switch and add a hint for LuCI two show the ports in order
matching the labels on the case
Add support for the second USB port and provide the volatage GPIOs. Use
GPIO#21 as PCIe reset pin. The lan led is connected to GPIO#38.
Name the rootfs partition ubi and remove the mtd/rootfs related kernel
bootargs to use the OpenWrt autoprobing based on the partition name.
Enable sysupgrade support to allow an upgrade from a running system.
Since sysupgrade wasn't supported till now, drop image build code which
was added to allow a sysupgrade from earlier OpenWrt versions.
Build images that allow an (initial) installation via EVA bootloader.
To install OpenWrt via Eva bootloader, within the first seconds after
power on a ftp connection need to be established to the FRITZ!Box at
192.168.178.1 and the the following ftp commands need to be run:
ftp> quote USER adam2
ftp> quote PASS adam2
ftp> binary
ftp> debug
ftp> passive
ftp> quote SETENV linux_fs_start 0
ftp> quote MEDIA FLSH
ftp> put /path/to/openwrt-lantiq-xrx200-FRITZ3370-eva-kernel.bin mtd1
ftp> put /path/to/openwrt-lantiq-xrx200-FRITZ3370-eva-filesystem.bin mtd0
Signed-off-by: Andrea Merello <andrea.merello@gmail.com>
Signed-off-by: Mathias Kresin <dev@kresin.me>
In case of error, the function devm_ioremap_resource() returns ERR_PTR()
and never returns NULL. The NULL test in the return value check should
be replaced with IS_ERR().
Fixes: f079b6406348 ("staging: mt7621-eth: add gigabit switch driver (GSW)")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In case of error, the function of_phy_connect() returns NULL pointer not
ERR_PTR(). The IS_ERR() test in the return value check should be
replaced with NULL test.
Fixes: e3cbf478f846 ("staging: mt7621-eth: add the drivers core files")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In case of error, the function devm_ioremap_resource() returns ERR_PTR()
and never returns NULL. The NULL test in the return value check should
be replaced with IS_ERR().
Fixes: e3cbf478f846 ("staging: mt7621-eth: add the drivers core files")
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is an amagalmation of two upstream commits dealing with whitespace
and dead code removal. I'm synching instead of having two separate commits
as they go out of order compared to previous commits here.
Tested on GnuBee PC1.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Mostly whitespace cleanups. Some unneeded code was removed.
MMC init was also moved to the probe function as in
6069bdd087
The cleanup commits are over 100, making it hard to do them individually.
Tested on GnuBee PC1 with an SD card being used as swap.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Both these devices have a wps and a reset button on
GPIO pins, which need kmod-gpio-button-hotplug
package to work.
Add this package to their default package config.
Troubleshooted and tested on a Viper.
Signed-off-by: Alberto Bursi <alberto.bursi@outlook.it>
1. compatible property in node gmac was wrong
2. ag71xx_setup_gmac_933x should use np of gmac-config and
not the pointer to gmac. gmac is only used for the reg address.
Signed-off-by: Johann Neuhauser <johann@it-neuhauser.de>
Linksys WRT32X (Venom) is identical in hardware to the WRT3200ACM
with a different flash layout and boots zImage rather than uImage.
Specification:
- Marvell Armada 385 88F6820 (2x 1.8GHz)
- 256MB of Flash
- 512MB of RAM
- 2.4GHz (bgn) and 5GHz (an+ac wave 2)
- 4x 1Gbps LAN + 1x 1Gbps WAN
- 1x USB 3.0 and 1x USB 2.0/eSATA (combo port)
Flash instruction:
Apply factory image via web-gui.
Signed-off-by: Michael Gray <michael.gray@lantisproject.com>
These resources are extracted from devicetree, so they aren't
needed here.
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Code currently defines:
and then compiles code only if they are defined.
We might want to disable some of these via devicetree one
day, but for now just remove the #defines and the
conditions - all the code for different ports is
easy to identify.
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
As the Interrupts for the PCI adapters are listed in
devicetree we shouldn't need to have them explicit in the code.
The simplest way to do this is to use of_irq_parse_and_map_pci()
and specify an interrupt-map which identifies the different
PCI hosts by bus/slot numbers.
This has the advantage that the hwirq number are mapped to virq
numbers for us, so the ugly hack can go.
Signed-off-by: NeilBrown <neil@brown.name>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove the owner assignment form the platform driver as
platform_driver_register() already initializes the owner.
Found using coccinelle.
