On Exynos4412 boards (Trats2, Odroid U3) after enabling L2 cache in
56b60b8b ("ARM: 8265/1: dts: exynos4: Add nodes for L2 cache
controller") the second suspend to RAM failed. First suspend worked fine
but the next one hang just after powering down of secondary CPUs (system
consumed energy as it would be running but was not responsive).

The issue was caused by enabling delayed reset assertion for CPU0 just
after issuing power down of cores. This was introduced for Exynos4 in
13cfa6c4 ("ARM: EXYNOS: Fix CPU idle clock down after CPU off").

The whole behavior is not well documented but after checking with vendor
code this should be done like this (on Exynos4):
1. Enable delayed reset assertion when system is running (for all CPUs).
2. Disable delayed reset assertion before suspending the system.
   This can be done after powering off secondary CPUs.
3. Re-enable the delayed reset assertion when system is resumed.

Fixes: 13cfa6c4 ("ARM: EXYNOS: Fix CPU idle clock down after CPU off")
Cc: <stable@vger.kernel.org>
Signed-off-by: NKrzysztof Kozlowski <k.kozlowski@samsung.com>
Tested-by: NBartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Tested-by: NChanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: NKukjin Kim <kgene@kernel.org>

Register cpuidle platform device on Exynos3250 SoC allowing EXYNOS
cpuidle driver usage on this SoC.

Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: NBartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Reviewed-by: NKrzysztof Kozlowski <k.kozlowski@samsung.com>
Acked-by: NKyungmin Park <kyungmin.park@samsung.com>
Tested-by: NChanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: NKukjin Kim <kgene@kernel.org>

Exynos has been (ab)using the gic_arch_extn to provide
wakeup from suspend, and it makes a lot of sense to convert
this code to use stacked domains instead.

This patch does just this, updating the DT files to actually
reflect what the HW provides.

BIG FAT WARNING: because the DTs were so far lying by not
exposing the fact that the PMU block is actually the first
interrupt controller in the chain for RTC, kernels with this patch
applied wont have any suspend-resume facility when booted
with old DTs, and old kernels with updated DTs may not even boot.

Also, I strongly suspect that there is more than two wake-up
interrupts on these platforms, but I leave it to the maintainers
to fix their mess.
Signed-off-by: NMarc Zyngier <marc.zyngier@arm.com>
Link: https://lkml.kernel.org/r/1426088693-15724-2-git-send-email-marc.zyngier@arm.com
[ jac: squash in maz's fixup from
  https://lkml.kernel.org/r/5506989D.9050703@arm.com ]
Signed-off-by: NJason Cooper <jason@lakedaemon.net>

The Exynos cpuidle driver has coupled cpuidle built-in so it cannot be
built without SMP:

arch/arm/mach-exynos/pm.c: In function 'exynos_cpu0_enter_aftr':
arch/arm/mach-exynos/pm.c:246:4: error: implicit declaration of function 'arch_send_wakeup_ipi_mask' [-Werror=implicit-function-declaration]
arch/arm/mach-exynos/built-in.o: In function 'exynos_pre_enter_aftr':
../arch/arm/mach-exynos/pm.c:300: undefined reference to 'cpu_boot_reg_base'
arch/arm/mach-exynos/built-in.o: In function 'exynos_cpu1_powerdown':
../arch/arm/mach-exynos/pm.c:282: undefined reference to 'exynos_cpu_power_down'

Fix it by adding missing checks for SMP.
Reported-by: NKrzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: NKukjin Kim <kgene@kernel.org>

The definition

	static const char *axxia_dt_match[] __initconst = {
		...

defines a changable array of constant strings. That is you must not do:

	*axxia_dt_match[0] = 'k';

but

	axxia_dt_match[0] = "different string";

is fine. So the annotation __initconst is wrong and yields a compiler
error when other really const variables are added with __initconst.

As the struct machine_desc member dt_compat is declared as

	const char *const *dt_compat;

making the arrays const is the better alternative over changing all
annotations to __initdata.
Signed-off-by: NUwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: NArnd Bergmann <arnd@arndb.de>

The following patch adds coupled cpuidle support for Exynos4210 to
an existing cpuidle-exynos driver.  As a result it enables AFTR mode
to be used by default on Exynos4210 without the need to hot unplug
CPU1 first.

