Add remove() methods for EDP and VOP so that U-Boot can shut down the
video on exit. This avoids leaving DMA running while booting Linux which
can cause problems if Linux uses the frame buffer for something else.

It also makes it clear what is needed to shut down video.

While we are here, make rkvop_enable() static.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Acked-by: NAnatolij Gustschin <agust@denx.de>
Squashed in 'rockchip: video: fix taking the VOP device out of standby':
Signed-off-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

Add a debug() statement so we can see when something goes wrong with the
regulator.
Signed-off-by: NSimon Glass <sjg@chromium.org>

This typo doesn't actually cause any problems, but is wrong. Fix it.
Signed-off-by: NSimon Glass <sjg@chromium.org>

At present if the
Signed-off-by: NSimon Glass <sjg@chromium.org>
Fixes: 874ee59a (rockchip: pwm: implement pwm_set_invert())

Most Chromebooks support chain-loading U-Boot but instructions are
somewhat scattered. Add a README to hold this information within the
U-Boot tree. Also add the standard developer keys to simplify the
instructions, since they are small.

For now this only supports nyan-big.
Signed-off-by: NSimon Glass <sjg@chromium.org>

Add a sample .its file for booting U-Boot on a nyan-big Chromebook.
Signed-off-by: NSimon Glass <sjg@chromium.org>

The clock fix-up for tegra is still present in the code. It causes a
divide-by-zero bug after relocation when chain-loading U-Boot from
coreboot. Fix this by adding a check.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Fixes: 74686766 (ARM: tegra: fix clock_get_periph_rate() for UART clocks)

At present CP15 init is disabled on tegra. Use the correct option so that
this init is performed on boot. This enables the instruction cache, for
example, which is critical to the machine running at full speed.
Signed-off-by: NSimon Glass <sjg@chromium.org>

If U-Boot is the secondary boot loader, or has been run from itself, the
SOR may already be powered up. Powering it up again causes a hang, so
detect this situation and skip it.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Acked-by: NAnatolij Gustschin <agust@denx.de>

Enable this so we can roughly measure CPU performance. Also enable the
cache command to allow for timing.
Signed-off-by: NSimon Glass <sjg@chromium.org>

For devices that need a delay between SPI transactions we seem to need an
additional delay before the first one if the CPU is running at full speed.
Add this, under control of the existing setting. At present it will only
be enabled with the Chrome OS EC.
Signed-off-by: NSimon Glass <sjg@chromium.org>

At present the interrupt does not work and the SPI bus runs much less
quickly than it should. Add settings to fix this.
Signed-off-by: NSimon Glass <sjg@chromium.org>

At present early clock init happens in SPL. If SPL did not run (because
for example U-Boot is chain-loaded from another boot loader) then the
clocks are not set as U-Boot expects.

Add a function to detect this and call the early clock init in U-Boot
proper.
Signed-off-by: NSimon Glass <sjg@chromium.org>

This file does not report a few possible errors and one message is missing
a newline. Fix these.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Reviewed-by: NJaehoon Chung <jh80.chung@samsung.com>

If CONFIG_ARMV7_LPAE is not defined we should make sure that the feature
is disabled. This can happen if U-Boot is chain-loaded from another boot
loader which does enable LPAE.
Signed-off-by: NSimon Glass <sjg@chromium.org>

At present if CONFIG_ARMV7_LPAE is defined then mmu_setup() will use
instructions which are invalid on ARMv4T. This happens on Tegra since it
has an ARMv4T boot CPU. Add a check for the architecture version to allow
the code to be built. It will not actually be executed by the boot CPU,
but needs to compile.
Signed-off-by: NSimon Glass <sjg@chromium.org>

This appears to be a typo. Fix it.
Signed-off-by: NSimon Glass <sjg@chromium.org>

This option allows skipping the call to lowlevel() while still performing
CP15 init. Support this on ARM720T so it can be used with Tegra.
Signed-off-by: NSimon Glass <sjg@chromium.org>

The existing test (patman --test) only covers basic checkpatch output.
We have had some problems with unicode processing and could use test
coverage for the various tags patman supports.

Add a new functional test which runs most of the patman flow on a few
test commits and checks that the results are correct.

See the documentation in the test for a description of what it does.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Tested-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

This is not a good variable name in Python because 'list' is a type. It
shows up highlighted in some editors. Rename it.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Tested-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

Allow the add_maintainers parameter to be a list of maintainers, thus
allowing us to simulate calling the script in tests without actually
needing it to work.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Tested-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

Add some unicode to the test patches to make sure that patman does the
right thing.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Tested-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

There is no need for this function to return the same object that was
passed in. Drop the return value.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Tested-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

Unicode characters may appear in input patches so we should not warn about
them. Drop this warning.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Tested-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

This is not a good variable name in Python because 'str' is a type. It
shows up highlighted in some editors. Rename it.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Tested-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

The communication filter reads data in blocks and converts each block to
unicode (if necessary) one at a time. In the unlikely event that a unicode
character in the input spans a block this will not work. We get an error
like:

UnicodeDecodeError: 'utf8' codec can't decode bytes in position 1022-1023:
   unexpected end of data

There is no need to change the input to unicode, so the easiest fix is to
drop this feature.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Tested-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

Don't mess with the email address when outputting them. Just make sure
they are encoded with utf-8.
Signed-off-by: NSimon Glass <sjg@chromium.org>
Tested-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

This change encodes the CC list to UTF-8 to avoid failures on
maintainer-addresses that include non-ASCII characters (observed on
Debian 7.11 with Python 2.7.3).