Signed-off-by: Christian Lütke-Stetzkamp <christian@lkamp.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Drop the "-x lzo" compression type override from MKUBIFS_OPTS since we
do not support it in mkfs.ubifs. Attempting to use LZO compression results
in "Error: bad compressor name".
The build system also appends an explicit "--compr=zlib" to the mkfs command
line, turning the "-x" flag into a no-op anyway.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
The subtarget configurations contained a lot of reduntant symbols, arbitrary
overrides and symbols set to =m.
Remove most of these redundant settings to shrink the subtarget configurations
to a manageable size.
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Fixes the following error spotted by the buildbot:
Package kmod-fs-squashfs is missing dependencies for the following libraries:
xz_dec.ko
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Make the call to of_mdiobus_register() dependant on CONFIG_OF to avoid
pulling in an unwanted dependency on of_mdio.ko.
Fixes: 3dc523f232 kernel: add OF support to rtl8355 driver
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
Fix ART offset (make it universal for 8/16 MB versions of the board) and
while at it, include also GPIO setup for h/w watchdog (EM6324QYSP5B).
Fixes: FS#1532
Signed-off-by: Piotr Dymacz <pepe2k@gmail.com>
The biggest (and the only?) disadventage of this is obviously an
increased image size.
For mips74k the size of vmlinux goes up from 4186684 B to 4701436 B.
Most devices use LZMA compressed kernel so probably more important is
vmlinux.lzma size which goes up from 1342945 B to the 1508498 B.
Still this isn't something that should stop target kernel bump. There
are various adventages of kernel 4.14. If kernel / image size is a
serious concern for anyone, it's perfectly possible to use previous
release which is pretty solid for the brcm47xx.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Cc: Paul Wassi <p.wassi@gmx.at>
Cc: Hauke Mehrtens <hauke@hauke-m.de>
Splitted out the dts file and create the new dts for the 256 MByte RAM and
the 512 MB RAM version.
Migrate both versions to the common board detection.
The install the 512 MByte Version on a board running the 256 MByte image,
a forceful sysupgrade with the -F flag is required.
Signed-off-by: Davide Ammirata <list@davidea.it>
The RavPower WD03 is a battery powered SD card reader and a USB port.
Specifications:
SOC: MediaTek MT7620N
BATTERY: 6000mah
WLAN: 802.11bgn
LAN: 1x 10/100 Mbps Ethernet
USB: 1x USB 2.0 (Type-A)
RAM: PM Tech PMD708416CTR-5CN 32 MB
FLASH: Holtek HT66F40 - 8 MB Flash
LED: Power button and 4 leds to indicate power level of the
battery (could not get control of that)
INPUT: Power, reset button
OTHER: USB SD-Card reader with card detect on GPIO#42
Tested and working:
- Ethernet
- 2.4 GHz WiFi (Correct MAC-address)
- installation from tftp
- OpenWRT sysupgrade (Preserving and non-preserving)
- LEDs
- Buttons
Installation:
- Download the sysupgrade image
- Place it in the root of a clean TFTP server running on your computer.
- Rename the image to "kernel" — be sure there is no file extension.
- Plug the WD03 into your computer via ethernet.
- Set your computer to use 10.10.10.254 as its IP address.
- With your WD03 shut down, hold down the power button until the first
white LED lights up.
- Push and hold the reset button and release the power button. Continue
holding the reset button for 30 seconds or until it begins searching
for files on your TFTP server, whichever comes first.
- The WD03 (10.10.10.128) will look for your computer at 10.10.10.254
and install the kernel file. Once it has finished installation of the
kernel file, it will search for a (nonexistent) rootfs file — when it
begins searching for this file, shut down the WD03 by holding the
power button normally.
- Start up your WD03 normally.
Signed-off-by: Matthias Badaire <mbadaire@gmail.com>
Backport patch accepted upstream which removes the reset asserts of the
xrx200 gphy driver on reboot/remove.
While doing a global software reset, these bits are not cleared and let
some bootloader fail to initialise the GPHYs. The bootloader don't expect
these bits to be set, as they aren't during power on.
The asserts were a workaround for a wrong syscon-reboot mask. With a mask
set which includes the GPHY resets of the first reset register, the
resets of the second reset register arn't required any more.
Signed-off-by: Mathias Kresin <dev@kresin.me>
Some of the new ath79 device trees were missing read-only tags on the
u-boot partitions. This adds the tags where previously missing.
Signed-off-by: Russell Senior <russell@personaltelco.net>
This new syntax is slightly better designed & uses "compatible" string.
