oweals/u-boot.git
7 years agoscripts: dtc: Add .gitignore
Bin Meng [Sun, 1 Oct 2017 05:56:22 +0000 (22:56 -0700)]
scripts: dtc: Add .gitignore

Ignore these generated files during the build of dtc.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agoMerge git://git.denx.de/u-boot-usb
Tom Rini [Sun, 1 Oct 2017 22:06:53 +0000 (18:06 -0400)]
Merge git://git.denx.de/u-boot-usb

7 years agoMerge git://git.denx.de/u-boot-rockchip
Tom Rini [Sun, 1 Oct 2017 17:05:53 +0000 (13:05 -0400)]
Merge git://git.denx.de/u-boot-rockchip

7 years agousb: dwc3: add UniPhier specific glue layer
Masahiro Yamada [Thu, 28 Sep 2017 13:01:00 +0000 (22:01 +0900)]
usb: dwc3: add UniPhier specific glue layer

Add UniPhier platform specific glue layer to support USB3 Host mode
on Synopsys DWC3 IP.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Marek Vasut <marex@denx.de>
7 years agousb: storage: Fix overwritten in usb_stor_set_max_xfer_blk()
Bin Meng [Thu, 28 Sep 2017 04:50:07 +0000 (21:50 -0700)]
usb: storage: Fix overwritten in usb_stor_set_max_xfer_blk()

The stored 'blk' value is overwritten to 'size / 512' before it can
be used in usb_stor_set_max_xfer_blk(). This is not what we want.
In fact, when 'size' exceeds the upper limit (USHRT_MAX * 512), we
should simply assign 'size' to the upper limit.

Reported-by: Coverity (CID: 167250)
Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: xhci: Set 'Average TRB Length' to 8 for control endpoints
Bin Meng [Mon, 18 Sep 2017 13:40:50 +0000 (06:40 -0700)]
usb: xhci: Set 'Average TRB Length' to 8 for control endpoints

Update the codes to conform with xHCI spec chapter 6.2.3.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: xhci: Set 'Error Count' to 0 for isoch endpoints
Bin Meng [Mon, 18 Sep 2017 13:40:49 +0000 (06:40 -0700)]
usb: xhci: Set 'Error Count' to 0 for isoch endpoints

Per xHCI spec, 'Error Count' should be set to 0 for isoch endpoints.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: xhci: Program max burst size for endpoint
Bin Meng [Mon, 18 Sep 2017 13:40:48 +0000 (06:40 -0700)]
usb: xhci: Program max burst size for endpoint

The 'Max Burst Size' indicates to the xHC the maximum number of
consecutive USB transactions that should be executed per scheduling
opportunity. This is a “zero-based” value, where 0 to 15 represents
burst sizes of 1 to 16, but at present this is always set to zero.
Let's program the required value according to real needs.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: xhci: Honor endpoint's interval
Bin Meng [Mon, 18 Sep 2017 13:40:47 +0000 (06:40 -0700)]
usb: xhci: Honor endpoint's interval

USB endpoint reports the period between consecutive requests to send
or receive data as bInverval in its endpoint descriptor. So far this
is ignored by xHCI driver and the 'Interval' field in xHC's endpoint
context is always programmed to zero which means 1ms for low speed
or full speed , or 125us for high speed or super speed. We should
honor the interval by getting it from endpoint descriptor.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: hub: Clear BH reset status change for a 3.0 hub
Bin Meng [Mon, 18 Sep 2017 13:40:46 +0000 (06:40 -0700)]
usb: hub: Clear BH reset status change for a 3.0 hub

USB 3.0 hubs report bit[5] in the port status change response as BH
reset. The hub shall set the C_BH_PORT_RESET field for this port.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: hub: Clear port reset before usb_hub_port_connect_change()
Bin Meng [Mon, 18 Sep 2017 13:40:45 +0000 (06:40 -0700)]
usb: hub: Clear port reset before usb_hub_port_connect_change()

During usb_hub_port_connect_change(), a port reset set feature
request is issued to the port, and later a port reset clear feature
is done to the same port before the function returns. However at
the end of usb_scan_port(), we attempt to clear port reset again
on a cached port status change variable, which should not be done.

Adjust the call to clear port reset to right before the call to
usb_hub_port_connect_change().

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: xhci: Fix max packet size for full speed device endpoint 0
Bin Meng [Mon, 18 Sep 2017 13:40:44 +0000 (06:40 -0700)]
usb: xhci: Fix max packet size for full speed device endpoint 0

In xhci_check_maxpacket(), the control endpoint 0 max packet size
is wrongly taken from the interface's endpoint descriptor. However
the default endpoint 0 does not come with an endpoint descriptor
hence is not included in the interface structure. Change to use
epmaxpacketin[0] instead.

The other bug in this routine is that when setting max packet size
to the xHC endpoint 0 context, it does not clear its previous value
at all before programming a new one.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: Read device descriptor after device is addressed for xHCI
Bin Meng [Mon, 18 Sep 2017 13:40:43 +0000 (06:40 -0700)]
usb: Read device descriptor after device is addressed for xHCI

For xHCI it is not possible to read a device descriptor before it
has been assigned an address. That's why usb_setup_descriptor()
was called with 'do_read' being false. But we really need try to
read the device descriptor before starting any real communication
with the default control endpoint.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: Only get 64 bytes device descriptor for full speed devices
Bin Meng [Mon, 18 Sep 2017 13:40:42 +0000 (06:40 -0700)]
usb: Only get 64 bytes device descriptor for full speed devices

Full speed device endpoint 0 can have 8/16/32/64 bMaxPacketSize0.
Other speed devices report fixed value per USB spec. So it only
makes sense if we send a get device descriptor with 64 bytes to
full speed devices.

