Travis-CI: Update PATH

We now can no longer re-use the provided path and need to give the
full PATH we want used now.

Signed-off-by: Tom Rini <[email protected]>

Merge git://git.denx.de/u-boot-dm

dm: test: replace dm_scan_dt() by of dm_extended_scan_fdt() in dm_do_test

This allows to scan the DT including all "clocks" node's sub-nodes
in which fixed-clock are defined.
All fixed-clock should be defined inside a clocks node which collect all
external oscillators. Until now, all clocks sub-nodes can't be binded except
if the "simple-bus" compatible string is added which is a hack.

Update test.dts by moving clk_fixed node inside clocks.

Signed-off-by: Patrice Chotard <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

dm: core: add clocks node scan

Currently, all fixed-clock declared in "clocks" node in device tree
can be binded by clk_fixed_rate.c driver only if each of them have
the "simple-bus" compatible string.
This constraint has been invoked here [1].

This patch offers a solution to avoid adding "simple-bus" compatible
string to nodes that are not busses.

[1] https://patchwork.ozlabs.org/patch/558837/

Signed-off-by: Patrice Chotard <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

patman: add support for omitting bouncing addresses

Add support for reading a list of bouncing addresses from a in-tree file
(doc/bounces) and from the ~/.patman config file. These addresses are
stripped from the Cc list.

Signed-off-by: Chris Packham <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Reviewed-by: Philipp Tomsich <[email protected] <mailto:[email protected]>>

moveconfig: Use fd.write() instead of print >>

Adjust this code so that it can work with Python 2 and 3.

Fixes: d73fcb1 (moveconfig: Support building a simple config database)
Reported-by: Chris Packham <[email protected]>
Signed-off-by: Simon Glass <[email protected]>

dm: core: Add ofnode_for_each_subnode()

Add a convenience macro to iterate over subnodes of a node. Make use of
this where appropriate in the code.

Signed-off-by: Simon Glass <[email protected]>

genconfig.py: Print defconfig next to warnings

At present we sometimes see warnings of the form:

/tmp/tmpMA89kB:36: warning: overriding the value of CMD_SPL.
Old value: "y", new value: "y".

This is not very useful as it does not show whch defconfig file it relates
to. Update the tool to show this.

Signed-off-by: Simon Glass <[email protected]>

dm: core: Show driver name with 'dm tree'

It is often useful to see which driver was actually selected for each
device. Add a new 'Driver' column to provide this information. Sample
output:

 Class       Probed   Driver     Name
----------------------------------------
 root        [ + ]    root_drive root_driver
 keyboard    [ + ]    i8042_kbd  |-- keyboard
 serial      [ + ]    ns16550_se |-- serial
 rtc         [   ]    rtc_mc1468 |-- rtc
 timer       [ + ]    tsc_timer  |-- tsc-timer
 syscon      [ + ]    ich6_pinct |-- pch_pinctrl
 pci         [ + ]    pci_x86    |-- pci
 northbridge [ + ]    bd82x6x_no |   |-- northbridge@0,0
 video       [ + ]    bd82x6x_vi |   |-- gma@2,0
 vidconsole0 [ + ]    vidconsole |   |   `-- gma@2,0.vidconsole0
...

Signed-off-by: Simon Glass <[email protected]>
Reviewed-by: Bin Meng <[email protected]>
Tested-by: Bin Meng <[email protected]>

dm: core: Drop use of strlcpy()

We can use printf() to limit the string width. Adjust the code to do this
instead of using strlcpy() which is a bit clumbsy.

Signed-off-by: Simon Glass <[email protected]>
Reviewed-by: Bin Meng <[email protected]>
Tested-by: Bin Meng <[email protected]>

sandbox: Convert SANDBOX_BITS_PER_LONG to Kconfig

Convert SANDBOX_BITS_PER_LONG to Kconfig and assign it a correct
number depending on which host we are going to build and run.

