platform/kernel/linux-starfive.git
9 years agoMerge branch 'socfpga/soc' into next/soc
Kevin Hilman [Thu, 11 Jun 2015 22:50:10 +0000 (15:50 -0700)]
Merge branch 'socfpga/soc' into next/soc

* socfpga/soc:
  ARM: socfpga: fix build error due to secondary_startup

9 years agoARM: socfpga: fix build error due to secondary_startup
Kevin Hilman [Thu, 11 Jun 2015 22:41:58 +0000 (15:41 -0700)]
ARM: socfpga: fix build error due to secondary_startup

After commit 02b4e2756e01 (ARM: v7 setup function should invalidate L1
cache) the soc specific secondary_startup is removed, causing build
failures:

../arch/arm/mach-socfpga/platsmp.c: In function 'socfpga_a10_boot_secondary':
../arch/arm/mach-socfpga/platsmp.c:66:140: error: 'socfpga_secondary_startup' undeclared (first use in this function)
../arch/arm/mach-socfpga/platsmp.c:66:140: note: each undeclared identifier is reported only once for each function it appears in

To fix, use the generic secondary_startup.

Cc: Dinh Nguyen <dinguyen@opensource.altera.com>
Signed-off-by: Kevin Hilman <khilman@linaro.org>
9 years agoMAINTAINERS: ARM64: EXYNOS: Extend entry for ARM64 DTS
Krzysztof Kozlowski [Sat, 6 Jun 2015 10:02:19 +0000 (19:02 +0900)]
MAINTAINERS: ARM64: EXYNOS: Extend entry for ARM64 DTS

Extend the Exynos entry to ARM64 device tree sources.

Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Kukjin Kim <kgene@kernel.org>
Cc: Kevin Hilman <khilman@kernel.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Olof Johansson <olof@lixom.net>
Cc: linux-samsung-soc@vger.kernel.org
Cc: linux-arm-kernel@lists.infradead.org
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Signed-off-by: Kevin Hilman <khilman@linaro.org>
9 years agoMerge tag 'samsung-mach-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene...
Kevin Hilman [Thu, 11 Jun 2015 21:44:21 +0000 (14:44 -0700)]
Merge tag 'samsung-mach-1' of git://git./linux/kernel/git/kgene/linux-samsung into next/soc

Samsung updates for v4.2

- add failure(exception) handling
  : of_iomap(), of_find_device_by_node() and kstrdup()

- add common poweroff to use PS_HOLD based for all of exynos SoCs
- add exnos_get/set_boot_addr() helper
- constify platform_device_id and irq_domain_ops
- get current parent clock for power domain on/off
- use core_initcall to register power domain driver
- make exynos_core_restart() less verbose

- add support coupled CPUidle for exynos3250

- fix exynos_boot_secondary() return value on timeout
- fix clk_enable() in s3c24xx adc
- fix missing of_node_put() for power domains

* tag 'samsung-mach-1' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung: (301 commits)
  ARM: EXYNOS: register power domain driver from core_initcall
  ARM: EXYNOS: use PS_HOLD based poweroff for all supported SoCs
  ARM: SAMSUNG: Constify platform_device_id
  ARM: EXYNOS: Constify irq_domain_ops
  ARM: EXYNOS: add coupled cpuidle support for Exynos3250
  ARM: EXYNOS: add exynos_get_boot_addr() helper
  ARM: EXYNOS: add exynos_set_boot_addr() helper
  ARM: EXYNOS: make exynos_core_restart() less verbose
  ARM: EXYNOS: fix exynos_boot_secondary() return value on timeout
  ARM: EXYNOS: Get current parent clock for power domain on/off
  ARM: SAMSUNG: fix clk_enable() WARNing in S3C24XX ADC
  ARM: EXYNOS: Add missing of_node_put() when parsing power domains
  ARM: EXYNOS: Handle of_find_device_by_node() and kstrdup() failures
  ARM: EXYNOS: Handle of of_iomap() failure
  Linux 4.1-rc4
  ....

9 years agoARM: ep93xx: simone: support for SPI-based MMC/SD cards
Mika Westerberg [Mon, 8 Jun 2015 13:22:39 +0000 (15:22 +0200)]
ARM: ep93xx: simone: support for SPI-based MMC/SD cards

This includes setting up EGPIOs 0 and 9 for card detection and
chip select respectively. This patch is needed to mount a root
filesystem on the SPI-based MMC card reader found on the Sim.One.

Signed-off-by: Mika Westerberg <mika.westerberg@iki.fi>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Signed-off-by: Kevin Hilman <khilman@linaro.org>
9 years agoMerge tag 'hi6220-soc-for-4.2' of git://github.com/hisilicon/linux-hisi into next/soc
Kevin Hilman [Thu, 11 Jun 2015 00:07:15 +0000 (17:07 -0700)]
Merge tag 'hi6220-soc-for-4.2' of git://github.com/hisilicon/linux-hisi into next/soc

ARM64: Hisilicon ARM64 SoC Updates for V4.2

- Added Hisilicon ARM64 SoC family support in Kconfig and defconfig

* tag 'hi6220-soc-for-4.2' of git://github.com/hisilicon/linux-hisi:
  arm64: Enable Hisilicon ARMv8 SoC family in Kconfig and defconfig

9 years agoMAINTAINERS: update Shawn's email to use kernel.org one
Shawn Guo [Wed, 3 Jun 2015 11:34:10 +0000 (19:34 +0800)]
MAINTAINERS: update Shawn's email to use  one

Update my mailbox to use kernel.org one for handling kernel community
maintenance traffic.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: Kevin Hilman <khilman@linaro.org>
9 years agoMerge tag 'imx-soc-4.2' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo...
Kevin Hilman [Wed, 10 Jun 2015 23:58:05 +0000 (16:58 -0700)]
Merge tag 'imx-soc-4.2' of git://git./linux/kernel/git/shawnguo/linux into next/soc

The i.MX SoC updates for 4.2:
 - Add new SoC i.MX7D support, which integrates two Cortex-A7 and one
   Cortex-M4 cores.
 - Support suspend from IRAM on i.MX53, so that DDR pins can be set to
   high impedance for more power saving during suspend.
 - Move i.MX clock drivers from arch/arm/mach-imx to drivers/clk/imx.
 - Move i.MX GPT timer driver from arch/arm/mach-imx into
   drivers/clocksource.
 - A couple of clock driver update for VF610 and i.MX6Q.
 - A few random code correction and improvement.

* tag 'imx-soc-4.2' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux: (44 commits)
  ARM: imx: imx7d requires anatop
  clocksource: timer-imx-gpt: remove include of <asm/mach/time.h>
  ARM: imx: move timer driver into drivers/clocksource
  ARM: imx: remove platform headers from timer driver
  ARM: imx: provide gpt device specific irq functions
  ARM: imx: get rid of variable timer_base
  ARM: imx: define gpt register offset per device type
  ARM: imx: move clock event variables into imx_timer
  ARM: imx: set up .set_next_event hook via imx_gpt_data
  ARM: imx: setup tctl register in device specific function
  ARM: imx: initialize gpt device type for DT boot
  ARM: imx: define an enum for gpt timer device type
  ARM: imx: move timer resources into a structure
  ARM: imx: use relaxed IO accessor in timer driver
  ARM: imx: make imx51/3 suspend optional
  ARM: clk-imx6q: refine sata's parent
  ARM: imx: clk-v610: Add clock for I2C2 and I2C3
  ARM: mach-imx: iomux-imx31: Use DECLARE_BITMAP
  ARM: imx: add imx7d clk tree support
  ARM: clk: imx: update pllv3 to support imx7
  ...