While we are here, update the comment block to be within 80 cols.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: xhci: Add interrupt transfer support
Bin Meng [Mon, 18 Sep 2017 13:40:41 +0000 (06:40 -0700)]
usb: xhci: Add interrupt transfer support

xHCI uses normal TRBs for both bulk and interrupt. This adds the
missing interrupt transfer support to xHCI so that devices like
USB keyboard that uses interrupt transfer can work.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: Handle audio extension endpoint descriptor in usb_parse_config()
Bin Meng [Mon, 18 Sep 2017 13:40:40 +0000 (06:40 -0700)]
usb: Handle audio extension endpoint descriptor in usb_parse_config()

Normal endpoint descriptor size is 7, but for audio extension it is
9. Handle that correctly when parsing endpoint descriptor.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: xhci: Don't assume LS/FS devices are always behind a HS hub
Bin Meng [Mon, 18 Sep 2017 13:40:39 +0000 (06:40 -0700)]
usb: xhci: Don't assume LS/FS devices are always behind a HS hub

At present xHCI driver assumes LS/FS devices are attached directly
to a HS hub. If they are connected to a LS/FS hub, the driver will
fail to perform the USB enumeration process on such devices.

This is fixed by looking from the device itself all the way up to
the HS hub where the TT that serves the device is located.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agotest: dm: usb: Update test cases for USB
Bin Meng [Sun, 1 Oct 2017 13:19:45 +0000 (06:19 -0700)]
test: dm: usb: Update test cases for USB

Now that we have changed to remove all devices under the root hub in
usb_stop(), and corrected the USB emulator select logic, it makes no
sense to do various tests based on 'usb tree' output since the order
of devices is no longer fixed. Remove these USB test cases related
to 'usb tree'.

For the USB remove test, ideally we should remove an emulator device
node from the device tree, but this is so far not working. Change to
test the 'usb stop' only.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agodm: usb: emul: Drop usb_emul_reset()
Bin Meng [Sun, 1 Oct 2017 13:19:44 +0000 (06:19 -0700)]
dm: usb: emul: Drop usb_emul_reset()

With the root hub unbinding in usb_stop(), there is no need to do
a Sandbox-specific reset operation. usb_emul_reset() is no longer
used anywhere, drop it.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agodm: usb: Remove no longer needed blk_unbind_all()
Bin Meng [Sun, 1 Oct 2017 13:19:43 +0000 (06:19 -0700)]
dm: usb: Remove no longer needed blk_unbind_all()

With the root hub unbinding in usb_stop(), there is no need to do
a blk uclass specific unbind operation.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agodm: usb: Fix broken usb_stop()
Bin Meng [Sun, 1 Oct 2017 13:19:42 +0000 (06:19 -0700)]
dm: usb: Fix broken usb_stop()

At present we only do device_remove() during usb stop. The DM API
device_remove() only marks the device state as inactivated, but
still keeps its USB topology (eg: parent, children, etc) in the DM
device structure. There is no issue if we only start USB subsystem
once and never stop it. But a big issue occurs when we do 'usb stop'
and 'usb start' multiple times.

Strange things may be observed with current implementation, like:
- the enumeration may report only 1 mass storage device is detected,
  but the total number of USB devices is correct.
- USB keyboard does not work anymore after a bunch of 'usb reset'
  even if 'usb tree' shows it is correctly identified.
- read/write flash drive via 'fatload usb' may complain "Bad device"

In fact, every time when USB host controller starts the enumeration
process, it takes random time for each USB port to show up online,
hence each USB device may appear in a different order from previous
enumeration, and gets assigned to a totally different USB address.
As a result, we end up using a stale USB topology in the DM device
structure which still reflects the previous enumeration result, and
it may create an exact same DM device name like generic_bus_0_dev_7
that is already in the DM device structure. And since the DM device
structure is there, there is no device_bind() call to bind driver to
the device during current enumeration process, eventually creating
an inconsistent software representation of the hardware topology, a
non-working USB subsystem.

The fix is to clear the unused USB topology in the usb_stop(), by
calling device_unbind() on each controller's root hub device, and
the unbinding will unbind all of its children automatically.

For Sandbox, we need scan the device tree each time when we start
the USB stack, in order to re-create the emulated USB devices and
bind drivers for them before we actually do the driver probe.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: emul: hub: Report the actual device speed of the emulation device
Bin Meng [Sun, 1 Oct 2017 13:19:41 +0000 (06:19 -0700)]
usb: emul: hub: Report the actual device speed of the emulation device

At present the usb hub emulator always reports its downstream port
speed as full speed. Actually it is high speed for sandbox-flash,
and low speed for sandbox-keyb. We can determine the device speed
by checking its device descriptor bcdUSB field, and do the proper
hub port status report based on that.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: emul: Expose find_descriptor() as a public API
Bin Meng [Sun, 1 Oct 2017 13:19:40 +0000 (06:19 -0700)]
usb: emul: Expose find_descriptor() as a public API

This can be useful outside of the sandbox usb emulation uclass
driver. Expose it as a public API with a proper prefix (usb_emul_).

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: sandbox: Fix emulator device select logic in usb_emul_find_devnum()
Bin Meng [Sun, 1 Oct 2017 13:19:39 +0000 (06:19 -0700)]
usb: sandbox: Fix emulator device select logic in usb_emul_find_devnum()

Current emulator select logic in usb_emul_find_devnum() is to test
the USB address. The USB address of the device being enumerated is
initialized to zero at the beginning of the enumeration process in
usb_setup_device(). At this point, the saved USB address in the
platform data has not been assigned to any valid USB address either.
This means: the logic will select an emulator device according to
its sequence of declaring order in the device tree. Take test.dts
for example, flash-stick@0 will be selected before flash-stick@1.
But unfortunately such logic is wrong.

In fact USB devices show up in a random order during the enumeration
which means usb_emul_find_devnum() may be called on port 3 for keyb@3
before on port 0 for flash-stick@0.

