GitHub/LineageOS/G12/android_kernel_amlogic_linux-4.9.git
10 years agoMerge tag 'sunxi-drivers-for-3.18' of git://git.kernel.org/pub/scm/linux/kernel/git...
Arnd Bergmann [Thu, 25 Sep 2014 16:15:34 +0000 (18:15 +0200)]
Merge tag 'sunxi-drivers-for-3.18' of git://git./linux/kernel/git/mripard/linux into next/drivers

Pull "Allwinner drivers additions for 3.18" from Maxime Ripard:

Nothing major, just handling the RTC driver changes needed for the A31/A23.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
* tag 'sunxi-drivers-for-3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/mripard/linux:
  rtc: sunxi: Depend on platforms sun4i/sun7i that actually have the rtc
  rtc: sun6i: Add sun6i RTC driver

10 years agodrivers/soc: ti: fix build break with modules
Olof Johansson [Wed, 24 Sep 2014 18:42:38 +0000 (11:42 -0700)]
drivers/soc: ti: fix build break with modules

Fixes below build break by not switching to stubs when the driver is a module:

drivers/soc/ti/knav_dma.c:418:7: error: redefinition of 'knav_dma_open_channel'
 void *knav_dma_open_channel(struct device *dev, const char *name,
       ^
In file included from drivers/soc/ti/knav_dma.c:26:0:
include/linux/soc/ti/knav_dma.h:165:21: note: previous definition of 'knav_dma_open_channel' was here
 static inline void *knav_dma_open_channel(struct device *dev, const char *name,
                     ^

Cc: Santosh Shilimkar <santosh.shilimkar@ti.com>
Signed-off-by: Olof Johansson <olof@lixom.net>
10 years agoMerge tag 'drivers-soc-ti-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/ssanto...
Olof Johansson [Wed, 24 Sep 2014 17:35:48 +0000 (10:35 -0700)]
Merge tag 'drivers-soc-ti-v2' of git://git./linux/kernel/git/ssantosh/linux-keystone into next/drivers

Merge "soc: Keystone SOC Navigator drivers for 3.18" from Santosh Shilimkar:

Keystone SOC Navigator drivers for 3.18

The Keystone Multi-core Navigator contains QMSS and packet DMA
subsystems which interwork together to form the Navigator cloud
used by various subsystems like NetCP, SRIO, SideBand Crypto
engines etc.

* tag 'drivers-soc-ti-v2' of git://git.kernel.org/pub/scm/linux/kernel/git/ssantosh/linux-keystone:
  MAINTAINERS: Add Keystone Multicore Navigator drivers entry
  soc: ti: add Keystone Navigator DMA support
  Documentation: dt: soc: add Keystone Navigator DMA bindings
  soc: ti: add Keystone Navigator QMSS driver
  Documentation: dt: soc: add Keystone Navigator QMSS bindings

Signed-off-by: Olof Johansson <olof@lixom.net>
10 years agoMAINTAINERS: Add Keystone Multicore Navigator drivers entry
Santosh Shilimkar [Thu, 10 Jul 2014 15:30:08 +0000 (11:30 -0400)]
MAINTAINERS: Add Keystone Multicore Navigator drivers entry

Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agosoc: ti: add Keystone Navigator DMA support
Santosh Shilimkar [Sun, 30 Mar 2014 21:29:04 +0000 (17:29 -0400)]
soc: ti: add Keystone Navigator DMA support

The Keystone Navigator DMA driver sets up the dma channels and flows for
the QMSS(Queue Manager SubSystem) who triggers the actual data movements
across clients using destination queues. Every client modules like
NETCP(Network Coprocessor), SRIO(Serial Rapid IO) and CRYPTO
Engines has its own instance of packet dma hardware. QMSS has also
an internal packet DMA module which is used as an infrastructure
DMA with zero copy.

Initially this driver was proposed as DMA engine driver but since the
hardware is not typical DMA engine and hence doesn't comply with typical
DMA engine driver needs, that approach was naked. Link to that
discussion -
https://lkml.org/lkml/2014/3/18/340

As aligned, now we pair the Navigator DMA with its companion Navigator
QMSS subsystem driver.

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Kumar Gala <galak@codeaurora.org>
Cc: Olof Johansson <olof@lixom.net>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Grant Likely <grant.likely@linaro.org>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Sandeep Nair <sandeep_n@ti.com>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoDocumentation: dt: soc: add Keystone Navigator DMA bindings
Santosh Shilimkar [Sun, 30 Mar 2014 21:29:04 +0000 (17:29 -0400)]
Documentation: dt: soc: add Keystone Navigator DMA bindings

The Keystone Navigator DMA driver sets up the dma channels and flows for
the QMSS(Queue Manager SubSystem) who triggers the actual data movements
across clients using destination queues. Every client modules like
NETCP(Network Coprocessor), SRIO(Serial Rapid IO) and CRYPTO
Engines has its own instance of packet dma hardware. QMSS has also
an internal packet DMA module which is used as an infrastructure
DMA with zero copy.

Initially this driver was proposed as DMA engine driver but since the
hardware is not typical DMA engine and hence doesn't comply with typical
DMA engine driver needs, that approach was naked. Link to that
discussion -
https://lkml.org/lkml/2014/3/18/340

As aligned, now we pair the Navigator DMA with its companion Navigator
QMSS subsystem driver.

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Kumar Gala <galak@codeaurora.org>
Cc: Olof Johansson <olof@lixom.net>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Grant Likely <grant.likely@linaro.org>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Sandeep Nair <sandeep_n@ti.com>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agosoc: ti: add Keystone Navigator QMSS driver
Sandeep Nair [Fri, 28 Feb 2014 15:47:50 +0000 (10:47 -0500)]
soc: ti: add Keystone Navigator QMSS driver

The QMSS (Queue Manager Sub System) found on Keystone SOCs is one of
the main hardware sub system which forms the backbone of the Keystone
Multi-core Navigator. QMSS consist of queue managers, packed-data structure
processors(PDSP), linking RAM, descriptor pools and infrastructure
Packet DMA.

The Queue Manager is a hardware module that is responsible for accelerating
management of the packet queues. Packets are queued/de-queued by writing or
reading descriptor address to a particular memory mapped location. The PDSPs
perform QMSS related functions like accumulation, QoS, or event management.
Linking RAM registers are used to link the descriptors which are stored in
descriptor RAM. Descriptor RAM is configurable as internal or external memory.