 Conflicts:
arch/arm/mach-imx/Kconfig

9 years agoMerge branch 'socfpga/soc' into next/soc
Kevin Hilman [Wed, 10 Jun 2015 23:04:48 +0000 (16:04 -0700)]
Merge branch 'socfpga/soc' into next/soc

* socfpga/soc:
  ARM: socfpga: support suspend to ram
  ARM: socfpga: add CPU_METHOD_OF_DECLARE for Arria 10
  ARM: socfpga: use CPU_METHOD_OF_DECLARE for socfpga_cyclone5

9 years agoARM: socfpga: support suspend to ram
Alan Tull [Fri, 5 Jun 2015 13:24:52 +0000 (08:24 -0500)]
ARM: socfpga: support suspend to ram

Add code that requests that the sdr controller go into
self-refresh mode.  This code is run from ocram.

Suspend-to-RAM and EDAC support are mutually exclusive on
SOCFPGA.  If the EDAC is enabled, it will prevent the
platform from going into suspend.

Example of how to request to suspend to ram:
 $ echo enabled > \
/sys/devices/soc/ffc02000.serial0/tty/ttyS0/power/wakeup

 $ echo -n mem > /sys/power/state

Signed-off-by: Alan Tull <atull@opensource.altera.com>
Signed-off-by: Dinh Nguyen <dinguyen@opensource.altera.com>
Signed-off-by: Kevin Hilman <khilman@linaro.org>
9 years agoARM: socfpga: add CPU_METHOD_OF_DECLARE for Arria 10
Dinh Nguyen [Wed, 3 Jun 2015 02:14:02 +0000 (21:14 -0500)]
ARM: socfpga: add CPU_METHOD_OF_DECLARE for Arria 10

Add boot_secondary implementation for the Arria10 platform. Bringing up
the secondary core on the Arria 10 platform is pretty similar to the
Cyclone/Arria 5 platform, with the exception of the following differences:

- Register offset to bringup CPU1 out of reset is different.
- The cpu1-start-addr for Arria10 contains an additional nibble.

Signed-off-by: Dinh Nguyen <dinguyen@opensource.altera.com>
Signed-off-by: Kevin Hilman <khilman@linaro.org>
9 years agoARM: socfpga: use CPU_METHOD_OF_DECLARE for socfpga_cyclone5
Dinh Nguyen [Wed, 3 Jun 2015 02:14:01 +0000 (21:14 -0500)]
ARM: socfpga: use CPU_METHOD_OF_DECLARE for socfpga_cyclone5

Convert cyclone5/arria5 to use CPU_METHOD_OF_DECLARE for smp operations.

Signed-off-by: Dinh Nguyen <dinguyen@opensource.altera.com>
Signed-off-by: Kevin Hilman <khilman@linaro.org>
9 years agoARM: EXYNOS: register power domain driver from core_initcall
Marek Szyprowski [Wed, 3 Jun 2015 23:09:27 +0000 (08:09 +0900)]
ARM: EXYNOS: register power domain driver from core_initcall

SYSMMU devices are registered very early in arch_initcall, so ensure
that they can get access to power domains by registering power domain
driver from earlier initcall. This change requires dropping usage of
the platform device associated with each power domain and replacing
clock calls with respective of_clk_* equivalents.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: EXYNOS: use PS_HOLD based poweroff for all supported SoCs
Marek Szyprowski [Mon, 9 Feb 2015 07:25:41 +0000 (08:25 +0100)]
ARM: EXYNOS: use PS_HOLD based poweroff for all supported SoCs

PS_HOLD based power off procedure is common for all Exynos SoCs,
so use it for every Exynos SoCs.

Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Tested-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Tested-by: Tobias Jakobi <tjakobi@math.uni-bielefeld.de>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: SAMSUNG: Constify platform_device_id
Krzysztof Kozlowski [Fri, 1 May 2015 15:06:35 +0000 (00:06 +0900)]
ARM: SAMSUNG: Constify platform_device_id

The platform_device_id is not modified by the driver and core
uses it as const.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: EXYNOS: Constify irq_domain_ops
Krzysztof Kozlowski [Mon, 27 Apr 2015 10:48:59 +0000 (19:48 +0900)]
ARM: EXYNOS: Constify irq_domain_ops

The irq_domain_ops are not modified by the driver and the irqdomain
core code accepts pointer to a const data.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: EXYNOS: add coupled cpuidle support for Exynos3250
Bartlomiej Zolnierkiewicz [Wed, 18 Mar 2015 13:09:57 +0000 (14:09 +0100)]
ARM: EXYNOS: add coupled cpuidle support for Exynos3250

The following patch adds coupled cpuidle support for Exynos3250 to
an existing cpuidle-exynos driver.  As a result it enables AFTR mode
to be used by default on Exynos3250 without the need to hot unplug
CPU1 first.

The detailed changelog:
- use exynos_[get,set]_boot_addr() in cpuidle-exynos.c and then make
  cpu_boot_reg_base() static
- use exynos_core_restart() in exynos_cpu0_enter_aftr()
- add missing smp_rmb() to exynos_cpu0_enter_aftr() (to make the code
  in-sync with the platform SMP code)
- add call_firmware_op(cpu_boot, 1) to exynos_cpu0_enter_aftr()
- use dsb_sev() instead of IPI wakeup for Exynos3250 in
  exynos_cpu0_enter_aftr()
- add CPU0 vs CPU1 synchronization based on S5P_PMU_SPARE2 register
  for Exynos3250 to cpuidle-exynos.c
- add flush_cache_all() for CPU1/0 before powerdown/AFTR for
  Exynos3250 to exynos_wfi_finisher()/exynos_do_idle()

Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: EXYNOS: add exynos_get_boot_addr() helper
Bartlomiej Zolnierkiewicz [Wed, 18 Mar 2015 13:09:56 +0000 (14:09 +0100)]
ARM: EXYNOS: add exynos_get_boot_addr() helper

Add get_cpu_boot_addr() firmware operation and then
exynos_get_boot_addr() helper.

This is a preparation for adding coupled cpuidle support
for Exynos3250 SoC.

There should be no functional changes caused by this patch.

Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: Russell King <rmk+kernel@arm.linux.org.uk>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: EXYNOS: add exynos_set_boot_addr() helper
Bartlomiej Zolnierkiewicz [Wed, 18 Mar 2015 13:09:55 +0000 (14:09 +0100)]
ARM: EXYNOS: add exynos_set_boot_addr() helper

Add exynos_set_boot_addr() helper and covert existing code
(exynos_boot_secondary() and exynos_smp_prepare_cpus()) to
use it.

There should be no functional changes caused by this patch.

Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: EXYNOS: make exynos_core_restart() less verbose
Bartlomiej Zolnierkiewicz [Wed, 18 Mar 2015 13:09:54 +0000 (14:09 +0100)]
ARM: EXYNOS: make exynos_core_restart() less verbose

There is a kernel message about secondary CPU bootup when
exynos_core_restart() is called through CPU hotplug code-path (the
only exynos_core_restart() user currently) so there is no need for
an extra info on Exynos3250 SoC about software reset.  This also
prepares exynos_core_restart() to be re-used in coupled cpuidle
code-path in the future.

Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: EXYNOS: fix exynos_boot_secondary() return value on timeout
Bartlomiej Zolnierkiewicz [Wed, 18 Mar 2015 13:09:53 +0000 (14:09 +0100)]
ARM: EXYNOS: fix exynos_boot_secondary() return value on timeout

exynos_boot_secondary() can erroneously return 0 or -ENOSYS even
when waiting on pen_release being set to -1 timeouts.  Fix it by
adjusting ret variable value to -ETIMEDOUT when necessary.

Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: EXYNOS: Get current parent clock for power domain on/off
Krzysztof Kozlowski [Fri, 3 Apr 2015 09:25:54 +0000 (11:25 +0200)]
ARM: EXYNOS: Get current parent clock for power domain on/off

Using a fixed (by DTS) parent for clocks when turning on the power
domain may introduce issues in other drivers. For example when such
driver changes the parent during runtime and expects that he is the
only place of such change.

Do not rely on DTS providing the fixed parent for such clocks. Instead
before switching domain off, grab a current parent of a clock with
clk_get_parent().

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Tested-by: Javier Martinez Canillas <javier.martinez@collabora.co.uk>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: SAMSUNG: fix clk_enable() WARNing in S3C24XX ADC
Sergiy Kibrik [Mon, 27 Apr 2015 05:29:44 +0000 (08:29 +0300)]
ARM: SAMSUNG: fix clk_enable() WARNing in S3C24XX ADC

Convert clk_enable/clk_disable to clk_prepare_enable/clk_disable_unprepare
calls as required by common clock framework. Removes this warning on probe:

 ------------[ cut here ]------------
 WARNING: CPU: 0 PID: 1 at drivers/clk/clk.c:889 __clk_enable+0x28/0x9c()
 Modules linked in:
 CPU: 0 PID: 1 Comm: swapper Not tainted 3.19.0+ #46
 [<c0013a98>] (unwind_backtrace) from [<c0010d8c>] (show_stack+0x10/0x14)
 [<c0010d8c>] (show_stack) from [<c001b018>] (warn_slowpath_common+0x7c/0xa4)
 [<c001b018>] (warn_slowpath_common) from [<c001b0d0>] (warn_slowpath_null+0x18/0x20)
 [<c001b0d0>] (warn_slowpath_null) from [<c01a5f80>] (__clk_enable+0x28/0x9c)
 [<c01a5f80>] (__clk_enable) from [<c01a600c>] (clk_enable+0x18/0x2c)
 [<c01a600c>] (clk_enable) from [<c001860c>] (s3c_adc_probe+0x11c/0x18c)
 [<c001860c>] (s3c_adc_probe) from [<c0153a10>] (platform_drv_probe+0x30/0x78)
 [<c0153a10>] (platform_drv_probe) from [<c01523e0>] (driver_probe_device+0xb0/0x1fc)
 [<c01523e0>] (driver_probe_device) from [<c01525dc>] (__driver_attach+0x68/0x88)
 [<c01525dc>] (__driver_attach) from [<c0150df0>] (bus_for_each_dev+0x70/0x94)
 [<c0150df0>] (bus_for_each_dev) from [<c0151c80>] (bus_add_driver+0xdc/0x1c4)
 [<c0151c80>] (bus_add_driver) from [<c0152ba4>] (driver_register+0x9c/0xe0)
 [<c0152ba4>] (driver_register) from [<c03041cc>] (adc_init+0x10/0x34)
 [<c03041cc>] (adc_init) from [<c00087fc>] (do_one_initcall+0x110/0x1cc)
 [<c00087fc>] (do_one_initcall) from [<c02ffccc>] (kernel_init_freeable+0xf4/0x1ac)
 [<c02ffccc>] (kernel_init_freeable) from [<c022f4fc>] (kernel_init+0x8/0xe0)
 [<c022f4fc>] (kernel_init) from [<c000e098>] (ret_from_fork+0x14/0x3c)
 ---[ end trace f4a1ea39a114fecf ]---

Signed-off-by: Sergiy Kibrik <sakib@meta.ua>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: EXYNOS: Add missing of_node_put() when parsing power domains
Krzysztof Kozlowski [Fri, 27 Mar 2015 12:21:28 +0000 (13:21 +0100)]
ARM: EXYNOS: Add missing of_node_put() when parsing power domains

Add missing of_node_put() to:
1. Error return path if allocating memory for exynos_pm_domain failed.
2. Second iteration over power domains if a child domain was not
   present or was incomplete.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Reported-by: Karol Wrona <k.wrona@samsung.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: EXYNOS: Handle of_find_device_by_node() and kstrdup() failures
Krzysztof Kozlowski [Fri, 27 Mar 2015 12:12:00 +0000 (13:12 +0100)]
ARM: EXYNOS: Handle of_find_device_by_node() and kstrdup() failures

Prevent possible NULL pointer dereference of pointer returned by
of_find_device_by_node(). Handle this by skipping such power domain.

Additionally fail the init on kstrdup() failure. Such case is actually
not fatal because the name for power domain allocated by kstrdup() is
used only in printk. Still as a precaution handle this as an error
condition.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoARM: EXYNOS: Handle of of_iomap() failure
Krzysztof Kozlowski [Fri, 27 Mar 2015 12:10:06 +0000 (13:10 +0100)]
ARM: EXYNOS: Handle of of_iomap() failure

Prevent possible NULL pointer dereference if of_iomap() fails.
Handle the error by skipping such power domain.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Kukjin Kim <kgene@kernel.org>
9 years agoarm64: Enable Hisilicon ARMv8 SoC family in Kconfig and defconfig
Bintian Wang [Tue, 6 Jan 2015 01:30:36 +0000 (09:30 +0800)]
arm64: Enable Hisilicon ARMv8 SoC family in Kconfig and defconfig

This patch introduces ARCH_HISI to enable Hisilicon SoC family in
Kconfig and defconfig.

Signed-off-by: Bintian Wang <bintian.wang@huawei.com>
Acked-by: Haojian Zhuang <haojian.zhuang@linaro.org>
Reviewed-by: Wei Xu <xuwei5@hisilicon.com>
Tested-by: Will Deacon <will.deacon@arm.com>
Tested-by: Tyler Baker <tyler.baker@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>
Acked-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Wei Xu <xuwei5@hisilicon.com>
9 years agoARM: imx: imx7d requires anatop
Arnd Bergmann [Fri, 29 May 2015 09:28:05 +0000 (11:28 +0200)]
ARM: imx: imx7d requires anatop

Like i.MX6, the i.MX7 code calls into the anatop driver, which fails
if that is disabled:

arch/arm/mach-imx/built-in.o: In function `imx7d_init_machine':
arch/arm/mach-imx/mach-imx7d.c:24: undefined reference to `imx_anatop_init'
arch/arm/mach-imx/built-in.o: In function `imx7d_init_irq':
arch/arm/mach-imx/mach-imx7d.c:29: undefined reference to `imx_init_revision_from_anatop'

This patch ensures that for an imx7-only build, we still get anatop
built-in, matching what we do for imx6. We also need to select
HAVE_IMX_MMDC, as that is needed by the anatop code.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoclocksource: timer-imx-gpt: remove include of <asm/mach/time.h>
Shawn Guo [Fri, 29 May 2015 13:02:36 +0000 (21:02 +0800)]
clocksource: timer-imx-gpt: remove include of <asm/mach/time.h>

The include of <asm/mach/time.h> is not needed at all, and causes build
error in some cases.  Remove it.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
9 years agoARM: imx: move timer driver into drivers/clocksource
Shawn Guo [Fri, 15 May 2015 07:41:00 +0000 (15:41 +0800)]
ARM: imx: move timer driver into drivers/clocksource

