efi_selftest: check installation of the device tree

The unit test checks if a device tree is installed. It requires that the
'compatible' property of the root node exists. If available it prints the
'serial-number' property.

The serial-number property is derived from the environment variable
'serial#'. This can be used to check if the image_setup_libfdt() function
is executed.

A Python test is supplied. It sets a value for serial# and checks that the
selftest shows this as serial-number.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: support device tree for bootefi selftest

The second argument of the bootefi command should always be usable to
specify a device tree. This was missing for bootefi selftest and
bootefi hello.

Proper error handling is added.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: check initialization of EFI subsystem is successful

Up to now errors in the initialization of the EFI subsystems was not
checked.

If any initialization fails, leave the bootefi command.

We do not retry initialization because this would require to undo all prior
initalization steps.

Suggested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: do_bootefi_exec should always return an EFI status code

The return type of do_bootefi_exec() is efi_status_t. So in case
of an error we should always return an EFI status code.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: efi_get_time_init should return status code

All EFI initialization functions should return a status code.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: exit status for efi_reset_system_init

efi_reset_system_init provides the architecture or board specific
initialization of the EFI subsystem. Errors should be caught and
signalled by a return code.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: simplify calling efi_init_obj_list

efi_init_obj_list() should be executed only once.

Rather than having the caller check this variable and the callee set it,
move all access to the variable inside the function. This reduces the
logic needed to call efi_init_obj_list().

Suggested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: consistently return efi_status_t efi_watchdog_register

efi_watchdog_register() should always return a status code and not
a boolean value.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: return efi_status_t from efi_net_register

Consistently return status codes form efi_net_register().

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: return efi_status_t from efi_gop_register

All initialization routines should return a status code instead of
a boolean.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
[agraf: Convert warnings to debug() prints]
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: efi_smbios_register should have a return value

Errors may occur inside efi_smbios_register().

- Return a status code.
- Remove unused variables.
- Use constants where applicable.

Suggested-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: provide new doc/README.uefi

Provides information about

- usage of the bootefi command
- overview of UEFI
- interaction between U-Boot and EFI drivers

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Leif Lindholm <leif.lindholm@linaro.org>
Acked-by: Leif Lindholm <leif.lindholm@linaro.org>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: delete doc/README.efi

Delete README.efi. It is replaced by a further patch.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: Initial EFI_DEVICE_PATH_UTILITIES_PROTOCOL

Not complete, but enough for Shell.efi and SCT.efi.  We'll implement the
rest as needed or once we have SCT running properly so there is a way to
validate the interface against the conformance test suite.

Initial skeleton written by Leif, and then implementation by Rob.

Signed-off-by: Leif Lindholm <leif.lindholm@linaro.org>
[Fill initial skeleton]
Signed-off-by: Rob Clark <robdclark@gmail.com>
[Rebase on v2018.03-rc1]
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: remove deprecated ConsoleControlProtocol

The console control protocol is not defined in the UEFI standard.

It exists in EDK2's EdkCompatiblityPkg package. But this package
is deprecated according to
https://github.com/tianocore/tianocore.github.io/wiki/Differences-between-EDK-and-EDK-II

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

efi_loader: parameter checks for LoadImage

Add parameter checks in efi_load_image().
Check memory allocation is successful in efi_load_image().

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Signed-off-by: Alexander Graf <agraf@suse.de>

Prepare v2018.05-rc1

Signed-off-by: Tom Rini <trini@konsulko.com>

Merge git://git.denx.de/u-boot-dm

image.h: add forward declaration of struct fdt_region

This header needs to know 'fdt_region' is a struct for the
fit_region_make_list() prototype.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

fdt_region: remove unneeded fdt_internal.h inclusion

fdt_region.c does not depend on anything in libfdt_internal.h

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

libfdt: move FDT_RAMDISK_OVERHEAD to image-fdt.c

This macro is locally referenced in common/image-fdt.c

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

core: ofnode: Fix translation for #size-cells == 0

Commit 286ede6 ("drivers: core: Add translation in live tree case") made
dev_get_addr always use proper bus translations for addresses read from
the device tree. But this leads to problems with certain busses, e.g.
I2C busses, which run into an error during translation, and hence stop
working.