Signed-off-by: Bin Meng <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

sandbox: Introduce Kconfig option for 32/64 bit host

It seems most of the time we are building and running sandbox on 64-bit host.
But we do support 32-bit host as well. Introduce Kconfig option for this.

Signed-off-by: Bin Meng <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

blk: dm: make blk_create_device() take a number of block instead of a size

There is an overflow problem when taking the size instead of the number
of blocks in blk_create_device(). This results in a wrong device size: the
device apparent size is its real size  modulo 4GB.
Using the number of blocks instead of the device size fixes the problem and
is more coherent with the internals of the block layer.

Signed-off-by: Jean-Jacques Hiblot <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

misc: Kconfig: Add SPL_I2C_EEPROM option

This option is an SPL-variant of the I2C_EEPROM option to enable
the driver for generic I2C-attached EEPROMs for SPL.

Signed-off-by: Wenyou Yang <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

board: sama5d4_xplained: Set mac address from eeprom

Add the code to set the ethernet mac address from eeprom by using
the common code from the common folder.

Signed-off-by: Wenyou Yang <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

board: sama5d2_xplained: Replace code to set mac address

Replace the code to set the ethernet mac address with the code from
the common folder.

Signed-off-by: Wenyou Yang <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

board: atmel: Create board/$(VENDOR)/common folder

Create board/$(VENDOR)/common folder to accommodate the common code
shared by other atmel boards, now put the code to set ethernet mac
address from eeprom, which uses the i2c eeprom driver.

Signed-off-by: Wenyou Yang <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

env: ti: boot: Select dtb name for X15 revC

Select dtb name for am57xx BeagleBoard-X15 revC

Signed-off-by: Lokesh Vutla <[email protected]>
Reviewed-by: Tom Rini <[email protected]>

board: ti: am57xx: Add dt support for BeagleBoard-X15 revC

Add support for selecting proper dtb for
am57xx BeagleBoard X15 revC u-boot from FIT

Signed-off-by: Lokesh Vutla <[email protected]>
Reviewed-by: Tom Rini <[email protected]>

board: ti: am571x: Add 666MHz support for AM571x IDK

AM571x supports DDR running at 666MHz. Right now it is
clocked at 532MHz which is lower than what is supported.
In order to have maximum performance on AM571-IDK,
switch DDR to 666MHz.

Signed-off-by: Steve Kipisz <[email protected]>
Signed-off-by: Lokesh Vutla <[email protected]>
Reviewed-by: Tom Rini <[email protected]>

board: ti: dra76: Add dt support

Add support for selecting proper dtb for
dra76x u-boot from FIT.

Signed-off-by: Lokesh Vutla <[email protected]>
Reviewed-by: Tom Rini <[email protected]>

arm: dts: Add u-boot specific compatibles

Separate out u-boot specific compatibles from dts files.
This will help in syncing dts files in future.
Also these will get deleted eventually once respective drivers
are capable of handling Linux dts files.

Signed-off-by: Lokesh Vutla <[email protected]>
Reviewed-by: Tom Rini <[email protected]>

arm: dts: dra7: sync DT with latest Linux

Sync all dra7* specific dts files with the upstream
kernel including changes queued for 4.14

https://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap.git/log/?h=omap-for-v4.14/dt-v3

Signed-off-by: Lokesh Vutla <[email protected]>
Reviewed-by: Tom Rini <[email protected]>

env: ti: boot: Select dtb name for dra76

Select dtb name for dra76-evm.

Reviewed-by: Tom Rini <[email protected]>
Signed-off-by: Lokesh Vutla <[email protected]>

board: ti: dra76-evm: Add support for powering on mmc ldo

ldo4 is used to poweron mmc on dra76-evm. Enable it.

Reviewed-by: Tom Rini <[email protected]>
Signed-off-by: Lokesh Vutla <[email protected]>

board: ti: dra76-evm: Add the pinmux data

Adding pinmux and IODELAY data for dra76-evm.

