]> git.sur5r.net Git - u-boot/log
u-boot
8 years agodm: core: Match compatible strings in order of priority
Paul Burton [Thu, 8 Sep 2016 06:47:36 +0000 (07:47 +0100)]
dm: core: Match compatible strings in order of priority

Device model drivers have previously been matched to FDT nodes by virtue
of being the first driver in the driver list to be compatible with the
node. This ignores the fact that compatible strings in the device tree
are listed in order of priority - that is, if we have a node with 2
compatible strings & a driver that matches each then we should always
probe the driver that matches the first compatible string.

Fix this by looping through the compatible strings for a node when
attempting to bind it in lists_bind_fdt and checking each driver for
a match of the first string, then each driver for a match of the second
string etc. Effectively this inverts the loops over compatible strings &
drivers.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agodm: regmap: Implement simple regmap_read & regmap_write
Paul Burton [Thu, 8 Sep 2016 06:47:35 +0000 (07:47 +0100)]
dm: regmap: Implement simple regmap_read & regmap_write

The regmap_read & regmap_write functions were previously declared in
regmap.h but not implemented anywhere. The regmap implementation &
commit message of 6f98b7504f70 ("dm: Add support for register maps
(regmap)") indicate that only memory mapped accesses are supported for
now, so providing simple implementations of regmap_read & regmap_write
is trivial. The access size is presumed to be 4 bytes & endianness is
presumed native, which are the defaults for the regmap code in Linux.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 years agonet: pch_gbe: Make 64 bit safe
Paul Burton [Thu, 8 Sep 2016 06:47:34 +0000 (07:47 +0100)]
net: pch_gbe: Make 64 bit safe

The pch_gbe driver previously casted pointers to & from unsigned 32 bit
integers in many locations. This breaks the driver on 64 bit systems,
producing streams of compiler warnings about mismatched pointer &
integer sizes and then failing to keep track of addresses correctly at
runtime.

Fix the driver for 64 bit systems by using unsigned longs in place of
the previously used 32 bit integers.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
8 years agonet: pch_gbe: Use dm_pci_map_bar to discover MMIO base
Paul Burton [Thu, 8 Sep 2016 06:47:33 +0000 (07:47 +0100)]
net: pch_gbe: Use dm_pci_map_bar to discover MMIO base

Reading the PCI BAR & converting the result to a physical address is not
safe across all architectures. For example on MIPS the virtual:physical
mapping is not 1:1, so we cannot directly make use of the physical
address.

Use the more generic BAR-mapping function dm_pci_map_bar to discover the
MMIO base address, which should work across architectures.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
8 years agopci: Flip condition for detecting non-PCI parent devices
Paul Burton [Thu, 8 Sep 2016 06:47:32 +0000 (07:47 +0100)]
pci: Flip condition for detecting non-PCI parent devices

In pci_uclass_pre_probe an attempt is made to detect whether the parent
of a device is a PCI device and that the device is thus a bridge. This
was being done by checking whether the parent of the device is of the
UCLASS_ROOT class. This causes problems if the PCI controller is a child
of some other non-PCI node, for example a simple-bus node.

For example, if the device tree contains something like the following
then pci_uclass_pre_probe would incorrectly believe that the PCI
controller is a bridge, with a PCI parent:

  / {
    some_child {
      compatible = "simple-bus";
      #address-cells = <1>;
      #size-cells = <1>;
      ranges = <>;

      pci_controller: pci@10000000 {
        compatible = "my-pci-controller";
        device_type = "pci";
        reg = <0x10000000 0x2000000>;
      };
    };
  };

Avoid this incorrect detection of bridges by instead checking whether
the parent devices class is UCLASS_PCI and treating a device as a bridge
when this is true, making use of device_is_on_pci_bus to perform this
test.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 years agopci: xilinx: Add a driver for Xilinx AXI to PCIe bridge
Paul Burton [Thu, 8 Sep 2016 06:47:31 +0000 (07:47 +0100)]
pci: xilinx: Add a driver for Xilinx AXI to PCIe bridge

This patch adds a driver for the Xilinx AXI bridge for PCI express, an
IP block which can be used on some generations of Xilinx FPGAs. This is
mostly a case of implementing PCIe ECAM specification, but with some
quirks about what devices are valid to access.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 years agodt-bindings: Add interrupt-controller/mips-gic.h header
Paul Burton [Thu, 8 Sep 2016 06:47:30 +0000 (07:47 +0100)]
dt-bindings: Add interrupt-controller/mips-gic.h header

Import a copy of the dt-bindings/interrupt-controller/mips-gic.h header
from Linux, such that we can use device trees which include it without
modification.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoserial: ns16550: Support clocks via phandle
Paul Burton [Thu, 8 Sep 2016 06:47:29 +0000 (07:47 +0100)]
serial: ns16550: Support clocks via phandle

Previously ns16550 compatible UARTs probed via device tree have needed
their device tree nodes to contain a clock-frequency property. An
alternative to this commonly used with Linux is to reference a clock via
a phandle. This patch allows U-Boot to support that, retrieving the
clock frequency by probing the appropriate clock device.