It turns out that of_translate_address() and fdt_translate_address()
stop the address translation with an error when they're asked to
translate addresses for busses where #size-cells == 0 (comment from
drivers/core/of_addr.c):

 * Note: We consider that crossing any level with #size-cells == 0 to mean
 * that translation is impossible (that is we are not dealing with a value
 * that can be mapped to a cpu physical address). This is not really specified
 * that way, but this is traditionally the way IBM at least do things

To fix this case, we check in both the live-tree and non-live tree-case,
whether the bus of the device whose address is about to be translated
has size-cell size zero. If this is the case, we just read the address
as a plain integer and return it, and only apply bus translations if the
size-cell size if greater than zero.

Signed-off-by: Mario Six <mario.six@gdsys.cc>
Signed-off-by: Martin Fuzzey <mfuzzey@parkeon.com>
Reported-by: Martin Fuzzey <mfuzzey@parkeon.com>
Fixes: 286ede6 ("drivers: core: Add translation in live tree case")
Reviewed-by: Simon Glass <sjg@chromium.org>

dm: core: make fixed-clock dt scan live dt compatible

dm_scan_fdt_node can't work when live dt is active,
we should use dm_scan_fdt_live instead.

Signed-off-by: Andy Yan <andy.yan@rock-chips.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

fs: cbfs: fix locating the cbfs header

The value at the end of the rom is not a pointer, it is an offset
relative to the end of rom.

Signed-off-by: Andre Heider <a.heider@gmail.com>

cmd: cbfs: fix reading the end_of_rom pointer for 64bit archs

The cast breaks the pointer on 64bit archs, so lets get rid of it.

Signed-off-by: Andre Heider <a.heider@gmail.com>
Reviewed-by: Alexander Graf <agraf@suse.de>

pinctrl-uclass: convert to use live dt

Use live dt interface for pinctrl_select_state_full()

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

core: add uclass_get_device_by_phandle_id() api

Add api for who can not get phandle from a device property.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

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

Merge git://git.denx.de/u-boot-x86

Merge git://git.denx.de/u-boot-riscv

arm64: a37xx: defconfigs: enable PCI_CMD and E1000 driver

Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: dts: enable pcie port

This patch enabled PCIe port on both devel-board
and espressobin board.

Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Wilson Ding <dingwei@marvell.com>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: defconfigs: enable aardvark pcie driver

Signed-off-by: Wilson Ding <dingwei@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: pci: add support for aardvark pcie driver

This patch introduced the Aardvark PCIe driver based
driver model.
The PCIe driver is supposed to work in Root Complex
mode. It only supports X1 lane width.

Signed-off-by: Wilson Ding <dingwei@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38725
Reviewed-by: Victor Gu <xigu@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
Tested-by: Hua Jing <jinghua@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: populate pcie memory region

This patch added a new region of 32MiB AT 0xe800.0000
to Armada37x0's memory map. This region is supposed to
be mapped in MMU in order to enable the access to the
PCI I/O or MEM resources.

Signed-off-by: Wilson Ding <dingwei@marvell.com>
Reviewed-on: http://vgitil04.il.marvell.com:8080/38724
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Victor Gu <xigu@marvell.com>
Signed-off-by: Ken Ma <make@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: remove old pinctrl implementation

Since the new pinctrl/gpio driver is used, so this patch removes
the old board specific pin control settings.

Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

Revert "arm64: a37xx: dts: Add pin control nodes to DT"

The commit "arm64: mvebu: Add pinctrl nodes for Armada 3700" has
added new pinctrl nodes.
This reverts commit f7cab0f95b05ec6a66fe4796b9ad44406d0cc864.

Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

doc: a37xx: Introduce pinctrl device tree binding

Reviewed-on: http://vgitil04.il.marvell.com:8080/43289
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Kostya Porotchkin <kostap@marvell.com>
Reviewed-by: Igal Liberman <igall@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: dts: Correct mpp definitions

This patch corrects below mpp definitions for armada 3720 DB board
and ESPRESSOBin board:
- "smi" pins group is added and "smi" function is set for eth0;
- Now pcie pins are used as gpio to implement PCIe function in
  hardware, so "pcie" group function is changed to "gpio".

