Merge branch 'for-next/at91'

Merge branch 'for-next/arm'

Release v2019.06.0 ...

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

mvebu: make the compiler inline mvebu_remap_registers ...

Without that gcc 7.3.1 (from OSELAS.Toolchain-2018.02.0) makes
mvebu_remap_registers a standalone function which in its prolog saves
two registers to the stack. However this function is called early during
startup when there is no stack available yet and sp still points into
the nirvana which makes barebox crash.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

serial_ns16550: readd support for mvebu armada 38x ...

The armada 38x device tree traditionally used "snps,dw-apb-uart" as
compatible for the internal UARTs. This was changed in Linux by commit
b7639b0b15dd ("serial: 8250_dw: Limit dw8250_tx_wait_empty quirk to
armada-38x devices") which was imported into barebox in e4067b75fb6c
("dts: update to v4.18-rc1") to "marvell,armada-38x-uart".

Teach the driver about the new compatible to make the console work
again.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

plathome-openblocks-a6: provide a kwbimage.cfg file ...

Since support for Openblocks A6 was introduced in 2014 and commit
906afb533186 ("ARM: mvebu: Add Plat'home's Kirkwood Openblocks A6 board
support") there is no kwbimage config. This patch guesses such a file
that at least makes barebox compile again with
CONFIG_MACH_PLATHOME_OPENBLOCKS_A6=y.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: Initial OP-TEE support ...

This adds initial support for OP-TEE, see https://www.op-tee.org/

barebox starts in secure mode as usual. When booting a kernel
the bootm code also loads the optee_os binary. Instead of jumping
into the kernel barebox jumps into the optee_os binary and puts
the kernel execution address into the lr register. OP-TEE then
jumps into the kernel in nonsecure mode.

The optee_os binary is passed with the -t option to bootm or
with global.bootm.tee.

Optionally OP-TEE can be compiled into barebox using the builtin firmware
feature. Enable the Kconfig option and place or link your tee binary as
optee.bin into the firmware directory.

The amount of SDRAM which is kept free for OP-TEE is configurable.

This patch was tested on a i.MX6 Nitrogen6x board.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Rouven Czerwinski <r.czerwinski@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: at91: correct __iomem 'qualifier' placement ...

__iomem applies to the pointee, not to the auto pointer
that points at it. Move the annotation to correct this.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

clk: at91: fix warning about missing const-safety ...

compiling clk-main.c and clk-generated.c results in:

warning: passing argument 1 of 'memcpy' discards 'const' qualifier from
pointer target type [-Wdiscarded-qualifiers]
  memcpy(clkmain->clk.parent_names, parent_names, parents_array_size);
           ~~~~~~~~~~~~^~~~~~~~~~~~~

Avoid this by replacing the xzalloc+memcpy pair with xmemdup.
Zero-initialization of the buffer isn't necessary, because
memcpy spans the whole buffer.

Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: at91: remove duplicate get_ddram_size code ...

Both at91_get_ddram_size and at91sama5_get_ddram_size are the
same if is_sdram == false and is_nb == true.

is_sdram is always false, because according to the sama5d{2,3,4}
datasheets, the lowest possible value for AT91_DDRSDRC_MD is 3
(i.e. none of them supports SDR SDRAM).

Therefore replace calls to at91sama5_get_ddram_size with
at91_get_ddram_size while is_nb == true and remove the duplicate code.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Tested-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: at91: fix at91sama5_get_ddram_size for sama5d4 ...

at91sama5_get_ddram_size() is despite the name specific to the
sama5d3 which it was added alongside of. sama5d4 board code continues
to use it, but accessing SAMA5D3_BASE_MPDDRC (0xffffea00) on
a sama5d4 should result in a Data Abort (Datasheet Figure 5-1)..

Fix this by giving at91sama5_get_ddram_size the mpddrc base
address as argument and migrate users to use one of two helpers
that specify the base address.

Tested-by: Sam Ravnborg <sam@ravnborg.org>
[afa: Sam didn't test this exact change, but tested the use
 of the critical part: accessing a different register]
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

arm: sama5d4: fix stack setup ...

The code that configure the stack uses sama5d3 constants.
Fix this to use the proper sama5d4 constants.

Boot tested on sama5d4_xplained.

Signed-off-by: Sam Ravnborg <sam@ravnborg.org>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: at91: remove #ifdefery around *_get_ddram_size helpers ...

c16bcbc644 ("ARM: at91: remove unused defines") removed the
conflicting defintions from these headers, so it's ok to
include them at the same time now. Do this and remove the
noise.

Suggested-by: Sascha Hauer <s.hauer@pengutronix.de>
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: at91: remove no-longer needed subarch #ifdefery ...

c16bcbc644 ("ARM: at91: remove unused defines") removed conflicting
definitions from the headers, so they can now be included all at
the same time. Do this.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: at91: remove references to non-existing CONFIG_ARCH_AT91* ...

These symbols aren't set anywhere in barebox and in some
cases they guard non-existing headers. Remove them.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: at91: debug_ll: make UART base address configurable ...

This is in line with other platforms such as i.MX, which allow
specifying a debug port. As we can't use port indices because
the UARTs aren't mapped consecutively, allow specifying a hex
base at configuration time.

A side effect of this patch is that sama5d4's HAVE_AT91_DBGU2
is now honored as well. Previously anything besides DBGU0
defaulted to DBGU1.

