oweals/u-boot.git
6 years agoarm: dts: sunxi: update H3 to new EMAC binding
Andre Przywara [Wed, 4 Apr 2018 00:31:18 +0000 (01:31 +0100)]
arm: dts: sunxi: update H3 to new EMAC binding

The U-Boot driver for the sun8i-emac was using some preliminary DT
binding. Now since Linux got its own driver in v4.15 and our driver
can now cope with both bindings, let's convert the DT nodes used by the
various H3 boards over to the new bindings used by the kernel.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
6 years agoarm: dts: sunxi: update A64 to new EMAC binding
Andre Przywara [Wed, 4 Apr 2018 00:31:17 +0000 (01:31 +0100)]
arm: dts: sunxi: update A64 to new EMAC binding

The U-Boot driver for the sun8i-emac was using some preliminary DT
binding. Now since Linux got its own driver in v4.15 and our driver
can now cope with both bindings, let's convert the DT nodes used for the
Pine64+ board over to the new bindings used by the kernel.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
6 years agonet: sun8i-emac: add support for new EMAC DT binding
Andre Przywara [Wed, 4 Apr 2018 00:31:16 +0000 (01:31 +0100)]
net: sun8i-emac: add support for new EMAC DT binding

The Ethernet MAC used in newer Allwinner SoCs (H3, A64, H5) got an
upstream Linux driver in v4.15.
This one uses a slightly different binding from the original one used
by the U-Boot driver.
The differences to the old binding are:
- The "syscon" address is held in a separate node, referenced via a
  phandle in the "syscon" property.
- The reference to the PHY is held in a property called "phy-handle",
  not "phy".
- The PHY register is at offset 0x30 in the syscon device, not at 0.
- The internal PHY is activated when the node, which phy-handle points
  to, is a child node of an "allwinner,sun8i-h3-mdio-internal" node.

Teach the U-Boot driver how to find its resources in a "new-style" DT,
so that we can use a Linux kernel compatible DT for U-Boot as well.
This keeps support for the old binding for now, to allow a smooth
transition.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
6 years agonet: sun8i-emac: support new pinctrl DT bindings
Andre Przywara [Wed, 4 Apr 2018 00:31:15 +0000 (01:31 +0100)]
net: sun8i-emac: support new pinctrl DT bindings

The Linux kernel driver for the Allwinner pin controller gained support
for generic properties, which are now also used in the DTs.
The sun8i-emac Ethernet driver for new Allwinner MACs reads the pins from
the DT, but so far only supported the old binding.
Update the parsing routine to cope with both the old and new bindings,
so that the newer DTs can be used with U-Boot and its Ethernet driver.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
6 years agosunxi: gpio: add missing compatible strings
Andre Przywara [Wed, 4 Apr 2018 00:31:14 +0000 (01:31 +0100)]
sunxi: gpio: add missing compatible strings

The sunxi GPIO driver is missing some compatible strings for recent
SoCs. While most of the sunxi GPIO code seems to not rely on this (and
so works anyway), the sunxi_name_to_gpio() function does and fails at
the moment (for instance when resolving the MMC CD pin name).
Add the compatible strings for the A64 and V3s, which were missing
from the list. This now covers all pinctrl nodes in our own DTs.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
6 years agosunxi: README.sunxi64: Add hint about non-debug of ARM Trusted Firmware
Andre Przywara [Wed, 4 Apr 2018 00:31:13 +0000 (01:31 +0100)]
sunxi: README.sunxi64: Add hint about non-debug of ARM Trusted Firmware

As we are running into issues where the final U-Boot FIT image file is
exceeding our size limit, add a hint to the README.sunxi64 file
to point out the possibility of building non-debug versions of the ATF
binary. These are about 12KB smaller than the standard debug build, and
so allow successful U-Boot builds for many boards with the Allwinner H5
SoC.
Please note that under normal circumstances the debug build is still
recommended, as it gives valuable clues in case something goes wrong in
the ATF.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
6 years agoconfigs: add NAND support for NES Classic
Miquel Raynal [Wed, 28 Feb 2018 19:52:02 +0000 (20:52 +0100)]
configs: add NAND support for NES Classic

Add NAND parameters to the Nintendo NES Classic configuration file which
features a Macronix NAND flash chip with 128kiB blocks of 2kiB pages
plus 64 OOB bytes.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: move the NAND parameters to Kconfig
Miquel Raynal [Wed, 28 Feb 2018 19:52:01 +0000 (20:52 +0100)]
sunxi: move the NAND parameters to Kconfig

Move the NAND parameters from defconfig files to Kconfig for SUNXI
architecture only. Fort now only the CHIP pro is migrated.

It would have been better to convert this defconfig entry to Kconfig for
all supported machines/architectures but it has been abandoned due to a
fairly high amount of errors reported by the moveconfig.py tool. This is
due to defines quite often being multiplications of values/other defines
not correctly handled.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: make NAND_SUNXI use ARCH_SUNXI as default in Kconfig
Miquel Raynal [Wed, 28 Feb 2018 19:52:00 +0000 (20:52 +0100)]
sunxi: make NAND_SUNXI use ARCH_SUNXI as default in Kconfig

Remove NAND_SUNXI from the CHIP pro defconfig to be automatically
selected depending on the state of ARCH_SUNXI.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: automatically select SPL_NAND_SUPPORT in Kconfig
Miquel Raynal [Wed, 28 Feb 2018 19:51:59 +0000 (20:51 +0100)]
sunxi: automatically select SPL_NAND_SUPPORT in Kconfig

Make SUNXI_NAND select SPL_NAND_SUPPORT in Kconfig, this limit the
number of entries to add in defconfig files when adding NAND support.

For now, the only board using it is the CHIP pro.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: dts: enable NAND on NES classic
Miquel Raynal [Wed, 28 Feb 2018 19:51:58 +0000 (20:51 +0100)]
sunxi: dts: enable NAND on NES classic

Let the Nintendo NES Classic use the Macronix NAND chip on it.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: allow NAND support to be compiled for sun8i platforms
Miquel Raynal [Wed, 28 Feb 2018 19:51:57 +0000 (20:51 +0100)]
sunxi: allow NAND support to be compiled for sun8i platforms

Add some clocks/PLL definitions as well as the dependency on MACH_SUN8I
in Kconfig.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: spl: remove DMA related settings of the NAND controller
Miquel Raynal [Wed, 28 Feb 2018 19:51:56 +0000 (20:51 +0100)]
sunxi: spl: remove DMA related settings of the NAND controller

Code has been changed to do not use DMA anymore with the NAND
controller, instead PIO is used. Then, DMA-specific initialization may
be dropped.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agospl: nand: sunxi: use PIO instead of DMA
Miquel Raynal [Wed, 28 Feb 2018 19:51:55 +0000 (20:51 +0100)]
spl: nand: sunxi: use PIO instead of DMA

SPL support was first written to support only the earlier generations of
Allwinner SoCs, and was only really enabled on the A13 / GR8. However,
those old SoCs had a DMA engine that has been replaced since the A31 by
another DMA controller that is no longer compatible.