Reviewed-on: http://vgitil04.il.marvell.com:8080/43287
Reviewed-by: Hua Jing <jinghua@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: pinctrl: Correct mpp definitions

This patch corrects below mpp definitions:
 - The sdio_sb group is composed of 6 pins and not 5;
 - The rgmii group contains pins mpp2[17:6] and not mpp2[19:6];
 - Pin of group "pmic0" is mpp1[6] but not mpp1[16];
 - Pin of group "pmic1" is mpp1[7] but not mpp1[17];
 - A new group "smi" is added in A0 with 2 pins - mpp2[19:18], its
   bitmask is bit4;
 - Group "pcie1" has 3 pins in A0 - mpp2[5:3], its bit mask is
   bit5 | bit9 | bit10 but not bit4;
 - Group "ptp" has 3 pins in A0 as Z1, but its bitmask is changed to
   bit11 | bit12 | bit13.

Reviewed-on: http://vgitil04.il.marvell.com:8080/43288
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: pinctrl: Fix gpio pin offset in register

For armada_37xx_update_reg(), the parameter offset should be pointer so
that it can be updated, otherwise offset will keep old value, and then
when offset is larger than or equal to 32 the mask calculated by
"BIT(offset)" will be 0 in gpio chip hook functions, it's an error,
this patch set offset parameter of armada_37xx_update_reg() as pointer.

Reviewed-on: http://vgitil04.il.marvell.com:8080/43287
Reviewed-by: Hua Jing <jinghua@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: pinctrl: Fix the pin 23 on south bridge

Pin 23 on South bridge does not belong to the rgmii group. It belongs to
a separate group which can have 3 functions.

Due to this the fix also have to update the way the functions are
managed. Until now each groups used NB_FUNCS(which was 2) functions. For
the mpp23, 3 functions are available but it is the only group which needs
it, so on the loop involving NB_FUNCS an extra test was added to handle
only the functions added.

The bug was visible when the gpio regulator used the gpio 23, the whole
rgmii group was setup to gpio which broke the Ethernet support on the
Armada 3720 DB board. Thanks to this patch, the UHS SD cards (which need
the vqmmc) _and_ the Ethernet work again.

Reviewed-on: http://vgitil04.il.marvell.com:8080/43284
Reviewed-by: Hua Jing <jinghua@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: dts: Fix the number of GPIO on south bridge

The number of pins in South Bridge is 30 and not 29. There is a fix for
the driver for the pinctrl, but a fix is also need at device tree level
for the GPIO.

Reviewed-on: http://vgitil04.il.marvell.com:8080/43286
Reviewed-by: Hua Jing <jinghua@marvell.com>
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: pinctrl: Fix number of pin in south bridge

On the south bridge we have pin from 0 to 29, so it gives 30 pins (and
not 29).

Reviewed-on: http://vgitil04.il.marvell.com:8080/43285
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Hua Jing <jinghua@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: dts: Add additional pinctrl definition

Add mmc pins, pcie pins and sdio pins definition and do these pins'
configuration for DB board and espressobin board;
Add uart2 pins configuration for DB board.

Reviewed-on: http://vgitil04.il.marvell.com:8080/40914
Reviewed-by: Wilson Ding <dingwei@marvell.com>
Tested-by: Wilson Ding <dingwei@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: dts: Add pinctrl configuration for ESPRESSOBin board

Reviewed-on: http://vgitil04.il.marvell.com:8080/40913
Reviewed-by: Wilson Ding <dingwei@marvell.com>
Tested-by: Wilson Ding <dingwei@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: defconfig: Enable PINCTRL and GPIO support for ESPRESSOBin board

This patch enable the PINCTRL and GPIO support, including the GPIO
command on the Armada 3720 espressobin board.

Reviewed-on: http://vgitil04.il.marvell.com:8080/40746
Tested-by: iSoC Platform CI <ykjenk@marvell.com>
Reviewed-by: Wilson Ding <dingwei@marvell.com>
Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: a37xx: dts: add gpio head file including

Cc: Simon Glass <sjg@chromium.org>
Cc: Stefan Roese <sr@denx.de>
Signed-off-by: Ken Ma <make@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

arm64: mvebu: a8k: Add distro boot support

The various load address values are taken from the a37xx configuration
and match the dowstream 'u-boot-2017.03-armada-17.10' release where
appropriate.

Signed-off-by: Mark Kettenis <kettenis@openbsd.org>
Signed-off-by: Stefan Roese <sr@denx.de>

kwbimage: Fix out of bounds access

The kwbimage format is reading beyond its header structure if it
misdetects a Xilinx Zynq image and tries to read it. Fix it by
sanity checking that the header we want to read fits inside our
file size.

Signed-off-by: Alexander Graf <agraf@suse.de>
Tested-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>

configs: clearfog: enable random random MAC address

This makes the network devices usable when booting a blank board over
UART, with no pre-configured MAC address stored in the environment area.

