The DDRSDR controller fails miserably to put LPDDR1 memories in
self-refresh. Force the controller to think it has DDR2 memories
during the self-refresh period, as the DDR2 self-refresh spec is
equivalent to LPDDR1, and is correctly implemented in the
controller.

Assume that the second controller has the same fault, but that is
untested.
Signed-off-by: NPeter Rosin <peda@axentia.se>
Acked-by: NNicolas Ferre <nicolas.ferre@atmel.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Removed timeout on XTAL, PLL lock and Master Clock Ready, hang if
something went wrong instead of continuing in unknown condition. There
is not much we can do if a PLL lock never ends, we are running in SRAM
and we will not be able to connect back the sdram or ddram in order to
be able to fire up a message or just panic.

As a bonus, not decounting the timeout register in slow clock mode
reduce cumulated suspend time and resume time from ~17ms to ~15ms.
Signed-off-by: NSylvain Rochet <sylvain.rochet@finsecur.com>
Acked-by: NWenyou.Yang <wenyou.yang@atmel.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

of_device_ids (i.e. compatible strings and the respective data) are not
supposed to change at runtime. All functions working with of_device_ids
provided by <linux/of.h> work with const of_device_ids. So mark the
non-const structs in arch/arm as const, too.

While at it also add some __initconst annotations.
Acked-by: NJason Cooper <jason@lakedameon.net>
Signed-off-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: NArnd Bergmann <arnd@arndb.de>

Commit 20e783e3 ("ARM: 8296/1: cache-l2x0: clean up aurora cache
handling") removed the only user of the Kconfig symbol CACHE_PL310.
Setting CACHE_PL310 is now pointless. Remove its Kconfig entry, and one
select of this symbol.
Signed-off-by: NPaul Bolle <pebolle@tiscali.nl>
Signed-off-by: NArnd Bergmann <arnd@arndb.de>

In a recent rearrangement of the at91 pm initialization code, a broken
set of declarations was added for the !CONFIG_PM-case, leading to
this link error:

arch/arm/mach-at91/board-dt-sama5.o: In function `at91_rm9200_pm_init':
arch/arm/mach-at91/generic.h:40: multiple definition of `at91_rm9200_pm_init'
arch/arm/mach-at91/setup.o:arch/arm/mach-at91/generic.h:40: first defined here
arch/arm/mach-at91/board-dt-sama5.o: In function `at91_sam9260_pm_init':
arch/arm/mach-at91/generic.h:41: multiple definition of `at91_sam9260_pm_init'
arch/arm/mach-at91/setup.o:arch/arm/mach-at91/generic.h:41: first defined here
arch/arm/mach-at91/board-dt-sama5.o: In function `at91_sam9g45_pm_init':
arch/arm/mach-at91/generic.h:42: multiple definition of `at91_sam9g45_pm_init'
arch/arm/mach-at91/setup.o:arch/arm/mach-at91/generic.h:42: first defined here

This adds the missing 'static inline' to the declarations to avoid
creating a copy of the functions in each file that includes the
header.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
Fixes: 4db0ba22 ("ARM: at91: pm: prepare for multiplatform")
Acked-by: NNicolas Ferre <nicolas.ferre@atmel.com>
Cc: Jean-Christophe Plagniol-Villard <plagnioj@jcrosoft.com>
Cc: Alexandre Belloni <alexandre.belloni@free-electrons.com>

Remove the string "(Device Tree)" after the machine name because all AT91
machines use the DT nowadays.
Also change some function names to unify following the convention:
- at91sam9xxx aren't named sam9, 9xxx but with the full name
- sama5 are the ones that don't have the at91 prefix anymore.
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Move the ramc initialization to pm.c as it is the only user left.
This allows us to get rid of at91_dt_initialize() that was the only one called
by the init_early() function pointer of struct machine_desc.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
[nicolas.ferre@atmel.com: adapt patch to newer series]
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

As board files are now DT only and can address all aspects of the SoC family,
we can rename them so that the mach-at91 directory looks cleaner.
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

at91_boot_soc and at91_init_soc structures are not used by any SoC, remove
their use. Also remove all the now empty SoC files.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
[nicolas.ferre@atmel.com: different organization of the patches]
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

For sama5d4, remove an indirection and the remaining need for at91_boot_soc
structure.
This will allow further cleanup.
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Only use SOC_AT91SAM9 for all the at91sam9 SoCs. It removes all the empty
at91sam9xxx.c SoC files. It also removes the useless at91_init_soc affectation
procedure and its "init" function pointer.
Only the SoC detection and display are kept for the at91sam9:
at91_soc_is_enabled() and at91_boot_soc.map_io() function calls are also
removed.