The QMSS driver manages the PDSP setups, linking RAM regions,
queue pool management (allocation, push, pop and notify) and descriptor
pool management. The specifics on the device tree bindings for
QMSS can be found in:
Documentation/devicetree/bindings/soc/keystone-navigator-qmss.txt

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Kumar Gala <galak@codeaurora.org>
Cc: Olof Johansson <olof@lixom.net>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Grant Likely <grant.likely@linaro.org>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Sandeep Nair <sandeep_n@ti.com>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoDocumentation: dt: soc: add Keystone Navigator QMSS bindings
Sandeep Nair [Fri, 28 Feb 2014 15:47:50 +0000 (10:47 -0500)]
Documentation: dt: soc: add Keystone Navigator QMSS bindings

The QMSS (Queue Manager Sub System) found on Keystone SOCs is one of
the main hardware sub system which forms the backbone of the Keystone
Multi-core Navigator. QMSS consist of queue managers, packed-data structure
processors(PDSP), linking RAM, descriptor pools and infrastructure
Packet DMA.

The Queue Manager is a hardware module that is responsible for accelerating
management of the packet queues. Packets are queued/de-queued by writing or
reading descriptor address to a particular memory mapped location. The PDSPs
perform QMSS related functions like accumulation, QoS, or event management.
Linking RAM registers are used to link the descriptors which are stored in
descriptor RAM. Descriptor RAM is configurable as internal or external memory.

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Kumar Gala <galak@codeaurora.org>
Cc: Olof Johansson <olof@lixom.net>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Grant Likely <grant.likely@linaro.org>
Cc: Rob Herring <robh+dt@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Sandeep Nair <sandeep_n@ti.com>
Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoMerge tag 'mailbox-for-v3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind...
Olof Johansson [Wed, 24 Sep 2014 05:10:18 +0000 (22:10 -0700)]
Merge tag 'mailbox-for-v3.18' of git://git./linux/kernel/git/tmlind/linux-omap into next/drivers

Mailbox related changes for omaps to get it to work with
device tree.

* tag 'mailbox-for-v3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap:
  mailbox/omap: add support for parsing dt devices
  Documentation: dt: add omap mailbox bindings

Signed-off-by: Olof Johansson <olof@lixom.net>
10 years agoMerge tag 'intc-for-v3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind...
Olof Johansson [Wed, 24 Sep 2014 05:08:40 +0000 (22:08 -0700)]
Merge tag 'intc-for-v3.18' of git://git./linux/kernel/git/tmlind/linux-omap into next/drivers

Merge "omap intc changes for v3.18 merge window" from Tony Lindgren:

Interrupt code related clean-up for omap2 and 3 to make
it ready to move to drivers/irqchip. Note that this series
does not yet move the interrupt code to drivers, that will
be posted separately as a follow-up series.

Note that this branch has a dependency to patches both
in fixes-v3.18-not-urgent and soc-for-v3.18 and is based on
a merge. Without doing the merge, off-idle would not work
properly for git bisect.

* tag 'intc-for-v3.18' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap: (325 commits)
  arm: omap: intc: switch over to linear irq domain
  arm: omap: irq: get rid of ifdef hack
  arm: omap: irq: introduce omap_nr_pending
  arm: omap: irq: remove nr_irqs argument
  arm: omap: irq: remove unnecessary header
  arm: omap: irq: drop omap2_intc_handle_irq()
  arm: omap: irq: drop omap3_intc_handle_irq()
  arm: omap: irq: call set_handle_irq() from .init_irq
  arm: omap: irq: move some more code around
  arm: boot: dts: omap2/3/am33xx: drop ti,intc-size
  arm: omap: irq: drop ti,intc-size support
  arm: boot: dts: am33xx/omap3: fix intc compatible flag
  arm: omap: irq: use compatible flag to figure out number of IRQ lines
  arm: omap: irq: add specific compatibles for omap3 and am33xx devices
  arm: omap: irq: drop .handle_irq and .init_irq fields
  arm: omap: irq: use IRQCHIP_DECLARE macro
  arm: omap: irq: call set_handle_irq() from intc_of_init
  arm: omap: irq: make intc_of_init static
  arm: omap: irq: reorganize code a little bit
  arm: omap: irq: always define omap3 support
  ...

Signed-off-by: Olof Johansson <olof@lixom.net>
10 years agoMerge tag 'at91-drivers2' of git://github.com/at91linux/linux-at91 into next/drivers
Olof Johansson [Wed, 24 Sep 2014 04:58:35 +0000 (21:58 -0700)]
Merge tag 'at91-drivers2' of git://github.com/at91linux/linux-at91 into next/drivers

Merge " Second drivers series for AT91/3.18" from Nicolas Ferre:

- move of the PIT (basic timer) from mach-at91 to its proper location:
  drivers/clocksource
- big cleanup of this driver along the way

* tag 'at91-drivers2' of git://github.com/at91linux/linux-at91:
  ARM: at91: PIT: Move the driver to drivers/clocksource
  ARM: at91: Give the PIT irq as an argument of at91sam926x_pit_init
  ARM: at91: Convert the boards to the init_time callback
  ARM: at91: soc: Add init_time callback
  ARM: at91: PIT: (Almost) remove the global variables
  ARM: at91: PIT: use request_irq instead of setup_irq
  ARM: at91: PIT: Use pr_fmt
  ARM: at91: PIT: Use consistent exit path in probe
  ARM: at91: dt: Remove init_time definitions
  ARM: at91: PIT: Rework probe functions
  ARM: at91: PIT: Use of_have_populated_dt instead of CONFIG_OF
  ARM: at91: PIT: Use DIV_ROUND_CLOSEST to compute the cycles
  ARM: at91: generic.h: Add include safe guards
  ARM: at91: PIT: Follow the general coding rules

Signed-off-by: Olof Johansson <olof@lixom.net>
10 years agortc: sunxi: Depend on platforms sun4i/sun7i that actually have the rtc
Chen-Yu Tsai [Tue, 26 Aug 2014 03:54:56 +0000 (11:54 +0800)]
rtc: sunxi: Depend on platforms sun4i/sun7i that actually have the rtc

Now that we have Kconfig options for individual sunxi platforms, let
the rtc-sunxi driver depend on the platforms that actually have this
hardware, sun4i and sun7i.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agortc: sun6i: Add sun6i RTC driver
Chen-Yu Tsai [Tue, 26 Aug 2014 03:54:55 +0000 (11:54 +0800)]
rtc: sun6i: Add sun6i RTC driver

This patch introduces the driver for the RTC in the Allwinner A31 and
A23 SoCs.

Unlike the RTC found in A10/A20 SoCs, which was part of the timer, the
RTC in A31/A23 are a separate hardware block, which also contain a few
controls for the RTC block hardware (a regulator and RTC block GPIO pin
latches), while also having separate interrupts for the alarms.