Without this, I get the following failure:
  Traceback (most recent call last):
    File "tools/patman/patman", line 159, in <module>
      options.add_maintainers)
    File "[snip]/u-boot/tools/patman/series.py", line 234, in MakeCcFile
      print(commit.patch, ', '.join(set(list)), file=fd)
  UnicodeEncodeError: 'ascii' codec can't encode character u'\xfc' in position 81: ordinal not in range(128)
from Heiko's email address:
  [..., u'"Heiko St\xfcbner" <heiko@sntech.de>', ...]

While with this change added this encodes to:
  "=?UTF-8?q?Heiko=20St=C3=BCbner?= <heiko@sntech.de>"
Signed-off-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: NSimon Glass <sjg@chromium.org>
Signed-off-by: NSimon Glass <sjg@chromium.org>
Tested-by: NSimon Glass <sjg@chromium.org>
Tested-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

In the case where a new build only decreases sizes and does not increase
any size we still want to report what functions have been dropped when
doing a bloat comparison.

Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: NTom Rini <trini@konsulko.com>
Reviewed-by: NSimon Glass <sjg@chromium.org>

In os_dirent_get_typename() we are checking that type falls within the
known values of the enum os_dirent_t.  With clang-3.8 testing this value
as being >= 0 results in a warning as it will always be true.  This
assumes of course that we are only given valid data.  Given that we want
to sanity check the input, we change this to check that it falls within
the range of the first to the last entry in the given enum.

Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: NTom Rini <trini@konsulko.com>
Reviewed-by: NSimon Glass <sjg@chromium.org>

Here is additional rk3368 and rk3399 support, rv1108 support,
refactoring HDMI video (brought in from Anatolij's tree to resolve
conflicts), some mkimage fixes and a few other things.

With video output support for the RK3399-Q7 (Puma) available, we want
CMD_BMP enabled and the support for 16bit, 24bit and 32bit BMPs
defined.
Signed-off-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

Version-changes: 2
- enable SYS_WHITE_ON_BLACK via defconfig
Reviewed-by: NSimon Glass <sjg@chromium.org>

The last set of rebases had dropped the 'grf' field from the common
rk_vop.  Add this back to un-break the build (and driver).
Signed-off-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>

This commit enables the RK3399 HDMI TX, which is very similar to the
one found on the RK3288.  As requested by Simon, this splits the HDMI
driver into a SOC-specific portion (rk3399_hdmi.c, rk3288_hdmi.c) and
a common portion (rk_hdmi.c).

Note that the I2C communication for reading the EDID works well with
the default settings, but does not with the alternate settings used on
the RK3288... this configuration aspect is reflected by the driverdata
for the RK3399 driver.

Having some sort of DTS-based configuration for the regulator
dependencies would be nice for the future, but for now we simply use
lists of regulator names (also via driverdata) that we probe.
Signed-off-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: NSimon Glass <sjg@chromium.org>

To prepare for the addition of RK3399 HDMI support, the HDMI driver is
refactored and broken into a chip-specific and a generic part.  This
change adds the internal interfaces, makes common/reusable functions
externally visible and splits the RK3288 driver into a separate file.

For the probing of regulators, we reuse the infrastructure created
during the VOP refactoring... i.e. we simply call into the helper
function defined for the VOP.
Signed-off-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: NSimon Glass <sjg@chromium.org>

The Linux driver now supports higher mpixelclock settings.
Add these to rockchip_phy_config[] and rockchip_mpll_cfg[].
Signed-off-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: NSimon Glass <sjg@chromium.org>

This commit adds a driver for the RK3399 VOPs capable and all the
necessary plumbing to feed the HDMI encoder. For the VOP-big, this
correctly tracks the ability to feed 10bit RGB data to the encoder.
Signed-off-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: NSimon Glass <sjg@chromium.org>

To prepare for adding the RK3399 VOP driver (which shares most of its
registers and config logic with the RK3228 VOP), this change refactors
the driver and splits the RK3288-specific driver off.

The changes in detail are:
- introduces a data-structure for chip-specific drivers to register
  features/callbacks with the common driver: at this time, this is
  limited to a callback for setting the pin polarities (between the
  VOP and the encoder modules) and a flag to signal 10bit RGB
  capability
- refactors the probing of regulators into a helper function that
  can take a list of regulator names to probe and autoset
- moves the priv data-structure into a (common) header file to be
  used by the chip-specific drivers to provide base addresses to
  the common driver
- uses a callback into the chip-specific driver to set pin polarities
  (replacing the direct register accesses previously used)
- splits enabling the output (towards an encoder) into a separate
  help function withint the common driver
Signed-off-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: NSimon Glass <sjg@chromium.org>

This introduces two new Kconfig options that configure the maximum
allowable framebuffer size (i.e. the memory reservation/allocation for
the framebuffer):
 - VIDEO_ROCKCHIP_MAX_XRES
 - VIDEO_ROCKCHIP_MAX_YRES
The resulting memory allocation will cover 4 byte per pixel for these
resolutions.
Signed-off-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: NSimon Glass <sjg@chromium.org>

For consistency sake (and as we are about to add new options to this
file), reformat the help for VIDEO_ROCKCHIP.
Signed-off-by: NPhilipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: NSimon Glass <sjg@chromium.org>