To fix this, we introduce a new emulator uclass specific platdata
to store the USB device's port number on its parent hub, and update
the logic to test the port number instead.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: sandbox: Initialize root hub's device speed to high speed
Bin Meng [Sun, 1 Oct 2017 13:19:38 +0000 (06:19 -0700)]
usb: sandbox: Initialize root hub's device speed to high speed

At present 'usb tree' shows that the root hub on the Sandbox USB
controller is at full speed. But its device descriptor says it's
USB 2.0, so let's report it as a high speed device.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: Fix comments of usb_emul_find_for_dev()
Bin Meng [Sun, 1 Oct 2017 13:19:37 +0000 (06:19 -0700)]
usb: Fix comments of usb_emul_find_for_dev()

There is no such a parameter called 'bus'.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agousb: emul: Remove maxpacketsize in usb_emul_setup_device()
Bin Meng [Sun, 1 Oct 2017 13:19:36 +0000 (06:19 -0700)]
usb: emul: Remove maxpacketsize in usb_emul_setup_device()

This parameter is never used.

Signed-off-by: Bin Meng <bmeng.cn@gmail.com>
7 years agoMerge tag 'signed-efi-next' of git://github.com/agraf/u-boot
Tom Rini [Sun, 1 Oct 2017 12:48:38 +0000 (08:48 -0400)]
Merge tag 'signed-efi-next' of git://github.com/agraf/u-boot

Patch queue for efi - 2017-10-01

Lots of new things this time. High level highlights are:

  - Shim support (to boot Fedora)
  - Initial set of unit tests
  - Preparations to support UEFI Shell

7 years agorockchip: enable TPL_SYSRESET for all rockchip SoCs
Kever Yang [Wed, 13 Sep 2017 01:39:06 +0000 (09:39 +0800)]
rockchip: enable TPL_SYSRESET for all rockchip SoCs

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

References: 87c16d4 "drivers: spl: consistently use the $(SPL_TPL_)
macro"

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agork3288: spl: Add dram_init_banksize
Jagan Teki [Wed, 27 Sep 2017 17:33:14 +0000 (23:03 +0530)]
rk3288: spl: Add dram_init_banksize

Falcon mode, is updating DDR dt node configuration through
spl_fixup_fdt() so add appropriate DDR base and size through
dram_init_banksize.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: defconfig: puma-rk3399: update for DM_REGULATOR support in SPL
Philipp Tomsich [Fri, 29 Sep 2017 17:28:03 +0000 (19:28 +0200)]
rockchip: defconfig: puma-rk3399: update for DM_REGULATOR support in SPL

The RK3399-Q7 requires DM regulator support in SPL, so we can use the
regulator framework to reenable the eMMC and SPI, if these had been
turned of by the BIOS_DISABLE signal.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agopower: spl: add SPL_DM_REGULATOR_FIXED in Kconfig
Philipp Tomsich [Fri, 29 Sep 2017 17:28:02 +0000 (19:28 +0200)]
power: spl: add SPL_DM_REGULATOR_FIXED in Kconfig

The Makefile already tests for SPL_DM_REGULATOR_FIXED, but Kconfig
does not provide it.  This adds SPL_DM_REGULATOR_FIXED to Kconfig.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agorockchip: rk3399-puma: add boot-on regulator to override BIOS_DISABLE
Philipp Tomsich [Fri, 29 Sep 2017 17:28:01 +0000 (19:28 +0200)]
rockchip: rk3399-puma: add boot-on regulator to override BIOS_DISABLE

The (Qseven) BIOS_DISABLE signal on the RK3399-Q7 (Puma) keeps the
eMMC and SPI in reset initially and we need to write a GPIO to turn
them on before continuing the boot-up.

This adds the DTS entries for the additional regulator and makes
pinctrl and gpio3 available during SPL.  It also adds a hook to the
spl_board_init() to ensure that the regulator gets probed and enabled.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agorockchip: puma-rk3399: update board_init()
Philipp Tomsich [Fri, 29 Sep 2017 17:28:00 +0000 (19:28 +0200)]
rockchip: puma-rk3399: update board_init()

The original initialisation code for board_init() was largely lifted
from the code on the EVB.  However, the RK3399-Q7 can do with a much
more concise init sequence.

This cleans up the board_init() by updating it to the essentials for
the RK3399-Q7 and getting rid of the accumulated cruft.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agorockchip: dts: rk3399-puma: add 'same-as-spl' to the front of the boot-order
Philipp Tomsich [Fri, 29 Sep 2017 17:27:59 +0000 (19:27 +0200)]
rockchip: dts: rk3399-puma: add 'same-as-spl' to the front of the boot-order

In the general case, we want to continue booting the full U-Boot
(contained in a discoverable FIT image) from the same device the SPL
stage was loaded from.  This prepends the 'same-as-spl' specifier to
our configurable boot-order to make this the default behaviour.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agorockchip: spl: rk3399: implement chip-specific board_spl_was_booted_from()
Philipp Tomsich [Fri, 29 Sep 2017 17:27:58 +0000 (19:27 +0200)]
rockchip: spl: rk3399: implement chip-specific board_spl_was_booted_from()

To support the new "same-as-spl" specifier in the boot-order on the
RK3399, this implements the chip-specific mapping from the information
obtainable from the BROM to a OF path name.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agorockchip: spl: support a 'same-as-spl'-specifier in the spl-boot-order
Philipp Tomsich [Fri, 29 Sep 2017 17:27:57 +0000 (19:27 +0200)]
rockchip: spl: support a 'same-as-spl'-specifier in the spl-boot-order

It is often desirable to configure the spl-boot-order (i.e. the order
that SPL probes devices to find the FIT image containing a full U-Boot)
such that it contains 'the same device the SPL stage was booted from'
early on.  To support this, we introduce the 'same-as-spl' specifier
for the spl-boot-order property.

This commit adds:
 - documentation for the new board_spl_was_booted_from() function that
   individual SoCs/boards should provide, if they can determine where
   the SPL was booted from
 - implements the new board_spl_was_booted_from() stub function
 - adds support for handling the 'same-as-spl' specifier and calling
   into the per-SoC/per-board support code.