The hardware is different enough to make a different driver for it.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
Reviewed-by: Varka Bhadram <varkabhadram@gmail.com>
Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
10 years agoARM: at91: PIT: Move the driver to drivers/clocksource
Maxime Ripard [Tue, 2 Sep 2014 16:12:35 +0000 (18:12 +0200)]
ARM: at91: PIT: Move the driver to drivers/clocksource

Now that we don't depend on anyting in the mach-at91 directory, we can just
move the driver to where it belongs.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Conflicts:
arch/arm/mach-at91/Kconfig
arch/arm/mach-at91/Makefile

10 years agoARM: at91: Give the PIT irq as an argument of at91sam926x_pit_init
Maxime Ripard [Mon, 15 Sep 2014 14:02:24 +0000 (16:02 +0200)]
ARM: at91: Give the PIT irq as an argument of at91sam926x_pit_init

This allows to remove the dependency of the timer driver on mach/hardware.h and
having an hardcoded interrupt number in the driver itself.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoarm: omap: intc: switch over to linear irq domain
Felipe Balbi [Tue, 9 Sep 2014 00:54:58 +0000 (17:54 -0700)]
arm: omap: intc: switch over to linear irq domain

now that we don't need to support legacy board-files,
we can completely switch over to a linear irq domain
and make use of irq_alloc_domain_generic_chips() to
allocate all generic irq chips for us.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: get rid of ifdef hack
Felipe Balbi [Tue, 9 Sep 2014 00:54:57 +0000 (17:54 -0700)]
arm: omap: irq: get rid of ifdef hack

we don't need the ifdef if we have omap_nr_pending
telling us how many pending registers we have
on current platform. This solves a possible
problem where we could try to handle bogus
interrupts on OMAP2 and OMAP3 if using single
zImage kernel, because we would end up reading
the following pending FIQ register.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: introduce omap_nr_pending
Felipe Balbi [Tue, 9 Sep 2014 00:54:57 +0000 (17:54 -0700)]
arm: omap: irq: introduce omap_nr_pending

that variable will tell us how many INTC_PENDING_IRQn
registers we have. It'll be used on a following patch
to cleanup omap_intc_handle_irq() a bit.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: remove nr_irqs argument
Felipe Balbi [Tue, 9 Sep 2014 00:54:55 +0000 (17:54 -0700)]
arm: omap: irq: remove nr_irqs argument

we can set our global omap_nr_irqs early on
and drop the extra argument to omap_init_irq().

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: remove unnecessary header
Felipe Balbi [Tue, 9 Sep 2014 00:54:54 +0000 (17:54 -0700)]
arm: omap: irq: remove unnecessary header

There's no need for that header to be included.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: drop omap2_intc_handle_irq()
Felipe Balbi [Tue, 9 Sep 2014 00:54:52 +0000 (17:54 -0700)]
arm: omap: irq: drop omap2_intc_handle_irq()

that was just a no-op wrapper around omap_intc_handle_irq
anyway.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: drop omap3_intc_handle_irq()
Felipe Balbi [Tue, 9 Sep 2014 00:54:52 +0000 (17:54 -0700)]
arm: omap: irq: drop omap3_intc_handle_irq()

now that we're calling set_handle_irq() from
init_irq(), we can safely drop all callers to
omap3_intc_handle_irq() and its definition.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: call set_handle_irq() from .init_irq
Felipe Balbi [Tue, 9 Sep 2014 00:54:52 +0000 (17:54 -0700)]
arm: omap: irq: call set_handle_irq() from .init_irq

the idea is that board-files won't need to set
.handle_irq on their machine_descs, which lets
us drop a little more pointless code.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: move some more code around
Felipe Balbi [Tue, 9 Sep 2014 00:54:51 +0000 (17:54 -0700)]
arm: omap: irq: move some more code around

We want .init_irq to call set_irq_handle() for
legacy platforms. Note that this code will also
be dropped once omap2/3 devices are completely
moved to DT.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: boot: dts: omap2/3/am33xx: drop ti,intc-size
Felipe Balbi [Tue, 9 Sep 2014 00:54:49 +0000 (17:54 -0700)]
arm: boot: dts: omap2/3/am33xx: drop ti,intc-size

we are now infering number of IRQ lines based
on correct compatible flag, which renders this
binding completely useless.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: drop ti,intc-size support
Felipe Balbi [Tue, 9 Sep 2014 00:54:48 +0000 (17:54 -0700)]
arm: omap: irq: drop ti,intc-size support

we don't need that anymore since specific
devices are passing correct compatible flags.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: boot: dts: am33xx/omap3: fix intc compatible flag
Felipe Balbi [Tue, 9 Sep 2014 00:54:48 +0000 (17:54 -0700)]
arm: boot: dts: am33xx/omap3: fix intc compatible flag

that way, our intc driver can figure out how
many IRQ lines INTC has.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: use compatible flag to figure out number of IRQ lines
Felipe Balbi [Tue, 9 Sep 2014 00:54:47 +0000 (17:54 -0700)]
arm: omap: irq: use compatible flag to figure out number of IRQ lines

so far, only am33xx has 128 lines, all other devices
have only 96.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: add specific compatibles for omap3 and am33xx devices
Felipe Balbi [Tue, 9 Sep 2014 00:54:46 +0000 (17:54 -0700)]
arm: omap: irq: add specific compatibles for omap3 and am33xx devices

with this, we can use a compatible flag to figure
out how many irq lines are wired up, no need for
our TI-specific ti,intc-size binding.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: drop .handle_irq and .init_irq fields
Felipe Balbi [Tue, 9 Sep 2014 00:54:45 +0000 (17:54 -0700)]
arm: omap: irq: drop .handle_irq and .init_irq fields

now we can safely drop those fields from our machine_desc.

While at that, also drop the now unused omap_intc_of_init()
definition.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: use IRQCHIP_DECLARE macro
Felipe Balbi [Tue, 9 Sep 2014 00:54:43 +0000 (17:54 -0700)]
arm: omap: irq: use IRQCHIP_DECLARE macro

IRQCHIP_DECLARE macro is used to declare the same
of_device_id structure for irqchips, it's just
a helper. No functional changes.

Note that we're temporarily including irqchip.h
with its full path, until we move this driver
to drivers/irqchip/.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: call set_handle_irq() from intc_of_init
Felipe Balbi [Tue, 9 Sep 2014 00:54:43 +0000 (17:54 -0700)]
arm: omap: irq: call set_handle_irq() from intc_of_init

this will let us drop .handle_irq and .init_irq fields
from our generic machine_descs.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: make intc_of_init static
Felipe Balbi [Tue, 9 Sep 2014 00:54:43 +0000 (17:54 -0700)]
arm: omap: irq: make intc_of_init static

nobody uses that function outside of this file,
so we don't need to expose it.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: reorganize code a little bit
Felipe Balbi [Tue, 9 Sep 2014 00:54:42 +0000 (17:54 -0700)]
arm: omap: irq: reorganize code a little bit

no functional changes, just moving code around.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: always define omap3 support
Felipe Balbi [Tue, 9 Sep 2014 00:54:40 +0000 (17:54 -0700)]
arm: omap: irq: always define omap3 support

remove ifdef around omap3 INTC support. This
will make it easier to reuse code for PM.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: rename omap3_intc_regs
Felipe Balbi [Tue, 9 Sep 2014 00:54:38 +0000 (17:54 -0700)]
arm: omap: irq: rename omap3_intc_regs

just to make it clearer that it can
be used on all omaps.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: remove unnecessary base_addr argument
Felipe Balbi [Tue, 9 Sep 2014 00:54:37 +0000 (17:54 -0700)]
arm: omap: irq: remove unnecessary base_addr argument

omap_intc_handle_irq now had an unnecessary
base_addr argument. Let's remove it and fix
all callers.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: switch over to intc_readl on omap_intc_handle_irq
Felipe Balbi [Tue, 9 Sep 2014 00:54:37 +0000 (17:54 -0700)]
arm: omap: irq: switch over to intc_readl on omap_intc_handle_irq

an almost blind conversion from readl_relaxed
to our newly introduced intc_readl().