Reviewed-by: Tom Rini <[email protected]>
Signed-off-by: Vignesh R <[email protected]>
Signed-off-by: Lokesh Vutla <[email protected]>

board: ti: dra76-evm: Add DDR data

dra76-evm has the ddr parts connectedi running at 666MHz:
EMIF1: MT41K512M16HA-125 AIT:A  x 2
EMIF2: MT41K512M8RH-125-AAT:E x 4
Add support for configuring the above DDR parts.

Reviewed-by: Tom Rini <[email protected]>
Signed-off-by: Lokesh Vutla <[email protected]>

board: ti: dra76-evm: Add the pmic data

dra76-evm uses lp8736 and tps65917 pmic for powering on
various peripherals. Add data for these pmics and register
for dra76-evm.

Reviewed-by: Tom Rini <[email protected]>
Signed-off-by: Keerthy <[email protected]>
Signed-off-by: Lokesh Vutla <[email protected]>

board: ti: dra76-evm: Add epprom support

The dra76-evm is a board based on TI's DRA76 processor
Add eeprom support

Reviewed-by: Tom Rini <[email protected]>
Signed-off-by: Lokesh Vutla <[email protected]>

arm: dra76: Add support for ES1.0 detection

dra76 family is a high-performance, infotainment application
device, based on OMAP architecture on a 28-nm technology.
This contains most of the subsystems, peripherals that are
available on dra74, dra72 family. This SoC mainly features
Subsystems:
- 2 x Cortex-A15 with max speed of 1.8GHz
- 2 X DSP
- 2 X Cortex-M4 IPU
- ISS
- CAL
- DSS
- VPE
- VIP
Connectivity peripherals:
- 1 USB3.0 and 3 USB2.0 subsystems
- 1 x SATA
- 2 x PCI Express Gen2
- 3-port Gigabit ethernet switch
- 2 x CAN
- MCAN

Adding CPU detection support for the dra76 ES1.0 soc
and update prcm, control module, dplls data.

Reviewed-by: Tom Rini <[email protected]>
Signed-off-by: Praneeth Bajjuri <[email protected]>
Signed-off-by: Lokesh Vutla <[email protected]>

arm: dra7: Kconfig: Select LP87565 related configs

LP87565 is present on dra76-evm. Select it for
TARGET_DRA7XX_EVM.

Signed-off-by: Lokesh Vutla <[email protected]>
Reviewed-by: Tom Rini <[email protected]>

palmas: Add support for powering different ldos

It is not necessary that ldo1 is used to power on mmc.
So, add support for passing ldo registers for powering on mmc.

Reviewed-by: Tom Rini <[email protected]>
Signed-off-by: Lokesh Vutla <[email protected]>

arm: omap5+: Add board specific ldo powering

It is not necessary all omap5+ based uses the same PMIC
to poweron mmc. So add support for enabling mmc based on board.

Reviewed-by: Tom Rini <[email protected]>
Signed-off-by: Lokesh Vutla <[email protected]>

Merge branch 'master' of git://git.denx.de/u-boot-video

Prepare v2017.09

Signed-off-by: Tom Rini <[email protected]>

MAINTAINERS: board: qcom: db410c: Maintainer changed

Replacing original author Mateusz Kulikowski
<[email protected]> as db410c maintainer

Signed-off-by: Jorge Ramirez-Ortiz <[email protected]>

ipu_common: Let the MX6 IPU clock be calculated in run-time

MX6Q/QP IPU operates at 264MHz and MX6DL IPU at 198MHz.

When running a SPL target, which supports multiple MX6 variants we cannot
properly setup the IPU clock frequency via CONFIG_IPUV3_CLK option as
such decision is done in build-time currently.

Remove the CONFIG_IPUV3_CLK option and let the IPU clock frequency be
configured in run-time on mx6.