After the cleanup on imx timer driver, now it's ready to be moved into
drivers/clocksource/.  Let's do it.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
9 years agoARM: imx: remove platform headers from timer driver
Shawn Guo [Fri, 15 May 2015 07:27:03 +0000 (15:27 +0800)]
ARM: imx: remove platform headers from timer driver

With the cleanup done before, the platform specific headers now can be
removed.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: provide gpt device specific irq functions
Shawn Guo [Fri, 22 May 2015 14:42:55 +0000 (22:42 +0800)]
ARM: imx: provide gpt device specific irq functions

It splits irq enable/disable/acknowledge operations into device specific
functions as the hooks in imx_gpt_data, so that we can save the use of
timer_is_xxx() and cpu_is_xxx() checking in these irq functions.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: get rid of variable timer_base
Shawn Guo [Fri, 22 May 2015 14:23:28 +0000 (22:23 +0800)]
ARM: imx: get rid of variable timer_base

We now have pointer to imx_timer structure available where timer base
address is needed, so we can just kill global timer_base by using
imxtm->base instead.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: define gpt register offset per device type
Shawn Guo [Fri, 22 May 2015 13:39:55 +0000 (21:39 +0800)]
ARM: imx: define gpt register offset per device type

It defines offset of gpt registers TSTAT, TCN and TCMP per device
type in imx_gpt_data, so that these registers can be accessed in an
way without timer_is_v2() checking.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: move clock event variables into imx_timer
Shawn Guo [Fri, 22 May 2015 08:38:49 +0000 (16:38 +0800)]
ARM: imx: move clock event variables into imx_timer

Since we now have imx_timer structure, it makes more sense to move those
clock event related variables into the structure, so that we can save
some global variables.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: set up .set_next_event hook via imx_gpt_data
Shawn Guo [Fri, 22 May 2015 07:51:41 +0000 (15:51 +0800)]
ARM: imx: set up .set_next_event hook via imx_gpt_data

Set up .set_next_event hook via imx_gpt_data, so that we can save the
use of timer_is_v2().

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: setup tctl register in device specific function
Shawn Guo [Fri, 15 May 2015 06:24:41 +0000 (14:24 +0800)]
ARM: imx: setup tctl register in device specific function

It creates a gpt device speicific data structure and adds function hook
gpt_setup_tctl in there to set up gpt TCTL register.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: initialize gpt device type for DT boot
Shawn Guo [Fri, 15 May 2015 05:38:20 +0000 (13:38 +0800)]
ARM: imx: initialize gpt device type for DT boot

Use different initialization function in CLOCKSOURCE_OF_DECLARE() to
initialize gpt device type for DT boot.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: define an enum for gpt timer device type
Shawn Guo [Fri, 15 May 2015 03:41:39 +0000 (11:41 +0800)]
ARM: imx: define an enum for gpt timer device type

Define an enum for gpt timer device type in include/soc/imx/timer.h to
tell the gpt block differences among SoCs.  Update non-DT users (clock
drivers) to pass the device type.

As we now have include/soc/imx/timer.h, the declaration of
mxc_timer_init() is moved into there as the best fit.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: move timer resources into a structure
Shawn Guo [Fri, 22 May 2015 05:53:45 +0000 (13:53 +0800)]
ARM: imx: move timer resources into a structure

Instead of passing around as individual argument, let's move timer
resources like irq and clocks together with base address into a data
structure, and pass pointer of the structure as argument to simplify
the function call interface.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: use relaxed IO accessor in timer driver
Shawn Guo [Tue, 19 May 2015 10:47:47 +0000 (18:47 +0800)]
ARM: imx: use relaxed IO accessor in timer driver

Replace the __raw_readl/__raw_writel with readl_relaxed/writel_relaxed
which is endian-safe, as a step of moving the driver code into folder
drivers/clocksource.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: make imx51/3 suspend optional
Arnd Bergmann [Thu, 21 May 2015 12:06:30 +0000 (14:06 +0200)]
ARM: imx: make imx51/3 suspend optional

A recent change to the imx53 power management caused a build
regression when CONFIG_SOC_IMX53 is disabled:

mach-imx/built-in.o:(.init.rodata+0x60): undefined reference to `imx53_suspend'
mach-imx/built-in.o:(.init.rodata+0x64): undefined reference to `imx53_suspend_sz'

This avoids the problem by compiling the code in question
conditionally on the presence of CONFIG_SOC_IMX53. For
consistency, I'm also changing the same thing for
CONFIG_SOC_IMX51.

An additional benefit of this approach is reduced code size
for kernels that only include support for one of the two
SoCs.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: 1579c7b9fe01 ("ARM: imx53: Set DDR pins to high impedance when in suspend to RAM.")
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: clk-imx6q: refine sata's parent
Sébastien Szymanski [Wed, 20 May 2015 14:30:37 +0000 (16:30 +0200)]
ARM: clk-imx6q: refine sata's parent

According to IMX6D/Q RM, table 18-3, sata clock's parent is ahb, not ipg.

Signed-off-by: Sébastien Szymanski <sebastien.szymanski@armadeus.com>
Reviewed-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: clk-v610: Add clock for I2C2 and I2C3
Mirza Krak [Wed, 20 May 2015 09:38:03 +0000 (11:38 +0200)]
ARM: imx: clk-v610: Add clock for I2C2 and I2C3

Add support for clock gating of I2C2 and I2C3.
We use I2C2 in a (not yet mainlined) device tree.

Signed-off-by: Mirza Krak <mirza.krak@hostmobility.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: mach-imx: iomux-imx31: Use DECLARE_BITMAP
Joe Perches [Wed, 20 May 2015 01:37:49 +0000 (18:37 -0700)]
ARM: mach-imx: iomux-imx31: Use DECLARE_BITMAP

Use the generic mechanism to declare a bitmap instead of unsigned long.

Signed-off-by: Joe Perches <joe@perches.com>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: add imx7d clk tree support
Frank Li [Mon, 18 May 2015 18:45:03 +0000 (02:45 +0800)]
ARM: imx: add imx7d clk tree support

Add i.MX7D clk tree support.

Enable all clock to bring up imx7.
Clock framework need be modified a little since imx7d
change clock design. otherwise system will halt and block the
other part upstream.

All clock refine need wait for Dong Aisheng's patch
clk: support clocks which requires parent clock on during operation
Or other solution ready.

Signed-off-by: Anson Huang <b20788@freescale.com>
Signed-off-by: Adrian Alonso <aalonso@freescale.com>
Signed-off-by: Frank Li <Frank.Li@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: clk: imx: update pllv3 to support imx7
Frank Li [Mon, 18 May 2015 18:45:02 +0000 (02:45 +0800)]
ARM: clk: imx: update pllv3 to support imx7

Add type IMX_PLLV3_ENET_IMX7

Signed-off-by: Frank Li <Frank.Li@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: clk-vf610: enable debug access port by default
Stefan Agner [Sun, 17 May 2015 22:13:33 +0000 (00:13 +0200)]
ARM: imx: clk-vf610: enable debug access port by default

Enabled DAP (debug access port) by default. This enables the hw-
breakpoint framework to make use of the breakpoints and watchpoints
supported by hardware.

[    0.215805] hw-breakpoint: found 2 (+1 reserved) breakpoint and 1 watchpoint registers.
[    0.224624] hw-breakpoint: maximum watchpoint size is 4 bytes.

