Signed-off-by: NJosh Wu <josh.wu@atmel.com>
Acked-by: NJean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Acked-by: NLudovic Desroches <ludovic.desroches@atmel.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

The PA24 pin is wrongly assigned to peripheral B.
In the current config there is 2 ETX3 pins (PA11 and PA24) and
no ETXER pin (PA22).
Signed-off-by: NBoris BREZILLON <b.brezillon@overkiz.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>
Cc: stable <stable@vger.kernel.org> # 3.8+

Signed-off-by: NJean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Commit d3f79584 ("ARM: cleanup undefined instruction entry code")
improved the register scheduling when handling undefined instructions.
A side effect of this is that r5 is now used as a temporary, whilst the
VFP probing code relies on r5 containing a non-zero value when VFP is
not supported.

This patch fixes the VFP detection code so that we don't rely on the
contents of r5. Without this patch, Linux dies loudly on CPUs without
VFP support.
Signed-off-by: NWill Deacon <will.deacon@arm.com>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

If one reads /proc/$PID/smaps, the mmap_sem belonging to the
address space of the task being examined is locked for reading.
All the pages of the vmas belonging to the task's address space
are then walked with this lock held.

If a gate_vma is present in the architecture, it too is examined
by the fs/proc/task_mmu.c code. As gate_vma doesn't belong to the
address space of the task though, its pages are not walked.

A recent cleanup (commit f6604efe) of the gate_vma initialisation
code set the vm_mm value to &init_mm. Unfortunately a non-NULL
vm_mm value in the gate_vma will cause the task_mmu code to attempt
to walk the pages of the gate_vma (with no mmap-sem lock held). If
one enables Transparent Huge Page support and vm debugging, this
will then cause OOPses as pmd_trans_huge_lock is called without
mmap_sem being locked.

This patch removes the .vm_mm value from gate_vma, restoring the
original behaviour of the task_mmu code.
Signed-off-by: NSteve Capper <steve.capper@linaro.org>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

Make the SHA1 asm code ABI conformant by making sure all stack
accesses occur above the stack pointer.

Origin:
http://git.openssl.org/gitweb/?p=openssl.git;a=commit;h=1a9d60d2Signed-off-by: NArd Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: NNicolas Pitre <nico@linaro.org>
Cc: stable@vger.kernel.org
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

This patch updates the marzen board code as if USB PHY isn't enabled
they phy will have a value set by ERR_PTR() rather than be NULL.

Without this patch a NULL pointer dereference  and kernel panic
occurs on initialisation of USB on marzen.

This resolves a regression introduced in 3.10-rc1 by
b7fa5c2a
("usb: phy: return -ENXIO when PHY layer isn't enabled").
Tested-by: NNguyen Hong Ky <nh-ky@jinso.co.jp>
Signed-off-by: NSimon Horman <horms+renesas@verge.net.au>

Before f7b861b7 ("arm: Use generic idle loop") ARM would kill the
CPU within the rcu idle section.  Now that the rcu_idle_enter()/exit()
pair have been pushed lower down in the idle loop this is no longer true
and so using RCU_NONIDLE here is no longer necessary and also harmful
because RCU is not actually idle at this point.

Cc: Russell King <linux@arm.linux.org.uk>
Acked-by: NPaul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: NStephen Boyd <sboyd@codeaurora.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

The AK8975 Kconfig option was renamed during the 3.10 merge window.
Adjust tegra_defconfig to enable the new name, so it's not missing
useful features.

Tegra DRM support used to be enabled in the default Tegra configuration,
but it now depends on CONFIG_TEGRA_HOST1X which is disabled by default.
Enable CONFIG_TEGRA_HOST1X so that DRM support is compiled in again.
Signed-off-by: NAlexandre Courbot <acourbot@nvidia.com>
[swarren, squashed Alex's and my changes together]
Signed-off-by: NStephen Warren <swarren@nvidia.com>
Signed-off-by: NOlof Johansson <olof@lixom.net>

The assignment of IRQ for the SMC91x ethernet adapter had two
problems making it non-working:

- It was not put into the ethernet device node. Let's do this
  by using the board-specific overlay, so we can make other
  overlays on other Nomadik boards.