While at that, also remove some hardcoded
register addresses.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: remove unused macro
Felipe Balbi [Tue, 9 Sep 2014 00:54:37 +0000 (17:54 -0700)]
arm: omap: irq: remove unused macro

no functional changes.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: remove rest of irq_banks usage
Felipe Balbi [Tue, 9 Sep 2014 00:54:35 +0000 (17:54 -0700)]
arm: omap: irq: remove rest of irq_banks usage

now we can finally remove the pointless irq_banks
array.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: add a global omap_nr_irqs variable
Felipe Balbi [Tue, 9 Sep 2014 00:54:34 +0000 (17:54 -0700)]
arm: omap: irq: add a global omap_nr_irqs variable

this will cache number of irqs. Also in preparation
for removal of irq_banks array.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: start to remove irq_banks array
Felipe Balbi [Tue, 9 Sep 2014 00:54:32 +0000 (17:54 -0700)]
arm: omap: irq: start to remove irq_banks array

We have a single bank in that array, this patch
is in preparation to remove that array. It just
shifts everything to a new set of functions
for register IO while also removing old ones.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: define INTC_ILR0 register
Felipe Balbi [Tue, 9 Sep 2014 00:54:32 +0000 (17:54 -0700)]
arm: omap: irq: define INTC_ILR0 register

this is currently used as a hardcoded 0x100
offset.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoarm: omap: irq: make omap_irq_base global
Felipe Balbi [Tue, 9 Sep 2014 00:54:31 +0000 (17:54 -0700)]
arm: omap: irq: make omap_irq_base global

This is in preparation for removing the pointless
irq_banks array.

Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoMerge branch 'omap-for-v3.18/fixes-not-urgent' into omap-for-v3.18/intc-v2
Tony Lindgren [Thu, 11 Sep 2014 20:03:25 +0000 (13:03 -0700)]
Merge branch 'omap-for-v3.18/fixes-not-urgent' into omap-for-v3.18/intc-v2

10 years agoARM: OMAP2+: make of_device_ids const
Uwe Kleine-König [Wed, 10 Sep 2014 08:26:17 +0000 (10:26 +0200)]
ARM: OMAP2+: make of_device_ids const

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 function parameters and structs for OMAP2+ as const, too.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: omap2: make arrays containing machine compatible strings const
Uwe Kleine-König [Thu, 11 Sep 2014 19:29:01 +0000 (21:29 +0200)]
ARM: omap2: make arrays containing machine compatible strings const

The definition

static const char *omap3_boards_compat[] __initconst = {

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

*omap3_boards_compat[0] = 'f';

but

omap3_boards_compat[0] = "another 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: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agomailbox/omap: add support for parsing dt devices
Suman Anna [Wed, 10 Sep 2014 19:20:59 +0000 (14:20 -0500)]
mailbox/omap: add support for parsing dt devices

Logic has been added to the OMAP2+ mailbox code to parse the
mailbox dt nodes and construct the different sub-mailboxes
associated with the instance. The DT representation of the
sub-mailbox devices is different from legacy platform data
representation to allow flexibility of interrupt configuration
between Tx and Rx fifos (to also possibly allow simplex devices
in the future). The DT representation gathers similar information
that was being passed previously through the platform data, except
for the interrupt type information, which is gathered through driver
compatible match data.

The non-DT support has to be maintained for now to not break
OMAP3 legacy boot, and the legacy-style code will be cleaned
up once OMAP3 is also converted to DT-boot only.

Cc: Jassi Brar <jassisinghbrar@gmail.com>
Cc: Rob Herring <robh+dt@kernel.org>
Signed-off-by: Suman Anna <s-anna@ti.com>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoDocumentation: dt: add omap mailbox bindings
Suman Anna [Wed, 10 Sep 2014 19:20:58 +0000 (14:20 -0500)]
Documentation: dt: add omap mailbox bindings

Add the device tree bindings document for OMAP2+ mailbox.

Cc: Rob Herring <robh+dt@kernel.org>
Cc: Pawel Moll <pawel.moll@arm.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Ian Campbell <ijc+devicetree@hellion.org.uk>
Cc: Kumar Gala <galak@codeaurora.org>
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoMerge tag 'socfpga_driver_for_v3.18' of git://git.rocketboards.org/linux-socfpga...
Arnd Bergmann [Tue, 9 Sep 2014 15:51:31 +0000 (17:51 +0200)]
Merge tag 'socfpga_driver_for_v3.18' of git://git.rocketboards.org/linux-socfpga-next into next/drivers

Pull "SOCFPGA driver update for v3.18" from Dinh Nguyen:

This is the EDAC driver for EDAC. Boris had given me permission to
take this patch together with it's DTS component. The DTS portion was in the
previous pull request.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
* tag 'socfpga_driver_for_v3.18' of git://git.rocketboards.org/linux-socfpga-next:
  edac: altera: Add Altera SDRAM EDAC support

10 years agoARM: dts: OMAP3+: Add PRM interrupt
Nishanth Menon [Fri, 22 Aug 2014 14:03:50 +0000 (09:03 -0500)]
ARM: dts: OMAP3+: Add PRM interrupt

Provide OMAP3, 4 and OMAP5 with interrupt number for PRM

And for DRA7, provide crossbar number for prm interrupt.

Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: omap: Remove stray ARCH_HAS_OPP references
Mark Brown [Sat, 6 Sep 2014 10:14:16 +0000 (11:14 +0100)]
ARM: omap: Remove stray ARCH_HAS_OPP references

OPP is now a normal kernel library selected by its users rather than a
feature that architectures need to enable so ARCH_HAS_OPP serves no
function any more - remove the selects.

Signed-off-by: Mark Brown <broonie@kernel.org>
Acked-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: DRA7: Add hook in SoC initcalls to enable pm initialization
Rajendra Nayak [Fri, 22 Aug 2014 14:02:34 +0000 (09:02 -0500)]
ARM: DRA7: Add hook in SoC initcalls to enable pm initialization

With consolidated code, now we can add the required hooks for
DRA7 to enable power management.