For example, a system might choose to provide the UART base clock as a
reference to a clock common to multiple devices:

  sys_clk: clock {
    compatible = "fixed-clock";
    #clock-cells = <0>;
    clock-frequency = <10000000>;
  };

  uart0: uart@10000000 {
    compatible = "ns16550a";
    reg = <0x10000000 0x1000>;
    clocks = <&sys_clk>;
  };

  uart1: uart@10000000 {
    compatible = "ns16550a";
    reg = <0x10001000 0x1000>;
    clocks = <&sys_clk>;
  };

This removes the need for the frequency information to be duplicated in
multiple nodes and allows the device tree to be more descriptive of the
system.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 years agoclk: Use dummy clk_get_by_* functions when CONFIG_CLK is disabled
Paul Burton [Thu, 8 Sep 2016 06:47:28 +0000 (07:47 +0100)]
clk: Use dummy clk_get_by_* functions when CONFIG_CLK is disabled

The implementations of clk_get_by_index & clk_get_by_name are only
available when CONFIG_CLK is enabled. Provide the dummies when this is
not the case in order to avoid build failures.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 years agoMIPS: Ensure cache ops complete in mips_cache_reset
Paul Burton [Wed, 21 Sep 2016 10:18:59 +0000 (11:18 +0100)]
MIPS: Ensure cache ops complete in mips_cache_reset

Ensure that cache operations complete before returning from
mips_cache_reset by placing a completion barrier (sync instruction)
before the return. Without this there is no guarantee that the cache ops
will complete before any subsequent memory accesses, since they are
indexed cache ops & thus not implicitly ordered with memory accesses.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoMIPS: Clear hazard between TagLo writes & cache ops
Paul Burton [Wed, 21 Sep 2016 10:18:58 +0000 (11:18 +0100)]
MIPS: Clear hazard between TagLo writes & cache ops

Writing to the coprocessor 0 TagLo registers introduces an execution
hazard in that we need that write to complete before any cache
instructions execute. Ensure that hazard is cleared by inserting an ehb
instruction between the TagLo writes & cache op loop.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoMIPS: Ensure Config.K0=2 applies before any memory accesses
Paul Burton [Wed, 21 Sep 2016 10:18:57 +0000 (11:18 +0100)]
MIPS: Ensure Config.K0=2 applies before any memory accesses

During boot we set Config.K0=2 (uncached) such that any accesses to the
kseg0 memory region are performed uncached before the caches are
initialised. This write to the Config register introduces an execution
hazard between it & any following memory accesses (such as the load of
_gp), which we need to clear in order to ensure those memory accesses
are actually performed uncached. Clear this execution hazard with the
insertion of an ehb execution hazard barrier instruction.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoMIPS: Malta: Enable CM & L2 support
Paul Burton [Wed, 21 Sep 2016 10:18:56 +0000 (11:18 +0100)]
MIPS: Malta: Enable CM & L2 support

Enable support for the MIPS Coherence Manager & L2 caches on the MIPS
Malta board, removing the need for us to attempt to bypass the L2 during
boot (which would fail with recent CPUs that expose L2 config via the CM
anyway).

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoMIPS: Join the coherent domain when a CM is present
Paul Burton [Wed, 21 Sep 2016 10:18:55 +0000 (11:18 +0100)]
MIPS: Join the coherent domain when a CM is present

MIPS Linux expects the bootloader to leave the boot CPU a member of the
coherent domain when running on a system with a CM, and we will need to
do so if we wish to make use of IOCUs to have cache-coherent DMA in
U-Boot (and on some systems there is no choice in that matter). When a
CM is present, join the coherent domain after completing cache
initialisation.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoMIPS: L2 cache support
Paul Burton [Wed, 21 Sep 2016 10:18:54 +0000 (11:18 +0100)]
MIPS: L2 cache support

This patch adds support for initialising & maintaining L2 caches on MIPS
systems. The L2 cache configuration may be advertised through either
coprocessor 0 or the MIPS Coherence Manager depending upon the system,
and support for both is included.