Signed-off-by: Baruch Siach <baruch@tkos.co.il>
Reviewed-by: Stefan Roese <sr@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>

dm: pcie: designware: add correct ATU handling

Currently, ATU (address translation unit) implementation doesn't
support translate addresses > 32 bits.

This patch allows to configure ATU correctly for different
memory accesses (memory, configuration and IO).
The same approach is used in Linux Kernel.

Signed-off-by: Igal Liberman <igall@marvell.com>
Signed-off-by: Stefan Roese <sr@denx.de>

x86: zImage: Pass working device tree data to the kernel

On x86 platforms, U-Boot does not pass Device Tree data to the kernel.
This prevents the kernel from using FDT loaded by U-Boot.

Read the working FDT address from the "fdtaddr" environment variable
and add a copy of the FDT data to the kernel setup_data list.

Signed-off-by: Ivan Gorinov <ivan.gorinov@intel.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
[bmeng: add #include <linux/libfdt.h> to zimage.c to fix build error]
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>

x86: mmc: Fix mapping of BAR memory

Use dm_pci_map_bar function for BAR mapping. This has the advantage
of clearing BAR flags and and only accepting mapped memory.

Signed-off-by: Bernhard Messerklinger <bernhard.messerklinger@br-automation.com>
Reviewed-by: Hannes Schmelzer <hannes.schmelzer@br-automation.com>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>

dm: pci: Avoid setting a PCI region that has 0 size

It makes no sense to set a PCI region that has 0 size.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

dm: pci: Check board information pointer in decode_regions()

PCI enumeration may happen very early on an x86 board. The board
information pointer should have been checked in decode_regions()
as its space may not be allocated yet.

With this commit, Intel Galileo board boots again.

Fixes: 664758c ("pci: Fix decode regions for memory banks")
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>

efi_stub: Fix GDT_NOTSYS check

The get_codeseg32() wants to know if a passed in descriptor has
flag GDT_NOTSYS set (desc & GDT_NOTSYS), not whether desc and
GDT_NOTSYS are not != 0 (desk && GDT_NOTSYS).

This is an obvious typo. Fix it up.

Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>

riscv: ae250: Support DT provided by the board at runtime

Enable CONFIG_OF_BOAD to support delivery dtb to u-boot
at run time instead of embedded.

There are two methods to delivery dtb.
 1 Pass from loader:
   When u-boot boot from RAM, gdb or loader can pass dtb
   via a2 to u-boot dynamically. Of course gdb or loader
   shall be in charge of dtb delivery.

 2 Configure CONFIG_SYS_FDT_BASE:
   It can be configured as RAM or ROM base statically,
   no mater u-boot boot from RAM or ROM.
   If it was configured as ROM base, dtb can be burned
   into ROM(spi flash) by spi driver.

Meanwhile remove CONFIG_SKIP_LOWLEVEL_INIT which is
useless in nx25-ae250 configuration.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

configs: Drop CONFIG_MMC_NDS32

Remove CONFIG_MMC_NDS32 from the three config
(adp-ae3xx_defconfig, adp-ag101p_defconfig, nx25-ae250_defconfig).

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

mmc: ftsdc010: Merge nds32_mmc to ftsdc010

nsd32_mmc was created to support ftsdc010 dm.
It is not necessary to separate both, so merge it
to ftsdc010.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

mmc: ftsdc010: Migrate CONFIG_FTSDC010_SDIO to Kconfig

Convert CONFIG_FTSDC010_SDIO to Kconfig.
So CONFIG_FTSDC010_SDIO can also be
removed from config_whitelist.txt.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

Drop CONFIG_FTSDC010_NUMBER

CONFIG_FTSDC010_NUMBER was not used anymore,
can be removed now.

So CONFIG_FTSDC010_NUMBER
can also be removed from config_whitelist.txt.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

Drop CONFIG_FTSDC010_BASE

After drop non-dm code of ftsdc010, the sd register
base definition can be droppped now.

