Add driver model support to this driver, while retaining support for the
legacy system. Driver model GPIO support is enabled with CONFIG_DM_GPIO
as usual.

Since gpio_is_valid() no longer exists, we can use the -EINVAL error
returned from gpio_request().
Signed-off-by: NSimon Glass <sjg@chromium.org>
Reviewed-by: NTom Rini <trini@ti.com>

Add a separate internal helper function to get a GPIO value, so that we
will be able to call it with the driver model version and avoid code
duplication.

Also move gpio_get_bank() and check_gpio() down below the helper functions
as these won't be needed with driver model.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Reviewed-by: NTom Rini <trini@ti.com>

Select serial0 as the console.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Acked-by: NTom Rini <trini@ti.com>

Try to use this option to select the correct uart for the console. This
mimics the behaviour of drivers/serial.c.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Reviewed-by: NTom Rini <trini@ti.com>

The probe logic sets up the pointer to the platform data in the device
tree decode method. It should be done in the probe() method, and anyway
the device tree decode method can't be used when CONFIG_OF_CONTROL is
not enabled.

Fix these two problems.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Reviewed-by: NTom Rini <trini@ti.com>

It is inconvenient to have to use casts when specifying platform data. Also
it is not strictly correct, since we should use map_sysmem() to convert an
address to a pointer.

Adjust the platform data to use an address.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Reviewed-by: NTom Rini <trini@ti.com>

There is a bug in the logic which checks for an available character. This
can cause invalid characters to be received - this was noticed on
beaglebone. Fix it.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Reviewed-by: NTom Rini <trini@ti.com>

We have moved the busy-wait loop out of drivers and into the uclass. This
means that we must reset the watchdog when busy-waiting.

Note: some drivers may still have a busy-wait even with driver model, as
a transition mechanism. Driver model will tolerate this, and is can be
cleaned up when all users of the driver use driver model. An example is
ns16550.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Reviewed-by: NTom Rini <trini@ti.com>

At present banks must be named and it is not possible to refer to GPIOs by
number in driver model. Some boards use numbering - e.g. OMAP. It is fairly
easy to support by detecting the absense of a bank name (which starts with
a letter).

Add support for numbered GPIOs in addition to the existing bank support.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Reviewed-by: NTom Rini <trini@ti.com>

Since the environment "VENDOR" is set in tcsh, it must be cleared in our
makefile.  Otherwise, boards without CONFIG_SYS_VENDOR fail to build:

  > make CROSS_COMPILE=arm-linux-gnueabi- wandboard_quad_defconfig all
    [ snip ]
    AR      arch/arm/lib/lib.a
    CC      arch/arm/lib/eabi_compat.o
  scripts/Makefile.build:55: /home/foo/u-boot/board/unknown/wandboard/ \
  Makefile: No such file or directory
  make[2]: *** No rule to make target `/home/foo/u-boot/board/unknown/ \
  wandboard/Makefile'.  Stop.
  make[1]: *** [board/unknown/wandboard] Error 2
  make: *** [__build_one_by_one] Error 2
Signed-off-by: NMasahiro Yamada <yamada.m@jp.panasonic.com>
Reported-by: NTom Everett <tom@khubla.com>
Reported-by: NJeroen Hofstee <jeroen@myspectrum.nl>

Without the private libgcc, we need a full multilib toolchain with
different libgcc or multiple toolchains to build all BE/LE and
hard-float/soft-float variants of MIPS boards.  That is not feasible.

This commit allows us to build all the MIPS boards with a single
kernel.org toolchain:

https://www.kernel.org/pub/tools/crosstool/files/bin/x86_64/4.9.0/
x86_64-gcc-4.9.0-nolibc_mips-linux.tar.xz

This change sounds reasonable for most users.  If necessary,
you can disable this option via "make menuconfig" or friends.
Signed-off-by: NMasahiro Yamada <yamada.m@jp.panasonic.com>
Acked-by: NDaniel Schwierzeck <daniel.schwierzeck@gmail.com>

When spl/.config is updated by "make spl/menuconfig" or friends,
spl/include/config/auto.conf, spl/include/generated/autoconf.h
and some other files must be updated by "make silentoldconfig".

There is no hook for SPL in the top Makefile, so this commit
touches .config when spl/.config is updated to invoke silentoldconfig.
Likewise for TPL.
Signed-off-by: NMasahiro Yamada <yamada.m@jp.panasonic.com>

cmd_u-boot-spl includes $(PLATFORM_LIBS) which changes
when CONFIG_USE_PRIVATE_GCC is updated.  The u-boot-spl image
should be re-linked if any prerequisite is newer than it
or the command line has changed.
$(call, if_changed,...) should be used instead of $(call cmd,...).
Signed-off-by: NMasahiro Yamada <yamada.m@jp.panasonic.com>

The private libgcc is supported only on ARM, MIPS, PowerPC, SH, x86.
Those architectures should "select" HAVE_PRIVATE_LIBGCC and
CONFIG_USE_PRIVATE_LIBGCC should depend on it.