Since the code directly uses that DMA controller, it cannot operate
properly on the later SoCs, while the NAND controller has not changed.

There's two paths forward, the first one would have been to add support
for that DMA controller too, the second to just remove the DMA usage
entirely and rely on PIO.

The later has been chosen because CPU overload at this stage is not an
issue and it makes the driver more generic, and easier to understand.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agospl: nand: sunxi: declare the ecc_bytes array globally
Miquel Raynal [Wed, 28 Feb 2018 19:51:54 +0000 (20:51 +0100)]
spl: nand: sunxi: declare the ecc_bytes array globally

Move the ecc_bytes array out of nand_max_ecc_strength() for future use
by nand_read_page().

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: spl: deassert the NAND controller reset line
Miquel Raynal [Wed, 28 Feb 2018 19:51:53 +0000 (20:51 +0100)]
sunxi: spl: deassert the NAND controller reset line

Ensure the NAND controller reset line is deasserted before use.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agospl: nand: sunxi: make the reset column helper more generic
Miquel Raynal [Wed, 28 Feb 2018 19:51:52 +0000 (20:51 +0100)]
spl: nand: sunxi: make the reset column helper more generic

Prepare the future use of an helper to move the data pointer (the
column) of the NAND chip by renaming nand_reset_column() to
nand_change_column(). Resetting the column is just a matter of giving 0
as argument.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agospl: nand: sunxi: ensure enough time has passed after changing the column
Miquel Raynal [Wed, 28 Feb 2018 19:51:51 +0000 (20:51 +0100)]
spl: nand: sunxi: ensure enough time has passed after changing the column

When changing the column, the ONFI specification states that a minimum
time of tCCS (Change Column Setup time) must elapse between the last
address cycle is asserted on the bus and the first data cycle is
clocked. An usual value for average NANDs is 500 nanoseconds. Round it
up to 1 microsecond to be safe.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agospl: nand: sunxi: create an helper to handle command execution
Miquel Raynal [Wed, 28 Feb 2018 19:51:50 +0000 (20:51 +0100)]
spl: nand: sunxi: create an helper to handle command execution

Executing a command is matter of always doing the following sequence:
  * Waiting for the FIFO to be empty so we can fill it with the new
    command.
  * Clearing the status register.
  * Writing the command in the FIFO.
  * Waiting for the command to finish.

Add a nand_exec_cmd() helper to handle this instead of repeating the
logic through the various functions.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agospl: nand: sunxi: add missing status clear
Miquel Raynal [Wed, 28 Feb 2018 19:51:49 +0000 (20:51 +0100)]
spl: nand: sunxi: add missing status clear

It is best practice to always clear the status register before executing
a command to be sure that the status read afterwards is relevant.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agospl: nand: sunxi: introduce the nand_wait_cmd_fifo_empty() helper
Miquel Raynal [Wed, 28 Feb 2018 19:51:48 +0000 (20:51 +0100)]
spl: nand: sunxi: introduce the nand_wait_cmd_fifo_empty() helper

One bit in the control registers indicates if the NAND controller is
ready to receive a new command. Otherwise, the command FIFO is full and
we should wait for this bit to flip. It then states that the last
command has been processed and the FIFO is now free to welcome another
command.

Add this sanity check before starting any new command.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agospl: nand: sunxi: introduce the nand_wait_int() helper
Miquel Raynal [Wed, 28 Feb 2018 19:51:47 +0000 (20:51 +0100)]
spl: nand: sunxi: introduce the nand_wait_int() helper

The pattern of polling on a status register until a bit is set or a
timeout occurs is repeated multiple times in the driver. Mutualize the
code by introducing the nand_wait_int() helper that does wait for the
bit to flip or returns an error in case of timeout.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agospl: nand: sunxi: fix typo on register name
Miquel Raynal [Wed, 28 Feb 2018 19:51:46 +0000 (20:51 +0100)]
spl: nand: sunxi: fix typo on register name

Change NFC_SEND_ADR to NFC_SEND_ADDR.

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agospl: nand: sunxi: fix second case of modulo by zero error
Miquel Raynal [Wed, 28 Feb 2018 19:51:45 +0000 (20:51 +0100)]
spl: nand: sunxi: fix second case of modulo by zero error

In the nand_read_buffer() step, the seed is calculated by doing a modulo
by conf->nseeds which is always zero when not using the randomizer (most
of SLC NANDs).

This situation turns out to lead to a run time freeze with certain
toolchains.

Derive this seed only when the randomizer is enabled (and conf->nseeds
logically not zero), exactly like what has been done before with an
identical situation, see commit ea3f750c73e3 ("nand: sunxi: Fix modulo
by zero error").

Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agomtd: nand: sunxi: fix ECC strength choice
Miquel Raynal [Wed, 28 Feb 2018 19:51:44 +0000 (20:51 +0100)]
mtd: nand: sunxi: fix ECC strength choice

When the requested ECC strength does not exactly match the strengths
supported by the ECC engine, the driver is selecting the closest
strength meeting the 'selected_strength > requested_strength'
constraint. Fix the fact that, in this particular case, ecc->strength
value was not updated to match the 'selected_strength'.

For instance, one can encounter this issue when no ECC requirement is
filled in the device tree while the NAND chip minimum requirement is not
a strength/step_size combo natively supported by the ECC engine.

Suggested-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Acked-by: Boris Brezillon <boris.brezillon@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agospl: fix binman_sym output check
Miquel Raynal [Wed, 28 Feb 2018 19:51:43 +0000 (20:51 +0100)]
spl: fix binman_sym output check

A previous commit introduced the use of binman in the SPL.

After the binman_sym call over the 'pos' symbol, the output value is
checked against BINMAN_SYM_MISSING (-1UL). According to the
documentation (tools/binman/README), when it comes to the 'pos'
attribute:

pos:
This sets the position of an entry within the image. The first
byte of the image is normally at position 0. If 'pos' is not
provided, binman sets it to the end of the previous region, or
the start of the image's entry area (normally 0) if there is no
previous region.

So instead of checking if the return value is BINMAN_SYM_MISSING, we
should also check if the value is not null.