- The IRQ number was actually completely wrong, this was the
  number for NHK8815, not S8815.

After this ethernet starts working on the USB S8815.
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Signed-off-by: NOlof Johansson <olof@lixom.net>

When I tried booting a stih415 Dual core A9 with multi_v7_defconfig, it
failed to boot. The issues seems to be changing by enabling or disabling
VT8550 platform. Having a quick look at dt_compat list, it seems to miss
a NULL terminator, which means of_flat_dt_match will compat check will
cross the boundary of dt_compat and fault at some point , which is what
was happening in my case.

Without this patch if we try to boot multi_v7_defconfig you might notice
that some of the platforms might fault if they fall after vt8500 in
machine-desc list. Other platforms which fall before vt8500 in mdesc list
will not fault.
Signed-off-by: NSrinivas Kandagatla <srinivas.kandagatla@st.com>
Acked-by: NTony Prisk <linux@prisktech.co.nz>
Signed-off-by: NOlof Johansson <olof@lixom.net>

During the conversion to the internal-regs' subnode, the L2-cache node
haven not been converted (due to a wrong choice made by myself during
the resolution of the merge conflict when I rebased the commit). This
leads to wrong address for L2 cache which prevent it to be used on
Armada 370. This commit fix the address reg of the e L2-cache node.
Signed-off-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: NJason Cooper <jason@lakedaemon.net>

When platform data were moved from arch/arm/mach-mv78xx0/common.c to
arch/arm/plat-orion/common.c with the commit "7e3819d8 ARM: orion:
Consolidate ethernet platform data", there were few typo made on
gigabit Ethernet interface ge10 and ge11. This commit writes back
their initial value, which allows to use this interfaces again.
Signed-off-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Acked-by: NAndrew Lunn <andrew@lunn.ch>
Cc: <stable@vger.kernel.org> # v3.0.x
Signed-off-by: NJason Cooper <jason@lakedaemon.net>

With the OMAP serial driver sysc cleanup patches in this series, we can
now remove the hwmod external apis for sysc fiddling.

While at this, also remove unused sysc auto idle api from hwmod code.
Tested-by: NVaibhav Bedia <vaibhav.bedia@ti.com>
Tested-by: NSourav Poddar <sourav.poddar@ti.com>
Signed-off-by: NRajendra nayak <rnayak@ti.com>
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Reviewed-by: NKevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>  # OMAP4/Panda
Signed-off-by: NPaul Walmsley <paul@pwsan.com>

UART IP idle handling now taken care by runtime pm backend(hwmod) indirectly
and OMAP serial driver is also cleaned up accordingly.

So remove the un-used slave idle platforms hooks now.
Tested-by: NVaibhav Bedia <vaibhav.bedia@ti.com>
Tested-by: NSourav Poddar <sourav.poddar@ti.com>
Signed-off-by: NRajendra nayak <rnayak@ti.com>
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Reviewed-by: NKevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>  # OMAP4/Panda
Signed-off-by: NPaul Walmsley <paul@pwsan.com>

OMAP UART IP needs software control for slave idle modes based on functional
state of the IP. i.e The IP slave idle settings should be set to 'noidle' when
being used and then put back to 'smart_idle' when unused. Currently this is
handled by the driver with function pointers implemented in platform code.
This however breaks in case of device tree because of missing idle handling
APIs.

Previous patches in this series added a flag HWMOD_SWSUP_SIDLE_ACTIVE which
takes care of the mentioned requirement. Hence add the flag for all UART IPs
to take advantage of feature supported by framework.

Subsequent patches removes the slave idle handling from driver code.
Tested-by: NVaibhav Bedia <vaibhav.bedia@ti.com>
Tested-by: NSourav Poddar <sourav.poddar@ti.com>
Signed-off-by: NRajendra Nayak <rnayak@ti.com>
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Reviewed-by: NKevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>  # OMAP4/Panda
Signed-off-by: NPaul Walmsley <paul@pwsan.com>

Some IPs (like UART) need the sidle mode to be controlled in SW only
while they are active. Once they go inactive, they need the IP to be
put back in HW control so they are also wakeup capable.

