With all the DT support preparation done, we are able to move Dove
to MVEBU easily. Legacy non-DT mach-dove is left untouched to rot
for a while before removal.

Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Convert the kirkwood t5325-setup.c to mostly device tree for
mach-mvebu. Part of the audio setup needs to remain in C for the
moment until suitable bindings are designed and implemented. So add
board code, triggered by the compatibility string.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

To prevent problems later with mvebu_v5_defconfig builds, we restrict
the scope of the changes this series makes.

Selecting ARCH_KIRKWOOD and MACH_KIRKWOOD is necessary during the
migration from mach-kirkwood to mach-mvebu.  Until the last four legacy
kirkwood board files have a DT equivalent, we need to support building
DT from both mach-kirkwood and mach-mvebu.

Reported-by: Kevin Hilman <khilman@linaro.org>
Acked-by: Kevin Hilman <khilman@linaro.org>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

The PCIe driver has been fully clock aware for quite a while.  Remove
the kexec code to enable the PCIe clock, since the PCIe driver will do
the right thing.

jac adds:
  [arnd]: fixes a build error when KEXEC is enabled and KIRKWOOD_LEGACY is not

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

During this release cycle, we're adding the new Armada 375, 380, and 385
SoCs.  We're also migrating DT kirkwood boards into mach-mvebu.  The
kirkwood changes make the different SoCs in mach-mvebu/ depend on
MULTI_V7 or MULTI_V5 as appropriate.

We add this dependency to the new SoCs so that when the branches are
merged, everything is as it should be.

Acked-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

kirkwood is very nearly fully DT. Remove most of the address
definitions from the header files and make it a local header file.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Add the Kirkwood PCIe compatibility string to mvebu-soc-id, so that it
can get the SoC ID and revision from the PCIe endpoints.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

The mvebu system controller already supports restarting orion
systems. Remove all the C code which will be replaced by the system
controller.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Move the kirkwood DT support into mach-mvebu, and make them part of
ARCH_MULTI_V5. Minimal changes have been made in order to make it
boot. Cleanup of the header files and integration with mvebu will
take place in following patches.

In order to help Debian transition between mach-kirkwood and
mach-mvebu, the DTS files are compiled for both, allowing Debian to
continue using mach-kirkwood until all remaining boards are supported
by mach-mvebu. Debian is then expected to simply swap from
mach-kirkwood to mach-mvebu and mach-kirkwood will be removed.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Kirkwood, which uses the Feroceon L2 cache controller will soon be
moving into mach-mvebu. Allow the cache controller to be built in this
situation.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

CPU_ARM926T should be selected if no other CPU is. Put the ! in the
right place so this works.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Fixes: 24e860fb

 ("ARM: multiplatform: always pick one CPU type")
Cc: stable@vger.kernel.org # v3.11+
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Instantiate the L2 cache from DT. Indicate in DT where the cache
control register is so that it is possible to enable/disable write
through on the CPU.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

With the gradual move to DT, kirkwood has become a lot less dependent
on plat-orion. cache-feroceon-l2.h is the last dependency. Move it out
so we can drop plat-orion when building DT only kirkwood boards.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Prepare mach-mvebu to house both ARCH_MULTI_V7 and ARCH_MULTI_V5
systems by adding ARCH_MULTI_V7 to the existing SOCs.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

To allow removal of the global map of registers, make the pm code
ioremap the registers it needs.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

With the move to mach-mvebu and MULTI_V5, the global iomap for all
registers will be going away. So explicitly map the CPU configuration
register before using it.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

In order to be able to move DT support into mach-mvebu, the DT code
needs to be cleanly separated from common and pcie code. Import the
needed bits of these files into board-dt.c. The "common" code then
becomes purely legacy, supporting non-DT boards, so reflect this in
the Makefile targets.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

This will be added back using the mach-mvebu equivalent once the move
has been made.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

The mv88f6281gtw_ge has a ethernet switch connected to the ethernet
port of the SoC. Convert the platform device instantiation to a DT
instantiation.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Cc: florian@openwrt.org
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

The pm code needs to be separated from common.h in order to split DT
and non-DT systems apart. Move the declarations into a header file of
its own and include it where needed.

Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Acked-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Tested-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com> (on kirkwood)
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

The previous name "Marvell SOCs with Device Tree support" is a bit
ambiguous and not too informative for users. Instead, use a more
appropriate name.

Signed-off-by: Ezequiel Garcia <ezequiel.garcia@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

This commit adds the basic support for the Armada 380 and Armada 385
SOCs. These SoCs share most of their IP with the Armada 370/XP
SoCs. The main difference is the use of a Cortex A9 CPU instead of the
PJ4B CPU. The Armada 380 is a single core Cortex-A9, while the Armada
385 is a dual-core Cortex-A9.

The support is introduced in board-v7.c, together with Armada 370/XP,
but a separate DT structure is added, because Armada 38x will need a
different set of SMP operations when the SMP support is introduced.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Early versions of Armada 375 SoC have a bug where the BootROM leaves
an external data abort pending. The kernel is hit by this data abort
as soon as it enters userspace, because it unmasks the data aborts at
this moment. We register a custom abort handler below to ignore the
first data abort to work around this problem.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

This commit adds the basic support for the Armada 375 SOCs. These SoCs
share most of their IP with the Armada 370/XP SoCs. The main
difference is the use of a Cortex A9 CPU instead of the PJ4B CPU. The
interrupt controller and the L2 cache controller are also different
they are respectively the GIC and the PL310.

The support is introduced in board-v7.c, together with Armada 370/XP,
but a separate DT structure is added, because Armada 375 will need a
different set of SMP operations when the SMP support is introduced.

Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

The system controller block in the Armada 375 has different register
offsets for the system reset and other related functions. Therefore,
this commit introduces the new "armada-375-system-controller"
compatible string to identify the Armada 375 variant of the system
controller.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Until now, the CPU_PJ4B Kconfig option was selected by
MACH_ARMADA_MVEBU, i.e for all Armada MVEBU SOCs. In preparation to
the introduction of Cortex-A9 based Armada MVEBU SOCs, this selection
is moved down to the Armada 370 and Armada XP specific options.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Due to a mistake made when merging Armada 370 and Armada XP DT machine
structures, the name of the structure was incorrectly chosen as being
ARMADA_XP_DT, while the structure also covers Armada 370. Therefore,
we rename the structure to ARMADA_370_XP_DT.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

In preparation to the introduction of the support of Armada 375 and
Armada 38x, this commit renames arch/arm/mach-mvebu/armada-370-xp.c to
arch/arm/mach-mvebu/board-v7.c. The board-v7.c name as we expect this
file to ultimately contain the DT_MACHINE_START definitions for all
ARMv7 Marvell EBU platforms (370, 375, 38x, XP and Dove as of today).

In relation to this file rename, this commit also:

 * Renames the hidden Kconfig symbol MACH_ARMADA_370_XP to
   MACH_MVEBU_V7. This hidden symbol is selected by the various
   per-SoC visible Kconfig options to trigger the build of board-v7.c.

 * Renames a certain number of functions in board-v7.c so that their
   armada_370_xp prefix is replaced by a mvebu prefix. The .dt_compat
   array keeps its armada_370_xp prefix because the new SOCs will be
   introduced with separate .dt_compat arrays, due to the need for
   different SMP operations.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

The ->map_io() implementation of Armada 370/XP simply calls
debug_ll_io_init(), which is exactly what the kernel does when
->map_io is NULL. Therefore, there is no need to have a specific
->map_io() implementation in Armada 370/XP.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Instead of the of_find_matching_node()/of_match_node() pair, which requires two
iterations through the match table, make use of of_find_matching_node_and_match(),
which only iterates through the table once.