This also updates the documentation for the 'u-boot,spl-boot-order'
property.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agorockchip: spl: add documentation for spl_node_to_boot_device()
Philipp Tomsich [Fri, 29 Sep 2017 17:27:56 +0000 (19:27 +0200)]
rockchip: spl: add documentation for spl_node_to_boot_device()

In the expectation that the spl-boot-order code will eventually
gain use outside of mach-rockchip: let's add documentation on the
spl_node_to_boot_device() function, which is likely to become a
publicly exported function.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agorockchip: bootrom: add definitions to retrieve BROM boot-source
Philipp Tomsich [Fri, 29 Sep 2017 17:27:55 +0000 (19:27 +0200)]
rockchip: bootrom: add definitions to retrieve BROM boot-source

The Rockchip BROM allows reading where it booted from from SRAM.
This adds the necessary definitions (as received from Kever) for
the location of this information in the RK3399's SRAM and naming
for the constants used.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agorockchip: rk3399: make spl_board_init board-specific
Philipp Tomsich [Fri, 29 Sep 2017 17:27:54 +0000 (19:27 +0200)]
rockchip: rk3399: make spl_board_init board-specific

The later-stage spl_board_init (as opposed to board_init_f) should set
up board-specific details: these differ between the EVB-RK3399 and the
RK3399-Q7 (Puma).

This moves spl_board_init back into the individual boards and removes
the unneeded functionality from Puma.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agoevb-rk3399: avoid redefinition of CONFIG_ENV_OFFSET
Heinrich Schuchardt [Wed, 27 Sep 2017 18:23:09 +0000 (20:23 +0200)]
evb-rk3399: avoid redefinition of CONFIG_ENV_OFFSET

CONFIG_ENV_OFFSET is already defined in
include/configs/rockchip-common.h

For CONFIG_ROCKCHIP_BACK_TO_BROM=y we redefine CONFIG_ENV_OFFSET
to an unsuitable value. We were lucky to get a compiler warning.

Remove the incorrect redefinition.

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>
7 years agorockchip: rk322x: update the sysreg number for boot mode
Kever Yang [Tue, 22 Aug 2017 07:34:58 +0000 (15:34 +0800)]
rockchip: rk322x: update the sysreg number for boot mode

The boot mode for rk322x is stored in sysreg 0, update it.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: dts: rk3328-evb: add i2c1 and rk805 nodes
Elaine Zhang [Wed, 27 Sep 2017 07:23:38 +0000 (15:23 +0800)]
rockchip: dts: rk3328-evb: add i2c1 and rk805 nodes

add i2c1 and rk805 nodes to support rk805 init setting.

Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agoconfigs: rk3328: add support for pmic rk8xx and regulator and i2c driver
Elaine Zhang [Wed, 27 Sep 2017 07:23:37 +0000 (15:23 +0800)]
configs: rk3328: add support for pmic rk8xx and regulator and i2c driver

Add defconfig for rk8xx and regulator and i2c controller.

Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: i2c: rk3328: support i2c for rk3328 SoC
Elaine Zhang [Wed, 27 Sep 2017 07:23:34 +0000 (15:23 +0800)]
rockchip: i2c: rk3328: support i2c for rk3328 SoC

Add compatible to support rk3328 i2c

Signed-off-by: Elaine Zhang <zhangqing@rock-chips.com>
Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Acked-by: Heiko Schocher<hs@denx.de>
7 years agorockchip: pinctrl: rk3368: add pinctrl for SDMMC0
Philipp Tomsich [Tue, 19 Sep 2017 13:49:14 +0000 (15:49 +0200)]
rockchip: pinctrl: rk3368: add pinctrl for SDMMC0

Apparently, our earlier assumption that the BROM will always set up
the iomux for SDcard communication does not always hold true: when
booting U-Boot from the on-module (on the RK3368-uQ7) eMMC, the SDcard
pins are not set up and need to be configured by the pinctrl driver to
allow SD card access.

This change implements support for setting up the SDMMC pins in
pinctrl for the RK3368.

Reported-by: Klaus Goger <klaus.goger@theobroma-systems.com>
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agorockchip: board: lion-rk3368: avoid ATF overwriting active SPL stage
Philipp Tomsich [Fri, 15 Sep 2017 13:57:15 +0000 (15:57 +0200)]
rockchip: board: lion-rk3368: avoid ATF overwriting active SPL stage

The RK3368-uQ7 ATF has been moved back to 0x100000 (1MB from the start
of DRAM) to avoid it overwriting the active SPL stage during FIT image
loading.  This change adapts the .its to match up (again) with our ATF
repository for the RK3368-uQ7.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agork3288: vyasa: Add falcon mode support
Jagan Teki [Wed, 27 Sep 2017 17:33:13 +0000 (23:03 +0530)]
rk3288: vyasa: Add falcon mode support

Add Falcon mode support in vyasa rk3288 board.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agork3288: vyasa: Add TPL support
Jagan Teki [Wed, 27 Sep 2017 17:33:12 +0000 (23:03 +0530)]
rk3288: vyasa: Add TPL support

Since the size of SPL can't be exceeded 0x8000 bytes in RK3288,
it is not possible add new SPL features like Falcon mode or etc.

So add TPL stage so-that adding new features to SPL is possible.
- TPL: DRAM init, clocks
- SPL: MMC, falcon, etc

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agoarmv7: rk3288: Move configure_l2ctlr to common
Jagan Teki [Wed, 27 Sep 2017 17:33:11 +0000 (23:03 +0530)]
armv7: rk3288: Move configure_l2ctlr to common

configure_l2ctlr will be shared between SPL and TPL so
move them into asm/arch/sys_proto.h

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agoarmv7: Move L2CTLR read/write to common
Jagan Teki [Wed, 27 Sep 2017 17:33:10 +0000 (23:03 +0530)]
armv7: Move L2CTLR read/write to common

L2CTLR read/write functions are common to armv7 so, move
them in to include/asm/armv7.h and use them where ever it need.