Signed-off-by: Rajendra Nayak <rnayak@ti.com>
[nm@ti.com: minor modifications]
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoMerge branch 'pull/v3.18/for-omap-soc' of https://github.com/nmenon/linux-2.6-playgro...
Tony Lindgren [Mon, 8 Sep 2014 22:20:15 +0000 (15:20 -0700)]
Merge branch 'pull/v3.18/for-omap-soc' of https://github.com/nmenon/linux-2.6-playground into omap-for-v3.18/soc

10 years agoMerge branch 'pull/v3.18/powerdomain-fixes' of https://github.com/nmenon/linux-2...
Tony Lindgren [Mon, 8 Sep 2014 22:04:24 +0000 (15:04 -0700)]
Merge branch 'pull/v3.18/powerdomain-fixes' of https://github.com/nmenon/linux-2.6-playground into omap-for-v3.18/fixes-not-urgent

10 years agoARM: OMAP5: Add hook in SoC initcalls to enable pm initialization
Santosh Shilimkar [Tue, 20 May 2014 21:19:23 +0000 (16:19 -0500)]
ARM: OMAP5: Add hook in SoC initcalls to enable pm initialization

With consolidated code, now we can add the required hooks for
OMAP5 to enable power management.

Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
[nm@ti.com: minor rebase updates]
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>
10 years agoARM: OMAP5 / DRA7: Enable CPU RET on suspend
Rajendra Nayak [Mon, 27 May 2013 10:16:44 +0000 (15:46 +0530)]
ARM: OMAP5 / DRA7: Enable CPU RET on suspend

On OMAP5 / DRA7, prevent a CPU powerdomain OFF and resulting MPU OSWR
and instead attempt a CPU RET and side effect, MPU RET in suspend.

NOTE: the hardware was originally designed to be capable of achieving
deep power states such as OFF and OSWR, however due to various issues
and risks, deepest valid state was determined to be CSWR - hence we use
the errata framework to handle this case.

Signed-off-by: Rajendra Nayak <rnayak@ti.com>
[nm@ti.com: updates]
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>
10 years agoARM: OMAP5 / DRA7: PM: Provide a dummy startup function for CPU hotplug
Santosh Shilimkar [Fri, 6 Jun 2014 22:30:43 +0000 (17:30 -0500)]
ARM: OMAP5 / DRA7: PM: Provide a dummy startup function for CPU hotplug

Dont assume that all OMAP4+ code will be able to use OMAP4 hotplug
logic. On OMAP5, DRA7, we do not need this in place yet, also,
currently the CPU startup pointer is located in omap4_cpu_pm_info
instead of cpu_pm_ops.

So, isolate the function to hotplug_restart pointer in cpu_pm_ops
where it should have belonged, initalize them as per valid startup
pointers for OMAP4430/60 as in current logic, however provide
dummy_cpu_resume to be the startup location as well.

Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
[nm@ti.com: split this out of original code and isolate it]
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>
10 years agoARM: OMAP5 / DRA7: PM: Avoid all SAR saves
Rajendra Nayak [Fri, 3 May 2013 10:04:40 +0000 (15:34 +0530)]
ARM: OMAP5 / DRA7: PM: Avoid all SAR saves

Get rid of all assumptions about always having a sar base on *all*
OMAP4+ platforms. We dont need one on DRA7 and it is not necessary at
this point for OMAP5 either.

Signed-off-by: Rajendra Nayak <rnayak@ti.com>
[nm@ti.com: Split and optimize]
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>
10 years agoARM: OMAP5 / DRA7: PM: Enable Mercury retention mode on CPUx powerdomains
Santosh Shilimkar [Thu, 12 Apr 2012 11:31:52 +0000 (17:01 +0530)]
ARM: OMAP5 / DRA7: PM: Enable Mercury retention mode on CPUx powerdomains

In addition to the standard power-management technique, the OMAP5 / DRA7
MPU subsystem also employs an SR3-APG (mercury) power management
technology to reduce leakage.

It allows for full logic and memories retention on MPU_C0 and MPU_C1 and
is controlled by the PRCM_MPU. Only "Fast-mode" is supported on the
OMAP5 and DRA7 family of processors.

Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
[nm@ti.com: minor consolidation]
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>
10 years agoARM: OMAP5 / DRA7: PM / wakeupgen: Enables ES2 PM mode by default
Santosh Shilimkar [Fri, 8 Feb 2013 11:37:31 +0000 (17:07 +0530)]
ARM: OMAP5 / DRA7: PM / wakeupgen: Enables ES2 PM mode by default

Enables MPUSS ES2 power management mode using ES2_PM_MODE in
AMBA_IF_MODE register.

0x0: OMAP5 ES1 behavior, CPU cores would enter and exit OFF mode together.
     Broken! Fortunately, we do not support this anymore.
0x1: OMAP5 ES2, DRA7 behavior, CPU cores are allowed to enter/exit OFF mode
     independently.

This is one time settings thanks to always ON domain.

Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
[nm@ti.com: minor conflict resolutions, consolidation for DRA7]
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>
10 years agoARM: OMAP5 / DRA7: PM: Set MPUSS-EMIF clock-domain static dependency
Santosh Shilimkar [Wed, 6 Feb 2013 10:21:45 +0000 (15:51 +0530)]
ARM: OMAP5 / DRA7: PM: Set MPUSS-EMIF clock-domain static dependency

With EMIF clock-domain put under hardware supervised control, memory
corruption and untraceable crashes are observed on OMAP5. Further
investigation revealed that there is a weakness in the PRCM on this
specific dynamic depedency.

The recommendation is to set MPUSS static dependency towards EMIF
clock-domain to avoid issues. This recommendation holds good for DRA7
family of devices as well.

Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
[rnayak@ti.com: DRA7]
Signed-off-by: Rajendra Nayak <rnayak@ti.com>
[nm@ti.com: conflict resolution, dra7]
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>
10 years agoARM: OMAP5 / DRA7: PM: Update CPU context register offset
Santosh Shilimkar [Wed, 6 Feb 2013 14:09:07 +0000 (19:39 +0530)]
ARM: OMAP5 / DRA7: PM: Update CPU context register offset

On OMAP5, RM_CPUi_CPUi_CONTEXT offset has changed. Update the code
so that same code works for OMAP4+ devices. DRA7 and OMAP5 have the same
context offset as well.

Signed-off-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
[rnayak@ti.com: for DRA7]
Signed-off-by: Rajendra Nayak <rnayak@ti.com>
[nm@ti.com: rebase, split/merge etc..]
Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>
10 years agoARM: AM437x: use pdata quirks for pinctrl information
Keerthy [Mon, 7 Apr 2014 06:24:49 +0000 (11:54 +0530)]
ARM: AM437x: use pdata quirks for pinctrl information

Provide pdata-quirks for Am437x processor family.

Signed-off-by: Keerthy <j-keerthy@ti.com>
10 years agoARM: DRA7: use pdata quirks for pinctrl information
Nishanth Menon [Fri, 23 May 2014 04:39:54 +0000 (23:39 -0500)]
ARM: DRA7: use pdata quirks for pinctrl information

Provide pdata-quirks for DRA7 processor family.