Reported-by: Eric Nelson <[email protected]>
Signed-off-by: Fabio Estevam <[email protected]>
Reviewed-by: Eric Nelson <[email protected]>
Reviewed-by: Stefano Babic <[email protected]>
[agust: fixed #endif in cgtqmx6eval.h]
Signed-off-by: Anatolij Gustschin <[email protected]>

cmd: ide: Make the first device the default one

At present the IDE device number is initialized to -1, which means
we cannot type "ide read" command before setting the device number
via "ide device #".

For convenience, let's set the first device as the default one.

Signed-off-by: Bin Meng <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

block: ide: Don't bother to create BLK device if no CDROM inserted

When there is no CDROM inserted, the block size is zero hence there
is no need to create a BLK device for it.

Signed-off-by: Bin Meng <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

block: ide: Fix block read/write with driver model

This converts the IDE driver to driver model so that block read and
write are fully functional.

Fixes: b7c6baef ("x86: Convert MMC to driver model")
Reported-by: Heinrich Schuchardt <[email protected]>
Signed-off-by: Bin Meng <[email protected]>
Reviewed-by: Simon Glass <[email protected]>
Tested-by: Heinrich Schuchardt <[email protected]>

blk: Use macros for block device vendor/product/rev string size

So far these are using magic numbers. Replace them with macros.

Signed-off-by: Bin Meng <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

edison: Disable CONFIG_USB_HOST_ETHER

This board does dwc3 gadget, not host, so we cannot have host support or
we will fail to link.

Signed-off-by: Tom Rini <[email protected]>

env: Migrate CONFIG_ENV_AES to Kconfig and deprecate

The underlying implementation for ENV_AES has security complications and
is not recommended for use.  Please see CVE-2017-3225 and CVE-2017-3226
for more details.  Mark this as deprecated now and delete this in the
medium term if no one comes forward to re-work the support.

Signed-off-by: Tom Rini <[email protected]>

usb: doc: update README.usb for new config options

A number of the config options for USB networking have been migrated to
Kconfig. Update README.usb to reflect this.

Signed-off-by: Chris Packham <[email protected]>

usb: net: migrate USB Ethernet adapters to Kconfig

This migrates ASIX, ASIX88179, MCS7830, RTL8152 and SMSC95XX to Kconfig.
Update defconfigs.

Signed-off-by: Chris Packham <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

usb: net: migrate CONFIG_USB_HOST_ETHER to Kconfig

CONFIG_USB_HOST_ETHER is the framework that the drivers are dependent on
USB_HOST_ETHER. Use this as a menu and move the existing LAN75XX and
LAN78XX options under new menu. Finally update the defconfigs that need
CONFIG_USB_HOST_ETHER.

Signed-off-by: Chris Packham <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

Kconfig: drop CONFIG_USB_ETHER_RNDIS

This is not a valid option. Drop it.

Signed-off-by: Chris Packham <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

Kconfig: drop CONFIG_USB_ETHER_DM9601

This is not a valid option. Drop it.

Signed-off-by: Chris Packham <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

README: Fix typo in description of CONFIG_OF_EMBED

The correct name is 'gd->fdt_blob', not 'gd->blob'.

Signed-off-by: Nobuhiro Iwamatsu <[email protected]>
CC: Simon Glass <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

configs: Resync with savedefconfig

Rsync all defconfig files using moveconfig.py

Signed-off-by: Tom Rini <[email protected]>

rockchip: enable SPL_SYSRESET config for all Rockchip SoCs

With Makefiles testing for $(SPL_TPL_)SYSRESET, we need SPL_SYSRESET
for do_reset() in SPL for Rockchip SoCs.

References: 87c16d4 "drivers: spl: consistently use the $(SPL_TPL_) macro"
Signed-off-by: Kever Yang <[email protected]>
Acked-by: Philipp Tomsich <[email protected]>
Reviewed-by: Philipp Tomsich <[email protected]>

rockchip: rk3328: fix syscon id table

syscon id table need a dummy member as NULL ending, or else system
will panic while try to match a compatible in this table as a list.