Cc: Tom Warren <twarren@nvidia.com>
Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
[Backed out the change to arch/arm/mach-tegra/cache.c:]
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: rk322x: fix pd_bus hclk/pclk
Kever Yang [Thu, 28 Sep 2017 10:24:03 +0000 (18:24 +0800)]
rockchip: rk322x: fix pd_bus hclk/pclk

The pd_bus hclk/pclk source is pd_bus aclk, not the PLL.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: mkimage: add support for rk3128 soc
Kever Yang [Wed, 27 Sep 2017 09:57:28 +0000 (17:57 +0800)]
rockchip: mkimage: add support for rk3128 soc

Add support for rk3128 package header in mkimage tool.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: rk3036: fix grf macro define
Kever Yang [Wed, 27 Sep 2017 09:02:58 +0000 (17:02 +0800)]
rockchip: rk3036: fix grf macro define

Some of macros definition are not correct, fix them according to TRM.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: clk: fix typo in rk322x clock driver
Kever Yang [Wed, 27 Sep 2017 08:59:56 +0000 (16:59 +0800)]
rockchip: clk: fix typo in rk322x clock driver

Fix typo RK322X/RK3036 in rk322x clock driver.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: rk3399: spl: convert to using BOOT_DEVICE_BOOTROM
Philipp Tomsich [Mon, 11 Sep 2017 10:48:12 +0000 (12:48 +0200)]
rockchip: rk3399: spl: convert to using BOOT_DEVICE_BOOTROM

Instead of directly calling into the back-to-bootrom code, the RK3399
common SPL implementation now uses BOOT_DEVICE_BOOTROM to trigger a
transfer back into the bootrom.

With this factored out, the spl_board_init function can not be
customised for each RK3399 board.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agorockchip: lion-rk3368: defconfig: remove duplicate CONFIG_ENV_IS_IN_MMC
Philipp Tomsich [Wed, 27 Sep 2017 11:27:38 +0000 (13:27 +0200)]
rockchip: lion-rk3368: defconfig: remove duplicate CONFIG_ENV_IS_IN_MMC

Remove a duplicate CONFIG_ENV_IS_IN_MMC assignment for the lion-rk3368
defconfig.

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: rk3188: ram: add support for 16bit row address
Kever Yang [Mon, 25 Sep 2017 08:33:22 +0000 (16:33 +0800)]
rockchip: rk3188: ram: add support for 16bit row address

RK3188 using the same ddr_conf for both 15 bit and 16 bit row address.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
[Fixed compile-error by declaring 'row':]
Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: dts: rk3229: remove dram channel info
Kever Yang [Wed, 27 Sep 2017 08:38:23 +0000 (16:38 +0800)]
rockchip: dts: rk3229: remove dram channel info

The dram channel info will be auto detect by the driver,
we do not need it.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
7 years agorockchip: rk322x: add sdram driver
Kever Yang [Wed, 27 Sep 2017 08:38:22 +0000 (16:38 +0800)]
rockchip: rk322x: add sdram driver

Add driver for rk322x to support sdram initialize in SPL.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: rk3328: move sdram driver to driver/ram
Kever Yang [Wed, 27 Sep 2017 08:11:33 +0000 (16:11 +0800)]
rockchip: rk3328: move sdram driver to driver/ram

Since we have CONFIG_RAM framwork and its driver folder, move the driver
into it.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: rk3288: move sdram driver to driver/ram
Kever Yang [Wed, 27 Sep 2017 08:11:32 +0000 (16:11 +0800)]
rockchip: rk3288: move sdram driver to driver/ram

Since we have CONFIG_RAM framwork and its driver folder, move the driver
into it.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: rk3188: move sdram driver to driver/ram
Kever Yang [Wed, 27 Sep 2017 08:11:31 +0000 (16:11 +0800)]
rockchip: rk3188: move sdram driver to driver/ram

Since we have CONFIG_RAM framwork and its driver folder, move the driver
into it.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: rk3399: move sdram driver to driver/ram
Kever Yang [Wed, 27 Sep 2017 08:11:30 +0000 (16:11 +0800)]
rockchip: rk3399: move sdram driver to driver/ram

Since we have CONFIG_RAM framwork and its driver folder, move the driver
into it.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: imply ADC and SARADC_ROCKCHIP for Rockchip SOCs
Philipp Tomsich [Wed, 20 Sep 2017 11:50:13 +0000 (13:50 +0200)]
rockchip: imply ADC and SARADC_ROCKCHIP for Rockchip SOCs

Enable the Rockchip SARADC driver for all Rockchip SoCs.

Note that the SARADC peripheral is available on all SoCs except the
RK3036 and RK3228.  However, as this is a DM-driver, enabling by
default will not cause any function problems (and can always be
changed from defconfig, if size is a concern).

Signed-off-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: David Wu <david.wu@rock-chips.com>
7 years agorockchip: dts: Enable SARADC for rk3399-evb
David Wu [Wed, 20 Sep 2017 06:47:19 +0000 (14:47 +0800)]
rockchip: dts: Enable SARADC for rk3399-evb

Enable the SARADC for download key pressed detect.

Signed-off-by: David Wu <david.wu@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: dts: Enable SARADC for rk3368-sheep
David Wu [Wed, 20 Sep 2017 06:46:19 +0000 (14:46 +0800)]
rockchip: dts: Enable SARADC for rk3368-sheep

Enable the SARADC for download key pressed detect.

Signed-off-by: David Wu <david.wu@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: dts: Enable SARADC for rk3368-px5-evb
David Wu [Wed, 20 Sep 2017 06:45:21 +0000 (14:45 +0800)]
rockchip: dts: Enable SARADC for rk3368-px5-evb

Enable the SARADC for download key pressed detect.

Signed-off-by: David Wu <david.wu@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: dts: Enable SARADC for rk3328-evb
David Wu [Wed, 20 Sep 2017 06:44:23 +0000 (14:44 +0800)]
rockchip: dts: Enable SARADC for rk3328-evb

Enable the SARADC for download key pressed detect.