It enables HAVE_AT91_SMD and HAVE_AT91_UTMI for all the sam9 SoCs but this only
represents 96 bytes of uncompressed kernel code.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
[nicolas.ferre@atmel.com: different organization of the patches]
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Setup arm_pm_idle and arm_pm_restart function pointers from
rm9200_dt_device_init() function to simplify the at91rm9200 initialization
process.
This same move is already done for the sam9s.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
[nicolas.ferre@atmel.com: adapt patch to newer series]
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

The SOC_AT91SAM9263 is being removed, stop using it.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

mach/system_rev.h is not used, remove it.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

In order to remove SOC_SAM9xxx options, stop using HAVE_AT91_DBGUx.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

The PM initialization needs internal SRAM for allocating a gen_pool and
use it to store its PM code. So we need to have of_platform_populate() before
this code.
Suggested-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

As all sam9 SoCs are setting arm_pm_idle to at91sam9_idle(), do it from
sam9_dt_device_init().
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Suggested-by: NArnd Bergmann <arnd@arndb.de>
[nicolas.ferre@atmel.com: adapt patch to newer series]
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

sam9n12 and sam9x5 don't set arm_pm_idle because of an oversight, fix that.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Suggested-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

As long as there is no other non-const variable marked __initdata in the
same compilation unit it doesn't hurt. If there were one however
compilation would fail with

error: $variablename causes a section type conflict

because a section containing const variables is marked read only and so
cannot contain non-const variables.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
[nicolas.ferre@atmel.com: update the paths after having re-arranged the patches]
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Newer SoCs: at91sam9x5, at91sam9n12, sama5d3 and sama5d4 embed a DDR controller
and have a different PMC status register layout than the at91sam9g45. Create
another at91_sam9x5_pm_init() function to match this compatibility.
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Move at91rm9200_idle() along with at91sam9_idle() in clk/at91/pmc.c.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: NMichael Turquette <mturquette@linaro.org>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

SRAM initialization is now done through the mmio-sram driver and
at91_init_sram() is not called anymore, remove it.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Now that the SRAM is initialized by the mmio-sram driver, .map_io is useless.
remove it.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Split at91_pm_init() in three variants that are called by the respective SoCs
.init_machine. This allows to remove the of_machine_is_compatible() calls and
move at91_pm_init() out of arch_initcall() which is required for multiplatform.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Check UDP and UHP on sam9x5, sam9n12 and the sama5 series.
Check UHP on the sam9g45.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Now that the SRAM is part of a genpool, use it to allocate memory to use for the
slowclock implementation.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Store SoC differences in a struct to remove cpu_is_* usage.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Since all rm9200 board files have been removed, there is no user of
at91rm9200_set_type() left. Remove it
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: NBoris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

at91rm9200_dt_initialize() is doing the same as at91_dt_initialize(), use that
one instead.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: NBoris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Move debug-macro.S from include/mach/ to include/debug where all other common
debug macros are.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: NBoris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

The irq fixup from at91_sysirq_mask_rtc and at91_sysirq_mask_rtt is now handled
by aic_common_rtc_irq_fixup and aic_common_rtt_irq_fixup. Remove those useless
functions.
Also remove the now unused mach/at91_rtt.h header.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: NBoris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Now that at91sam9 SoCs are only supported through DT, remove
CONFIG_MACH_AT91SAM9_DT and use CONFIG_SOC_AT91SAM9 instead.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: NBoris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Now that rm9200 is only supported through DT, remove CONFIG_MACH_AT91RM9200_DT
and use CONFIG_SOC_AT91RM9200 instead.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: NBoris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

CONFIG_NEED_MACH_MEMORY_H is not set by any at91 platform, remove mach/memory.h
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: NBoris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Since removal of !DT boards, asm/irq.h inclusion is not needed in these product
files.
Signed-off-by: NBoris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

These GPIO pin descriptions are now moved with the pinctrl driver. We can
safely remove this useless header file.
Reported-by: NArnd Bergmann <arnd@arndb.de>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Appearance: On some SAMA5D4EK boards, after power up, the Eth1 doesn't work.

Reason: The PIOE2 pin is connected to the NAND_Tree# of KSZ8081,
But it outputs LOW during the reset period, which cause the NAND_Tree# enabled.

Add phy_fixup() to disable NAND_Tree by overriding the Operation
Mode Strap Override register(i.e. Register 16h) to clear the NAND_Tree bit.
Signed-off-by: NWenyou Yang <wenyou.yang@atmel.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

All functions declared in this file are gone.
Signed-off-by: NArnd Bergmann <arnd@arndb.de>
[nicolas.ferre@atmel.com: re-order patches so modify board-dt-sam9]
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

These files were left behind with no reason. Remove them.
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>