Signed-off-by: Nishanth Menon <nm@ti.com>
10 years agoARM: OMAP5: use pdata quirks for pinctrl information
Nishanth Menon [Thu, 22 May 2014 20:19:29 +0000 (15:19 -0500)]
ARM: OMAP5: use pdata quirks for pinctrl information

Provide pdata-quirks for OMAP5 processor family.

Signed-off-by: Nishanth Menon <nm@ti.com>
10 years agoARM: OMAP4+: PM: Use only valid low power state for CPU hotplug
Nishanth Menon [Thu, 24 Jul 2014 15:24:19 +0000 (10:24 -0500)]
ARM: OMAP4+: PM: Use only valid low power state for CPU hotplug

Not all SoCs support OFF mode - for example DRA74/72. So, use valid
power state during CPU hotplug.

Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoARM: OMAP4+: PM: use only valid low power state for suspend
Nishanth Menon [Fri, 6 Jun 2014 06:17:37 +0000 (01:17 -0500)]
ARM: OMAP4+: PM: use only valid low power state for suspend

We are using power domain state as RET and logic state as OFF. This
state is OSWR. This may not always be supported on ALL power domains. In
fact, on certain power domains, this might result in a hang on certain
platforms. Instead, depend on powerdomain data to provide accurate
information about the supported powerdomain states and use the
appropriate function to query and use it as part of suspend path.

Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoARM: OMAP4+: PM: Make logic state programmable
Nishanth Menon [Fri, 6 Jun 2014 02:40:39 +0000 (21:40 -0500)]
ARM: OMAP4+: PM: Make logic state programmable

Move the logic state as different for each power domain. This allows us
to customize the deepest power state we should target over all for each
powerdomain in the follow on patches.

Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoARM: OMAP2+: powerdomain: introduce logic for finding valid power domain
Nishanth Menon [Fri, 6 Jun 2014 06:04:20 +0000 (01:04 -0500)]
ARM: OMAP2+: powerdomain: introduce logic for finding valid power domain

powerdomain configuration in OMAP is done using PWRSTCTRL register for
each power domain. However, PRCM lets us write any value we'd like to
the logic and power domain target states, however the SoC integration
tends to actually function only at a few discrete states. These valid
states are already in our powerdomains_xxx_data.c file.

So, provide a function to easily query valid low power state that the
power domain is allowed to go to.

Based on work originally done by Jean Pihet <j-pihet@ti.com>
https://patchwork.kernel.org/patch/1325091/ . There is no attempt to
create a new powerdomain solution here, except fixing issues seen
attempting invalid programming attempts. Future consolidation to the
generic powerdomain framework should consider this requirement as
well.

Similar solutions have been done in product kernels in the past such
as:
https://android.googlesource.com/kernel/omap.git/+blame/android-omap-panda-3.0/arch/arm/mach-omap2/pm44xx.c

Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoARM: OMAP2+: powerdomain: pwrdm_for_each_clkdm iterate only valid clkdms
Nishanth Menon [Sat, 7 Jun 2014 00:36:41 +0000 (19:36 -0500)]
ARM: OMAP2+: powerdomain: pwrdm_for_each_clkdm iterate only valid clkdms

No need to invoke callback when the clkdm pointer is NULL.

Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoARM: OMAP5: powerdomain data: fix powerdomain powerstate
Nishanth Menon [Wed, 20 Aug 2014 13:13:16 +0000 (08:13 -0500)]
ARM: OMAP5: powerdomain data: fix powerdomain powerstate

Update the power domain power states for final production chip
capability. OFF mode, OSWR etc have been descoped for various domains.

Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoARM: OMAP: DRA7: powerdomain data: fix powerdomain powerstate
Nishanth Menon [Fri, 6 Jun 2014 06:21:51 +0000 (01:21 -0500)]
ARM: OMAP: DRA7: powerdomain data: fix powerdomain powerstate

DRA7 supports only CSWR for CPU, MPU power domains. Core power domain
supports upto INA.

Signed-off-by: Nishanth Menon <nm@ti.com>
Reviewed-by: Kevin Hilman <khilman@linaro.org>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoARM: OMAP3+: PRM: register interrupt information from DT
Nishanth Menon [Tue, 12 Aug 2014 12:05:19 +0000 (07:05 -0500)]
ARM: OMAP3+: PRM: register interrupt information from DT

Allow the PRM interrupt information to be picked up from device tree.
OMAP3 may use legacy boot and needs to be compatible with old dtbs
(without interrupt populated), for these, we use the value which is
pre-populated.

Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoARM: OMAP4+: PRM: Enable wakeup capability for OMAP5, DRA7
Nishanth Menon [Mon, 11 Aug 2014 19:58:23 +0000 (14:58 -0500)]
ARM: OMAP4+: PRM: Enable wakeup capability for OMAP5, DRA7

OMAP5 and DRA7 can now use pinctrl based I/O daisychain wakeup
capability. So, enable the support.

Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoARM: OMAP4+: PRM: remove "wkup" event
Nishanth Menon [Fri, 23 May 2014 05:20:37 +0000 (00:20 -0500)]
ARM: OMAP4+: PRM: remove "wkup" event

"wkup" event at bit offset 0 exists only on OMAP3.
OMAP4430/60 PRM_IRQSTATUS_A9, OMAP5/DRA7 PRM_IRQSTATUS_MPU

register bit 0 is DPLL_CORE_RECAL_ST not wakeup event like OMAP3.

The same applies to AM437x as well.

Remove the wrong definition.

Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoARM: OMAP4+: PRM: register interrupt information from DT
Nishanth Menon [Thu, 22 May 2014 20:19:29 +0000 (15:19 -0500)]
ARM: OMAP4+: PRM: register interrupt information from DT

Allow the PRM interrupt information to be picked up from device tree.
the only exception is for OMAP4 which uses values pre-populated and allows
compatibility with older dtb.

Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoARM: OMAP4: PRM: use the generic prm_inst to allow logic to be abstracted
Nishanth Menon [Thu, 22 May 2014 20:00:55 +0000 (15:00 -0500)]
ARM: OMAP4: PRM: use the generic prm_inst to allow logic to be abstracted

use the generic function to pick up the prm_instance for a generic logic
which can be reused from OMAP4+

Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoARM: OMAP4+: prminst: provide function to find prm_dev instance offset
Nishanth Menon [Thu, 22 May 2014 19:53:54 +0000 (14:53 -0500)]
ARM: OMAP4+: prminst: provide function to find prm_dev instance offset

PRM device instance can vary depending on SoC. We already handle the
same during reset of the device, However, this is also needed
for other logic instances. So, first abstract this out to a generic
function.