Currently, this option is enabled on Tegra boards and x86 architecture.
Move the definition from header files to Kconfig.
Signed-off-by: NMasahiro Yamada <yamada.m@jp.panasonic.com>
Tested-by: NSimon Glass <sjg@chromium.org>
Acked-by: NSimon Glass <sjg@chromium.org>
Cc: Stephen Warren <swarren@nvidia.com>
Cc: Tom Warren <twarren@nvidia.com>

Now CONFIG_USE_PRIVATE_LIBGCC is only used as a boolean macro.
Remove CONFIG_USE_PRIVATE_LIBGCC=path/to/libgcc syntax.
Signed-off-by: NMasahiro Yamada <yamada.m@jp.panasonic.com>
Tested-by: NSimon Glass <sjg@chromium.org>
Acked-by: NSimon Glass <sjg@chromium.org>

The motivation of this commit is to change CONFIG_USE_PRIVATE_LIBGCC
to a boolean macro so we can move it to Kconfig.

In the current implementation, there are two forms of syntax
for this macro:

  - CONFIG_USE_PRIVATE_LIBGCC=y
  - CONFIG_USE_PRIVATE_LIBGCC=path/to/private/libgcc

The latter is only used by x86 architecture.
With a little bit refactoring, it can be converted to the former.
Signed-off-by: NMasahiro Yamada <yamada.m@jp.panasonic.com>
Tested-by: NSimon Glass <sjg@chromium.org>
Acked-by: NSimon Glass <sjg@chromium.org>

CONFIG_SYS_HZ is always defined as 1000 in config_fallbacks.h
(but some boards still have redundant definitions).

This commit moves the definition and the document in README to
Kconfig.  Since lib/Kconfig can assure that CONFIG_SYS_HZ is 1000,
the sanity check in lib/time.c should be removed.
Signed-off-by: NMasahiro Yamada <yamada.m@jp.panasonic.com>
Reviewed-by: NMarek Vasut <marex@denx.de>

Add target to build it automatically upon "make" / MAKEALL. This can/should
be set by board / cpu specific headers if a special U-Boot image is
required for this SoC / board.

E.g. used by Marvell Armada XP to automatically build the u-boot.kwb
target.
Signed-off-by: NStefan Roese <sr@denx.de>
Cc: Masahiro Yamada <yamada.m@jp.panasonic.com>

This patch integrates the Barebox version of this kwbimage.c file into
U-Boot. As this version supports the image version 1 type for the
Armada XP / 370 SoCs.

It was easier to integrate the existing and known to be working Barebox
source than to update the current U-Boot version to support this
v1 image header format. Now all Marvell MVEBU SoCs are supported:

Image type 0: Kirkwood & Dove
Image type 1: Armada 370 & Armada XP

Please note that the current v1 support has this restuction (same as
has Barebox version):

Not implemented: support for the register headers and secure headers
in v1 images

Tested on Marvell DB-78460-BP eval board.
Signed-off-by: NStefan Roese <sr@denx.de>
Tested-by: NLuka Perkov <luka@openwrt.org>

Signed-off-by: NStefan Roese <sr@denx.de>
Tested-by: NLuka Perkov <luka@openwrt.org>

The barebox version of the kwboot tool has evolved a bit. To support
Armada XP and Dove. Additionally a few minor fixes have been applied.
So lets sync with the latest barebox version.

Please note that the main difference between both versions now is, that
the U-Boot version still supports the -p option, to dynamically patch
an image for UART boot mode. I didn't test it now though.
Signed-off-by: NStefan Roese <sr@denx.de>
Tested-by: NLuka Perkov <luka@openwrt.org>

All those functions removed with this patch are not accessed at all. So lets
remove them.
Signed-off-by: NStefan Roese <sr@denx.de>

The maxBCM board is equipped with the Marvell Armada-XP MV78460 SoC. It
integrates an SPI NOR flash and an Marvell 88E6185 switch.
Signed-off-by: NStefan Roese <sr@denx.de>

This patch adds basic support for the Marvell DB-MV784MP-GP evaulation
board. This is the first board that uses the recently created
Armada XP 78460 SoC support.
Signed-off-by: NStefan Roese <sr@denx.de>
Tested-by: NLuka Perkov <luka@openwrt.org>

This basic support for the Marvell Armada XP is base on the existing kirkwood
support. Which has been generatized by moving some common files into
common marvell locations.