The failure happens when using both the SPL file and the U-Boot file
independently instead of the concatenated file (SPL + padding + U-Boot).
This is because the U-Boot binary file alone does not have the U-Boot
header while it is present in the concatenation file. Not having the
header forces the SPL to discover where it should load U-Boot. The
binman_sym call is supposed to do that but fails. Because of the wrong
check, the destination address was set to 0 while it should have been
somewhere in RAM. This, obviously, stalls the board.

Fixes: 8bee2d251afb ("binman: Add binman symbol support to SPL")
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Signed-off-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: Add DRAM_SUN8I_A83T kconfig entry
Jagan Teki [Wed, 10 Jan 2018 10:50:26 +0000 (16:20 +0530)]
sunxi: Add DRAM_SUN8I_A83T kconfig entry

Add proper and simple kconfig option for dram_sun8i_a83t.c
instead of using MACH type on Makefile.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: Add DRAM_SUN8I_A33 kconfig entry
Jagan Teki [Wed, 10 Jan 2018 10:47:39 +0000 (16:17 +0530)]
sunxi: Add DRAM_SUN8I_A33 kconfig entry

Add proper and simple kconfig option for dram_sun8i_a33.c
instead of using MACH type on Makefile.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: Add DRAM_SUN8I_A23 kconfig entry
Jagan Teki [Wed, 10 Jan 2018 10:45:14 +0000 (16:15 +0530)]
sunxi: Add DRAM_SUN8I_A23 kconfig entry

Add proper and simple kconfig option for dram_sun8i_a23.c
instead of using MACH type on Makefile.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: Add DRAM_SUN9I kconfig entry
Jagan Teki [Fri, 16 Mar 2018 18:48:01 +0000 (00:18 +0530)]
sunxi: Add DRAM_SUN9I kconfig entry

Add proper and simple kconfig option for dram_sun9i.c
instead of using MACH type on Makefile.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: Add DRAM_SUN4I kconfig entry
Jagan Teki [Wed, 10 Jan 2018 10:33:34 +0000 (16:03 +0530)]
sunxi: Add DRAM_SUN4I kconfig entry

Add proper and simple kconfig option for dram_sun4i.c
instead of using MACH type on Makefile.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: add DRAM_SUN6I kconfig
Jagan Teki [Fri, 16 Mar 2018 18:46:36 +0000 (00:16 +0530)]
sunxi: add DRAM_SUN6I kconfig

Add proper and simple kconfig option for dram_sun6i.c
instead of using MACH type on Makefile.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: Add AXP_PMIC_BUS kconfig entry
Jagan Teki [Wed, 14 Feb 2018 16:58:30 +0000 (22:28 +0530)]
sunxi: Add AXP_PMIC_BUS kconfig entry

Add simple and meaningful kconfig option for pmic_bus.c
instead of using MACH type on Makefile.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: Add SUN8I_RSB kconfig entry
Jagan Teki [Thu, 11 Jan 2018 07:53:52 +0000 (13:23 +0530)]
sunxi: Add SUN8I_RSB kconfig entry

Add simple and meaningful kconfig option for rsb.c
instead of using MACH type on Makefile.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: Use SUN6I_PRCM if used
Jagan Teki [Thu, 11 Jan 2018 07:53:02 +0000 (13:23 +0530)]
sunxi: Use SUN6I_PRCM if used

SUN6I_PRCM is also used for SUN8I and SUN9I, so
select the same on respective MACH types.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: Add SUN6I_P2WI kconfig entry
Jagan Teki [Thu, 11 Jan 2018 07:51:58 +0000 (13:21 +0530)]
sunxi: Add SUN6I_P2WI kconfig entry

Add simple and meaningful kconfig option for p2wi.c
instead of using MACH type on Makefile.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agosunxi: Add SUN6I_PRCM kconfig entry
Jagan Teki [Thu, 11 Jan 2018 07:51:15 +0000 (13:21 +0530)]
sunxi: Add SUN6I_PRCM kconfig entry

Add simple and meaningful kconfig option for prcm.c
instead of using MACH type on Makefile.

PRCM (Power/Reset/Clock Management) is considered as a
Multi-Functional Device, so used the same on Kconfig definition.

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
6 years agocommon: move init_helpers.h prototypes in init.h
Patrick Delaunay [Tue, 13 Mar 2018 12:57:04 +0000 (13:57 +0100)]
common: move init_helpers.h prototypes in init.h

Merge init_helpers.h in the new file init.h
with only prototypes for init_cache_f_r
used in common/board_f.c

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
6 years agocommon: move board_info.c prototypes in init.h
Patrick Delaunay [Tue, 13 Mar 2018 12:57:03 +0000 (13:57 +0100)]
common: move board_info.c prototypes in init.h

Move function prototype for common/init/board_init.c
from common.h to init.h

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
6 years agocommon: move board_r.c prototypes in init.h
Patrick Delaunay [Tue, 13 Mar 2018 12:57:02 +0000 (13:57 +0100)]
common: move board_r.c prototypes in init.h

Move function prototypes used in common/board_r.c
from common.h to init.h

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
6 years agocommon: move board_init.c prototypes in init.h
Patrick Delaunay [Tue, 13 Mar 2018 12:57:01 +0000 (13:57 +0100)]
common: move board_init.c prototypes in init.h

Move function prototypes for common/init/board_init.c
from common.h to init.h

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
6 years agocommon: move board_f.c prototypes in init.h
Patrick Delaunay [Tue, 13 Mar 2018 12:57:00 +0000 (13:57 +0100)]
common: move board_f.c prototypes in init.h

Move prototypes for function used in common/board_f.c
from common.h to init.h
Remove weak for arch_reserve_stacks in prototype
(checkpatch issue)

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
6 years agoppc: Remove orphan xilinx_irq.h file
Michal Simek [Tue, 13 Mar 2018 10:17:07 +0000 (11:17 +0100)]
ppc: Remove orphan xilinx_irq.h file

This file is completely unused and it should be removed as the part of
ppc405/ppc440 xilinx platform removal.