So CONFIG_FTSDC010_BASE and CONFIG_FTSDC010_BASE_LIST both
can be removed from config_whitelist.txt

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

board: Drop ftsdc010 non-dm code

Remove board_mmc_init() in adp-ag101p, adp-ae3xx
and nx25-ae250 boards.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

mmc: ftsdc010: Drop non-dm code

Only three defconfig(adp-ag101p_defconfig,
adp-ae3xx_defconfig, nx25-ae250_defconfig)
set CONFIG_FTSDC010=y. And they all also
enable CONFIG_DM_MMC. So the non-dm code
of ftsdc010 can be dropped now.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>
Cc: Greentime Hu <green.hu@gmail.com>

nds32: dts: AG101P support sd High-Speed mode

Enable High-Speed mode with cap-sd-highspeed in dts

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>

riscv: dts: AE250 support sd High-Speed mode

Enable High-Speed mode with cap-sd-highspeed in dts.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>

mmc: ftsdc010: Support High-Speed mode

ftsdc010 dm driver has been disable High-Speed mode
as default to work around Andes AE3XX platform's problem,
because of it does not support High-Speed mode in
commit id 73cd56b2df213c629191139e5c6705e069b6214f.

But other platforms or SoCs maybe support this function.
So High-Speed mode can be enabled from dts with
cap-mmc-highspeed or cap-sd-highspeed property.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>

doc: ae250: Describe riscv-linux booting via u-boot

Simply record riscv-linux booting steps and messages
from bbl via u-boot on QEMU in README.ae250.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>

tools: mkimage: Support RISC-V arch

Add riscv uimage arch to support riscv-linux booting.
It can Convert riscv-linux to image which can be
booted by bootm command.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>

riscv: bootm: Remove ATAGS

ATAGS is not supported and will be replaced
by DT in riscv-linux. So can be removed now.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>

riscv: bootm: Support to boot riscv-linux

riscv-linux should use BBL (Berkeley bootloader) for
loading the Linux kernel.

U-Boot can play as FSBL(first stage bootloader)
to boot BBL and riscv-linux.

In BBL's init_first_hart(), it will pass dtb with a1.
So implement bootm to pass arguments to BBL correctly.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>

riscv: checkpatch: Fix static const char * array declarations

It is reported by checkpatch.pl
WARNING: static const char * array
should probably be static const char * const

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>

riscv: checkpatch: Fix missing a blank line after declarations

It is reported by checkpatch.pl
WARNING: Missing a blank line after declarations.

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>

riscv: checkpatch: Fix alignment should match open parenthesis

It is reported by checkpatch.pl.
CHECK: Alignment should match open parenthesis

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>

riscv: checkpatch: Fix use of volatile

It is reported by checkpatch.pl
WARNING: Use of volatile is usually wrong: see
Documentation/process/volatile-considered-harmful.rst

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>

riscv: checkpatch: Fix Macro argument reuse

It is CHECK reported by checkpatch.pl
CHECK: Macro argument reuse 'PTE' - possible side-effects?

Signed-off-by: Rick Chen <rick@andestech.com>
Signed-off-by: Rick Chen <rickchen36@gmail.com>

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

rockchip: sdhci: support 8bit bus-width

The Rockchip-specific SDHCI wrapper does not process the 'bus-width'
property in the SDHCI node. Consequently, the bus is always kept in
4bit mode, even if 8bit wide operation is available, supported and
requested in the DTS.

This change adds processing of the 'bus-width' property and sets the
host capability flag for an 8bit wide bus, if set to 8. As the logic
in sdhci.c does not support clearing the 4bit capability, we assume
that 4bit operation is always supported.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

rockchip: mmc: update MAINTAINERS

The Rockchip-specific wrappers to the DW-MMC and the SDHCI driver
were not covered as part of what's maintained by the architecture
maintainers.  Add them here.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

rockchip: defconfig: puma-rk3399: enable DMA for SDHCI controller

For the RK3399-Q7, we have a fast eMMC connected in an 8 bit wide
configuration to the SDHCI controller (sdhci@fe330000).  Enable DMA
within the SDHCI driver to get the best performance out of it.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

rockchip: rk3399: set fdtfile

Signed-off-by: Mark Kettenis <kettenis@openbsd.org>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Tested-by: Klaus Goger <klaus.goger@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

rockchip: clk: rk3188: update dpll settings to make EMAC work

The patch set dpll settings for 300MHz to values used by binary
blob[1]. With new values dpll still generate 300MHz clock, but
EMAC work. Probably with new values dpll generate more stable clock.

dpll on rk3188 provide clocks to DDR and EMAC. With current
dpll settings EMAC doesn't work on radxa rock. EMAC sends packets
to network, but it doesn't receive anything. ifconfig shows a lot
of framing errors.