If the L2 can be bypassed then we bypass it early in boot & initialise
the L1 caches first, such that we can start making use of the L1
instruction cache as early as possible. Otherwise we initialise the L2
first such that the L1s have no opportunity to generate access to the
uninitialised L2.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoMIPS: Map CM Global Control Registers
Paul Burton [Wed, 21 Sep 2016 10:18:53 +0000 (11:18 +0100)]
MIPS: Map CM Global Control Registers

Map the Global Control Registers (GCRs) provided by the MIPS Coherence
Manager (CM) in preparation for using some of them in later patches.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoMIPS: Define register names for cache init
Paul Burton [Wed, 21 Sep 2016 10:18:52 +0000 (11:18 +0100)]
MIPS: Define register names for cache init

Define names for registers holding cache sizes throughout
mips_cache_reset, in order to make the code easier to read & allow for
changing register assignments more easily.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoMIPS: If we don't need DDR for cache init, init cache first
Paul Burton [Wed, 21 Sep 2016 10:18:51 +0000 (11:18 +0100)]
MIPS: If we don't need DDR for cache init, init cache first

On systems where cache initialisation doesn't require zeroed memory (ie.
systems where CONFIG_SYS_MIPS_CACHE_INIT_RAM_LOAD is not defined)
perform cache initialisation prior to lowlevel_init & DDR
initialisation. This allows for DDR initialisation code to run cached &
thus significantly faster.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoMIPS: Preserve Config implementation-defined bits
Paul Burton [Wed, 21 Sep 2016 10:18:50 +0000 (11:18 +0100)]
MIPS: Preserve Config implementation-defined bits

The coprocessor 0 Config register includes 9 implementation defined
bits, which in some processors do things like enable write combining or
other functionality. We ought not to wipe them to 0 during boot. Rather
than doing so, preserve their value & only clear the bits standardised
by the MIPS architecture.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoMIPS: Enable use of the instruction cache earlier
Paul Burton [Wed, 21 Sep 2016 10:18:49 +0000 (11:18 +0100)]
MIPS: Enable use of the instruction cache earlier

Enable use of the instruction cache immediately after it has been
initialised. This will only take effect if U-Boot was linked to run from
kseg0 rather than kseg1, but when this is the case the data cache
initialisation code will run cached & thus significantly faster.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoMIPS: Probe cache line sizes once during boot
Paul Burton [Wed, 21 Sep 2016 10:18:48 +0000 (11:18 +0100)]
MIPS: Probe cache line sizes once during boot

Rather than probing the cache line sizes on every call of any cache
maintenance function, probe them once during boot & store the values in
the global data structure for later use. This will reduce the overhead
of the cache maintenance functions, which isn't a big deal yet but
becomes more important once L2 caches which may expose their properties
via coprocessor 2 or the CM are supported.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoMIPS: ath79: Use mach_cpu_init instead of arch_cpu_init
Paul Burton [Wed, 21 Sep 2016 10:18:47 +0000 (11:18 +0100)]
MIPS: ath79: Use mach_cpu_init instead of arch_cpu_init

In order to prepare for MIPS arch code making use of arch_cpu_init in a
later patch, stop using it from ath79 SoC code & instead use the new
mach_cpu_init which is provided for this purpose.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
8 years agoboard_f: Add a mach_cpu_init callback
Paul Burton [Wed, 21 Sep 2016 10:18:46 +0000 (11:18 +0100)]
board_f: Add a mach_cpu_init callback

Currently we have a mismash of architectures which use arch_cpu_init
from architecture-wide code (arc, avr32, blackfin, mips, nios2, xtensa)
and architectures which use arch_cpu_init from machine/SoC level code
(arm, x86).

In order to clean this mess up & allow for both use cases, introduce a
new mach_cpu_init callback which is run immediately after arch_cpu_init.
This will allow for architectures to have arch-wide code without needing
individual machines to all implement their own arch_cpu_init with a call
to some common function.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 years agomips: Add MIPSfpga platform support
Zubair Lutfullah Kakakhel [Fri, 29 Jul 2016 14:11:20 +0000 (15:11 +0100)]
mips: Add MIPSfpga platform support

MIPSfpga is an FPGA based dev platform.

In a nutshell, its a microAptiv cpu core with lots of Xilinx IP blocks

The FPGA dev board used is the Nexys4DDR board by Digilent.

For more information, check the Readme file in board/imgtec/xilfpga

Signed-off-by: Zubair Lutfullah Kakakhel <Zubair.Kakakhel@imgtec.com>
Reviewed-by: Paul Burton <paul.burton@imgtec.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
8 years agomips: xilfpga: Add device tree files
Zubair Lutfullah Kakakhel [Fri, 29 Jul 2016 14:11:19 +0000 (15:11 +0100)]
mips: xilfpga: Add device tree files

Mostly the same as the Kernel upstream device tree file except for

- alias for the serial console node
- ethernet node as the ethernet stuff isn't upstream on kernel.org yet
- uart clock-frequency passed directly in the node

Signed-off-by: Zubair Lutfullah Kakakhel <Zubair.Kakakhel@imgtec.com>
Reviewed-by: Paul Burton <paul.burton@imgtec.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
8 years agonet: emaclite: Enable driver for MIPS
Zubair Lutfullah Kakakhel [Wed, 27 Jul 2016 11:25:09 +0000 (12:25 +0100)]
net: emaclite: Enable driver for MIPS