Signed-off-by: Kever Yang <[email protected]>
Acked-by: Philipp Tomsich <[email protected]>
Reviewed-by: Philipp Tomsich <[email protected]>

ARM: rmobile: Disable CONFIG_ARCH_FIXUP_FDT_MEMORY

Disable CONFIG_ARCH_FIXUP_FDT_MEMORY to prevent U-Boot from modifying
the memory {} nodes in the DT passed to the Linux kernel. The R8A779x
DT contains multiple memory {} nodes, while U-Boot only modifies the
first one and stuffs all the memory entries into it, which is wrong.
Disabling CONFIG_ARCH_FIXUP_FDT_MEMORY is the least intrusive way to
fix the issue this close to the release, while the real fix is to
extend the fdt_fixup_memory_banks() to handle multiple memory nodes
in DT.

Signed-off-by: Marek Vasut <[email protected]>
Cc: Nobuhiro Iwamatsu <[email protected]>

ARM: rmobile: Fix SD divider settings on Gen3

On RCar M3 and on RCar H3 newer than and not including ES1.0, the SD clock
must be divided by 4 rather than 2 because a hardware workaround present
only in the H3 ES1.0 has been removed from these chips. U-Boot currently
only supports M3 and H3 ES 2.0 and newer, so configure the SD pre-divider
to 4 to prevent SD instability.

Signed-off-by: Marek Vasut <[email protected]>
Cc: Nobuhiro Iwamatsu <[email protected]>

Merge git://git.denx.de/u-boot-net

net: fix typos

%s/Desriptor/Descriptor/g

Signed-off-by: Heinrich Schuchardt <[email protected]>
Acked-by: Joe Hershberger <[email protected]>

armv8: ls1043/ls1046aqds: add support for RGMII_TXID

Signed-off-by: Madalin Bucur <[email protected]>
Acked-by: Joe Hershberger <[email protected]>
Acked-by: York Sun <[email protected]>

net: nfs: Drop CONFIG_NFS_READ_SIZE

In the general case, CONFIG_NFS_READ_SIZE is unchanged from the default
of 1024.  There are in fact no in-tree users that increase this size.
Adjust the comment to reflect what could be done in the future in
conjunction with CONFIG_IP_DEFRAG.

Cc: Joe Hershberger <[email protected]>
Signed-off-by: Tom Rini <[email protected]>
Acked-by: Joe Hershberger <[email protected]>

net: phy: realtek: fix enabling of the TX-delay for RTL8211F

The old logic always enabled the TX-delay when the phy-mode was set to
PHY_INTERFACE_MODE_RGMII. With this patch we enable the TX delay for
PHY_INTERFACE_MODE_RGMII_ID and PHY_INTERFACE_MODE_RGMII_TXID and
disable it for PHY_INTERFACE_MODE_RGMII.

Based on a similar change made in the Linux Realtek PHY driver
by Martin Blumenstingl <[email protected]>.

Signed-off-by: Madalin Bucur <[email protected]>
Acked-by: Joe Hershberger <[email protected]>
Acked-by: York Sun <[email protected]>

spl: typo fix for SPL_ATF_SUPPORT description

Delete one redundant 'which' for SPL_ATF_SUPPORT description.

Signed-off-by: Kever Yang <[email protected]>

linker_lists: remove incorrect comment

Remove a comment line refering to a non-existent file.

Signed-off-by: Heinrich Schuchardt <[email protected]>

vsprintf: vsprintf does not have parameter size

The inline documentation of vsprintf mentions a parameter size
which does not exist in the function declaration.

int vsprintf(char *buf, const char *fmt, va_list args);

Signed-off-by: Heinrich Schuchardt <[email protected]>

Revert "dm: arm: rpi: Drop CONFIG_OF_EMBED"

This reverts commit 25877d4e4c45451c5398aec3de50e0d5befe0e9f.

This is a workaround for Raspberry Pi boot failures seen when passing on
the device tree provided by the Raspberry Pi firmware at boot. Without
CONFIG_OF_EMBED, we just get stuck at "Starting kernel ..." when we try
to boot Linux with this device tree.