The flag HWMOD_SWSUP_SIDLE takes care of IPs which need the sidle
mode to be *always* controlled in SWSUP. We now have a need to control
IPs sidle mode in SWSUP only while its active.

So define a new flag 'HWMOD_SWSUP_SIDLE_ACT' to help the framework
know about these new IP requirements.
Tested-by: NVaibhav Bedia <vaibhav.bedia@ti.com>
Tested-by: NSourav Poddar <sourav.poddar@ti.com>
Signed-off-by: NRajendra Nayak <rnayak@ti.com>
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Reviewed-by: NKevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>  # OMAP4/Panda
Signed-off-by: NPaul Walmsley <paul@pwsan.com>

_enable_sysc() and _idle_sysc() handle the midle mode programming correctly
and program HWMOD_IDLEMODE_SMART or HWMOD_IDLEMODE_SMART_WKUP respectively
for supported IPs (The ones which support hardware controlled midle modes)

However the same programming logic is missing when it comes to sidle mode
programming. Here they seem to just set HWMOD_IDLEMODE_SMART (Again for the
ones which support hardware controlled sidle modes)

This problem was hidden due to the fact that a call to _enable_wakeup()
in those same functions would overwrite the idlemodes and program them
correctly (to HWMOD_IDLEMODE_SMART_WKUP in the supported cases)

So fix the sidlemode handling correctly in these functions and handle the
_enable_wakeup() for SIDLEMODE supported IPs same as the way its handled
for MIDLEMODE supported ones.
Tested-by: NVaibhav Bedia <vaibhav.bedia@ti.com>
Tested-by: NSourav Poddar <sourav.poddar@ti.com>
Signed-off-by: NRajendra Nayak <rnayak@ti.com>
Signed-off-by: NSantosh Shilimkar <santosh.shilimkar@ti.com>
Reviewed-by: NKevin Hilman <khilman@linaro.org>
Tested-by: Kevin Hilman <khilman@linaro.org>  # OMAP4/Panda
Signed-off-by: NPaul Walmsley <paul@pwsan.com>

Since 82a68267 ('ARM: dts: mvebu: Convert all the mvebu files to use
the range property') all the device nodes of Armada 370/XP are under a
common 'ranges' property that translates the device register addresses
into their absolute address, thanks to the base address of the
internal register space.

However, beyond just the register areas, there are also PCIe I/O and
memory regions, whose addresses should be properly translated. This
patch fixes the Armada 370 and XP ranges property to take PCIe into
account properly.
Signed-off-by: NThomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: NJason Cooper <jason@lakedaemon.net>

Initially ARCH_REQUIRE_GPIOLIB was part of Thomas Petazzoni series
when he introduced the gpiolib support for mvebu:
93a59cf3 arm: mvebu: use GPIO support now that a driver is available

This commit was written to be applied for the ARCH_MVEBU which was
located in arch/arm/KConfig and was merged in 3.7.

In the same time Rob Herring moved the ARCH_MVEBU block to
arch/arm/mach-mvebu/Kconfig with this commit and also merged in 3.7:
387798b3 ARM: initial multiplatform support

Unfortunately the ARCH_REQUIRE_GPIOLIB have been lost during this
migration. This was not noticed until the v3.10-rc1, because mvebu as
part of ARCH_MULTIPLATFORM was always selected with ARCH_VEXPRESS, and
this architect selected ARCH_REQUIRE_GPIOLIB.

Since the following commit from Arnd: "883a106b ARM: default machine
descriptor for multiplatform", ARCH_VEXPRESS was then no more selected
by default with ARCH_MVEBU and it made appeared the lack of
ARCH_REQUIRE_GPIOLIB for mvebu. This commit added back the selection
of ARCH_REQUIRE_GPIOLIB for ARCH_MVEBU.
Signed-off-by: NGregory CLEMENT <gregory.clement@free-electrons.com>
Signed-off-by: NJason Cooper <jason@lakedaemon.net>

devm_ioremap_resource does sanity checks on the given resource. No need to
duplicate this in the driver.
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>

devm_ioremap_resource does sanity checks on the given resource. No need to
duplicate this in the driver.
Signed-off-by: NWolfram Sang <wsa@the-dreams.de>
Acked-by: NStephen Warren <swarren@nvidia.com>

It is required to enable respective clock-domain before
enabling any clock/module inside that clock-domain.