For details see Documentation/devicetree/bindings/mtd/partition.txt .
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
This commit adds support for the AVM Fritz!Box 4020 WiFi-router.
SoC: Qualcomm Atheros QCA9561 (Dragonfly) 750MHz
RAM: Winbond W971GG6KB-25
FLASH: Macronix MX25L12835F
WiFi: QCA9561 b/g/n 3x3 450Mbit/s
USB: 1x USB 2.0
IN: WPS button, WiFi button
OUT: Power LED green, Internet LED green, WLAN LED green,
LAN LED green, INFO LED green, INFO LED red
UART: Header Next to Black metal shield
Pinout is 3.3V - RX - TX - GND (Square Pad is 3.3V)
The Serial setting is 115200-8-N-1.
Tested and working:
- Ethernet (LAN + WAN)
- WiFi (correct MAC)
- Installation via EVA bootloader
- OpenWRT sysupgrade
- Buttons
- LEDs
Not working:
- USB port
Installation via EVA:
In the first seconds after Power is connected, the bootloader will
listen for FTP connections on 169.254.157.1 (Might also be 192.168.178.1). Firmware can be uploaded
like following:
ftp> quote USER adam2
ftp> quote PASS adam2
ftp> binary
ftp> debug
ftp> passive
ftp> quote MEDIA FLSH
ftp> put openwrt-sysupgrade.bin mtd1
Note that this procedure might take up to two minutes. After transfer is
complete you need to powercycle the device to boot OpenWRT.
Signed-off-by: David Bauer <mail@david-bauer.net>
Some maintainers prefer DTS files licensed under permissive license like
MIT / BSD. As all DT bindings should be OS independent and DTS files are
pretty separated from Linux code it probably makes sense to share them
across projects.
The safest solution is to use dual licensing: that way it stays clear
these files can be used in GPL projects without depending on current
belief of licenses compatibility.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
Acked-by: John Crispin <john@phrozen.org>
This target aims to replace ar71xx mid-term. The big part that is still
missing is making the MMIO/AHB wifi work using OF. NAND and mikrotik
subtargets will follow.
Signed-off-by: John Crispin <john@phrozen.org>
This switches bcm53xx from spi-bcm53xx to the spi-bcm-qspi driver. The
later one was developed by Broadcom and is more advanced one. It
supports more modes, setting a speed, setting bits per word and uses
IRQs instead of polling.
This increases kernel size from 1808120 B to the 1811160 B (by 3040 B).
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
This adds an interrim patch for v4.14 based on an
upstream commit to get ethernet working on D-Link DNS-313
(probably also on the Wiliboards)
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Since the D-Link devices boot from hard disk we need to
add the following changes to Gemini:
- Supply a bootarg in the device tree so we can boot from
the right partition (/dev/sda4 on DNS-313)
- Disable forced command line in config so the kernel picks
up the right bootargs from the device tree
- Enable EXT4 in the config as this is used for rootfs else
we get nowhere, we cannot load this as a module because,
well, it is supposed to be loaded from the root partition
(chicken and egg problem).
- Enable jbd2 and mbcache (needed by ext4)
Also clean out the premature attempts to dynamically
modify the command line in the Image makefile: we should
pass this with the device tree bootargs instead, it works
much better.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Roman Yeryomin <roman@advem.lv>
Disable video and input drivers - they are needed only for one
board and can be enabled via modules if/when UI will be available.
Use LED drivers from packages.
Cleanup other things unnecessarily overriding defaults.
Signed-off-by: Roman Yeryomin <roman@advem.lv>
This adds the patches to get fairly complete Gemini support
using kernel v4.14. It is mainly a backport of patches from
kernel v4.16 with omissions of things like graphics that require
substantial changes and will be better handled once we move
to the v4.16 kernel proper.
On top of this are some WIP patches for USB support.
Tested on Raidsonic NAS4220B and D-link DNS-313.
ChangeLog v4->v5:
- Fix ethernet single gmac usecase
- Fix USB reset (patch from Hans)
- Fix Raidsonic ethernet skew delay
- Fix kernel config (bridge, squashfs, jffs2, usb)
- Disable second usb port on Raidsonic board until fotg210_hcd is fixed
ChangeLog v3->v4:
- Make sure to use tabs rather than spaces in base-files.
- Use the dns313 image tool from the firmware-utils.
- Break out the addition of the v4.14 patches and the removal
of the v4.4 patches to separate (big) patches.