Signed-off-by: Paul Barker <[email protected]>
Cc: Simon Glass <[email protected]>
Tested-by: Jonathan Gray <[email protected]>
Reviewed-by: Simon Glass <[email protected]>

tools: .gitignore: Add libfdt related files

Some files are generated during libfdt build. Ignore them.

This was wrongly put in the .gitignore in the root directory before.
Now let's remove entries there and put them in the right place.

Fixes: 34e2c285 ("gitignore: add intermediates from libfdt build")
Signed-off-by: Bin Meng <[email protected]>

tools: .gitignore: Sort in alphabetical order

These are currently out of alphabetical order.

Signed-off-by: Bin Meng <[email protected]>

doc: remove duplicate text in README.gpt

Signed-off-by: Alison Chaiken <[email protected]>

part: mac: Suppress the error message after reading ddb

Change to use 'debug' to output the error message if it fails to
read the driver descriptor block.

Signed-off-by: Bin Meng <[email protected]>

env: Add 'envtools' target to 'make help'

The target is not currently mentioned anywhere. Add it to the help so
people can find it.

Signed-off-by: Simon Glass <[email protected]>

env: Fix operation of 'make environ'

This was broken by the recent environment refactoring. Specifically:

$  make environ
scripts/Makefile.build:59: tools/environ/Makefile: No such file or directory
make[1]: *** No rule to make target 'tools/environ/Makefile'.  Stop.
make: *** [Makefile:1469: environ] Error 2

Fix this by updating the Makefile and adjusting the #include filesnames in
two C files.

Fixes: ec74f5f (Makefile: Rename 'env' target to 'environ')
Reported-by: Måns Rullgård <[email protected]>
Signed-off-by: Simon Glass <[email protected]>

usb: gadget: g_dnl: Sync internal SN variable with env

Since commit 842778a09104 ("usb: gadget: g_dnl: only set iSerialNumber
if we have a serial#") "fastboot devices" stopped to show correct device
serial number for TI boards, showing this line instead:

    ???????????? fastboot

This is because serial# env variable could be set after g_dnl gadget was
initialized (e.g. by using env_set() in the board file).

To fix this, let's update internal serial number variable (g_dnl_serial)
when "serial#" env var is changed.

Signed-off-by: Sam Protsenko <[email protected]>
Cc: Felipe Balbi <[email protected]>
Tested-by: Heiko Schocher <[email protected]>
Acked-by: Łukasz Majewski <[email protected]>
Reviewed-by: Marek Vasut <[email protected]>

env: Fix out of tree building of tools-all

With the move of environment code from common/ to env/ a number of
changes needed to be made to various make targets.  We missed updating
some of the files required for out of tree builds of the tools.  Correct
the 'environ' target to know that we need to work under tools/env/ still
(not tools/environ/) and then update the wrappers in env_attr.c and
env_flags.c to point to the new correct file.

Reported-by: Peter Robinson <[email protected]>
Tested-by: Peter Robinson <[email protected]>
Signed-off-by: Tom Rini <[email protected]>

Merge git://git.denx.de/u-boot-video

Merge git://git.denx.de/u-boot-rockchip

MAINTAINERS: add lcd files to video section

LCD drivers are video related, so add them to video section.

Signed-off-by: Anatolij Gustschin <[email protected]>

rockchip: i2c: fix >32 byte reads

The hw can read up to 32 bytes at a time. If we need
more than one chunk, we have to enter the plain RX mode.

Signed-off-by: Wadim Egorov <[email protected]>
Acked-by: Philipp Tomsich <[email protected]>

rockchip: firefly-rk3399: enable ATF and dwmmc

enable the ATF option and SDCard in defconfig.

Signed-off-by: Kever Yang <[email protected]>
Acked-by: Philipp Tomsich <[email protected]>
Reviewed-by: Philipp Tomsich <[email protected]>

rockchip: timer: fix U_BOOT_DRIVER name

When I originally added this driver, I did some careless (and in
retrospect: mindless) copy & paste for the U_BOOT_DRIVER structure
skeletion... unfortunately, the 'arc_timer' string was committed
and slipped through all reviews.