Signed-off-by: Zubair Lutfullah Kakakhel <Zubair.Kakakhel@imgtec.com>
Reviewed-by: Paul Burton <paul.burton@imgtec.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
8 years agonet: emaclite: use __raw_readl/writel instead of weird define
Zubair Lutfullah Kakakhel [Wed, 27 Jul 2016 11:25:08 +0000 (12:25 +0100)]
net: emaclite: use __raw_readl/writel instead of weird define

out_be32 and in_be32 are actually #defined to little endian
writel/readl in arch/microblaze.

Just use __raw_writel/readl instead. That is also what is used
in the Linux kernel driver for this IP block

Tested on MIPSfpga. Can tftp a kernel.

Signed-off-by: Zubair Lutfullah Kakakhel <Zubair.Kakakhel@imgtec.com>
Reviewed-by: Paul Burton <paul.burton@imgtec.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
8 years agonet: emaclite: Use ioremap_nocache
Zubair Lutfullah Kakakhel [Wed, 27 Jul 2016 11:25:07 +0000 (12:25 +0100)]
net: emaclite: Use ioremap_nocache

Virtual to physical mapping isn't necessarily 1:1 for all architectures

Using ioremap_nocache allows for the arch code to translate the
physical address to a virtual address.

Signed-off-by: Zubair Lutfullah Kakakhel <Zubair.Kakakhel@imgtec.com>
Reviewed-by: Paul Burton <paul.burton@imgtec.com>
Reviewed-by: Daniel Schwierzeck <daniel.schwierzeck@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
8 years agoRevert "Increase default of CONFIG_SYS_MALLOC_F_LEN for SPL_OF_CONTROL"
Masahiro Yamada [Mon, 19 Sep 2016 12:40:26 +0000 (21:40 +0900)]
Revert "Increase default of CONFIG_SYS_MALLOC_F_LEN for SPL_OF_CONTROL"

This reverts commit 90c08d9e08c7a108ab904f3bbdeb558081757892.

I took a closer look at this after the commit was applied, and found
CONFIG_SYS_MALLOC_F_LEN=0x2000 was too much.  8KB memory for SPL is
actually too big for some boards.  Perhaps 0x800 is enough, but the
situation varies board by board.

Let's postpone our decision until we come up with a better idea.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
8 years agoA20-OLinuXino-Lime2: Enable USB gadget support
Tom Rini [Mon, 19 Sep 2016 14:03:32 +0000 (10:03 -0400)]
A20-OLinuXino-Lime2: Enable USB gadget support

Based on A13-OLinuXino, enable DFU and UMS support.

Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
8 years agoMerge branch 'master' of git://git.denx.de/u-boot-uniphier
Tom Rini [Sun, 18 Sep 2016 16:12:04 +0000 (12:12 -0400)]
Merge branch 'master' of git://git.denx.de/u-boot-uniphier

8 years agoMerge branch 'master' of git://www.denx.de/git/u-boot-sunxi
Tom Rini [Sun, 18 Sep 2016 16:11:50 +0000 (12:11 -0400)]
Merge branch 'master' of git://www.denx.de/git/u-boot-sunxi

8 years agoMerge branch 'master' of git://www.denx.de/git/u-boot-arc
Tom Rini [Sun, 18 Sep 2016 16:10:54 +0000 (12:10 -0400)]
Merge branch 'master' of git://www.denx.de/git/u-boot-arc

8 years agoARM: uniphier: update DRAM init code for LD20 SoC
Masahiro Yamada [Fri, 16 Sep 2016 18:33:12 +0000 (03:33 +0900)]
ARM: uniphier: update DRAM init code for LD20 SoC

Import the latest version from the Diag software.

  - Support LD21 SoC (including DDR chips in the package)
  - Per-board granule adjustment for both reference and TV boards
  - Misc cleanups

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: add PLL init code for LD20 SoC
Masahiro Yamada [Fri, 16 Sep 2016 18:33:11 +0000 (03:33 +0900)]
ARM: uniphier: add PLL init code for LD20 SoC

Initialize the DPLL (PLL for DRAM) in SPL, and others in U-Boot
proper.  Split the common code into pll-base-ld20.c for easier
re-use.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: collect clock/PLL init code into a single directory
Masahiro Yamada [Fri, 16 Sep 2016 18:33:10 +0000 (03:33 +0900)]
ARM: uniphier: collect clock/PLL init code into a single directory