Signed-off-by: Michal Simek <michal.simek@xilinx.com>
6 years agoMAINTAINERS: TI SYSTEM SECURITY: remove invalid file
Heinrich Schuchardt [Mon, 12 Mar 2018 22:58:33 +0000 (23:58 +0100)]
MAINTAINERS: TI SYSTEM SECURITY: remove invalid file

Remove a link to a non-existent file.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Andrew F. Davis <afd@ti.com>
6 years agoMAINTAINERS: ARM TI: remove invalid paths
Heinrich Schuchardt [Mon, 12 Mar 2018 22:40:29 +0000 (23:40 +0100)]
MAINTAINERS: ARM TI: remove invalid paths

Remove non-existing paths.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
6 years agoMAINTAINERS: ARM SAMSUNG: remove invalid paths
Heinrich Schuchardt [Mon, 12 Mar 2018 22:17:11 +0000 (23:17 +0100)]
MAINTAINERS: ARM SAMSUNG: remove invalid paths

Remove non-existing directories.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Acked-by: Minkyu Kang <mk7.kang@samsung.com>
6 years agoMAINTAINERS: FLATTENED DEVICE TREE: correct maintained path
Heinrich Schuchardt [Mon, 12 Mar 2018 22:08:09 +0000 (23:08 +0100)]
MAINTAINERS: FLATTENED DEVICE TREE: correct maintained path

Change due to commit b08c8c487083 ("libfdt: move headers to
<linux/libfdt.h> and <linux/libfdt_env.h>")

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
6 years agoMAINTAINERS: ARM HISILICON: correct maintained path
Heinrich Schuchardt [Mon, 12 Mar 2018 22:02:17 +0000 (23:02 +0100)]
MAINTAINERS: ARM HISILICON: correct maintained path

Fix an incorrect path.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
6 years agoMAINTAINERS: ARM FREESCALE IMX: remove invalid path
Heinrich Schuchardt [Mon, 12 Mar 2018 21:56:18 +0000 (22:56 +0100)]
MAINTAINERS: ARM FREESCALE IMX: remove invalid path

arch/arm/cpu/armv7/mx*/ does not relate to any existing directory.

Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Fabio Estevam <fabio.estevam@nxp.com>
6 years agopost: remove reference to power functions
Patrick Delaunay [Mon, 12 Mar 2018 08:22:44 +0000 (09:22 +0100)]
post: remove reference to power functions

The 2 functions board_power_mode and board_poweroff
are no more existing in U-Boot code (check with grep)
This patch updates the documentation and removes
the unnecessary prototypes.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
6 years agocommon: remove unused prototypes
Patrick Delaunay [Mon, 12 Mar 2018 08:19:22 +0000 (09:19 +0100)]
common: remove unused prototypes

Remove the prototypes for 4 functions no more existing
in U-Boot code (checked by grep):
- checkflash
- checkdram
- mdm_init
- setup_board_extra

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
6 years agocommon: add a prototype for mach_cpu_init()
Patrick Delaunay [Fri, 9 Mar 2018 17:28:12 +0000 (18:28 +0100)]
common: add a prototype for mach_cpu_init()

Add a new file init.h with the prototype for arch_cpu_init
Add a prototype for mach_cpu_init() to avoid a warning:
no previous prototype for ‘mach_cpu_init’

It is a first step to move all the functions prototype
used during U-Boot initialization (board_f.c / board_r.c)
from common.h to init.h

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
Reviewed-by: Masahiro Yamada <yamada.masahiro@socionext.com>
6 years agoConvert all of CONFIG_CONS_INDEX to Kconfig
Tom Rini [Wed, 7 Mar 2018 00:02:27 +0000 (19:02 -0500)]
Convert all of CONFIG_CONS_INDEX to Kconfig

This converts the following to Kconfig:
   CONFIG_CONS_INDEX

We have existing entries for this option in a number of places, with
different guards on them.  They're also sometimes used for things not
directly inside of the serial driver.  First, introduce a new symbol to
guard the use of CONFIG_CONS_INDEX, so that in the case where we don't
need this for the serial driver, but for some other use, we can still do
it.  Next, consolidate all of these into the single entry in
drivers/serial/Kconfig.  Finally, introduce CONS_INDEX_[023456] so that
we can imply a correct value here to make the defconfig side of this
smaller.

Signed-off-by: Adam Ford <aford173@gmail.com>
[trini: Rework a lot of the logic here, such that I took authorship from
Adam, but kept his S-o-B line]
Signed-off-by: Tom Rini <trini@konsulko.com>
6 years agodra7xx: Always enable UART1 and UART3 in early_padconf
Tom Rini [Wed, 7 Mar 2018 00:02:26 +0000 (19:02 -0500)]
dra7xx: Always enable UART1 and UART3 in early_padconf

It is safe to always setup the pinmux for UART1 and UART3 to be used in
early_padconf and then if these pins are needed later on, they will be
re-muxed.  This allows us to drop the usage of CONFIG_CONS_INDEX here.

Cc: Lokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>
6 years agodm: MIGRATION: Add migration plan for DM_SPI & DM_SPI_FLASH
Jagan Teki [Wed, 14 Mar 2018 13:28:03 +0000 (18:58 +0530)]
dm: MIGRATION: Add migration plan for DM_SPI & DM_SPI_FLASH

Due to adding various new functionalities with SPI and SPI-FLASH
subsystems which are rounding in Mailing list for year these long
term supporting spi driver which never seen any update with
driver-model conversion.

So added migration plan for DM_SPI and DM_SPI_FLASH with
deadline v2018.09

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
6 years agodefconfig: k2l_hs_evm: Add k2l_hs_evm_defconfig
Madan Srinivas [Tue, 6 Mar 2018 16:10:10 +0000 (11:10 -0500)]
defconfig: k2l_hs_evm: Add k2l_hs_evm_defconfig

Adds a dedicated defconfig to build TI K2L secure
devices and updates MAINTAINERS.

k2l_hs_evm_defconfig is created from the k2l_evm_defconfig
and removes support for SPL, as SPL is not supported on K2
HS devices. Corrects SYS_TEXT_BASE for HS devices.

Also adds TI_SECURE_DEVICE and FIT_IMAGE_POST_PROCESS to
include support for secure image creation and authentication

Signed-off-by: Madan Srinivas <madans@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
Acked-by: Andrew F. Davis <afd@ti.com>
6 years agoconfigs: k2l: Updates u-boot env to install secure boot monitor
Madan Srinivas [Tue, 6 Mar 2018 16:10:09 +0000 (11:10 -0500)]
configs: k2l: Updates u-boot env to install secure boot monitor

This patch updates the default u-boot env for K2L HS devices
to install the secure boot monitor and load the fitImage during
boot.