[1] https://github.com/linux-rockchip/u-boot-rockchip/blob/u-boot-rk3288/
    tools/rk_tools/3188_LPDDR2_300MHz_DDR3_300MHz_20130830.bin

Signed-off-by: Alexander Kochetkov <al.kochet@gmail.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

video: rk3288_mipi: Combine NULL check into IS_ERR_OR_NULL()

Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

video: rk3288_mipi: check in rk_mipi_ofdata_to_platdata

The error checking should also catch the case that no range has beeen
defined.
syscon_get_first_range() returns NULL if no range is defined.
cf. rk3399_mipi.c

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

video: rk3399_mipi: Combine NULL check into IS_ERR_OR_NULL()

Signed-off-by: Punit Agrawal <punit.agrawal@arm.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

video: rk3399_mipi: correct error checking

Pointers are never negative. Use macro IS_ERR() for checking.
cf. rk3288_mipi.c

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

usb: rockchip: remove duplicate assignement.

Assigning f_rkusb->reboot_flag twice doesn't make sense.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

driver: ram: rockchip: rk3399: missing counter increment

If we want to check the duration we need to increment the counter.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

rockchip: pinctrl: rk3188: Move the iomux definitions into pinctrl-driver

Clean the iomux definitions at grf_rk3188.h, and move them into
pinctrl-driver for resolving the compiling error of redefinition.

Signed-off-by: Alexander Kochetkov <al.kochet@gmail.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

rockchip: pinctrl: rk3036: Move the iomux definitions into pinctrl-driver

Clean the iomux definitions at grf_rk3036.h, and move them into
pinctrl-driver for resolving the compiling error of redefinition.

Signed-off-by: Alexander Kochetkov <al.kochet@gmail.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

rockchip: pinctrl: rk3399: add support for I2C[123467]

This adds support for the (to date unsupported) I2C controllers 1~4
and 6~7 (i.e. now all controllers except I2C5, which is not accessible
on the RK3399-Q7, are supported by pinctrl).

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Tested-by: Klaus Goger <klaus.goger@theobroma-systems.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

rockchip: pinctrl: rk3399: fix GPIO2B1 and GPIO2B2 shift value

The shift values for GPIO2B1 and GPIO2B2 had in fact referred to
GPIO2B0 and GPIO2B1, respectively.  This substitutes the correct
values.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>

Merge git://git.denx.de/u-boot-ubi

Merge git://git.denx.de/u-boot-spi

Merge git://git.denx.de/u-boot-i2c

Merge tag 'xilinx-for-v2018.05' of git://git.denx.de/u-boot-microblaze

Xilinx changes for v2018.05

- Fix mkimage recognition
- Update all my fragments

ZynqMP:
- Use clk driver
- Support loading elfs in el1
- Various DTS and defconfig changes
- Enable newer pmufw versions
- Support more clocks
- Remove ep108
- Secure image support
- Fix memtest setup

Zynq:
- Enabling watchdog driver
- Support more clocks
- defconfig changes

fpga:
- Simplify error path

net:
- GMII case update

Convert CONFIG_UBI_SILENCE_MSG to Kconfig

Signed-off-by: Petr Vorel <petr.vorel@gmail.com>

Convert CONFIG_UBIFS_SILENCE_MSG to Kconfig

Introduce another difference from upstream (kernel) source in
fs/ubifs/super.c: adding preprocessor condition as y variable in
mount_ubifs() depends on CONFIG_UBIFS_SILENCE_MSG:
fs/ubifs/super.c:1337:15: error: variable ?y? set but not used [-Werror=unused-but-set-variable]
  long long x, y;

Not setting CONFIG_UBIFS_SILENCE_MSG in am335x_igep003x_defconfig and
igep0032_defconfig. Although it was defined in their config headers, it
depends on CMD_UBIFS which is not set for them.

Signed-off-by: Petr Vorel <petr.vorel@gmail.com>
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Heiko Schocher <hs@denx.de>

ubifs: Reintroduce using CONFIG_UBIFS_SILENCE_MSG

Use of CONFIG_UBIFS_SILENCE_MSG was added in
147162dac6 ("ubi: ubifs: Turn off verbose prints")

Then it was removed in
ff94bc40af ("mtd, ubi, ubifs: resync with Linux-3.14")

Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Heiko Schocher <hs@denx.de>
Signed-off-by: Petr Vorel <petr.vorel@gmail.com>