Signed-off-by: David Wu <david.wu@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: dts: Enable SARADC for rk3288-popmetal
David Wu [Wed, 20 Sep 2017 06:43:08 +0000 (14:43 +0800)]
rockchip: dts: Enable SARADC for rk3288-popmetal

Enable the SARADC for download key pressed detect.

Signed-off-by: David Wu <david.wu@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: dts: Enable SARADC for rv1108-evb
David Wu [Wed, 20 Sep 2017 06:41:38 +0000 (14:41 +0800)]
rockchip: dts: Enable SARADC for rv1108-evb

Enable the SARADC for download key pressed detect.

Signed-off-by: David Wu <david.wu@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: dts: rv1108: Add SARADC node at dtsi level
David Wu [Wed, 20 Sep 2017 06:40:11 +0000 (14:40 +0800)]
rockchip: dts: rv1108: Add SARADC node at dtsi level

Signed-off-by: David Wu <david.wu@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: clk: Add rk3399 SARADC clock support
David Wu [Wed, 20 Sep 2017 06:38:58 +0000 (14:38 +0800)]
rockchip: clk: Add rk3399 SARADC clock support

The clk_saradc is dividing from the 24M, clk_saradc=24MHz/(saradc_div_con+1).
SARADC integer divider control register is 8-bits width.

Signed-off-by: David Wu <david.wu@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: clk: Add rk3368 SARADC clock support
David Wu [Wed, 20 Sep 2017 06:37:50 +0000 (14:37 +0800)]
rockchip: clk: Add rk3368 SARADC clock support

The clk_saradc is dividing from the 24M, clk_saradc=24MHz/(saradc_div_con+1).
SARADC integer divider control register is 8-bits width.

Signed-off-by: David Wu <david.wu@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: clk: Add rk3328 SARADC clock support
David Wu [Wed, 20 Sep 2017 06:35:44 +0000 (14:35 +0800)]
rockchip: clk: Add rk3328 SARADC clock support

The clk_saradc is dividing from the 24M, clk_saradc=24MHz/(saradc_div_con+1).
SARADC integer divider control register is 10-bits width.

Signed-off-by: David Wu <david.wu@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: clk: Add SARADC clock support for rk3288
David Wu [Wed, 20 Sep 2017 06:28:19 +0000 (14:28 +0800)]
rockchip: clk: Add SARADC clock support for rk3288

The clk_saradc is dividing from the 24M, clk_saradc=24MHz/(saradc_div_con+1).
SARADC integer divider control register is 8-bits width.

Signed-off-by: David Wu <david.wu@rock-chips.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agorockchip: clk: Add rv1108 SARADC clock support
David Wu [Wed, 20 Sep 2017 06:28:18 +0000 (14:28 +0800)]
rockchip: clk: Add rv1108 SARADC clock support

The clk_saradc is dividing from the 24M, clk_saradc=24MHz/(saradc_div_con+1).
SARADC integer divider control register is 10-bits width.

Signed-off-by: David Wu <david.wu@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agodm: adc: Add driver for Rockchip SARADC
David Wu [Wed, 20 Sep 2017 06:28:16 +0000 (14:28 +0800)]
dm: adc: Add driver for Rockchip SARADC

The ADC can support some channels signal-ended some bits Successive Approximation
Register (SAR) A/D Converter, like 6-channel and 10-bit. It converts the analog
input signal into some bits binary digital codes.

Signed-off-by: David Wu <david.wu@rock-chips.com>
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
7 years agoenv: ti: boot: Show boot status information
Sam Protsenko [Thu, 21 Sep 2017 19:37:59 +0000 (22:37 +0300)]
env: ti: boot: Show boot status information

Add tracing printings to Linux/Android boot commands, so that we can see
what's going on. Helps to trace possible bugs on early stages and
improves the output for user (which is especially useful, because we
have a bunch of boot commands executing one by one).

Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
7 years agoenv: ti: boot: Extract command for eMMC Linux boot
Sam Protsenko [Thu, 21 Sep 2017 19:37:58 +0000 (22:37 +0300)]
env: ti: boot: Extract command for eMMC Linux boot

Extract commands for booting Linux from eMMC to separate command. It
seems more logical that way, and allows us to run the whole command set
from U-Boot shell with only one command.

Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
7 years agoarm: dra7xx: Fix Linux boot from eMMC
Sam Protsenko [Thu, 21 Sep 2017 19:37:57 +0000 (22:37 +0300)]
arm: dra7xx: Fix Linux boot from eMMC

Right now on OMAP5-based boards we have only one partition defined for
Linux boot, which is rootfs. That doesn't work with bootpart=1:2 (that
is defined in include/environment/ti/boot.h). To fix Linux boot we may
either:

 1. Change bootpart to be 1:1
 2. Or add preceding partition, so that rootfs is actually 1:2

Second choice seems more reasonable, as DFU is already using similar
partition table and can rely on bootpart to be 1:2.

This patch adds "bootloader" partition. So now eMMC layout for Linux
boot looks like this:

offset               content                 size          partition
(KiB)                                        (KiB)
Reviewed-by: Tom Rini <trini@konsulko.com>
=======================================================================

0       +--------------------------------+
        | MBR/GPT header                 |   128           -
128     +--------------------------------+
        | MLO                            |   256           -
384     +--------------------------------+
        | u-boot.img                     |   1792          bootloader
2176    +--------------------------------+
        | //////////// hole //////////// |   256           -
2432    +--------------------------------+
        | U-Boot environment             |   128           -
2560    +--------------------------------+
        | U-Boot environment (redundant) |   128           -
2688    +--------------------------------+
        | rootfs                         |   remaining     rootfs
end     +--------------------------------+

=======================================================================

Guard hole appears because U-Boot environment offset was calculated for
Android partition table, which has two additional partitions in place of
that hole ("environment" and "misc" partitions).