Now PLLs for DRAM controller are initialized in SPL, and the others
in U-Boot proper.  Setting up all of them in a single directory will
be helpful when we want to share code between SPL and U-Boot proper.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: move PLL init code to U-Boot proper where possible
Masahiro Yamada [Fri, 16 Sep 2016 18:33:09 +0000 (03:33 +0900)]
ARM: uniphier: move PLL init code to U-Boot proper where possible

The PLL for the DRAM interface must be initialized in SPL, but the
others can be delayed until U-Boot proper.  Move them from SPL to
U-Boot proper to save the precious SPL memory footprint.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: rename CONFIG_DPLL_SSC_RATE_1PER
Masahiro Yamada [Fri, 16 Sep 2016 18:33:08 +0000 (03:33 +0900)]
ARM: uniphier: rename CONFIG_DPLL_SSC_RATE_1PER

Basically, this should not be configured by users.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: move XIRQ pin-mux settings of LD11/LD20
Masahiro Yamada [Fri, 16 Sep 2016 18:33:07 +0000 (03:33 +0900)]
ARM: uniphier: move XIRQ pin-mux settings of LD11/LD20

This is the last code in the mach-uniphier/pinctrl/ directory.
Push the remaining code out to delete the directory entirely.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: consolidate System Bus pin-mux settings for LD11/LD20
Masahiro Yamada [Fri, 16 Sep 2016 18:33:06 +0000 (03:33 +0900)]
ARM: uniphier: consolidate System Bus pin-mux settings for LD11/LD20

Use the pin-mux data in the pinctrl drivers by directly calling
pinctrl_generic_set_state().

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: dts: uniphier: include System Bus pin group node in SPL DT
Masahiro Yamada [Fri, 16 Sep 2016 18:33:05 +0000 (03:33 +0900)]
ARM: dts: uniphier: include System Bus pin group node in SPL DT

This will be needed for setting up the System Bus pin-mux via the
LD11/LD20 pinctrl driver.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: consolidate NAND pin-mux settings
Masahiro Yamada [Fri, 16 Sep 2016 18:33:04 +0000 (03:33 +0900)]
ARM: uniphier: consolidate NAND pin-mux settings

The NAND subsystem has not supported the Driver Model yet, but the
NAND pin-mux data are already in the pinctrl drivers.  Use them by
calling pinctrl_generic_set_state() directly.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: remove ad-hoc pin-mux code for sLD3
Masahiro Yamada [Fri, 16 Sep 2016 18:33:03 +0000 (03:33 +0900)]
ARM: uniphier: remove ad-hoc pin-mux code for sLD3

These settings are nicely cared by the pinctrl driver now.  Remove.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: remove redundant pin-muxing for EA24 pin of sLD3 SoC
Masahiro Yamada [Fri, 16 Sep 2016 18:33:02 +0000 (03:33 +0900)]
ARM: uniphier: remove redundant pin-muxing for EA24 pin of sLD3 SoC

This is enabled by default for all the supported boot modes.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: uniphier: select PINCTRL and SPL_PINCTRL
Masahiro Yamada [Fri, 16 Sep 2016 18:33:01 +0000 (03:33 +0900)]
ARM: uniphier: select PINCTRL and SPL_PINCTRL

Now all UniPhier SoCs support a pinctrl driver.  Select (SPL_)PINCTRL
since it is mandatory even for base use.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agoARM: dts: uniphier: add pinctrl device node and pinctrl properties
Masahiro Yamada [Fri, 16 Sep 2016 18:33:00 +0000 (03:33 +0900)]
ARM: dts: uniphier: add pinctrl device node and pinctrl properties

DT-side updates to make pinctrl on sLD3 SoC really available.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agopinctrl: uniphier: add UniPhier sLD3 pinctrl driver
Masahiro Yamada [Fri, 16 Sep 2016 18:32:59 +0000 (03:32 +0900)]
pinctrl: uniphier: add UniPhier sLD3 pinctrl driver

Add pin-mux support for UniPhier sLD3 SoC.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agopinctrl: uniphier: support 4bit-width pin-mux register capability
Masahiro Yamada [Fri, 16 Sep 2016 18:32:58 +0000 (03:32 +0900)]
pinctrl: uniphier: support 4bit-width pin-mux register capability

On LD4 SoC or later, the pin-mux registers are 8bit wide, while 4bit
wide on sLD3 SoC.  Support it for the sLD3 pinctrl driver.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
8 years agosunxi: Enable USB gadget support for Sinlinx SinA33
Chen-Yu Tsai [Wed, 14 Sep 2016 02:26:36 +0000 (10:26 +0800)]
sunxi: Enable USB gadget support for Sinlinx SinA33

Sinlinx SinA33 has a USB OTG port, but VBUS is controlled manually from
a jumper pad.