While we're here, mark the of_system_controller table const.

Signed-off-by: Josh Cartwright <joshc@codeaurora.org>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Some objects depend on CONFIG_ARCH_MVEBU whereas this whole Makefile
depends on the same symbol. Moreover CONFIG_ARCH_MVEBU can't be
selected as a module. So we can simplify this Makefile by moving all
the object from obj-$(CONFIG_ARCH_MVEBU) to obj-y.

Signed-off-by: Gregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: Jason Cooper <jason@lakedaemon.net>

Add support for the flexible mmap memory layout (as described in
http://lwn.net/Articles/91829

). This is especially very interesting on
parisc since we currently only support 32bit userspace (even with a
64bit Linux kernel).

Signed-off-by: Helge Deller <deller@gmx.de>

On Linux, only parisc uses a different value for EWOULDBLOCK which
causes a lot of troubles for applications not checking for both values.
Since the hpux compat is long dead, make EWOULDBLOCK behave the same as
all other architectures.

Signed-off-by: Guy Martin <gmsoft@tuxicoman.be>
Signed-off-by: Helge Deller <deller@gmx.de>

The stat.h header file is exported to userspace. Some userspace
applications failed to compile due to missing/unknown types, so we
better convert it to use native types only (like it's done on other
architectures too).

Signed-off-by: Helge Deller <deller@gmx.de>

Signed-off-by: Helge Deller <deller@gmx.de>

This commit:
f8dae006

: parisc: Ensure full cache coherency for kmap/kunmap
caused negative caching side-effects, e.g. hanging processes with expect and
too many inequivalent alias messages from flush_dcache_page() on Debian 5 systems.

This patch now partly reverts it and has been in production use on our debian buildd
makeservers since a week without any major problems.

Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: John David Anglin <dave.anglin@bell.net>
Cc: stable@vger.kernel.org # v3.9+
Signed-off-by: Helge Deller <deller@gmx.de>

With commit d8d14bd0

 ("fs/compat: fix lookup_dcookie() parameter
handling") I changed the type of the len parameter of the
lookup_dcookie() syscall.

However I missed that there was still a stale declaration in
arch/tile/..  which now causes a compile error on tile:

  In file included from fs/dcookies.c:28:0:
  include/linux/compat.h:425:17: error: conflicting types for 'compat_sys_lookup_dcookie'
  fs/dcookies.c:207:1: error: conflicting types for 'compat_sys_lookup_dcookie'

Simply remove the declaration in the tile architecture, which is only a
leftover from before the different compat lookup_dcookie() versions have
been merged.  The correct declaration is now in include/linux/compat.h

The build error was reported by Fenguang's build bot.

Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Acked-by: Chris Metcalf <cmetcalf@tilera.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

TI_EDMA fell out of automatically selected options in the multi_v7
defconfig due to a select being removed from the davinci Kconfig entry. So
we need to re-enable explicitly to not regress some platforms.

The rest is just the result of running 'make multi_v7_defconfig + make
savedefconfig' to remove entries that are no longer needed due to changed
dependencies/selects or defaults.

Signed-off-by: Olof Johansson <olof@lixom.net>

Recent boot farm testing has highlighted some issues with mvebu and
multiplatform kernels.  Increase the test coverage so we can discover
these issues earlier.

Signed-off-by: Jason Cooper <jason@lakedaemon.net>
Signed-off-by: Olof Johansson <olof@lixom.net>

The board schematic states that the "SD_CARD_DET_N gets pulled to GND
when card is inserted" so the polarity has been updated to active low.

Polarity is now specified with a GPIO define instead of a magic number.

Signed-off-by: Tim Kryger <tim.kryger@linaro.org>
Reviewed-by: Matt Porter <matt.porter@linaro.org>
Signed-off-by: Olof Johansson <olof@lixom.net>