The patch is heavily based on earlier cpuidle-exynos4210 driver from
Daniel Lezcano:

http://www.spinics.net/lists/linux-samsung-soc/msg28134.html

Changes from Daniel's code include:
- porting code to current kernels
- fixing it to work on my setup (by using S5P_INFORM register
  instead of S5P_VA_SYSRAM one on Revison 1.1 and retrying poking
  CPU1 out of the BOOT ROM if necessary)
- fixing rare lockup caused by waiting for CPU1 to get stuck in
  the BOOT ROM (CPU hotplug code in arch/arm/mach-exynos/platsmp.c
  doesn't require this and works fine)
- moving Exynos specific code to arch/arm/mach-exynos/pm.c
- using cpu_boot_reg_base() helper instead of BOOT_VECTOR macro
- using exynos_cpu_*() helpers instead of accessing registers
  directly
- using arch_send_wakeup_ipi_mask() instead of dsb_sev()
  (this matches CPU hotplug code in arch/arm/mach-exynos/platsmp.c)
- integrating separate exynos4210-cpuidle driver into existing
  exynos-cpuidle one

Cc: Colin Cross <ccross@google.com>
Cc: Kukjin Kim <kgene.kim@samsung.com>
Cc: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Cc: Tomasz Figa <tomasz.figa@gmail.com>
Signed-off-by: NBartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Signed-off-by: NDaniel Lezcano <daniel.lezcano@linaro.org>
Acked-by: NKyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: NKukjin Kim <kgene@kernel.org>

Also setup memory allocations for version 8 of the MFC as present in
Samsung Exynos 5422/5800 SoCs
Signed-off-by: NSjoerd Simons <sjoerd.simons@collabora.co.uk>
Signed-off-by: NKukjin Kim <kgene@kernel.org>

As all these code has been moved into i2c driver, now we can
safely remove them from machine files.

CC: Russell King <linux@arm.linux.org.uk>
Signed-off-by: NPankaj Dubey <pankaj.dubey@samsung.com>
Signed-off-by: NKukjin Kim <kgene@kernel.org>

Let's register restart handler from PMU driver for restart
functionality. So that we can remove restart hooks from
machine specific file, and thus moving ahead when PMU moved
to driver folder, this functionality can be reused for ARM64
based Exynos SoC's.
Signed-off-by: NPankaj Dubey <pankaj.dubey@samsung.com>
Acked-by: NGuenter Roeck <linux@roeck-us.net>
Tested-by: NVivek Gautam <gautam.vivek@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Let's register restart handler for Exynos5440 from it's clock driver
for restart functionality. So that we can cleanup restart hooks from
machine specific file.

CC: Sylwester Nawrocki <s.nawrocki@samsung.com>
CC: Tomasz Figa <tomasz.figa@gmail.com>
Signed-off-by: NPankaj Dubey <pankaj.dubey@samsung.com>
Acked-by: NGuenter Roeck <linux@roeck-us.net>
Acked-by: NSylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

This patch adds support for Exynos4415 SoC. Exynos4415 is based on
the 32-bit RISC processor for Smartphone. Exynos4415 has Cortex A9
quad-cores and has a target speed of 1.6GHz and provides 8.5GB/s
memory bandwidth.
Signed-off-by: NChanwoo Choi <cw00.choi@samsung.com>
Acked-by: NKyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

This patch removes all unused static iomapping from exynos4/5_iodesc
table, and at the same time removes related macros from mach/map.h and
plat/map-s5p.h. All such mappings are present in exynos.c but not
currently there are no users of these mappings, so it is safe to remove
these.
Signed-off-by: NPankaj Dubey <pankaj.dubey@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Register cpuidle platform device on Exynos4x12 SoCs allowing EXYNOS
cpuidle driver usage on these SoCs.

AFTR mode reduces power consumption on Trats2 board (Exynos4412 SoC
with secure firmware enabled) by ~12% when EXYNOS cpuidle driver is
enabled (in both cases the default exynos_defconfig config is used
and CPU1-3 are offlined).

Currently Exynos4412 SoC support is limited to Trats2 board.
Signed-off-by: NBartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Acked-by: NKyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Fix building of exynos_defconfig with disabled CONFIG_PM_SLEEP by
adding checking whether Exynos cpuidle support is enabled before
accessing exynos_enter_aftr.