Without this clock, the hw-breakpoint driver claims an undefined
instruction during initialization:
[    0.227380] hw-breakpoint: Debug register access (0xee003e17) caused undefined instruction on CPU 0
[    0.227519] hw-breakpoint: CPU 0 failed to disable vector catch

Signed-off-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx53: Set DDR pins to high impedance when in suspend to RAM.
Martin Fuzzey [Tue, 12 May 2015 13:31:03 +0000 (15:31 +0200)]
ARM: imx53: Set DDR pins to high impedance when in suspend to RAM.

In order to save power the DDR pins should be put into high
impedance when in suspend to RAM.

This requires manually requesting self refresh (rather than using the
automatic mode implemented by the CCM / ESDCTL), followed by
reconfiguring the IOMUXC.

Of course the code to do this cannot itself run from DDR so the
code is copied to and executed from internal memory.

In my tests using a custom i.MX53 board with LPDDR2 RAM
this reduced the suspend power consumption from 200mW to 60mW.

Signed-off-by: Martin Fuzzey <mfuzzey@parkeon.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: add msl support for imx7d
Anson Huang [Thu, 7 May 2015 17:35:55 +0000 (01:35 +0800)]
ARM: imx: add msl support for imx7d

Add i.MX7D MSL support.

Signed-off-by: Anson Huang <b20788@freescale.com>
Signed-off-by: Frank Li <Frank.Li@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx7d: add low level debug uart support
Anson Huang [Wed, 6 May 2015 15:16:07 +0000 (23:16 +0800)]
ARM: imx7d: add low level debug uart support

Add low level uart debug support for imx7d

Signed-off-by: Frank Li <Frank.Li@freescale.com>
Signed-off-by: Bai Ping <b51503@freescale.com>
Signed-off-by: Anson Huang <b20788@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: mmdc: Include "common.h" header file
Fabio Estevam [Thu, 2 Apr 2015 22:23:32 +0000 (19:23 -0300)]
ARM: imx: mmdc: Include "common.h" header file

Include the "common.h" header file to fix the following sparse warning:

arch/arm/mach-imx/mmdc.c:66:5: warning: symbol 'imx_mmdc_get_ddr_type' was not declared. Should it be static?

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoclk: imx: clk-cpu: Include "clk.h" header file
Fabio Estevam [Wed, 29 Apr 2015 21:34:42 +0000 (18:34 -0300)]
clk: imx: clk-cpu: Include "clk.h" header file

Include the "clk.h" header file to fix the following sparse warning:

drivers/clk/imx/clk-cpu.c:77:12: warning: symbol 'imx_clk_cpu' was not declared. Should it be static?

Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: Correct the comments in time.c
Shenwei Wang [Wed, 29 Apr 2015 21:40:27 +0000 (16:40 -0500)]
ARM: imx: Correct the comments in time.c

The comments were corrected as the following to reflect
the real situation of Freescale MXC timer IP block.
There are totally 4 version of the timer on Freescale i.MX SoCs.

Signed-off-by: Shenwei Wang <shenwei.wang@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: Remove the duplicated function declaration
Shenwei Wang [Wed, 29 Apr 2015 21:18:37 +0000 (16:18 -0500)]
ARM: imx: Remove the duplicated function declaration

Removed the duplicated function declaration of mxc_timer_init
which was already declared in drivers/clk/imx/clk.h.

Signed-off-by: Shenwei Wang <shenwei.wang@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: Constify irq_domain_ops
Krzysztof Kozlowski [Mon, 27 Apr 2015 12:51:39 +0000 (21:51 +0900)]
ARM: imx: Constify irq_domain_ops

The irq_domain_ops are not modified by the driver and the irqdomain core
code accepts pointer to a const data.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski.k@gmail.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: using unsigned variable for do_div
Anson Huang [Thu, 7 May 2015 16:16:51 +0000 (00:16 +0800)]
ARM: imx: using unsigned variable for do_div

The definition of do_div uses unsigned long long
variable as its first parameter, better to pass
a u64 variable as first parameter when calling
do_div function.

Signed-off-by: Anson Huang <b20788@freescale.com>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx6: initialize CCM_CLPCR_LPM into RUN mode earlier
Shawn Guo [Wed, 29 Apr 2015 05:07:03 +0000 (13:07 +0800)]
ARM: imx6: initialize CCM_CLPCR_LPM into RUN mode earlier

Commit 4631960d26da ("ARM: imx6: set initial power mode in pm function")
moves imx6_set_lpm() from clock init function into
imx6_pm_common_init().  This causes a hang when cpuidle support is
enabled.  The reason for that is ARM core clock is shut down
unexpectedly by WAIT mode.  It happens with the following call stack:

    cpuidle_register_governor()
        cpuidle_switch_governor()
            cpuidle_uninstall_idle_handler()
                synchronize_sched()
                    wait_rcu_gp()
                        wait_for_completion()

When wait_for_completion() is called as above, all cores are idle/WFI.
Hence, the reset value of CCM_CLPCR_LPM - WAIT mode, will trigger a
hardware shutdown of the ARM core clock.

To fix the regression, we need to ensure that CCM_CLPCR_LPM is
initialized into RUN mode earlier than cpuidle governor registration,
which is a postcore_initcall.  This patch creates function
imx6_pm_ccm_init() to map CCM block and initialize CCM_CLPCR_LPM into
RUN mode, and have the function called from machine .init_irq hook,
which should be early enough.

Reported-by: Kevin Hilman <khilman@kernel.org>
Fixes: 8fb76a07e2cb ("ARM: imx6: set initial power mode in pm function")
Tested-by: Kevin Hilman <khilman@linaro.org>
Tested-by: Tyler Baker <tyler.baker@linaro.org>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: drop epit timer initialization from imx35 clock driver
Shawn Guo [Tue, 28 Apr 2015 01:19:02 +0000 (09:19 +0800)]
ARM: imx: drop epit timer initialization from imx35 clock driver

EPIT provides another timer implementation besides the default GPT
timer.  The imx35 clock driver will use EPIT timer when option
CONFIG_MXC_USE_EPIT is enabled.  However, initializing timers from
clock driver is a workaround solution and causes problem when we move
clock drivers into driver/clk.

Let's simply drop the EPIT initialization from there.  If people really
want this EPIT option, EPIT timer driver needs to be reworked to do the
initialization in a standard way - use CLOCKSOURCE_OF_DECLARE() with
device tree support.

Reported-by: kbuild test robot <fengguang.wu@intel.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoMAINTAINERS: add new folders into IMX entry
Shawn Guo [Sun, 26 Apr 2015 13:58:12 +0000 (21:58 +0800)]
MAINTAINERS: add new folders into IMX entry

Add new created folders drivers/clk/imx/ and include/soc/imx/ into IMX
entry.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: move clock drivers into drivers/clk
Shawn Guo [Sun, 26 Apr 2015 13:54:29 +0000 (21:54 +0800)]
ARM: imx: move clock drivers into drivers/clk

After the cleanup on clock drivers, they are now ready to be moved into
drivers/clk.  Let's move them into drivers/clk/imx folder.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
Acked-by: Stephen Boyd <sboyd@codeaurora.org>
9 years agoARM: imx: remove inclusions of platform headers
Shawn Guo [Sat, 25 Apr 2015 10:43:45 +0000 (18:43 +0800)]
ARM: imx: remove inclusions of platform headers

With the cleanup done before, we now can simply define base address and
irq as needed in clock driver, to get those platform header inclusions
removed.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: add clk-pllv1 type support
Shawn Guo [Sun, 26 Apr 2015 05:33:39 +0000 (13:33 +0800)]
ARM: imx: add clk-pllv1 type support