Enable OTG in gadget mode, as well as the download gadget and related
functions.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
8 years agosunxi: Enable USB host support for Sinlinx SinA33
Chen-Yu Tsai [Wed, 14 Sep 2016 02:26:35 +0000 (10:26 +0800)]
sunxi: Enable USB host support for Sinlinx SinA33

Sinlinx SinA33 has 1 USB host port. Enable EHCI_HCD support for it.
Also enable USB mass storage support so we can access USB sticks.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
8 years agosunxi: Add mmc0 card detect pin for Sinlinx SinA33
Chen-Yu Tsai [Wed, 14 Sep 2016 02:26:34 +0000 (10:26 +0800)]
sunxi: Add mmc0 card detect pin for Sinlinx SinA33

Sinlinx SinA33 uses PB4 for mmc0 card detect.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
8 years agosunxi: Add defconfig and dts for the NanoPi NEO
Jelle van der Waa [Tue, 13 Sep 2016 16:03:23 +0000 (18:03 +0200)]
sunxi: Add defconfig and dts for the NanoPi NEO

The NanoPi NEO is a simple h3 board with 512MB RAM, ethernet, one usb
and one usb OTG connector.

Signed-off-by: Jelle van der Waa <jelle@vdwaa.nl>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
8 years agosunxi: musb: Re-init musb controller on repeated probe calls
Hans de Goede [Sat, 17 Sep 2016 14:02:38 +0000 (16:02 +0200)]
sunxi: musb: Re-init musb controller on repeated probe calls

With sunxi-musb musb_lowlevel_init() can fail when a charger; or no cable
is plugged into the otg port.

To avoid leaking the struct musb allocated by musb_init_controller()
on repeated musb_usb_probe() calls, we were caching its result.
But musb_init_controller() does more, such as calling sunxi_musb_init()
which enables the clocks.

Not calling sunxi_musb_init() causes the musb controller to stop working
after a "usb reset" since that calls musb_usb_remove() which disables the
clocks.

This commit fixes this by removing the caching of the struct returned
from musb_init_controller(), it replaces this by free-ing the allocated
memory in musb_usb_remove() and calling musb_usb_remove() on
musb_usb_probe() errors to ensure proper cleanup.

While at it also make musb_usb_probe() and musb_usb_remove() static.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
8 years agosunxi: musb: Power off OTG port VBUS when disabled
Chen-Yu Tsai [Wed, 7 Sep 2016 06:25:21 +0000 (14:25 +0800)]
sunxi: musb: Power off OTG port VBUS when disabled

The Linux kernel musb driver expects VBUS to be off while initializing
musb. Having it on results in a repeating string of warnings, followed
by an unusable peripheral. The peripheral is only usable after
physically removing the OTG adapter, letting musb reset its state.

This partially reverts commit c9f8947e6604 ("sunxi: usb-phy: Never
power off the usb ports")

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
8 years agosunxi: axp2xx: disable ldoio0/1 at boot
Hans de Goede [Mon, 12 Sep 2016 07:52:52 +0000 (09:52 +0200)]
sunxi: axp2xx: disable ldoio0/1 at boot

When cold-booting the ldoio0/1 regulators are always off / the
gpios are always at tristate. But when re-booting from android these
are sometimes on. Disable them at axp_init time (iow as early as possible)
to remove this difference between a cold boot and a reboot.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
8 years agoMakefile: Give a build error if ad-hoc CONFIG options are added
Simon Glass [Wed, 14 Sep 2016 03:44:07 +0000 (21:44 -0600)]
Makefile: Give a build error if ad-hoc CONFIG options are added

New CONFIG options should be added via Kconfig. To help prevent new ad-hoc
CONFIGs from being added, give a build error when these are detected.

Signed-off-by: Simon Glass <sjg@chromium.org>
Acked-by: Stephen Warren <swarren@nvidia.com>
Tested-by: Stephen Warren <swarren@nvidia.com>
8 years agoKconfig: Add a whitelist of ad-hoc CONFIG options
Simon Glass [Wed, 14 Sep 2016 03:44:06 +0000 (21:44 -0600)]
Kconfig: Add a whitelist of ad-hoc CONFIG options

Add a list of ad-hoc CONFIG options that don't use Kconfig. This can be used
to check that new ones are not being added.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_YMODEM_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:19:03 +0000 (23:19 -0600)]
Convert CONFIG_SPL_YMODEM_SUPPORT to Kconfig