The build error message:
arch/arm/mach-exynos/built-in.o:(.data+0x74): undefined reference to `exynos_enter_aftr'
make: *** [vmlinux] Error 1

This patch has been tested on Exynos4210 based Origen board.
Signed-off-by: NBartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Acked-by: NKyungmin Park <kyungmin.park@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Since genpd at late init, will try to disable unused PM domains we
don't need to do it from here as well.
Signed-off-by: NUlf Hansson <ulf.hansson@linaro.org>
Reviewed-by: NKevin Hilman <khilman@linaro.org>
Signed-off-by: NRafael J. Wysocki <rafael.j.wysocki@intel.com>

Commit 1c2f87c2
(ARM: 8025/1: Get rid of meminfo) dropped the upper bound on
the number of memory banks that can be added as there was no
technical need in the kernel. It turns out though, some bootloaders
(specifically the arndale-octa exynos boards) may pass invalid memory
information and rely on the kernel to not parse this data. This is a
bug in the bootloader but we still need to work around this.
Work around this by introducing a dt_fixup function. This function
gets called before the flattened devicetree is scanned for memory
and the like. In this fixup function for exynos, limit the maximum
number of memory regions in the devicetree.
Signed-off-by: NLaura Abbott <lauraa@codeaurora.org>
Tested-by: NAndreas Färber <afaerber@suse.de>
[glikely: Added a comment and fixed up function name]
Signed-off-by: NGrant Likely <grant.likely@linaro.org>

Exynos initialisation code now relies on obtaining the PMU address via
DT, so add the exynos5260 PMU compatible string to DT match table.
Signed-off-by: NVikas Sajjan <vikas.sajjan@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Exynos initialization code now relies on obtaining the PMU address,
so add the new 5410 value to the list of compatible string matches.
This unbreaks booting on 5410 based boards.

Fixes: fce9e5bb ("ARM: EXYNOS: Add support for mapping PMU base
address via DT")
Signed-off-by: NAndreas Faerber <afaerber@suse.de>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

As exynos_cpuidle_init() and exynos_cpufreq_init() functions have just
one line of code for registering platform devices. So we can move them
to exynos_dt_machine_init() and remove exynos_cpuidle_init() and
exynos_cpufreq_init(). This will help in reducing lines of code in
exynos.c, making it more clean.
Suggested-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NPankaj Dubey <pankaj.dubey@samsung.com>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Under "arm/mach-exynos" many files are using PMU register offsets.
Since we have added support for accessing PMU base address via DT,
now we can remove PMU mapping from exynosX_iodesc. Let's convert
all these access using iomapped address.
This will help us in removing static mapping of PMU base address
as well as help in reducing dependency over machine header files.
Thus helping for migration of PMU implementation from machine to
driver folder which can be reused for ARM64 based SoC.

Also as we have removed static mappings from "regs-pmu.h" it does
not need map.h anymore. But "platsmp.c" needed this and till now it
got included indirectly. So lets move header inclusion of
"mach/map.h" from "regs-pmu.h" to "platsmp.c".
Signed-off-by: NPankaj Dubey <pankaj.dubey@samsung.com>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Exynos5420 is big.Little Soc. It uses cpuidle-big-litle generic cpuidle driver.
Hence do not allow exynos cpuidle driver registration for Exynos5420.
Signed-off-by: NChander Kashyap <chander.kashyap@linaro.org>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Acked-by: NDaniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Add support for mapping Samsung Power Management Unit (PMU)
base address from device tree.
Signed-off-by: NPankaj Dubey <pankaj.dubey@samsung.com>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

While making PMU implementation to be device tree based, there are
few register offsets related with SYSREG present in regs-pmu.h, so
let's make a new header file "regs-sys.h" to keep all such SYSREG
related register offsets and remove them from "regs-pmu.h"
Signed-off-by: NPankaj Dubey <pankaj.dubey@samsung.com>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

As machine function ops are used only in this file let's make
them static. Also remove unused and unwanted declarations from
common.h.
Signed-off-by: NPankaj Dubey <pankaj.dubey@samsung.com>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Currently, the exynos cpuidle driver works correctly only on exynos4210
and 5250. Trying to use it with just one CPU online on any other exynos
SoCs will lead to system failure, due to unsupported AFTR mode on other
SoCs. This patch fixes the problem by registering the driver only on
supported SoCs and letting others simply use default WFI mode until
support for them is added.
Signed-off-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Commit 1754c42e("ARM: exynos: move sysram info to exynos.c") missed
out the CONFIG_ prefix causing exynos_sysram_init() to get called twice
for SMP configurations.
Signed-off-by: NAbhilash Kesavan <a.kesavan@samsung.com>
Reviewed-by: NSachin Kamat <sachin.kamat@samsug.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