Signed-off-by: Madan Srinivas <madans@ti.com>
Acked-by: Andrew F. Davis <afd@ti.com>
Reviewed-by: Tom Rini <trini@konsulko.com>
6 years agowatchdog: omap_wdt: improve watchdog reset path
Ruslan Bilovol [Thu, 1 Mar 2018 01:15:48 +0000 (03:15 +0200)]
watchdog: omap_wdt: improve watchdog reset path

Remove busy looping during watchdog reset.
Each polling of W_PEND_WTGR bit ("finish posted
write") after watchdog reset takes 120-140us
on BeagleBone Black board. Current U-Boot code
has watchdog resets in random places and often
there is situation when watchdog is reset
few times in a row in nested functions.
This adds extra delays and slows the whole system.

Instead of polling W_PEND_WTGR bit, we skip
watchdog reset if the bit is set. Anyway, watchdog
is in the middle of reset *right now*, so we can
just return.

This noticeably increases performance of the
system. Below are some measurements on BBB:
 - DFU upload over USB                 15% faster
 - fastboot image upload               3x times faster
 - USB ep0 transfers with 4k packets   20% faster

Signed-off-by: Ruslan Bilovol <ruslan.bilovol@gmail.com>
Tested-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Lokesh Vutla <lokeshvutla@ti.com>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
Tested-by: Alex Kiernan <alex.kiernan@gmail.com>
6 years agoboard: am335x: Set serial# variable
Sam Protsenko [Tue, 27 Feb 2018 22:26:15 +0000 (00:26 +0200)]
board: am335x: Set serial# variable

serial# variable is needed to show the device correctly in "fastboot
devices" output. It's useful when we have several devices (in fastboot
mode) connected to single host and want to choose which one to flash.

We can't use omap_die_id_serial() for this, because AM335x lacks
DIE_ID, as can be seen from AM335x TRM. Let's do next:
 - reuse board_serial variable (obtained from EEPROM in
   set_board_info_env() function) to set serial#
 - if board_serial is "unknown", reuse ethaddr variable to set serial#

Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
6 years agoclk: at91: clk-system: add set/get_rate operations
Wenyou Yang [Fri, 9 Feb 2018 03:34:52 +0000 (11:34 +0800)]
clk: at91: clk-system: add set/get_rate operations

To support set/get the clock rate, add set/get_rate operations.

Signed-off-by: Wenyou Yang <wenyou.yang@microchip.com>
6 years agoclk: at91: add PLLADIV driver
Wenyou Yang [Fri, 9 Feb 2018 03:34:51 +0000 (11:34 +0800)]
clk: at91: add PLLADIV driver

As said in the SAMA5D2 datasheet, the PLLA clock must be divided
by 2 by writing the PLLADIV2 bit in PMC_MCKR, if the ratio between
PCK and MCK is 3 (MDIV = 3). This is the purpose of the driver.

Signed-off-by: Wenyou Yang <wenyou.yang@microchip.com>
6 years agoclk: at91: add USB Host clock driver
Wenyou Yang [Fri, 9 Feb 2018 03:34:50 +0000 (11:34 +0800)]
clk: at91: add USB Host clock driver

Add USB clock driver to configure the input clock and the divider
in the PMC_USB register to generate a 48MHz and a 12MHz signal to
the USB Host OHCI.

Signed-off-by: Wenyou Yang <wenyou.yang@microchip.com>
6 years agoMerge git://git.denx.de/u-boot-uniphier
Tom Rini [Thu, 15 Mar 2018 13:58:30 +0000 (09:58 -0400)]
Merge git://git.denx.de/u-boot-uniphier

6 years agoARM: dts: uniphier: sync with Linux 4.16-rc5
Masahiro Yamada [Thu, 15 Mar 2018 02:43:03 +0000 (11:43 +0900)]
ARM: dts: uniphier: sync with Linux 4.16-rc5

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
6 years agoARM: uniphier: change load addresses for bigger kernel image
Masahiro Yamada [Thu, 22 Feb 2018 10:00:08 +0000 (19:00 +0900)]
ARM: uniphier: change load addresses for bigger kernel image

The kernel size is getting bigger and bigger.  Because the Android
common kernel is even bigger than the vanilla one, so images overlap
in the current memory layout.

Adjust the load address for a bigger kernel image.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
6 years agoARM: uniphier: increase the firmware size in update command environment
Masahiro Yamada [Fri, 2 Feb 2018 05:16:36 +0000 (14:16 +0900)]
ARM: uniphier: increase the firmware size in update command environment

If a bigger EL32 is used, fip.bin can be larger than the currently
assumed boot firmware size.  Adjust the update commands.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
6 years agoMerge git://git.denx.de/u-boot-spi
Tom Rini [Thu, 15 Mar 2018 12:27:27 +0000 (08:27 -0400)]
Merge git://git.denx.de/u-boot-spi

6 years ago.travis.yml: Add lzop
Tom Rini [Wed, 14 Mar 2018 18:02:49 +0000 (14:02 -0400)]
.travis.yml: Add lzop

We need lzop now in order to make some FIT images that use LZO
compression on the contents.

Signed-off-by: Tom Rini <trini@konsulko.com>
6 years agoomap3_logic: Enable DM_SPI
Adam Ford [Wed, 7 Mar 2018 11:23:10 +0000 (05:23 -0600)]
omap3_logic: Enable DM_SPI

With the introduction of ("Boards,Need,to,Switch,DM] spi: omap3_spi:
Full dm conversion"), Driver Model for McSPI is now available on the
omap3.  This enables the config by default on omap3_logic boards.

Signed-off-by: Adam Ford <aford173@gmail.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
6 years agospi: omap3: Fix redeclared error
Jagan Teki [Wed, 14 Mar 2018 17:37:31 +0000 (23:07 +0530)]
spi: omap3: Fix redeclared error

omap3_spi_set_speed|mode redeclared bus symbol, fix the same.

error:
drivers/spi/omap3_spi.c: In function ‘omap3_spi_set_speed’:
drivers/spi/omap3_spi.c:650:18: error: ‘bus’ redeclared as different kind of symbol
  struct udevice *bus = dev->parent;

Signed-off-by: Jagan Teki <jagan@amarulasolutions.com>
6 years agoMerge git://git.denx.de/u-boot-sunxi
Tom Rini [Wed, 14 Mar 2018 17:27:14 +0000 (13:27 -0400)]
Merge git://git.denx.de/u-boot-sunxi

6 years agosunxi: move CONFIG_SYS_TEXT_BASE out of defconfigs
Andre Przywara [Sat, 3 Mar 2018 23:49:02 +0000 (23:49 +0000)]
sunxi: move CONFIG_SYS_TEXT_BASE out of defconfigs

Commit 278b90ce786f ("configs: Migrate CONFIG_SYS_TEXT_BASE") made
CONFIG_SYS_TEXT_BASE a proper Kconfig variable, with the consequence
of moving the common definition shared by almost every sunxi board
into 123 individual defconfig files. But the U-Boot start address
for Allwinner boards is a platform decision which has been around for
ages, so defining it in each *board* config file seems a bit over the
top.
Define the standard values (160MB into DRAM for most SoCs, with two
SoC exceptions) if ARCH_SUNXI is selected, and delete the lines from
the individual defconfigs.