Convert CONFIG_SPL_YMODEM_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_WATCHDOG_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:19:02 +0000 (23:19 -0600)]
Convert CONFIG_SPL_WATCHDOG_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_USB_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:19:01 +0000 (23:19 -0600)]
Convert CONFIG_SPL_USB_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_USB_HOST_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:19:00 +0000 (23:19 -0600)]
Convert CONFIG_SPL_USB_HOST_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_USBETH_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:59 +0000 (23:18 -0600)]
Convert CONFIG_SPL_USBETH_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_SPI_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:58 +0000 (23:18 -0600)]
Convert CONFIG_SPL_SPI_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_SPI_FLASH_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:57 +0000 (23:18 -0600)]
Convert CONFIG_SPL_SPI_FLASH_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_SERIAL_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:56 +0000 (23:18 -0600)]
Convert CONFIG_SPL_SERIAL_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_SATA_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:55 +0000 (23:18 -0600)]
Convert CONFIG_SPL_SATA_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_POWER_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:54 +0000 (23:18 -0600)]
Convert CONFIG_SPL_POWER_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoRemove CONFIG_SPL_PINCTRL_SUPPORT
Simon Glass [Tue, 13 Sep 2016 13:05:50 +0000 (07:05 -0600)]
Remove CONFIG_SPL_PINCTRL_SUPPORT

This option is not used. Drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_ONENAND_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:52 +0000 (23:18 -0600)]
Convert CONFIG_SPL_ONENAND_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_NOR_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:51 +0000 (23:18 -0600)]
Convert CONFIG_SPL_NOR_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_NET_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:50 +0000 (23:18 -0600)]
Convert CONFIG_SPL_NET_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_NET_VCI_STRING to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:49 +0000 (23:18 -0600)]
Convert CONFIG_SPL_NET_VCI_STRING to Kconfig

This converts the following to Kconfig:
   CONFIG_SPL_NET_VCI_STRING

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_NAND_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:48 +0000 (23:18 -0600)]
Convert CONFIG_SPL_NAND_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_MUSB_NEW_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:47 +0000 (23:18 -0600)]
Convert CONFIG_SPL_MUSB_NEW_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_MTD_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:46 +0000 (23:18 -0600)]
Convert CONFIG_SPL_MTD_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_MPC8XXX_INIT_DDR_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:45 +0000 (23:18 -0600)]
Convert CONFIG_SPL_MPC8XXX_INIT_DDR_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_MMC_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:44 +0000 (23:18 -0600)]
Convert CONFIG_SPL_MMC_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_LIBGENERIC_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:43 +0000 (23:18 -0600)]
Convert CONFIG_SPL_LIBGENERIC_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_LIBDISK_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:42 +0000 (23:18 -0600)]
Convert CONFIG_SPL_LIBDISK_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_LIBCOMMON_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:41 +0000 (23:18 -0600)]
Convert CONFIG_SPL_LIBCOMMON_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_I2C_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:40 +0000 (23:18 -0600)]
Convert CONFIG_SPL_I2C_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_GPIO_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:39 +0000 (23:18 -0600)]
Convert CONFIG_SPL_GPIO_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_FAT_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:38 +0000 (23:18 -0600)]
Convert CONFIG_SPL_FAT_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_EXT_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:37 +0000 (23:18 -0600)]
Convert CONFIG_SPL_EXT_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_ETH_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:36 +0000 (23:18 -0600)]
Convert CONFIG_SPL_ETH_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_ENV_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:35 +0000 (23:18 -0600)]
Convert CONFIG_SPL_ENV_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_DRIVERS_MISC_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:34 +0000 (23:18 -0600)]
Convert CONFIG_SPL_DRIVERS_MISC_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_DMA_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:33 +0000 (23:18 -0600)]
Convert CONFIG_SPL_DMA_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_HASH_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:32 +0000 (23:18 -0600)]
Convert CONFIG_SPL_HASH_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoConvert CONFIG_SPL_CRYPTO_SUPPORT to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:31 +0000 (23:18 -0600)]
Convert CONFIG_SPL_CRYPTO_SUPPORT to Kconfig

Move this option to Kconfig and tidy up existing uses.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agospear: Use upper case for CONFIG options
Simon Glass [Tue, 13 Sep 2016 05:18:30 +0000 (23:18 -0600)]
spear: Use upper case for CONFIG options

There are a few options which use lower case. We should use upper case for
all CONFIG options.

Signed-off-by: Simon Glass <sjg@chromium.org>
[trini: Add usbtty/nand hunk to include/configs/spear3xx_evb.h]
Signed-off-by: Tom Rini <trini@konsulko.com>
8 years agoMove existing use of CONFIG_SPL_RSA to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:29 +0000 (23:18 -0600)]
Move existing use of CONFIG_SPL_RSA to Kconfig

A few boards define this in a header file which is incorrect. It means that
Kconfig options that rely on this cannot be used. Move it.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoMove existing use of CONFIG_SPL_DM to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:28 +0000 (23:18 -0600)]
Move existing use of CONFIG_SPL_DM to Kconfig

A few boards define this in a header file which is incorrect. It means that
Kconfig options that rely on this cannot be used. Move it.