This fixes the U_BOOT_DRIVER name to read 'rockchip_rk3368_timer'
(as originally intended).

Signed-off-by: Philipp Tomsich <[email protected]>
Reported-by: Artturi Alm <[email protected]>

rockchip: rk3288: Add reset reason detection

Sometimes it's helpful to know the reset reason caused in the SoC.
Add reset reason detection for the RK3288 SoC.
This will set an environment variable which represents the reset reason.

Signed-off-by: Wadim Egorov <[email protected]>
Acked-by: Philipp Tomsich <[email protected]>
Reviewed-by: Philipp Tomsich <[email protected]>

rockchip: phycore: Read configuration EEPROM & set ethaddr in late init

Read SoM information from EEPROM and set ethaddr in late init.

Signed-off-by: Wadim Egorov <[email protected]>
Acked-by: Philipp Tomsich <[email protected]>
Reviewed-by: Philipp Tomsich <[email protected]>

Merge branch 'rmobile' of git://git.denx.de/u-boot-sh

ARM: rmobile: Add missing IPSR18 bits to R8A7795 PFC

The IPSR18 register bits were missing from the R8A7795 ES2.0+ PFC
tables, which triggered a BUG() in sh_pfc driver. This is because
of an out-of-bounds access to the pinmux_gpios[] array in the PFC
tables, which was too short due to the missing IPSR18 bits.

Signed-off-by: Marek Vasut <[email protected]>
Cc: Nobuhiro Iwamatsu <[email protected]>
Signed-off-by: Nobuhiro Iwamatsu <[email protected]>

ARM: rmobile: Drop board_mmc_init

The board_mmc_init() is no longer invoked when DM is used, so move
all the pinmux config into board_init() instead.

Signed-off-by: Marek Vasut <[email protected]>
Cc: Nobuhiro Iwamatsu <[email protected]>
Signed-off-by: Nobuhiro Iwamatsu <[email protected]>

ARM: rmobile: Replace CONFIG_RAVB with CONFIG_RENESAS_RAVB in boards

The configuration option name is the later, so replace usage of
CONFIG_RAVB in board files with CONFIG_RENESAS_RAVB , otherwise
the RAVB pinmux is not set and ethernet does not work.

Signed-off-by: Marek Vasut <[email protected]>
Cc: Nobuhiro Iwamatsu <[email protected]>
Signed-off-by: Nobuhiro Iwamatsu <[email protected]>

splash_source: Verify FIT magic

Before reading entire FIT image, add sanity check by testing image
header against FDT_MAGIC. This should help avoid problems in situations
where FIT is not yet available from storage device, for example when
performing initial programming of device.

Cc: Anatolij Gustschin <[email protected]>
Acked-by: Tomas Melin <[email protected]>

video: ipuv3_fb: skip IPU shutdown if IPU was not enabled before

Boards can skip display interface init using board_video_skip().
If display interface was not initialized (e.g. no ipuv3 framebuffer
registered or IPU clock disabled), booting Linux stops due to the
crash in IPU shutdown function, when accessing IPU registers.
Check IPU clock and skip shutdown if clock is not enabled.

Signed-off-by: Anatolij Gustschin <[email protected]>

Merge git://git.denx.de/u-boot-dm

Prepare v2017.09-rc4

Signed-off-by: Tom Rini <[email protected]>

crypto/fsl: fix obj-yy in Makefile

When enabling CONFIG_CMD_BLOB and/or CONFIG_CMD_DEKBLOB, the build fails
with a linker error:
  ...
  LD      u-boot
arch/arm/mach-imx/built-in.o: In function `blob_encap_dek':
/home/clemens/dev/u-boot/arch/arm/mach-imx/cmd_dek.c:46: undefined
reference to `blob_dek'

This is due to an error in the Makefile, resulting in obj-yy/obj-yn/..
and fsl_blob.o is therefore not linked.