Signed-off-by: Andre Przywara <andre.przywara@arm.com>
Acked-by: Maxime Ripard <maxime.ripard@bootlin.com>
Reviewed-by: Jagan Teki <jagan@openedev.com>
6 years agocommon/memsize.c: restore content of the base address
Patrick Delaunay [Thu, 25 Jan 2018 17:07:46 +0000 (18:07 +0100)]
common/memsize.c: restore content of the base address

For the last return of function get_ram_size(),
when size=maxsize, restore the base address (*base)
content.

Add comment for the remaining case to avoid regression:
this case is already correctly handled.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
6 years agocommon/memsize.c: prepare get_ram_size update
Patrick Delaunay [Thu, 25 Jan 2018 17:07:45 +0000 (18:07 +0100)]
common/memsize.c: prepare get_ram_size update

Save the content of the base address in a new variable
in stack (save_base) to prepare restore this content.

This patch don't modified code behavior and stack usage.

Signed-off-by: Patrick Delaunay <patrick.delaunay@st.com>
6 years agoconfigs: am335x_boneblack: Enable fastboot flash capability
Sam Protsenko [Tue, 27 Feb 2018 21:45:04 +0000 (23:45 +0200)]
configs: am335x_boneblack: Enable fastboot flash capability

Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Acked-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Tom Rini <trini@konsulko.com>
6 years agoomap: Fix AM335x build with enabled fastboot flash
Sam Protsenko [Tue, 27 Feb 2018 21:45:03 +0000 (23:45 +0200)]
omap: Fix AM335x build with enabled fastboot flash

When enabling CONFIG_FASTBOOT_FLASH in am335x_boneblack_defconfig, next
build errors and warnings occur:

    arch/arm/mach-omap2/utils.c: In function ‘omap_set_fastboot_cpu’:
    arch/arm/mach-omap2/utils.c:26:16: warning: implicit declaration of
              function ‘omap_revision’ [-Wimplicit-function-declaration]
              u32 cpu_rev = omap_revision();
                            ^~~~~~~~~~~~~
    arch/arm/mach-omap2/utils.c:29:7: error: ‘DRA762_ES1_0’ undeclared
              (first use in this function)

Include asm/omap_common.h explicitly to avoid those.

Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Acked-by: Lukasz Majewski <lukma@denx.de>
Reviewed-by: Tom Rini <trini@konsulko.com>
6 years agoARM: dts: dra7x: Make pinctrl and IOdelays for MMC2 available in SPL
Jean-Jacques Hiblot [Tue, 27 Feb 2018 16:05:50 +0000 (17:05 +0100)]
ARM: dts: dra7x: Make pinctrl and IOdelays for MMC2 available in SPL

The SPL can't use high speed MMC modes if the associated pinctrl and
IOdelays are described in the DTS.
Make them available in SPL by tagging the nodes with 'u-boot,dm-spl;'

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
6 years agoconfigs: dra7xx_evm: Enable support for the HS200 mmc mode in the SPL
Jean-Jacques Hiblot [Tue, 27 Feb 2018 16:05:49 +0000 (17:05 +0100)]
configs: dra7xx_evm: Enable support for the HS200 mmc mode in the SPL

Beside enabling the support for HS200 in mmc core, enabling the HS200
support in the SPL requires multi-dtb support in the SPL because pinctrl
and IOdelays vary across SOCs.

Also we need to make sure that the pinctrl properties arenot removed from
the dts by setting CONFIG_OF_SPL_REMOVE_PROPS to remove only clocks and
interrupts properties.

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
6 years agoARM: dts: dra76-evm: shift to using common IOdelay data
Jean-Jacques Hiblot [Tue, 27 Feb 2018 16:05:48 +0000 (17:05 +0100)]
ARM: dts: dra76-evm: shift to using common IOdelay data

Now that we have a device-tree include file with common
MMC/SD IOdelay data for DRA76x SoC, shift the EVM device-tree
file to using that.
Also fix the name of the IO voltage regulator for mmc1.

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
6 years agoARM: dts: dra76x: create a common file with MMC/SD IOdelay data
Jean-Jacques Hiblot [Tue, 27 Feb 2018 16:05:47 +0000 (17:05 +0100)]
ARM: dts: dra76x: create a common file with MMC/SD IOdelay data

Add a common device-tree include file with MMC/SD IOdelay data
for DRA76x SoC based on the linux DTSI file.

In the most common case, IOdelay data available in datamanual
can directly be used. This file caters to that common case.

Data is based on DRA76x datamanual, SPRS993A, revised July 2017.

Signed-off-by: Jean-Jacques Hiblot <jjhiblot@ti.com>
6 years agoARM: omap3: evm: Sync with omap3_beagle configuration
Derald D. Woods [Mon, 26 Feb 2018 23:52:59 +0000 (17:52 -0600)]
ARM: omap3: evm: Sync with omap3_beagle configuration

This commit does the following for OMAP3 EVM:

- Track omap3_beagle changes where possible
- Remove CONFIG_SYS_MPUCLK and CONFIG_MTD which are not needed for the
  default board configuration to work.
- Remove CONFIG_SPL_SYS_MALLOC_SIMPLE and CONFIG_SYS_MALLOC_F_LEN from
  default config. CONFIG_SYS_MALLOC_F_LEN=0x400 is set by configuration
  automatically. [allows board to boot with SERIAL_SEARCH_ALL=y]
- Use updated NAND layout:

device nand0 <omap2-nand.0>, # parts = 6

#: name size offset
-------------------------------------------------
0: spl 0x00080000(512k) 0x00000000
1: u-boot 0x001e0000(1920k) 0x00080000
2: u-boot-env 0x00020000(128k) 0x00260000
3: dtb 0x00020000(128k) 0x00280000
4: kernel 0x00600000(6m) 0x002a0000
5: rootfs 0x0f760000(-) 0x008a0000
-------------------------------------------------

Signed-off-by: Derald D. Woods <woods.technical@gmail.com>
6 years agoARM: omap3: beagle: Enable DM_SERIAL, update distro usage and NAND layout
Derald D. Woods [Mon, 26 Feb 2018 23:52:58 +0000 (17:52 -0600)]
ARM: omap3: beagle: Enable DM_SERIAL, update distro usage and NAND layout