Instead of calling cpu_is_xxx() in clk-pllv1 driver, let's add clk-pllv1
type support to handle the difference/quirk in particular SoC designs.
Doing so will help get clk-pllv1 driver ready for being moved out of
arch/arm/mach-imx folder.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx6: do not use cpu_is_xxx() in clock driver
Shawn Guo [Sun, 26 Apr 2015 02:43:52 +0000 (10:43 +0800)]
ARM: imx6: do not use cpu_is_xxx() in clock driver

As we're about to move clock drivers out of arch/arm/mach-imx,
cpu_is_xxx() shouldn't be used any more.  Let's avoid the call by
looking at the device tree machine compatible string to determine
which SoC the clock driver is running on.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx6: let pm code map CCM block on its own
Shawn Guo [Sat, 25 Apr 2015 15:37:12 +0000 (23:37 +0800)]
ARM: imx6: let pm code map CCM block on its own

We are about to move imx6 clock driver into drivers/clk, so let's get
imx6 pm code map CCM block on its own rather than relying on clock
driver to do the mapping.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx6: set initial power mode in pm function
Shawn Guo [Sat, 25 Apr 2015 14:59:19 +0000 (22:59 +0800)]
ARM: imx6: set initial power mode in pm function

Rather than setting initial low-power mode in every single i.MX6 clock
initialization function, we should really do that in pm code.  Let's
move imx6q_set_lpm(WAIT_CLOCKED) call into imx6_pm_common_init().

While at it, let's rename the function to imx6_set_lpm() since it's
actually common for all i.MX6 SoCs.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx5: let pm code map CCM block on its own
Shawn Guo [Sat, 25 Apr 2015 14:38:19 +0000 (22:38 +0800)]
ARM: imx5: let pm code map CCM block on its own

We are about to move imx5 clock driver into drivers/clk, so let's get
imx5 pm code map CCM block on its own rather than relying on clock
driver to do the mapping.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: move revision definitions and declarations into a header
Shawn Guo [Sat, 25 Apr 2015 13:03:15 +0000 (21:03 +0800)]
ARM: imx: move revision definitions and declarations into a header

The revision definitions and declarations are widely used by clock
drivers.  As a step of moving clock drivers out of arch/arm/mach-imx,
let's create header include/soc/imx/revision.h to accommodate them.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: use dynamic mapping for CCM
Shawn Guo [Sat, 25 Apr 2015 08:02:53 +0000 (16:02 +0800)]
ARM: imx: use dynamic mapping for CCM

Replace the static mapping of CCM block in clock drivers with dynamic
mapping.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: imx: use dynamic mapping for timer
Shawn Guo [Sat, 25 Apr 2015 07:44:10 +0000 (15:44 +0800)]
ARM: imx: use dynamic mapping for timer

Pass physical address of timer block to mxc_timer_init() call, which in
turn does dynamic mapping within the function.  Thus, we can avoid using
static mapping in clock drivers.

Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agodt-bindings: add imx7d clock ID definitions
Frank Li [Wed, 3 Jun 2015 06:41:29 +0000 (14:41 +0800)]
dt-bindings: add imx7d clock ID definitions

It adds the imx7d clock ID definitions which will be used by both imx7d
clock driver and device tree.

Signed-off-by: Frank Li <Frank.Li@freescale.com>
Signed-off-by: Shawn Guo <shawn.guo@linaro.org>
9 years agoARM: Kconfig: Select clocksource in STM32 entry
Maxime Coquelin [Fri, 22 May 2015 21:50:52 +0000 (23:50 +0200)]
ARM: Kconfig: Select clocksource in STM32 entry

STM32 clocksource driver needs to be selected if ARCH_STM32.

Signed-off-by: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
9 years agoMerge tag 'arm-soc/for-4.2/soc-part2' of http://github.com/broadcom/stblinux into...
Arnd Bergmann [Mon, 1 Jun 2015 15:54:31 +0000 (17:54 +0200)]
Merge tag 'arm-soc/for-4.2/soc-part2' of github.com/broadcom/stblinux into next/soc

Merge "changes for Broadcom SoCs":

- Dan fixes an error path in the BCM63xx SMP code

- Ray adds the relevant Kconfig selects to enable the Broadcom NAND driver on Cygnus

- Kevin provides a change to the Broadcom GISB arbiter driver to make it work with
  MIPS-based big-endian STB SoCs (this was a long-standing change that had dependencies on
  code in drivers/of/*)

- Gregory enables the use of GPIOLIB for brcmstb SoCs and bumps the number of GPIOs for
  these platforms

* tag 'arm-soc/for-4.2/soc-part2' of http://github.com/broadcom/stblinux:
  ARM: brcmstb: Add default gpio number
  ARM: brcmstb: Select ARCH_WANT_OPTIONAL_GPIOLIB
  bus: brcmstb_gisb: Honor the "big-endian" and "native-endian" DT properties
  ARM: BCM: Enable NAND support for iProc SoCs
  ARM: BCM63xx: fix an error path in bcm63xx_pmb_power_on_cpu()

9 years agoARM: brcmstb: Add default gpio number
Gregory Fong [Fri, 29 May 2015 02:14:10 +0000 (19:14 -0700)]
ARM: brcmstb: Add default gpio number

Out of the brcmstb SoCs that I know, BCM3390 has the largest numbers
of GPIOs, with its
- 320 "peripheral" GPIOs
- 5*32 = 160 UPG GPIOs (counting unused lines, which do get counted)
- 2*32 = 64 UPG AON GPIOs (counting unused lines)
Total: 544

I suspect that the upper limit will only need to be higher in the
future, so set it to 1024.

Signed-off-by: Gregory Fong <gregory.0xf0@gmail.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
9 years agoARM: brcmstb: Select ARCH_WANT_OPTIONAL_GPIOLIB
Gregory Fong [Fri, 29 May 2015 02:14:09 +0000 (19:14 -0700)]
ARM: brcmstb: Select ARCH_WANT_OPTIONAL_GPIOLIB

Select ARCH_WANT_OPTIONAL_GPIOLIB from BRCMSTB to allow GPIOLIB and
GPIO_BRCMSTB to be enabled.

Signed-off-by: Gregory Fong <gregory.0xf0@gmail.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
9 years agoMerge tag 'renesas-soc-for-v4.2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Arnd Bergmann [Fri, 29 May 2015 13:00:02 +0000 (15:00 +0200)]
Merge tag 'renesas-soc-for-v4.2' of git://git./linux/kernel/git/horms/renesas into next/soc

Merge "Renesas ARM Based SoC Updates for v4.2" from Simon Horman:

* Only select sound drivers that build

* tag 'renesas-soc-for-v4.2' of git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas:
  ARM: shmobile: only select sound drivers that build

9 years agobus: brcmstb_gisb: Honor the "big-endian" and "native-endian" DT properties
Kevin Cernekee [Wed, 26 Nov 2014 00:49:53 +0000 (16:49 -0800)]
bus: brcmstb_gisb: Honor the "big-endian" and "native-endian" DT properties

On chips strapped for BE, we'll need to use ioread32be/iowrite32be instead of
ioread32/iowrite32.