of_get_flat_dt_prop return type is now const.
Fixes the following compilation warning introduced by commit 9d0c4dfe
("of/fdt: update of_get_flat_dt_prop in prep for libfdt")

arch/arm/mach-exynos/exynos.c:259:6: warning:
assignment discards ‘const’ qualifier from pointer target type [enabled by default]
Signed-off-by: NSachin Kamat <sachin.kamat@linaro.org>
Reviewed-by: NTushar Behera <tushar.behera@linaro.org>
Cc: Rob Herring <robh@kernel.org>
Signed-off-by: NArnd Bergmann <arnd@arndb.de>

This solves a problem with building with CONFIG_SMP=n due to missing
sysram_base_addr (or sysram_ns_base_addr) variables.

The new setup method is more awkward than I'd like for it to be, but
it can't be done in init_early() since ioremap is not yet available,
but it needs to happen before SMP.
Reported-by: NRussell King <linux@arm.linux.org.uk>
Cc: Kukjin Kim <kgene.kim@samsung.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: NOlof Johansson <olof@lixom.net>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NArnd Bergmann <arnd@arndb.de>

This was caught by a panic on Broadcom mobile platforms.

Note that this code is all going away with the pending l2x0 cleanup
series from Russell, but we need this here until that's landed so we
can enable exynos multiplatform.
Signed-off-by: NOlof Johansson <olof@lixom.net>

This patch add Exynos3250's SoC ID. Exynos 3250 is SoC that
is based on the 32-bit RISC processor for Smartphone.
Exynos3250 uses Cortex-A7 dual cores and has a target speed
of 1.0GHz.
Signed-off-by: NChanwoo Choi <cw00.choi@samsung.com>
Acked-by: NKyungmin Park <kyungmin.park@samsung.com>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

This patch add basic arch side support for exynos5260 SoC.
Note that this is required to enable build for clock driver.
Signed-off-by: NPankaj Dubey <pankaj.dubey@samsung.com>
Signed-off-by: NRahul Sharma <rahul.sharma@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

exynos was unconditionally calling the L2 cache initialisation from an
early_initcall.  This breaks multiplatform kernels.  Thankfully,
converting to generic l2c initialisation fixes this.
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

The cache size should already be present in the L2 cache auxiliary
control register: it is part of the integration process to configure
the hardware IP.  Most platforms get this right, yet still many
cargo-cult program, and assume that they always need specifying to
the L2 cache code.  Remove them so we can find out which really need
this.
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Since we now automatically enable early BRESP in core L2C-310 code when
we detect a Cortex-A9, we don't need platforms/SoCs to set this bit
explicitly.  Instead, they should seek to preserve the value of bit 30
in the auxiliary control register.
Acked-by: NTony Lindgren <tony@atomide.com>
Acked-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

No more dependency on the arch code. The platform_data field is used to set the
PM callback as the other cpuidle drivers.
Signed-off-by: NDaniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-by: NViresh Kumar <viresh.kumar@linaro.org>
Reviewed-by: NBartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

There is no point to register the cpuidle driver for the 5440 as it has only
one WFI state which is the default idle function when the cpuidle driver is
disabled.

By disabling cpuidle we prevent to enter to the governor computation for
nothing, thus saving a lot of processing time.

The only drawback is the statistic via sysfs on this state which is lost but
it is meaningless and it could be retrieved from the ftrace easily.
Signed-off-by: NDaniel Lezcano <daniel.lezcano@linaro.org>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Acked-by: NAmit Kucheria <amit.kucheria@linaro.org>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Exynos specific macros and declarations have been moved to
mach-exynos. Inclusion of plat/cpu.h is no more necessary.
Signed-off-by: NSachin Kamat <sachin.kamat@linaro.org>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

'exynos_subsys' has no users. Remove this code.
Signed-off-by: NSachin Kamat <sachin.kamat@linaro.org>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>

Instead of hardcoding the SYSRAM details for each SoC,
pass this information through device tree (DT) and make
the code SoC agnostic. Generic DT SRAM bindings are
used for achieving this.
Signed-off-by: NSachin Kamat <sachin.kamat@linaro.org>
Acked-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NHeiko Stuebner <heiko@sntech.de>
Reviewed-by: NTomasz Figa <t.figa@samsung.com>
Signed-off-by: NKukjin Kim <kgene.kim@samsung.com>