Fix it by splitting it up into two obj-y lines.

Signed-off-by: Clemens Gruber <[email protected]>
Reviewed-by: Fabio Estevam <[email protected]>

git: mailrc: Update e-mail address

Signed-off-by: Lukasz Majewski <[email protected]>

dm: core: Add livetree documentation

Add some documentation for the live device tree support in U-Boot. This
was missing from the initial series.

Signed-off-by: Simon Glass <[email protected]>
Suggested-by: Lukasz Majewski <[email protected]>
Reviewed-by: Łukasz Majewski <[email protected]>

gitignore: add intermediates from libfdt build

Since ee95d10 (fdt: Build the new python libfdt module), a number of
additional files are auto-generated/installed into the tools directory.
List these in .gitignore to suppress having them listed as untracked.

Signed-off-by: Philipp Tomsich <[email protected]>
References: ee95d10 (fdt: Build the new python libfdt module)
Reviewed-by: Simon Glass <[email protected]>

fdt: fix 'prop (...) not found!' error in 'fdt set' command

This commit brings things back to the well known working state of the
command.
-
With commit 9620d87259572ef21f0df60988d9a932ca673779
(cmd/fdt: support single value replacement within an array)

there was an error introduced modifying (inserting) a property to a
device-tree node.
fdt_getprop(...) returnes a len with -1 for a non-existing property, but
a memcpy with len -1 isn't a good idea and things went wrong (crash).
-
Some times later Tom did repair this
with commit 99bb38e2cce9d99238458e0f6d1880c6d2e80a4d
(fdt: Check for NULL return from fdt_getprop in 'fdt set')

This repairs the crash but the behaviour of the command isn't like
before, it makes it impossible to insert a property.
-

Signed-off-by: Hannes Schmelzer <[email protected]>
Acked-by: Simon Glass <[email protected]>

dm: Add migration plan for CONFIG_BLK

The CONFIG_BLK conversion involves quite invasive changes in the U-Boot
code, with #ifdefs and different code paths. We should try to move over to
this soon so we can drop the old code.

Set a deadline of 9 months for this work, rounded up to the next release.

Signed-off-by: Simon Glass <[email protected]>
Reviewed-by: Tom Rini <[email protected]>

Makefile: Suppress output of python libfdt build command

This should not be printed by default. Prefix it with $(Q).

Fixes ee95d10b: ("fdt: Build the new python libfdt module")
Signed-off-by: Bin Meng <[email protected]>

Makefile: Quiesce libfdt build

Since commit 3809e302 "Makefile: honor PYTHON configuration properly",
the build commands of libfdt are printed while previously were not.

This adds the missing '--quiet' back.

Signed-off-by: Bin Meng <[email protected]>

omap3: evm: Fixes for CONFIG_NAND, SPL_OS_BOOT, USB, and environment

- Pass MTDPARTS kernel arguments to kernel
- Use Kconfig CONFIG_NAND instead of CONFIG_SYS_EXTRA_OPTIONS="NAND"
- Call 'usb_stop' on kernel start
- Update Falcon mode setup to match other OMAP3 boards
- Use "uEnv.txt" as boot script instead of "boot.scr"

block: Drop the ftide020 driver

This is not used in U-Boot.

Signed-off-by: Bin Meng <[email protected]>

block: ide: Drop CONFIG_IDE_LED

This is actually not used. Drop it.

Signed-off-by: Bin Meng <[email protected]>

block: ide: Drop CONFIG_IDE_INIT_POSTRESET

This is not referenced anywhere. Drop it.

Signed-off-by: Bin Meng <[email protected]>

nvme: Remove dead codes in nvme_setup_io_queues()

Execution cannot reach this statement: "nr_io_queues = result;"

Reported-by: Coverity (CID: 166731)
Signed-off-by: Bin Meng <[email protected]>
Reviewed-by: Tom Rini <[email protected]>