Signed-off-by: Kevin Cernekee <cernekee@gmail.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
9 years agoARM: BCM: Enable NAND support for iProc SoCs
Ray Jui [Fri, 29 May 2015 01:09:14 +0000 (18:09 -0700)]
ARM: BCM: Enable NAND support for iProc SoCs

Select CONFIG_MTD_NAND_BRCMNAND for all iProc SoCs

Signed-off-by: Ray Jui <rjui@broadcom.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
9 years agoARM: BCM63xx: fix an error path in bcm63xx_pmb_power_on_cpu()
Dan Carpenter [Wed, 27 May 2015 08:25:12 +0000 (11:25 +0300)]
ARM: BCM63xx: fix an error path in bcm63xx_pmb_power_on_cpu()

We need to unlock and unmap some resourses before returning.

Fixes: 3f2a43c98d72 ('ARM: BCM63xx: Add secondary CPU PMB initialization sequence')
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
9 years agoARM: shmobile: only select sound drivers that build
Arnd Bergmann [Tue, 19 May 2015 13:40:37 +0000 (15:40 +0200)]
ARM: shmobile: only select sound drivers that build

A couple of codec drivers are selected by shmobile platform code,
but depend on I2C, which results in a build error:

sound/soc/codecs/ak4642.c:638:1: warning: data definition has no type or storage class
 module_i2c_driver(ak4642_i2c_driver);
 ^
sound/soc/codecs/ak4642.c:638:1: error: type defaults to 'int' in declaration of 'module_i2c_driver' [-Werror=implicit-int]
sound/soc/codecs/ak4642.c:638:1: warning: parameter names (without types) in function declaration
sound/soc/codecs/ak4642.c:627:26: warning: 'ak4642_i2c_driver' defined but not used [-Wunused-variable]

This ensures that we do not enable the respective drivers when I2C
is disabled.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Simon Horman <horms+renesas@verge.net.au>
9 years agoARM: use ARM_SINGLE_ARMV7M for ARMv7-M platforms
Stefan Agner [Wed, 20 May 2015 22:35:44 +0000 (00:35 +0200)]
ARM: use ARM_SINGLE_ARMV7M for ARMv7-M platforms

Use the new config symbol ARM_SINGLE_ARMV7M which groups config
symbols used by modern ARMv7-M platforms. This allows supporting
multiple ARMv7-M platforms in one kernel image. However, a common
kernel image requires the combined platforms to share the same
main memory layout to be bootable.

Signed-off-by: Stefan Agner <stefan@agner.ch>
Acked-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Acked-by: Joachim Eastwood <manabian@gmail.com>
Acked-by: Maxime Coquelin <maxime.coquelin@st.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
9 years agoARM: zx: fix building with CONFIG_THUMB2_KERNEL
Arnd Bergmann [Thu, 21 May 2015 08:36:42 +0000 (10:36 +0200)]
ARM: zx: fix building with CONFIG_THUMB2_KERNEL

The newly added zx platform causes a build error when
CONFIG_THUMB2_KERNEL is enabled:

arch/arm/mach-zx/headsmp.S:16: Error: invalid immediate for address calculation (value = 0x00000004)

I'm assuming that the ROM code that is calling these entry
points runs in ARM mode, so there would be another problem
in the same file, and we can solve both problems at once
by adding a '.arm' statement that will make zx_resume_jump
and zx_secondary_startup both be built as ARM code.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Jun Nie <jun.nie@linaro.org>
Tested-by: Jun Nie <jun.nie@linaro.org>
9 years agoMerge tag 'omap-for-v4.2/omap1-v2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Arnd Bergmann [Fri, 22 May 2015 14:34:06 +0000 (16:34 +0200)]
Merge tag 'omap-for-v4.2/omap1-v2' of git://git./linux/kernel/git/tmlind/linux-omap into next/soc

Merge fixed up omap1 sparse irq support for v4.2 from Tony Lindgren:

Add support for CONFIG_SPARSE_IRQ for omap1. This takes us a bit closer
to making omap1 support multiarch. After this series we still need to
make omap1 use the common clock framework and fix up the drivers to not
rely on includes from mach and plat directories.

Note that this branch depends on a GPIO driver fix in v4.1-rc3
d2d05c65c40e ("gpio: omap: Fix regression for MPUIO interrupts").

* tag 'omap-for-v4.2/omap1-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap:
  ARM: OMAP1: Fix section mismatch warnings for omap_cfg_reg
  ARM: OMAP1: Fix randconfig builds if ARCH_OMAP15XX not selected
  ARM: OMAP1: Change interrupt numbering for sparse IRQ
  ARM: omap1: Switch to use MULTI_IRQ
  ARM: OMAP1: Switch to use generic irqchip in preparation for sparse IRQ
  ARM: OMAP1: Move UART defines to prepare for sparse IRQ

9 years agoMerge tag 'arm-soc/for-4.2/soc-take2' of http://github.com/broadcom/stblinux into...
Arnd Bergmann [Fri, 22 May 2015 14:32:02 +0000 (16:32 +0200)]
Merge tag 'arm-soc/for-4.2/soc-take2' of github.com/broadcom/stblinux into next/soc

Merge mach-bcm changes from Florian Fainelli:

This pull request contains the following changes:

- Rafal adds an additional fault code to be ignored by the kernel on BCM5301X SoC

- BCM63138 SMP support which:
* common code to control the PMB bus, to be shared with a reset
  controller driver in drivers/reset
* secondary CPU initialization sequence using PMB helpers
* small changes suggested by Russell King to allow platforms to disable VFP

* tag 'arm-soc/for-4.2/soc-take2' of http://github.com/broadcom/stblinux:
  ARM: BCM63xx: Add SMP support for BCM63138
  ARM: vfp: Add vfp_disable for problematic platforms
  ARM: vfp: Add include guards
  ARM: BCM63xx: Add secondary CPU PMB initialization sequence
  ARM: BCM63xx: Add Broadcom BCM63xx PMB controller helpers
  ARM: BCM5301X: Ignore another (BCM4709 specific) fault code

9 years agoARM: OMAP1: Fix section mismatch warnings for omap_cfg_reg
Tony Lindgren [Thu, 21 May 2015 21:50:23 +0000 (14:50 -0700)]
ARM: OMAP1: Fix section mismatch warnings for omap_cfg_reg

This is cleary used after init time too for example for
configuring UART wake-up events during runtime. This fixes
section mismatch warnings for randconfig builds that happen
because __init_or_module.

Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: OMAP1: Fix randconfig builds if ARCH_OMAP15XX not selected
Tony Lindgren [Thu, 21 May 2015 21:50:23 +0000 (14:50 -0700)]
ARM: OMAP1: Fix randconfig builds if ARCH_OMAP15XX not selected

With the omap1 SPARSE_IRQ changes mach/irqs.h is no longer
automatically included. Turns out now we rely on ARCH_OMAP15XX
including hardware.h from memory.h, so without ARCH_OMAP15XX
we get build failures.

As we have legacy drivers still relying on these indirect
includes, let's not add more mach includes to the drivers.
Those have to be removed anyways for multiplatform support.

Let's fix up mach-omap1 to include soc.h where cpu_is_omap
checks are done, and common.h for board-*.c files.

But let's keep the indirect memory.h include for now to avoid
unnecessary churn in the drivers.

Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: BCM63xx: Add SMP support for BCM63138
Florian Fainelli [Fri, 15 Aug 2014 02:37:45 +0000 (19:37 -0700)]
ARM: BCM63xx: Add SMP support for BCM63138

Add support for booting the secondary CPU on BCM63138, this involves:

- locating the bootlut to write the reset vector
- powering up the second CPU when we need to using the DT-supplied PMB
  references
- disabling VFP when enabled such that we can keep having SMP

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
9 years agoARM: vfp: Add vfp_disable for problematic platforms
Florian Fainelli [Fri, 17 Apr 2015 23:53:58 +0000 (16:53 -0700)]
ARM: vfp: Add vfp_disable for problematic platforms

Some platforms might not be able to fully utilize VFP when e.g: one CPU
out of two in a SMP complex lacks a VFP unit. Adding code to migrate
task to the CPU which has a VFP unit would be cumbersome and not
performant, instead, just add the ability to disable VFP.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
9 years agoARM: vfp: Add include guards
Florian Fainelli [Fri, 17 Apr 2015 23:53:02 +0000 (16:53 -0700)]
ARM: vfp: Add include guards

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
9 years agoARM: BCM63xx: Add secondary CPU PMB initialization sequence
Florian Fainelli [Fri, 17 Apr 2015 17:50:40 +0000 (10:50 -0700)]
ARM: BCM63xx: Add secondary CPU PMB initialization sequence

The sequence to initialize a secondary CPU using the BCM63138 PMB is
extremely specific and represents much more code than any other on-chip
peripheral (AHCI, USB 3.0 or integrated Ethernet switch), as such we
keep that code local and utilize Device Tree to lookup all the resources
we need from the CPU device tree node.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
9 years agoARM: BCM63xx: Add Broadcom BCM63xx PMB controller helpers
Florian Fainelli [Fri, 17 Apr 2015 17:45:55 +0000 (10:45 -0700)]
ARM: BCM63xx: Add Broadcom BCM63xx PMB controller helpers

This patch adds both common register definitions and helper functions
used to issue read/write commands to the Broadcom BCM63xx PMB controller
which is used to power on and release from reset internal on-chip
peripherals such as the integrated Ethernet switch, AHCI, USB, as well
as the secondary CPU core.

This is going to be utilized by the BCM63138 SMP code, as well as by the
BCM63138 reset controller later.

Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
9 years agoARM: vf610: enable Cortex-M4 configuration on Vybrid SoC
Stefan Agner [Tue, 19 May 2015 22:03:52 +0000 (00:03 +0200)]
ARM: vf610: enable Cortex-M4 configuration on Vybrid SoC

This patch allows to build the Kernel for Vybrid (VF6xx) SoC
when ARMv7-M CPU is selected. The resulting image runs on the
secondary Cortex-M4 core. This core has equally access to all
peripherals as the main Cortex-A5 core. However, there is no
resource control mechanism, hence when both cores are used
simultaneously, orthogonal device tree's are required.

The boot CPU is dependent on the SoC variant. The available
boards use mostly variants where the Cortex-A5 is the primary
and hence the boot CPU. Booting the secondary Cortex-M4 CPU
needs SoC specific registers written. There is no in kernel
support for this right now, a external userspace utility
called "m4boot" can be used to boot the kernel:

m4boot xipImage initramfs.cpio.lzo vf610m4-colibri.dtb

Signed-off-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
9 years agoARM: introduce ARM_SINGLE_ARMV7M for ARMv7-M platforms
Stefan Agner [Tue, 19 May 2015 22:03:51 +0000 (00:03 +0200)]
ARM: introduce ARM_SINGLE_ARMV7M for ARMv7-M platforms

This introduces a new top level config symbol ARM_SINGLE_ARMV7M
for non-MMU, ARMv7-M platforms. It also support multiple ARMv7-M
platforms in one kernel image since the cores share the same
basic memory layout and interrupt controller. However, this works
only if the combined platforms also have a similar (main) memory
layout.

Signed-off-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
9 years agoARM: unify MMU/!MMU addruart calls
Stefan Agner [Tue, 19 May 2015 22:03:50 +0000 (00:03 +0200)]
ARM: unify MMU/!MMU addruart calls

Remove the needless differences between MMU/!MMU addruart calls.
This allows to use the same addruart macro on SoC level. Useful
for SoC consisting of multiple CPUs with and without MMU such as
Freescale Vybrid.

Signed-off-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
9 years agoMerge branch 'irq/for-arm' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip...
Arnd Bergmann [Wed, 20 May 2015 21:09:12 +0000 (23:09 +0200)]
Merge branch 'irq/for-arm' of git://git./linux/kernel/git/tip/tip into next/soc

* 'irq/for-arm' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  irqchip: vf610-mscm: Support NVIC parent chip
  irqchip: nvic: Support hierarchy irq domain
  genirq: generic chip: Support hierarchy domain
  genirq: Add irq_chip_(enable/disable)_parent
  irqdomain: Add non-hierarchy helper irq_domain_set_info

9 years agoARM: BCM5301X: Ignore another (BCM4709 specific) fault code
Rafał Miłecki [Wed, 11 Feb 2015 15:40:58 +0000 (16:40 +0100)]
ARM: BCM5301X: Ignore another (BCM4709 specific) fault code

Broadcom ARM devices seem to generate some fault once per boot. We
already have an ignoring handler for BCM4707/BCM4708, but BCM4709
generates different code.

Signed-off-by: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
9 years agoMerge tag 'arm-soc/for-4.2/maintainers' of http://github.com/broadcom/stblinux into...
Arnd Bergmann [Wed, 20 May 2015 20:37:26 +0000 (22:37 +0200)]
Merge tag 'arm-soc/for-4.2/maintainers' of github.com/broadcom/stblinux into next/soc

This pull request for the MAINTAINERS file contains the following changes:

- Brian adds a general "brcmstb" regexp to catch Broadcom Set Top Box related
  changes throughout the Linux tree

* tag 'arm-soc/for-4.2/maintainers' of http://github.com/broadcom/stblinux:
  MAINTAINERS: add brcmstb regex

9 years agoARM: OMAP1: Change interrupt numbering for sparse IRQ
Tony Lindgren [Wed, 20 May 2015 16:01:21 +0000 (09:01 -0700)]
ARM: OMAP1: Change interrupt numbering for sparse IRQ

Change interrupt numbering for sparse IRQ. We do this using
a fixed offset until we can drop irqs.h once all it's users
have been updated.

Note that this depends on the GPIO fix for the MPUIO IRQs
"gpio: omap: Fix regression for MPUIO interrupts".

Also note that this patch adds some extra irq alloc warnings
that will go away when we stop calling irq_alloc_descs
in gpio-omap.c with a follow-up patch.

Cc: Aaro Koskinen <aaro.koskinen@iki.fi>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: omap1: Switch to use MULTI_IRQ
Tony Lindgren [Wed, 20 May 2015 16:01:21 +0000 (09:01 -0700)]
ARM: omap1: Switch to use MULTI_IRQ

This allows us to get a bit further with SPARSE_IRQ and
MULTIARCH support.

Note that we now also rename omap_irq_flags to omap_l2_irq
as that's the omap_irq_flags naming is confusing. It just
contains the interrupt number for the l2 irq.

Cc: Aaro Koskinen <aaro.koskinen@iki.fi>
Signed-off-by: Tony Lindgren <tony@atomide.com>
9 years agoARM: OMAP1: Switch to use generic irqchip in preparation for sparse IRQ
Tony Lindgren [Wed, 20 May 2015 16:01:21 +0000 (09:01 -0700)]
ARM: OMAP1: Switch to use generic irqchip in preparation for sparse IRQ

Let's set up things ready for enabling sparse IRQ and remove the
omap_read/write usage.

Cc: Aaro Koskinen <aaro.koskinen@iki.fi>
Signed-off-by: Tony Lindgren <tony@atomide.com>