This is in preparation for the upcoming Armada XP MV78460 support.
Signed-off-by: NStefan Roese <sr@denx.de>
Tested-by: NLuka Perkov <luka@openwrt.org>

To support the Armada XP SoC, we just need to include the correct header.
Signed-off-by: NStefan Roese <sr@denx.de>
Acked-by: NHeiko Schocher <hs@denx.de>
Tested-by: NLuka Perkov <luka@openwrt.org>

The Marvell MV78460 eval board DB-78460-BP seems to need a longer
PHY autonegotiation timeout than the "standard" 4 seconds. So lets
make this timeout configurable. If not defined in the board config
header the original 4000ms is used.
Signed-off-by: NStefan Roese <sr@denx.de>
Cc: Joe Hershberger <joe.hershberger@gmail.com>

This patch adds support for the NETA ethernet controller which is integrated
in the Marvell Armada XP SoC's. This port is based on the Linux driver which
has been stripped of the in U-Boot unused portions.

Tested on the Marvell MV78460 eval board db-78460-bp.
Signed-off-by: NStefan Roese <sr@denx.de>
Cc: Joe Hershberger <joe.hershberger@gmail.com>
Tested-by: NLuka Perkov <luka@openwrt.org>

Additionally the SDRAM address decoding register address is not hard coded
in the C code any more. A define is introduced for this base address.

This makes is possible to use those gpio functions from other MVEBU SoC's
as well.
Signed-off-by: NStefan Roese <sr@denx.de>
Tested-by: NLuka Perkov <luka@openwrt.org>
Acked-by: NPrafulla Wadaskar <prafulla@marvell.com>

This makes is possible to use this SPI driver from other MVEBU SoC's as well.
As the upcoming Armada XP support will do.
Signed-off-by: NStefan Roese <sr@denx.de>
Reviewed-by: NJagannadha Sutradharudu Teki <jaganna@xilinx.com>
Tested-by: NLuka Perkov <luka@openwrt.org>
Acked-by: NPrafulla Wadaskar <prafulla@marvell.com>

This makes is possible to use those gpio functions from other MVEBU SoC's as well.
Signed-off-by: NStefan Roese <sr@denx.de>
Tested-by: NLuka Perkov <luka@openwrt.org>
Acked-by: NPrafulla Wadaskar <prafulla@marvell.com>

Compile the pin multiplexing only on Kirkwood platforms. As the
Armada XP doesn't need it.
Signed-off-by: NStefan Roese <sr@denx.de>
Reviewed-by: NJagannadha Sutradharudu Teki <jaganna@xilinx.com>
Tested-by: NLuka Perkov <luka@openwrt.org>
Acked-by: NPrafulla Wadaskar <prafulla@marvell.com>

These mbus functions are ported from Barebox. The Barebox version is
ported from Linux. These functions will be first used by the upcoming
Armada XP support. Later other Marvell SoC's will be adopted to use
these functions as well (Kirkwood, Orion).
Signed-off-by: NStefan Roese <sr@denx.de>
Tested-by: NLuka Perkov <luka@openwrt.org>

This patch does the following:
- Rename defines and registers to not use kirkwood
- Remove unused defines
- Use clrsetbits() accessor functions
- Coding style cleanup
- Clear 25MHZ bit in timer controller register init for Armada XP

There is no functional change for kirkwood. At least not intentionally.

This will be used by the upcoming Armada XP support.
Signed-off-by: NStefan Roese <sr@denx.de>
Tested-by: NLuka Perkov <luka@openwrt.org>
Acked-by: NPrafulla Wadaskar <prafulla@marvell.com>

This move makes it possible to use this kirkwood SPI driver from other
MVEBU platforms as well. This will be used by the upcoming Armada XP
support.
Signed-off-by: NStefan Roese <sr@denx.de>
Reviewed-by: NJagannadha Sutradharudu Teki <jaganna@xilinx.com>
Tested-by: NLuka Perkov <luka@openwrt.org>
Acked-by: NPrafulla Wadaskar <prafulla@marvell.com>

This move makes is possible to use this header not only from kirkwood
platforms but from all Marvell mvebu platforms.
Signed-off-by: NStefan Roese <sr@denx.de>
Tested-by: NLuka Perkov <luka@openwrt.org>
Acked-by: NPrafulla Wadaskar <prafulla@marvell.com>

By moving some kirkwood files into a Marvell common directory, those files
can be used by other Marvell platforms as well. The name mvebu is taken
from the Linux kernel source tree. It has been chosen there to represent
the SoC's from the Marvell EBU (Engineering Business Unit). Those SoC's
currently are:

Armada 370/375/XP, Dove, mv78xx0, Kirkwood, Orion5x

This will be used by the upcoming Armada XP (MV78460) platform support.
Signed-off-by: NStefan Roese <sr@denx.de>
Tested-by: NLuka Perkov <luka@openwrt.org>
Acked-by: NPrafulla Wadaskar <prafulla@marvell.com>