This patch also changes rootfs offset from 2 MiB further to 2688 KiB,
so that there won't be any collisions with U-Boot environment when we
flash rootfs.

Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
7 years agoarm: am57xx: Fix Linux boot from eMMC
Sam Protsenko [Thu, 21 Sep 2017 19:37:56 +0000 (22:37 +0300)]
arm: am57xx: Fix Linux boot from eMMC

Right now on OMAP5-based boards we have only one partition defined for
Linux boot, which is rootfs. That doesn't work with bootpart=1:2 (that
is defined in include/environment/ti/boot.h). To fix Linux boot we may
either:

 1. Change bootpart to be 1:1
 2. Or add preceding partition, so that rootfs is actually 1:2

Second choice seems more reasonable, as DFU is already using similar
partition table and can rely on bootpart to be 1:2.

This patch adds "bootloader" partition. So now eMMC layout for Linux
boot looks like this:

offset               content                 size          partition
(KiB)                                        (KiB)
Reviewed-by: Tom Rini <trini@konsulko.com>
=======================================================================

0       +--------------------------------+
        | MBR/GPT header                 |   128           -
128     +--------------------------------+
        | MLO                            |   256           -
384     +--------------------------------+
        | u-boot.img                     |   1792          bootloader
2176    +--------------------------------+
        | //////////// hole //////////// |   256           -
2432    +--------------------------------+
        | U-Boot environment             |   128           -
2560    +--------------------------------+
        | U-Boot environment (redundant) |   128           -
2688    +--------------------------------+
        | rootfs                         |   remaining     rootfs
end     +--------------------------------+

=======================================================================

Guard hole appears because U-Boot environment offset was calculated for
Android partition table, which has two additional partitions in place of
that hole ("environment" and "misc" partitions).

This patch also changes rootfs offset from 2 MiB further to 2688 KiB,
so that there won't be any collisions with U-Boot environment when we
flash rootfs.

Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
7 years agoarm: dts: omap3: sync DTS with Linux 4.14-rc1
Karthik Tummala [Thu, 21 Sep 2017 08:40:35 +0000 (14:10 +0530)]
arm: dts: omap3: sync DTS with Linux 4.14-rc1

This re-syncs omap3 DTS file with current file from
Linux v4.14-rc1 to ensure a consistent configuration. Upstream
Linux removed the redundant Interrupt-parent property from usbhsohci,
usbhsehci, ssi_port1 and ssi_port2 sub nodes.

Signed-off-by: Karthik Tummala <karthik@techveda.org>
7 years agodm: clk: add missing .priv_auto_alloc_size() for stm32f7
Patrice Chotard [Thu, 21 Sep 2017 08:08:09 +0000 (10:08 +0200)]
dm: clk: add missing .priv_auto_alloc_size() for stm32f7

Add missing .priv_auto_alloc_size() callback.
Previously private struct stm32_clk was not allocate
which leads to unpredictable behaviour

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
7 years agoARM: dts: OMAP5+: Enable all gpios in SPL
Lokesh Vutla [Thu, 21 Sep 2017 02:52:35 +0000 (08:22 +0530)]
ARM: dts: OMAP5+: Enable all gpios in SPL

With DM enabled, gpio numbering is assigned based on the
probed order of gpios, irrespective of the gpio base. So enable
all necessary gpios in SPL.

Fixes: edf25d94d55c (“ARM: dts: OMAP5+: Enable gpio in SPL”)
Reported-by: Gou, Hongmei <h-gou@ti.com>
Tested-by: Aparna Balasubramanian <aparnab@ti.com>
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoconfigs: keystone2: env: Fix burn_uboot_spi command
Lokesh Vutla [Thu, 21 Sep 2017 02:01:13 +0000 (07:31 +0530)]
configs: keystone2: env: Fix burn_uboot_spi command

Now the u-boot spi image is greater than 0x80000, increase the same
in env during spi erase.

Reported-by: Yan Liu <yan-liu@ti.com>
Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com>
7 years agoconfigs: k2g_evm: Make findfdt command populate fdtfile variable
Andrew F. Davis [Wed, 20 Sep 2017 19:45:09 +0000 (14:45 -0500)]
configs: k2g_evm: Make findfdt command populate fdtfile variable

On all other platforms the command 'findfdt' populates the variable
'fdtfile', but on K2G we only populate 'name_fdt'. The generic boot
and automation scripts fail when 'findfdt' is not populated, fix
this for K2G.

Signed-off-by: Andrew F. Davis <afd@ti.com>
7 years agoconfigs: am43xx_evm: Avoid relocation onto firewall at the end of DRAM
Andrew F. Davis [Wed, 20 Sep 2017 19:38:50 +0000 (14:38 -0500)]
configs: am43xx_evm: Avoid relocation onto firewall at the end of DRAM

On secure devices the initial secure software may install a firewall at
the end of DRAM, define protected RAM to avoid space.

Signed-off-by: Andrew F. Davis <afd@ti.com>
7 years agoarm: am33xx: security: adds auth support for encrypted images
Madan Srinivas [Wed, 20 Sep 2017 19:37:36 +0000 (14:37 -0500)]
arm: am33xx: security: adds auth support for encrypted images

This patch adds support for authentication of both plain
text and encrypted binaries. A new SECDEV package is needed
to enable encryption of binaries by default for AM3x.

The ROM authentication API detects encrypted images at
runtime and automatically decrypts the image if the
signature verification passes.

Addition of encryption on AM3x results in a change in the
image format. On AM4x, AM5x and, on AM3x devices signing
clear test images, the signature is appended to the end of the
binary.

On AM3x, when the SECDEV package is used to create signed
and encrypted images, the signature is added as a header
to the start of the binary. So the binary size calculation
has been updated to reflect this change.

The signing tools and encrypted image format for AM3x
cannot be changed to behave like AM4x and AM5x to
maintain backward compatibility with older Sitara
M-Shield releases.