Signed-off-by: Nishanth Menon <nm@ti.com>
Acked-by: Santosh Shilimkar <santosh.shilimkar@ti.com>
10 years agoMerge tag 'at91-drivers' of git://github.com/at91linux/linux-at91 into next/drivers
Arnd Bergmann [Fri, 5 Sep 2014 20:26:40 +0000 (22:26 +0200)]
Merge tag 'at91-drivers' of git://github.com/at91linux/linux-at91 into next/drivers

Merge "First batch of AT91 drivers for 3.18" from Nicolas Ferre:

- reset, poweroff and ram drivers are moved to their proper
  location instead of being in mach-at91 directory. They now use
  the appropriate frameworks.
- big amount of removal of these machine specific drivers and use
  of the newly created drivers. This lead to an overhaul of the setup.c AT91
  startup code.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
* tag 'at91-drivers' of git://github.com/at91linux/linux-at91: (31 commits)
  power: reset: at91-poweroff: fix wakeup status register index
  ARM: at91/power/reset: fix Kconfig "depends on" directive
  ARM: at91: fix ramc standby function registration
  ARM: at91: Remove rstc and shdwc headers
  ARM: at91: Remove rstc and shdwnc global base addresses
  ARM: at91/pm: Remove show_reset_status function
  ARM: at91: Remove poweroff code
  ARM: at91: Register the poweroff driver
  ARM: at91: Remove poweroff DT probing
  ARM: at91: Remove reset code from the machine code
  ARM: at91: Call at91_register_devices in the board files
  ARM: at91: Probe the reset driver
  ARM: at91/soc: Introduce register_devices callback
  ARM: at91: Remove the old-style reset probing
  ARM: at91: Rework ramc mapping code
  ARM: at91: setup: Switch to pr_fmt
  ARM: at91: remove old irq material
  ARM: at91: make use of the new AIC driver for dt enabled boards
  ARM: at91: enclose at91_aic_xx calls in IS_ENABLED(CONFIG_OLD_IRQ_AT91) blocks
  ARM: at91: introduce OLD_IRQ_AT91 Kconfig option
  ...

10 years agoedac: altera: Add Altera SDRAM EDAC support
Thor Thayer [Wed, 3 Sep 2014 15:27:54 +0000 (10:27 -0500)]
edac: altera: Add Altera SDRAM EDAC support

This patch adds support for the CycloneV and ArriaV SDRAM controllers.
Correction and reporting of SBEs, Panic on DBEs.

There was a discussion thread on whether this driver should be an mfd driver
or just make use of syscon, which is already a mfd. Ultimately, the
decision to use a simple syscon interface was reached.[1]

[1] https://lkml.org/lkml/2014/7/30/514

[dinguyen] Fixed Kconfig to have EDAC_ALTERA_MC as a tristate to prevent a
build failure for allmodconfig.

Signed-off-by: Thor Thayer <tthayer@opensource.altera.com>
Acked-by: Borislav Petkov <bp@suse.de>
[dinguyen] cleaned up commit message
Signed-off-by: Dinh Nguyen <dinguyen@opensource.altera.com>
10 years agoARM: dts: am335x-bone*: Fix model name and update compatibility information
Nishanth Menon [Wed, 3 Sep 2014 16:03:10 +0000 (11:03 -0500)]
ARM: dts: am335x-bone*: Fix model name and update compatibility information

Beaglebone white and beaglebone black differ in tiny little aspects.
This is the reason why we maintain seperate dts for these platforms.
However, there is no real way to decode from dtb which platform it is
since compatible and model name are the same for both platforms.

Fix this so that beaglebone black and beaglebone are identifiable,
while maintaining compatibility for older zImages which might use old
beaglebone compatible flag for black as well.

Reported-by: Tom Rini <trini@ti.com>
Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: dts: omap4-panda: Fix model and SoC family details
Nishanth Menon [Thu, 28 Aug 2014 20:45:03 +0000 (15:45 -0500)]
ARM: dts: omap4-panda: Fix model and SoC family details

Currently we claim that omap4-panda and omap4-panda-es are essentially
the same, but they are not since PandaBoard-ES uses OMAP4460 and
PandaBoard uses OMAP4430.

So, split the common definition and make the model name available.

Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
10 years agoARM: at91: Convert the boards to the init_time callback
Maxime Ripard [Tue, 1 Jul 2014 09:33:25 +0000 (11:33 +0200)]
ARM: at91: Convert the boards to the init_time callback

Now that we have the init_time callback in the at91_init_soc structure, convert
all the boards and SoC to this.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoARM: at91: soc: Add init_time callback
Maxime Ripard [Tue, 2 Sep 2014 16:05:04 +0000 (18:05 +0200)]
ARM: at91: soc: Add init_time callback

Introduce an init_time callback to the at91_init_soc structure to be able to
tweak the init_time machine callback on a per-soc basis, instead of having to
rely on a global one.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Conflicts:
arch/arm/mach-at91/setup.c

10 years agoARM: at91: PIT: (Almost) remove the global variables
Maxime Ripard [Tue, 1 Jul 2014 09:33:23 +0000 (11:33 +0200)]
ARM: at91: PIT: (Almost) remove the global variables

The timer driver is using some global variables to define some variables it has
to use in most of its functions, like the base address.

Use some container_of calls to have a single dynamic (and local) variable to
hold this content.

The only exception is in the !DT case, where the call chain to
at91sam926x_ioremap_pit and then at91sam926x_pit_init as init_time makes it
hard for the moment to pass the physical address of the timer.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoARM: at91: PIT: use request_irq instead of setup_irq
Maxime Ripard [Tue, 1 Jul 2014 09:33:22 +0000 (11:33 +0200)]
ARM: at91: PIT: use request_irq instead of setup_irq

We can use the generic request_irq now to register a timer interrupt handler,
instead of the more complex setup_irq.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoARM: at91: PIT: Use pr_fmt
Maxime Ripard [Tue, 1 Jul 2014 09:33:21 +0000 (11:33 +0200)]
ARM: at91: PIT: Use pr_fmt

All the panic messages hardcode the same prefix. Define the pr_fmt macro to
unify its definition.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoARM: at91: PIT: Use consistent exit path in probe
Maxime Ripard [Tue, 1 Jul 2014 09:33:20 +0000 (11:33 +0200)]
ARM: at91: PIT: Use consistent exit path in probe

The current probe code is, whenever it fails to retrieve a resource, either
panicing, silently returning, or returning and displaying an error.

Make the code consistent by always panicing whenever it fails to grab any
resource.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoARM: at91: dt: Remove init_time definitions
Maxime Ripard [Tue, 2 Sep 2014 16:02:26 +0000 (18:02 +0200)]
ARM: at91: dt: Remove init_time definitions

The current AT91 DT boards have a completely generic init_time definition.
Remove them from the machine declaration.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
Conflicts:
arch/arm/mach-at91/board-dt-sam9.c
arch/arm/mach-at91/board-dt-sama5.c

10 years agoARM: at91: PIT: Rework probe functions
Maxime Ripard [Tue, 1 Jul 2014 09:33:18 +0000 (11:33 +0200)]
ARM: at91: PIT: Rework probe functions

The PIT timer driver until now had a single probe function, disregarding wether
it was probed through DT or in the old-style way. This code later on was
calling some DT function to retrieve the proper values for its base address,
interrupts and clocks.