ChangeLog v2->v3:
- Update the kernel config as indicated by Hauke Martens:
- Regenerate again after rebasing using kernel_oldconfig
dropping a few optimization settings that are now generic
- Drop CFG80211 stuff (module)
- Drop CIFS stuff (module)
- Drop MAC80211 (module)
- Drop wireless drivers (module)
- Enabled OverlayFS
- Added proper DNS-313 boot image generation with the special
file header tool.
- Disable CMA in the kernel
- Enable LZMA compression of the kernel
- Consequently name the nas4220b images nas4220b
- Update preinit MAC detection script to handle also DNS-313
- Add board.d/03_hdparm to set the disk to spin down after
1 minute by default, if we have the hdparm tool installed
ChangeLog v1->v2:
- Processed config through kernel_oldconfig
- Processed patches through make target/linux/{clean,refresh} V=99
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Roman Yeryomin <roman@advem.lv>
Add target device as at91-sama5d2_ptc_ek in SAMA5D2 subtarget and
build images for SAMA5D2 PTC Ek board.
Signed-off-by: Sandeep Sheriker Mallikarjun <sandeepsheriker.mallikarjun@microchip.com>
reorganizing at91 subtargets based on sama5 soc features and this fix
below problems.
1. able to set neon flags to sama5d2 & sama5d4 subtargets.
2. fix the make clean which removes all the subtargets in bin folder.
3. able to configure kernel specific to subtarget.
4. able to set vfpu4 flags to samad3 subtargets.
Signed-off-by: Sandeep Sheriker Mallikarjun <sandeepsheriker.mallikarjun@microchip.com>
creating sdcard image with ext4 rootfs only and ignoring creating
other filesystem in sdcard image.
Signed-off-by: Sandeep Sheriker Mallikarjun <sandeepsheriker.mallikarjun@microchip.com>
Add support for SAMA5D27 SOM1 with target device as at91-sama5d27_som1_ek
in SAMA5 subtarget and build images for SAMA5D27 SOM1 Ek board.
Signed-off-by: Sandeep Sheriker Mallikarjun <sandeepsheriker.mallikarjun@microchip.com>
copying at91bootstrap and uboot binaries to sdcard image is done
based on device name instead of dts name and this avoids creating
multiple u-boot and at91bootstrap binaries.
Signed-off-by: Sandeep Sheriker Mallikarjun <sandeepsheriker.mallikarjun@microchip.com>
In typical OpenWrt setups, there are no other users that have a shell spawned for them by default.
This can be overriden by the kernel.dmesg_output syssctl.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
sama5d2 & samad4 have neon feature and sama5d3 does not have neon
feature due to which sama5d3 boot fails with error message Kernel
panic - not syncing: Attempted to kill init! exitcode=0x00000004.
removing neon & VFPv4 support to fix this
Signed-off-by: Sandeep Sheriker Mallikarjun <sandeepsheriker.mallikarjun@microchip.com>
when external kernel is selected from menuconfig, device wb50n is not
avaliable and build fails. As a fix adding checks for external kernel.
Signed-off-by: Sandeep Sheriker Mallikarjun <sandeepsheriker.mallikarjun@microchip.com>
Kernel support already present.
Patch adds userspace for the 'iLS' suffix model of the RB2011 family.
Enables correct initial switch settings, sysupgrade, etc.
https://mikrotik.com/product/RB2011iLS-IN
Signed-off-by: Simon Paterson <simon.paterson.nz@gmail.com>
The harden branch predictor was backported for arm64 with 4.9.92-96.
Fixes: 9aa196e0f2 ("kernel: bump 4.9 to bump 4.9.96")
Signed-off-by: Mathias Kresin <dev@kresin.me>
This is mainly a bug fix for multi-core MIPS systems where L1 caches besides the primary do not get flushed.
The most obvious problem is data corruption on SATA and USB devices where read requests are typically larger than the cacheline size.
This may also fix ar71xx systems that suffer from similar data corruption but I have not tested if it does.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Some Ubiquiti U-boot versions, in particular the "U-Boot 1.1.4.2-s956
(Jun 10 2015 - 10:54:50)" found with AirOS 5.6, do not correctly flush the
caches for the whole kernel address range after decompressing the kernel
image, leading to hard to debug boot failures, depending on kernel version
and configuration.
As a workaround, prepend the relocate-kernels loader, which will invalidate
the caches after moving the kernel to the correct load address.
Reported-by: Andreas Ziegler <dev@andreas-ziegler.de>
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