Note that quite a few boards defined this options but do not appear to
actually use SPL:

BSC9132QDS_NOR_DDRCLK100_SECURE
BSC9132QDS_NOR_DDRCLK133_SECURE
BSC9132QDS_SDCARD_DDRCLK100_SECURE
BSC9132QDS_SDCARD_DDRCLK133_SECURE
BSC9132QDS_SPIFLASH_DDRCLK100_SECURE
BSC9132QDS_SPIFLASH_DDRCLK133_SECURE
C29XPCIE_NOR_SECBOOT
P1010RDB-PA_36BIT_NAND_SECBOOT
P1010RDB-PA_36BIT_SPIFLASH_SECBOOT
P1010RDB-PA_NAND_SECBOOT
P1010RDB-PA_NOR_SECBOOT
P1010RDB-PB_36BIT_NOR_SECBOOT
P1010RDB-PB_36BIT_SPIFLASH_SECBOOT
P1010RDB-PB_NAND_SECBOOT
P1010RDB-PB_NOR_SECBOOT
P3041DS_SECURE_BOOT
P4080DS_SECURE_BOOT
P5020DS_NAND_SECURE_BOOT
P5040DS_SECURE_BOOT
T1023RDB_SECURE_BOOT
T1024QDS_DDR4_SECURE_BOOT
T1024QDS_SECURE_BOOT
T1024RDB_SECURE_BOOT
T1040RDB_SECURE_BOOT
T1042D4RDB_SECURE_BOOT
T1042RDB_SECURE_BOOT
T2080QDS_SECURE_BOOT
T2080RDB_SECURE_BOOT
T4160QDS_SECURE_BOOT
T4240QDS_SECURE_BOOT
ls1021aqds_nor_SECURE_BOOT
ls1021atwr_nor_SECURE_BOOT
ls1043ardb_SECURE_BOOT

For these boards CONFIG_SPL_DM will no-longer be defined in SPL. But since
they apparently don't have an SPL, this should not matter.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoKconfig: tpl: Add some TPL support options to Kconfig
Simon Glass [Tue, 13 Sep 2016 05:18:27 +0000 (23:18 -0600)]
Kconfig: tpl: Add some TPL support options to Kconfig

Some of the SPL options have TPL equivalents. Add these to Kconfig so that
we can convert these options over to work from Kconfig.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoKconfig: spl: Add SPL support options to Kconfig
Simon Glass [Tue, 13 Sep 2016 13:05:23 +0000 (07:05 -0600)]
Kconfig: spl: Add SPL support options to Kconfig

There are a lot of SPL options in U-Boot to enable various features and
drivers. Currently these do not use Kconfig. Add them to Kconfig along
with suitable help, and drop them from the README.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoUse separate options for TPL support
Simon Glass [Tue, 13 Sep 2016 05:18:25 +0000 (23:18 -0600)]
Use separate options for TPL support

At present TPL uses the same options as SPL support. In a few cases the board
config enables or disables the SPL options depending on whether
CONFIG_TPL_BUILD is defined.

With the move to Kconfig, options are determined for the whole build and
(without a hack like an #undef in a header file) cannot be controlled in this
way.

Create new TPL options for these and update users. This will allow Kconfig
conversion to proceed for these boards.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoDrop CONFIG_SPL_RAM_SUPPORT
Simon Glass [Tue, 13 Sep 2016 05:18:24 +0000 (23:18 -0600)]
Drop CONFIG_SPL_RAM_SUPPORT

This option does not exist in U-Boot. Drop it.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoarm: fsl: Adjust ordering of #ifndef CONFIG_SPL_BUILD
Simon Glass [Tue, 13 Sep 2016 05:18:23 +0000 (23:18 -0600)]
arm: fsl: Adjust ordering of #ifndef CONFIG_SPL_BUILD

The secure boot header files incorrectly define SPL options only if
CONFIG_SPL_BUILD is defined. This means that the options are only enabled
in an SPL build, and not with a normal 'make xxx_defconfig'. This means
that moveconfig.py cannot work, since it sees the options as disabled even
when they may be manually enabled in an SPL build.

Fix this by changing the order.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agoKconfig: Move SPL settings into their own file
Simon Glass [Tue, 13 Sep 2016 05:18:22 +0000 (23:18 -0600)]
Kconfig: Move SPL settings into their own file

Move the SPL settings into common/spl where most of the SPL code is kept.

Signed-off-by: Simon Glass <sjg@chromium.org>
8 years agomoveconfig: Add an option to commit changes
Simon Glass [Tue, 13 Sep 2016 05:18:21 +0000 (23:18 -0600)]
moveconfig: Add an option to commit changes

The moveconfig tool is quite clever and generally produces results that
are suitable for sending as a patch without further work. The main required
step is to add the changes to a commit.

Add an option to do this automatically. This allows moveconfig to be used
from a script to convert multiple CONFIG options, once per commit.

Signed-off-by: Simon Glass <sjg@chromium.org>