While this was working, it was preventing the usage of CLOCKSOURCE_OF_DECLARE,
and the two different probe path were not as clearly separated as they could
be.

Rework the probe path to take this into account, and switch to
CLOCKSOURCE_OF_DECLARE.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoARM: at91: PIT: Use of_have_populated_dt instead of CONFIG_OF
Maxime Ripard [Tue, 1 Jul 2014 09:33:17 +0000 (11:33 +0200)]
ARM: at91: PIT: Use of_have_populated_dt instead of CONFIG_OF

Until now, the machines, even when CONFIG_OF was enabled, were calling
at91sam926x_ioremap_pit to try to map the PIT address using the defined
physical address.

Obviously, with DT, it's not appropriate anymore, and some code was added to
the function to deal with this case.

Unfortunately, this code was conditionned on CONFIG_OF, which can be enabled,
even though no DT was actually used, which would result in such a case, to this
code being executed, without any reason.

Moreover, the logic that was here before to bail out of the function just check
in the DT to see if the PIT node is there, which is the case in all our DTSI.

All this can be made much more straightforward just by using
of_have_populated_dt to bail out.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoARM: at91: PIT: Use DIV_ROUND_CLOSEST to compute the cycles
Maxime Ripard [Tue, 1 Jul 2014 09:33:16 +0000 (11:33 +0200)]
ARM: at91: PIT: Use DIV_ROUND_CLOSEST to compute the cycles

Until now, the pit_cycle computation was dividing the rate by HZ, rounding to
the closest integer, but without using the appropriate macro.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoARM: at91: generic.h: Add include safe guards
Maxime Ripard [Tue, 1 Jul 2014 09:33:15 +0000 (11:33 +0200)]
ARM: at91: generic.h: Add include safe guards

The generic.h header file doesn't have any safe guards against multiple
inclusion. It only worked so far because all the symbols defined in it were
extern, but this is a rather fragile assumption.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoARM: at91: PIT: Follow the general coding rules
Maxime Ripard [Tue, 1 Jul 2014 09:33:14 +0000 (11:33 +0200)]
ARM: at91: PIT: Follow the general coding rules

Replace all masks and bits definitions by matching calls to BIT and GENMASK.
While we're at it, also fix a few style issues.

Signed-off-by: Maxime Ripard <maxime.ripard@free-electrons.com>
Acked-by: Boris BREZILLON <boris.brezillon@free-electrons.com>
Acked-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agopower: reset: at91-poweroff: fix wakeup status register index
Nicolas Ferre [Mon, 1 Sep 2014 14:11:19 +0000 (16:11 +0200)]
power: reset: at91-poweroff: fix wakeup status register index

The wakeup status is read from Shutdown Status Register and not the
SHDW_CR which is the one at address 0.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoARM: at91/power/reset: fix Kconfig "depends on" directive
Nicolas Ferre [Mon, 1 Sep 2014 08:26:04 +0000 (10:26 +0200)]
ARM: at91/power/reset: fix Kconfig "depends on" directive

MACH_AT91 doesn't exist so we can't "depend" on it. Fix the typo
by using the proper ARCH_AT91 config option.

Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoARM: at91: fix ramc standby function registration
Alexandre Belloni [Wed, 27 Aug 2014 16:06:51 +0000 (18:06 +0200)]
ARM: at91: fix ramc standby function registration

After the for_each_matching_node loop, we end up with a null value for np. Then,
of_match_node() is not matching anything and we can't register the standby
function and "ramc no standby function available" is printed.

Fix that by selecting the first available standby function. For now,
at91_pm_set_standby doesn't support multiple different standby functions and no
existing SoCs have different RAM controllers.

Signed-off-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
10 years agoLinux 3.17-rc3
Linus Torvalds [Mon, 1 Sep 2014 01:23:04 +0000 (18:23 -0700)]
Linux 3.17-rc3

10 years agoMerge tag 'xtensa-20140830' of git://github.com/czankel/xtensa-linux
Linus Torvalds [Mon, 1 Sep 2014 00:08:42 +0000 (17:08 -0700)]
Merge tag 'xtensa-20140830' of git://github.com/czankel/xtensa-linux

Pull Xtensa updates from Chris Zankel:
 "Xtensa improvements for 3.17:
   - support highmem on cores with aliasing data cache.  Enable highmem
     on kc705 by default
   - simplify addition of new core variants (no need to modify Kconfig /
     Makefiles)
   - improve robustness of unaligned access handler and its interaction
     with window overflow/underflow exception handlers
   - deprecate atomic and spill registers syscalls
   - clean up Kconfig: remove orphan MATH_EMULATION, sort 'select'
     statements
   - wire up renameat2 syscall.

  Various fixes:
   - fix address checks in dma_{alloc,free}_coherent (runtime BUG)
   - fix access to THREAD_RA/THREAD_SP/THREAD_DS (debug build breakage)
   - fix TLBTEMP_BASE_2 region handling in fast_second_level_miss
     (runtime unrecoverable exception)
   - fix a6 and a7 handling in fast_syscall_xtensa (runtime userspace
     register clobbering)
   - fix kernel/user jump out of fast_unaligned (potential runtime
     unrecoverabl exception)
   - replace termios IOCTL code definitions with constants (userspace
     build breakage)"

* tag 'xtensa-20140830' of git://github.com/czankel/xtensa-linux: (25 commits)
  xtensa: deprecate fast_xtensa and fast_spill_registers syscalls
  xtensa: don't allow overflow/underflow on unaligned stack
  xtensa: fix a6 and a7 handling in fast_syscall_xtensa
  xtensa: allow single-stepping through unaligned load/store
  xtensa: move invalid unaligned instruction handler closer to its users
  xtensa: make fast_unaligned store restartable
  xtensa: add double exception fixup handler for fast_unaligned
  xtensa: fix kernel/user jump out of fast_unaligned
  xtensa: configure kc705 for highmem
  xtensa: support highmem in aliasing cache flushing code
  xtensa: support aliasing cache in kmap
  xtensa: support aliasing cache in k[un]map_atomic
  xtensa: implement clear_user_highpage and copy_user_highpage
  xtensa: fix TLBTEMP_BASE_2 region handling in fast_second_level_miss
  xtensa: allow fixmap and kmap span more than one page table
  xtensa: make fixmap region addressing grow with index
  xtensa: fix access to THREAD_RA/THREAD_SP/THREAD_DS
  xtensa: add renameat2 syscall
  xtensa: fix address checks in dma_{alloc,free}_coherent
  xtensa: replace IOCTL code definitions with constants
  ...