During common-clock migration, .clkdm_name field got missed
for "clkdiv32k_ick" clock, which leaves "clk_24mhz_clkdm"
unused; so it will be disabled even if childs of this clock-domain
is enabled, which keeps child modules in idle mode.

This fixes the kernel crash observed on AM335xEVM-SK platform,
where clkdiv32_ick clock is being used as a gpio debounce clock
and since clkdiv32k_ick is in idle mode it leads to below crash -

Crash Log:
==========
[    2.598347] Unhandled fault: external abort on non-linefetch (0x1028) at
0xfa1ac150
[    2.606434] Internal error: : 1028 [#1] SMP ARM
[    2.611207] Modules linked in:
[    2.614449] CPU: 0    Not tainted  (3.8.4-01382-g1f449cd-dirty #4)
[    2.620973] PC is at _set_gpio_debounce+0x60/0x104
[    2.626025] LR is at clk_enable+0x30/0x3c

Cc: stable@vger.kernel.org # v3.9
Signed-off-by: NVaibhav Hiremath <hvaibhav@ti.com>
Cc: Rajendra Nayak <rnayak@ti.com>
Acked-by: NPaul Walmsley <paul@pwsan.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

With the latest device tree reorganization which introduced the
'internal-reg' node, now the only region translated is the internal register's.
This makes the description of the hardware incomplete, for it lacks the
Device Bus childs address space.

In order to fix this, it's required to add a 'ranges' entry with a suitable
address space to map Device Bus childs, on a per-board basis.
This patch fixes the ranges property on the Armada XP GP board.
Signed-off-by: NEzequiel Garcia <ezequiel.garcia@free-electrons.com>
Signed-off-by: NJason Cooper <jason@lakedaemon.net>

omap2plus_defconfig is missing CONFIG_ARCH_MULTI_V6.  This results in
undefined instruction traps in u-boot (and boot failures) on OMAP2xxx
SoCs, which are ARM11-based.  Fix by setting CONFIG_ARCH_MULTI_V6.
Signed-off-by: NPaul Walmsley <paul@pwsan.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

Fix to return -ENOMEM in the d->chan alloc error handling
case instead of 0, as done elsewhere in this function.
Signed-off-by: NWei Yongjun <yongjun_wei@trendmicro.com.cn>
Signed-off-by: NTony Lindgren <tony@atomide.com>

_enable_preprogram is marked as __init, but is called from _enable
which is not. Without this patch, the board oopses after init. Tested
on custom hardware and on beagle board xM. Otherwise we can get:

Unable to handle kernel paging request at virtual address 000b0012
pgd = cf968000
*pgd=8fb06831, *pte=00000000, *ppte=00000000
PREEMPT ARM
Modules linked in:
CPU: 0    Not tainted  (3.9.0 #2)
PC is at _enable_preprogram+0x1c/0x24
LR is at omap_hwmod_enable+0x34/0x60
   psr: 80000093
sp : cf95de08  ip : 00002de5  fp : bec33d4c
r10: 00000000  r9 : 00000002  r8 : b6dd2c78
r7 : 00000004  r6 : 00000000  r5 : a0000013  r4 : cf95c000
r3 : 00000000  r2 : b6dd2c7c  r1 : 00000000  r0 : 000b0012
Flags: Nzcv  IRQs off  FIQs on  Mode SVC_32  ISA ARM  Segment user
Control: 10c5387d  Table: 8f968019  DAC: 00000015
Process otpcmd (pid: 607, stack limit = 0xcf95c230)
Stack: (0xcf95de08 to 0xcf95e000)
de00:                   00000001 cf91f840 00000000 c001d6fc 00000002 cf91f840
de20: cf8f7e10 c001de54 cf8f7e10 c001de78 c001de68 c01d5e80 00000000 cf8f7e10
de40: cf8f7e10 c01d5f28 cf8f7e10 c0530d30 00000000 c01d6f28 00000000 c0088664
de60: b6ea1000 cfb05284 cf95c000 00000001 cf95c000 60000013 00000001 cf95dee4
de80: cf870050 c01d7308 cf870010 cf870050 00000001 c0278b14 c0526f28 00000000
dea0: cf870050 ffff8e18 00000001 cf95dee4 00000000 c0274f7c cf870050 00000001
dec0: cf95dee4 cf1d8484 000000e0 c0276464 00000008 cf9c0000 00000007 c0276980
dee0: cf9c0000 00000064 00000008 cf1d8404 cf1d8400 c01cc05c 0000270a cf1d8504
df00: 00000023 cf1d8484 00000007 c01cc670 00000bdd 00000001 00000000 cf449e60
df20: cf1dde70 cf1d8400 bec33d18 cf1d8504 c0246f00 00000003 cf95c000 00000000
df40: bec33d4c c01cd078 00000003 cf1d8504 00000081 c01cbcb8 bec33d18 00000003
df60: bec33d18 c00a9034 00002000 c00a9c68 cf92fe00 00000003 c0246f00 cf92fe00
df80: 00000000 c00a9cb0 00000003 00000000 00008e70 00000000 b6f17000 00000036
dfa0: c000e484 c000e300 00008e70 00000000 00000003 c0246f00 bec33d18 bec33d18
dfc0: 00008e70 00000000 b6f17000 00000036 00000000 00000000 b6f6d000 bec33d4c
dfe0: b6ea1bd0 bec33d0c 00008c9c b6ea1bdc 60000010 00000003 00000000 00000000
(_omap_device_enable_hwmods+0x20/0x34)
(omap_device_enable+0x3c/0x50)
(_od_runtime_resume+0x10/0x1c)
(__rpm_callback+0x54/0x98)
(rpm_callback+0x64/0x7c)
(rpm_resume+0x434/0x554)
(__pm_runtime_resume+0x48/0x74)
(omap_i2c_xfer+0x28/0xe8)
(__i2c_transfer+0x3c/0x78)
(i2c_transfer+0x6c/0xc0)
(i2c_master_send+0x38/0x48)
(sha204p_send_command+0x60/0x9c)
(sha204c_send_and_receive+0x5c/0x1e0)
(sha204m_read+0x94/0xa0)
(otp_do_read+0x50/0xa4)
(vfs_ioctl+0x24/0x40)
(do_vfs_ioctl+0x1b0/0x1c0)
(sys_ioctl+0x38/0x54)
(ret_fast_syscall+0x0/0x30)
Code: e1a08002 ea000009 e598003c e592c05c (e7904003)

Cc: stable@vger.kernel.org
Signed-off-by: NJean-Philippe Fran=C3=A7ois <jp.francois@cynove.com>
Acked-by: NKevin Hilman <khilman@linaro.org>
[tony@atomide.com: updated description with oops]
Signed-off-by: NTony Lindgren <tony@atomide.com>

Commit ad871c10 (ARM: dts: OMAP: Add usb_otg and glue data to
OMAP3+ boards) added support for MUSB on omap3 for device tree,
but added the interrupts the wrong way probably as they were
copied from the omap4.dtsi file. On omap3 we have TI specific
interrupt controller, not GIC.

Fix this by specifying the interrupt following the TI INTC
binding.

Without this fix MUSB won't work as it is trying to use
irq0 instead of irq92.
Signed-off-by: NTony Lindgren <tony@atomide.com>

Commit 84ebc102 (USB: remove
CONFIG_USB_SUSPEND option) failed to remove all of the usages of
USB_SUSPEND throughout the kernel.  This patch (as1677) removes the
remaining instances of that symbol.
Signed-off-by: NAlan Stern <stern@rowland.harvard.edu>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

MACH_MOP500 selects REGULATOR_FIXED_VOLTAGE but not REGULATOR. This
patch makes it select the latter too.