Fixes: 06a0773ee31 ("ARM: at91: add sama5d4 soc support #2")
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Tested-by: Sam Ravnborg <sam@ravnborg.org>
[afa: moved base address defaults to common/Kconfig]
Signed-off-by: Ahmad Fatoum <afa@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: at91: replace __raw_{readl, writel} of peripherals with readl, writel ...

Use the potentially endianness-changing readl, writel and siblings
directly. They looks prettier and are the correct thing to do, as
even if the CPU is in big-endian mode, the peripherals are little-endian.

Unlike Linux, barebox readl,writel are the same Linux'
{readl,writel}_relaxed (they don't imply memory barriers)
and thus there shouldn't be any functional change.

Patch was generated by a mass search and replace. I looked it over,
adjust some whitespace and further verified by reviewing the output of

	git diff HEAD~1 --word-diff | \
		perl -pe 's/\[-(.*?)__raw_/{+$1/; s/-\]\{\+/+}{+/;' \
		-e 's/(\{\+.*?\+\})\1/__ALL_IS_WELL__/' | grep '+}{+'

which filters out the common case of lines where a single
__raw_{readT,writeT} had its __raw_ prefix stripped without any
further changes.

Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Tested-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

clk: at91: delete dead i2s/audio code ...

They aren't used anywhere, thus drop them.

Suggested-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

clk: at91: fix compilation errors in sama5d2.c ...

sama5d2.c was added alongside the update to the remaining clk drivers
to support the new upstream device tree bindings, but wasn't wired in
anywhere.

To prepare for usage in future sama5d2 support, fix compilation errors
related to absence of locks and unavailability of audio/i2s clock/pll
handling.

With this change, there are no users for the few definition-less at91
audio/i2s registration function declarations. So remove those as well.

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

lib: select BASE64 if BLOBGEN is selected ...

Blobgen requires the BASE64 libraries, select them to provide them if
BLOBGEN is selected.

Signed-off-by: Rouven Czerwinski <r.czerwinski@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

Documentation: imx8mq-evk: collapse BMOD switch superposition ...

Table 3. Boot mode setting in the user's guide[1] lists
0010 as 'Boot from EMMC'. Adjust diagram accordingly.

[1]: https://www.mouser.com/ds/2/302/IMX8MDQLQEVKHUG-1280333.pdf

Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: boards: kindle mx50: adapt MCI name to DTS ...

The MCI instances got aliases in the DTS from linux upstream which changed
the eMMC devicename e.g. from disk0 to mmc2. Adapt to this.

Signed-off-by: Alexander Kurz <akurz@blala.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: include: dma: Zero out DMA coherent memory in no-MMU case ...

Add code to match the behavior of dma_alloc_coherent() when CONFIG_MMU
is selected.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

LICENSES: remove empty other, exceptions directories ...

These directories were originally added as they were required for
spdxcheck.py, which upstream checkpatch.pl invokes.

As git doesn't index empty directories, they contained empty
.gitignore files. As these files were 0 bytes, make distclean now
removes them.

We already patch checkpatch.pl, so we will just patch spdxcheck.py
when we add it and get rid of these directories.
They can be added back should the need arise.

Reported-by: Geordan Neukum <gneukum1@gmail.com>
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

commands: spi: fix NULL pointer dereference ...

The renaming of "master" to "controller" done in
    812a9dd
    spi: Generalize SPI "master" to "controller"

breaks the spi command, because the spi.master pointer is not assigned.
The pointer is kept for backwards compatability.

Assign it, so that spi drivers that are still using "spi->master" can be
used.

Signed-off-by: Steffen Trumtrar <s.trumtrar@pengutronix.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

serdev: Do not call .receive_buf() callback recursively ...

Code implementing .receive_buf() callback can potentially call
serdev_device_write(), which will call serdev_device_poller(). We need
to make sure that such a call is a no-op in order to prevent
corrupting shared data buffer as well as breaking .receive_buf
callback that most likely does not expect that to happen.

Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Tested-by: Cory Tusar <cory.tusar@zii.aero>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

spi: Kconfig: remove dependency list from Freescale QSPI controller ...

The dependency list for the Freescale QSPI controller doesn't match
the SoCs the hardware is in at all. Remove the dependency entirely as
the driver builds on all machines.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: Layerscape: Add PPA firmware support ...

The "Primary Protected Application" (PPA) is a PSCI compliant firmware
distributed by NXP. It is needed to start the secondary cores on
Layerscape SoCs. Without it Linux will be started in EL3 and doesn't
work properly. The precompiled firmware images can be found on
https://github.com/NXP/qoriq-ppa-binary and are not included in
barebox.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: psci: factor out of_psci_fixup() to separate file ...

of_psci_fixup() can be used by code which doesn't use the barebox
psci implementation, but provides its own PSCI compatible firmware.
Factor it out to a separate file to compile it independently of the
barebox PSCI implementation.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>

ARM: unify asm/arm-smccc.h and linux/arm-smccc.h ...

Since 2c20674f0c ("ARM: Add code to support SMCCC on AArch64") we have
both arch/arm/include/asm/arm-smccc.h and include/linux/arm-smccc.h.
The latter was introduced with Kernel commit 680a0873e ("arm: kernel:
Add SMC structure parameter") included, so we have to add the parameter
for ARM32 aswell.

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>