Signed-off-by: Madan Srinivas <madans@ti.com>
Signed-off-by: Andrew F. Davis <afd@ti.com>
7 years agodefconfigs: Add a config for AM335x High Security EVM with UART Boot support
Andrew F. Davis [Wed, 20 Sep 2017 19:33:51 +0000 (14:33 -0500)]
defconfigs: Add a config for AM335x High Security EVM with UART Boot support

Add a new defconfig file for the AM335x High Security EVM. This config
is specific for the case of UART booting

Signed-off-by: Andrew F. Davis <afd@ti.com>
7 years agoARM: omap3: am3517-evm: Add device tree and DM support
Adam Ford [Wed, 20 Sep 2017 01:32:11 +0000 (20:32 -0500)]
ARM: omap3: am3517-evm: Add device tree and DM support

With the device tree ported from Linux 4.13, this enables
Driver Model and Device Tree support for the am3517-evm

Signed-off-by: Adam Ford <aford173@gmail.com>
Tested-by: Derald D. Woods <woods.technical@gmail.com>
7 years agoarm: dts: am3517_evm: Sync DTS files with Linux 4.13-RC5
Adam Ford [Wed, 20 Sep 2017 01:32:10 +0000 (20:32 -0500)]
arm: dts: am3517_evm: Sync DTS files with Linux 4.13-RC5

To keep the DTS and DTSI files clean and in sync with Linux, new
u-boot.dtsi files are added.

There are some spacing issues in the patch, but they appear to be
present in the Linux source files.  I'll try to get to fixing them there,
and do a future re-sync at a later date.

Signed-off-by: Adam Ford <aford173@gmail.com>
Tested-by: Derald D. Woods <woods.technical@gmail.com>
7 years agokconfiglib: update with 'imply' support
Ulf Magnusson [Tue, 19 Sep 2017 10:52:55 +0000 (12:52 +0200)]
kconfiglib: update with 'imply' support

Corresponds to 375506d (File writing nit) from upstream
(https://github.com/ulfalizer/Kconfiglib).

Adds proper 'imply' support and fixes a few minor issues, one of which
previously triggered the following weird warning:

  configs/taurus_defconfig: /tmp/tmpisI45S:6: warning: assignment to SPL_LDSCRIPT changes mode of containing choice from "arch/$(ARCH)/cpu/u-boot-spl.lds" to "y"

The change in 8639f69 (genconfig.py: Print defconfig next to warnings)
was reapplied.

tools/moveconfig.py previously depended on a hack that merged 'select's
with 'imply's. It was modified to look at the union of
Symbol.get_selected_symbols() and Symbol.get_implied_symbols(), which
should give the same behavior.

tools/genboardscfg.py was verified to produce identical board.cfg's
before and after the change.

Signed-off-by: Ulf Magnusson <ulfalizer@gmail.com>
7 years agoscripts/checkpatch: add missing spelling.txt
Heinrich Schuchardt [Tue, 19 Sep 2017 17:28:32 +0000 (19:28 +0200)]
scripts/checkpatch: add missing spelling.txt

Checkpatch can use two dictionaries for finding typos:
* scripts/spelling.txt
* /usr/share/codespell/dictionary.txt
  which comes with package codespell and is selected by parameter
  --codespell

If the file scripts/spelling.txt is missing, a warning is shown.

Add the spelling.txt file from the Linux kernel 4.13 source.

We can later add our own typo corrections.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Dan Murphy <dmurphy@ti.com>
7 years agotest/py: fix anchors in HTML status report
Stephen Warren [Mon, 18 Sep 2017 17:50:47 +0000 (11:50 -0600)]
test/py: fix anchors in HTML status report

The current code wraps a pre tag inside an a tag. For some reason, this
causes at least Firefox to attempt to drag the pre section content when
using a mouse drag to select text. Re-order the tags so that the text can
be selected using the mouse, at least if you start the drag outside the
text (after the end of the line, for example).

Signed-off-by: Stephen Warren <swarren@nvidia.com>
7 years agotest/py: add skip marker for reliance on tools
Stephen Warren [Mon, 18 Sep 2017 17:11:49 +0000 (11:11 -0600)]
test/py: add skip marker for reliance on tools

Some tests use external tools (executables) during their operation. Add
a test.py mark to indicate this. This allows those tests to be skipped if
the required tool is not present.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
7 years agotest/py: provide more information about test skip reason
Stephen Warren [Mon, 18 Sep 2017 17:11:48 +0000 (11:11 -0600)]
test/py: provide more information about test skip reason

When skipping tests, explicitly mention the board type or config option
that caused the skip. This will help people understand/fix any issues.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
7 years agotest/py: Document required tools/packages
Stephen Warren [Mon, 18 Sep 2017 17:11:47 +0000 (11:11 -0600)]
test/py: Document required tools/packages

Some tests rely on external tools. Mention these in the test/py README.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
7 years agoTI: ARCH_OMAP2PLUS: Enable SPL_STACK_R and provide default value
Tom Rini [Sun, 17 Sep 2017 15:44:49 +0000 (11:44 -0400)]
TI: ARCH_OMAP2PLUS: Enable SPL_STACK_R and provide default value

On ARCH_OMAP2PLUS platforms we know what the DDR layout is going to be,
and that it is safe to use SPL_STACK_R and provide a default value for
it.  select this and re-sync the defconfigs.

Signed-off-by: Tom Rini <trini@konsulko.com>
7 years agoARM: da850-evm: Enable DM_I2C
Adam Ford [Mon, 18 Sep 2017 01:43:48 +0000 (20:43 -0500)]
ARM: da850-evm: Enable DM_I2C

With DM now enabled with the device tree pulled from Linux, we can
enable DM_I2C in U-Boot.

Reviewed-by: Jagan Teki <jagan@openedev.com>
Signed-off-by: Adam Ford <aford173@gmail.com>
[trini: Add DM_I2C_COMPAT to da850_am18xxevm to fix warning]
Signed-off-by: Tom Rini <trini@konsulko.com>