This commit does the following for BeagleBoard{-xM}:

- Enable DM_SERIAL which also enables SPL_DM_SERIAL
- Misc. config updates in support of DM_SERIAL
- Use updated NAND layout (BeagleBoard):

device nand0 <omap2-nand.0>, # parts = 6

#: name size offset
-------------------------------------------------
0: spl 0x00080000(512k) 0x00000000
1: u-boot 0x001e0000(1920k) 0x00080000
2: u-boot-env 0x00020000(128k) 0x00260000
3: dtb 0x00020000(128k) 0x00280000
4: kernel 0x00600000(6m) 0x002a0000
5: rootfs 0x0f760000(-) 0x008a0000
-------------------------------------------------

- Use MMC and UBIFS support from config_distro_bootcmd.h
- Use LEGACY_MMC and NAND support defined in omap3_beagle.h

[MMC(0:1)/extlinux/extlinux.conf]
---8<-------------------------------------------------------------------
default omap3-beagle-mmc-fat
label omap3-beagle-mmc-fat
kernel /zImage
fdt /omap3-beagle.dtb
append console=ttyO2,115200n8 root=/dev/mmcblk0p2 rw rootfstype=ext4 rootwait
---8<-------------------------------------------------------------------

[LEGACY_MMC(0:2)/boot/extlinux/extlinux.conf]
---8<-------------------------------------------------------------------
default omap3-beagle-mmc-ext4
label omap3-beagle-mmc-ext4
kernel /boot/zImage
fdt /boot/omap3-beagle.dtb
append console=ttyO2,115200n8 root=/dev/mmcblk0p2 rw rootfstype=ext4 rootwait
---8<-------------------------------------------------------------------

[UBIFS(ubi0:rootfs)/boot/extlinux/extlinux.conf]
---8<-------------------------------------------------------------------
default omap3-beagle-nand-ubifs
label omap3-beagle-nand-ubifs
kernel /boot/zImage
fdt /boot/omap3-beagle.dtb
append console=ttyO2,115200n8 root=ubi0:rootfs rw ubi.mtd=rootfs rootfstype=ubifs rootwait
---8<-------------------------------------------------------------------

The board's 'bootcmd' can also be overridden with:

CONFIG_BOOTCOMMAND="run findfdt; run mmcboot"
CONFIG_BOOTCOMMAND="run findfdt; run nandboot"
CONFIG_BOOTCOMMAND="run findfdt; run ramboot"

or

CONFIG_BOOTCOMMAND="run findfdt; run somethingelse"

The use of 'run findfdt' is required.

Signed-off-by: Derald D. Woods <woods.technical@gmail.com>
6 years agoARM: omap3: beagle: Set 'mtdids' and 'mtdparts' in board file
Derald D. Woods [Mon, 26 Feb 2018 23:52:57 +0000 (17:52 -0600)]
ARM: omap3: beagle: Set 'mtdids' and 'mtdparts' in board file

BeagleBoard 'xM' does not really have NAND. CONFIG_MTDIDS_DEFAULT and
CONFIG_MTDPARTS_DEFAULT can/should be empty for 'xM'. This commit sets
the defined values if they exist.

Signed-off-by: Derald D. Woods <woods.technical@gmail.com>
6 years agoenv: ti: boot: Get rid of magic numbers
Sam Protsenko [Mon, 26 Feb 2018 21:18:01 +0000 (23:18 +0200)]
env: ti: boot: Get rid of magic numbers

Get the start address and the size of partitions using partition names
rather than partition numbers. This way we can change the partition
table further without changing the boot code.

Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
6 years agocmd: part: Extract common code to separate function
Sam Protsenko [Mon, 26 Feb 2018 21:18:00 +0000 (23:18 +0200)]
cmd: part: Extract common code to separate function

Refactor the code for "part start" and "part size" commands to avoid
code duplication.

Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
6 years agocmd: part: Allow passing partition name to start and size
Sam Protsenko [Mon, 26 Feb 2018 21:17:59 +0000 (23:17 +0200)]
cmd: part: Allow passing partition name to start and size

Allow passing the partition name to "part start" and "part size"
commands, so we can avoid magic numbers in the environment.

Consider one real use-case: in include/environment/ti/boot.h we have
commands like these:

    setenv boot_part 9
    part start mmc ${mmcdev} ${boot_part} boot_start
    part size mmc ${mmcdev} ${boot_part} boot_size
    mmc read ${loadaddr} ${boot_start} ${boot_size}

Now suppose that we have changed the partition table and boot_part now
is 10. We will need to fix commands above. And anyone who relies on
these boot commands, will need to change them accordingly, too (this was
an actual case in our lab while testing Linux boot on Android
environment).

By providing the option to pass partition name instead, we fix mentioned
issue, by eliminating the necessity to use magic numbers.

Signed-off-by: Sam Protsenko <semen.protsenko@linaro.org>
Reviewed-by: Lukasz Majewski <lukma@denx.de>
6 years agotest/py: highlight warnings in the log summary
Stephen Warren [Tue, 20 Feb 2018 19:51:55 +0000 (12:51 -0700)]
test/py: highlight warnings in the log summary

Currently, if a test emits a warning message but otherwise passes, there's
no indication of this in the log summary, which can lead to warnings being
missed. Enhance the test logic to explicitly mention warnings in otherwise
passing tests, and not to collapse the log sections for tests with
warnings, so that they're more easily seen when scanning the log.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
6 years agotest/py: add MMC/SD block read test
Stephen Warren [Tue, 20 Feb 2018 19:51:54 +0000 (12:51 -0700)]
test/py: add MMC/SD block read test

Add a standalone MMC block read test. This allows direct testing of MMC
access rather than relying on doing so as a side-effect of e.g. DFU or
UMS testing, which may not be enabled on all platforms.

Signed-off-by: Stephen Warren <swarren@nvidia.com>
6 years agodisk: part: use common api to lookup part driver
Kever Yang [Sat, 10 Feb 2018 09:55:38 +0000 (17:55 +0800)]
disk: part: use common api to lookup part driver

Do not need to scan disk every time when we get part info
by name.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
[trini: Fix build in !CONFIG_HAVE_BLOCK_DEVICE case]
Signed-off-by: Tom Rini <trini@konsulko.com>
6 years agodisk: part: scan the disk if the part_type is unknown
Kever Yang [Sat, 10 Feb 2018 09:55:37 +0000 (17:55 +0800)]
disk: part: scan the disk if the part_type is unknown

If a DUT do not have partition table, and we write one with 'gpt write'
cmd, we should able to list the partition with 'part list' cmd.
It's reasonable to scan the disk again if the initial part_type is
unknown in case we just write a new one into disk.