Seen as:
    warning: (MACH_DOVE_DT && ARCH_KIRKWOOD_DT && MACH_AMS_DELTA &&
    MACH_MOP500 && TPS6105X) selects REGULATOR_FIXED_VOLTAGE which has
    unmet direct dependencies (REGULATOR)
Signed-off-by: NEmilio López <emilio@elopez.com.ar>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

since commit 838a2ae8
Author: Shawn Guo <shawn.guo@linaro.org>
Date:   Sat Jan 12 11:50:05 2013 +0000

ARM: use clockevents_config_and_register() where possible

The timer is wrongly configured and result in a nice crash
so revert it on rm9200 timer.
Signed-off-by: NJean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Cc: stable <stable@vger.kernel.org> # 3.8+
Cc: Shawn Guo <shawn.guo@linaro.org>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

CONFIG_ARCH_SUNXI currently doesn't enable a gpiolib, which causes build
problems when building a kernel with only the sunxi platform enabled.
Select ARCH_REQUIRE_GPIOLIB to solve this.
Signed-off-by: NMaxime Ripard <maxime.ripard@free-electrons.com>
Tested-by: NEmilio López <emilio@elopez.com.ar>

Instead of requesting all available spi cs-gpios, request only the ones used on
the board, in our case on the cpu module.
Signed-off-by: NLudovic Desroches <ludovic.desroches@atmel.com>
Acked-by: NJean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Signed-off-by: NJean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>
Cc: stable <stable@vger.kernel.org> # 3.8+

First Ethernet device has a ".0" appended onto the device name.
Since on a non-DT boot the ethernet will be named "smsc911x.0"
and since the clocks are not converted to device tree these
names need to be matched when providing the name from
auxdata.
Signed-off-by: NLee Jones <lee.jones@linaro.org>
[edited commit message]
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

A recent move to rid header files which were hindering multiplatform
support forced address allocations out of the headers and into the
files which were using them. We also lost some useful macros such as
IO_ADDRESS(), so physical -> virtual addressing has been carried out
manually in this case. Unfortunately the incorrect value was converted.
This patch rectifies the error and ensures earlyprintk works again.
Signed-off-by: NLee Jones <lee.jones@linaro.org>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

Since: "05ec260e mfd: db8500-prcmu: update resource passing", the AB8500's
platform data 'ab8500_platdata' is passed directly as an attribute to
'db8500_prcmu_pdata', so there's no requirement to assign it a second
time. In fact, it's only due to an ordering issue that the entire
'db8500_prcmu_pdata' data structure isn't completely over-written by the
assignment in u8500_init_devices().
Signed-off-by: NLee Jones <lee.jones@linaro.org>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>

The implementation of cmpxchg64() for the ARM v6 and v7 architecture
casts parameter 2 and 3 (the old and new 64bit values) to an unsigned
long before calling the atomic_cmpxchg64() function. This clears
the top 32 bits of the old and new values, resulting in the wrong
values being compare-exchanged. Luckily, this only appears to be used
for 64-bit sched_clock, which we don't (yet) have on ARM.

This bug was introduced by commit 3e0f5a15 ("ARM: 7404/1: cmpxchg64:
use atomic64 and local64 routines for cmpxchg64").

Cc: <stable@vger.kernel.org>
Acked-by: NWill Deacon <will.deacon@arm.com>
Signed-off-by: NJaccon Bastiaansen <jaccon.bastiaansen@gmail.com>
Signed-off-by: NRussell King <rmk+kernel@arm.linux.org.uk>

The patch:

  387870f2 mm: dmapool: use provided gfp flags for all dma_alloc_coherent() calls

makes these calls on Kirkwood and Orion5x redundant. The drivers are
not making atomic requests for coherent memory and hence the default
pool size is now sufficient.

Jason Cooper added mach-mvebu/ hunk, and corrected minor typos in commit
message.
Signed-off-by: NAndrew Lunn <andrew@lunn.ch>
Signed-off-by: NJason Cooper <jason@lakedaemon.net>