Signed-off-by: Kever Yang <kever.yang@rock-chips.com>
6 years agomach-stm32: Use default memory map as background region
Patrice Chotard [Wed, 28 Feb 2018 16:15:00 +0000 (17:15 +0100)]
mach-stm32: Use default memory map as background region

On linux kernel side, on STM32F7 and STM32H7 SoCs, DMA requires
uncachable regions. These regions are defined in DT.
Since kernel linux v4.15, on ARMv7-M Cortex, kernel is able
to configure MPU regions depending on DT settings.

As kernel is able to configure MPU, this allows to remove
MPU region settings in bootloader.

On Cortex M processors, MPU allows to use a default memory map.
(see B3.5.4 MPU Control Register, MPU_CTRL in
https://developer.arm.com/products/architecture/m-profile/docs/ddi0403/latest/armv7-m-architecture-reference-manual)
Use the default memory map as background region for all STM32 SoCs
family with an additional MPU region corresponding to the SDRAM area.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
6 years agoconfigs: stm32: Enable DOS_PARTITION for STM32F4/F7 boards
Patrice Chotard [Wed, 28 Feb 2018 15:49:55 +0000 (16:49 +0100)]
configs: stm32: Enable DOS_PARTITION for STM32F4/F7 boards

Enable DOS_PARTITION for boards STM32F429-eval, STM32F469-disco
and STM32F746-disco.
This allows to read FAT partition on mmc.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
6 years agoARM: dts: Add support for stm32f746-evaluation board support
Patrice Chotard [Fri, 16 Feb 2018 12:27:03 +0000 (13:27 +0100)]
ARM: dts: Add support for stm32f746-evaluation board support

This board offers:
  _ STM32F746NGH6 microcontroller with 1 Mbyte Flash and 320+4 Kbytes RAM
  _ Six 5 V power supply options:
        Power jack
        ST-LINK/V2-1 USB connector
        User USB HS connector
        User USB FS1 connector
        User USB FS2 connector
        Daughterboard
  _ SAI Audio DAC, stereo audio jack which supports headset with microphone
  _ Stereo digital microphone, audio jack connector used to connect
    external speakers
  _ 2 Gbytes (or more) SDMMC interface microSD card
  _ RF-EEPROM on I2C compatible serial interface
  _ RS-232 communication
  _ IrDA transceiver
  _ JTAG/SWD and ETM trace debug support, ST-LINK/V2-1 embedded
  _ IEEE-802.3-2002 compliant Ethernet connector
  _ Camera module
  _ 8Mx32 bit SDRAM, 1Mx16 bit SRAM & 8Mx16 bit Nor Flash
  _ 512 Mbits QuadSPI Nor Flash
  _ 5.7 inch 640x480 pixel TFT color LCD with capacitive touch panel
  _ Joystick with 4-direction control and selector
  _ Reset, WakeUp/Tamper or key button
  _ 4 color user LEDs
  _ Extension connectors & memory connectors for daughterboard or
    wrapping board
  _ USB OTG HS and FS with Micro-AB connectors
  _ RTC with backup battery
  _ CAN 2.0A/B compliant connection
  _ Potentiometer
  _ Motor control connector

More detailed information are available here :
http://www.st.com/en/evaluation-tools/stm32746g-eval.html

To compile stm32f746-eval board, use same defconfig as
stm32f746-disco, the only difference is to pass
"DEVICE_TREE=stm32746g-eval".

Signed-off-by: Christophe Priouzeau <christophe.priouzeau@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
Acked-by: Vikas Manocha <vikas.manocha@st.com>
6 years agoarch-stm32: Clean arch-stm32f7/syscfg.h
Patrice Chotard [Fri, 9 Feb 2018 12:09:57 +0000 (13:09 +0100)]
arch-stm32: Clean arch-stm32f7/syscfg.h

Remove all unused defines

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
6 years agoarch-stm32: Remove stm32_periph.h
Patrice Chotard [Fri, 9 Feb 2018 12:09:56 +0000 (13:09 +0100)]
arch-stm32: Remove stm32_periph.h

Remove arch/arm/include/asm/arch-stm32fx/stm32_periph.h
as all defines or enums are no more used.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
6 years agoarch-stm32: Factorize stm32.h for STM32F4 and F7
Patrice Chotard [Fri, 9 Feb 2018 12:09:55 +0000 (13:09 +0100)]
arch-stm32: Factorize stm32.h for STM32F4 and F7

For STM32F4 and F7 SoCx family, a specific stm32.h file exists.
Some common defines are duplicated or even unused in each of
these stm32.h.

Factorize all common definition in arch/arm/include/asm/stm32f.h and keep
specific definitions in each arch/arm/include/asm/arch-stm32fx/stm32.h.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
6 years agoarch-stm32: Move gpio.h for STM32 SoCs in include/asm/
Patrice Chotard [Fri, 9 Feb 2018 12:09:54 +0000 (13:09 +0100)]
arch-stm32: Move gpio.h for STM32 SoCs in include/asm/

Instead to have 3 identical gpio.h for all STM32 SoCs,
migrate them in one file in include/asm.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
6 years agoarch-stm32f4: Remove fmc.h file
Patrice Chotard [Fri, 9 Feb 2018 12:09:53 +0000 (13:09 +0100)]
arch-stm32f4: Remove fmc.h file

fmc.h file is no more used, remove it.
All FMC related defines are declared in drivers/ram/stm32_sdram.c
which is common to all STM32 SoCs.

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
6 years agoARM: dts: stm32: limit sdio frequency to 14Mhz for stm32f429i-eval
Patrice Chotard [Thu, 8 Feb 2018 16:25:19 +0000 (17:25 +0100)]
ARM: dts: stm32: limit sdio frequency to 14Mhz for stm32f429i-eval

This avoids the following errors while reading on mmc:
  Read data bytes CRC error: 0x2
  switch to partitions #0, OK
  mmc0 is current device
  Read data bytes CRC error: 0x82002
  ** Unrecognized filesystem type **

Signed-off-by: Patrice Chotard <patrice.chotard@st.com>
6 years agoclk: clk_stm32f: Add DSI clock support
Patrice Chotard [Thu, 8 Feb 2018 16:20:51 +0000 (17:20 +0100)]
clk: clk_stm32f: Add DSI clock support

DSI clock is available on STM32F769-disco and
STM32F469-disco board.

Signed-off-by: Yannick Fertre <yannick.fertre@st.com>
Signed-off-by: Patrice Chotard <patrice.chotard@st.com>