It used to be an ad-hoc hack defined by the x86 version of
<asm/bitops.h> that enabled a couple of library routines to know whether
an integer multiply is faster than repeated shifts and additions.

This just makes it use the real Kconfig system instead, and makes x86
(which was the only architecture that did this) select the option.

NOTE! Even for x86, this really is kind of wrong.  If we cared, we would
probably not enable this for builds optimized for netburst (P4), where
shifts-and-adds are generally faster than multiplies.  This patch does
*not* change that kind of logic, though, it is purely a syntactic change
with no code changes.

This was triggered by the fact that we have other places that really
want to know "do I want to expand multiples by constants by hand or
not", particularly the hash generation code.
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

Remove the rbtree used to keep track of machine to physical mappings:
the frontend can grant the same page multiple times, leading to errors
inserting or removing entries from the mach_to_phys tree.

Linux only needed to know the physical address corresponding to a given
machine address in swiotlb-xen. Now that swiotlb-xen can call the
xen_dma_* functions passing the machine address directly, we can remove
it.
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>
Tested-by: NDenis Schneider <v1ne2go@gmail.com>

xen_dma_unmap_page, xen_dma_sync_single_for_cpu and
xen_dma_sync_single_for_device are currently implemented by calling into
the corresponding generic ARM implementation of these functions. In
order to do this, firstly the dma_addr_t handle, that on Xen is a
machine address, needs to be translated into a physical address.  The
operation is expensive and inaccurate, given that a single machine
address can correspond to multiple physical addresses in one domain,
because the same page can be granted multiple times by the frontend.

To avoid this problem, we introduce a Xen specific implementation of
xen_dma_unmap_page, xen_dma_sync_single_for_cpu and
xen_dma_sync_single_for_device, that can operate on machine addresses
directly.

The new implementation relies on the fact that the hypervisor creates a
second p2m mapping of any grant pages at physical address == machine
address of the page for dom0. Therefore we can access memory at physical
address == dma_addr_r handle and perform the cache flushing there. Some
cache maintenance operations require a virtual address. Instead of using
ioremap_cache, that is not safe in interrupt context, we allocate a
per-cpu PAGE_KERNEL scratch page and we manually update the pte for it.

arm64 doesn't need cache maintenance operations on unmap for now.
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>
Tested-by: NDenis Schneider <v1ne2go@gmail.com>

The flag tells us that the hypervisor maps a grant page to guest
physical address == machine address of the page in addition to the
normal grant mapping address. It is needed to properly issue cache
maintenance operation at the completion of a DMA operation involving a
foreign grant.
Signed-off-by: NStefano Stabellini <stefano.stabellini@eu.citrix.com>
Tested-by: NDenis Schneider <v1ne2go@gmail.com>

Nathan reports that we leak TLS information from the parent context
during an exec, as we don't clear the TLS registers when flushing the
thread state.

This patch updates the flushing code so that we:

  (1) Unconditionally zero the tpidr_el0 register (since this is fully
      context switched for native tasks and zeroed for compat tasks)

  (2) Zero the tp_value state in thread_info before clearing the
      tpidrr0_el0 register for compat tasks (since this is only writable
      by the set_tls compat syscall and therefore not fully switched).

A missing compiler barrier is also added to the compat set_tls syscall.

Cc: <stable@vger.kernel.org>
Acked-by: NNathan Lynch <Nathan_Lynch@mentor.com>
Reported-by: NNathan Lynch <Nathan_Lynch@mentor.com>
Signed-off-by: NWill Deacon <will.deacon@arm.com>

This patch avoids fuse hangs on sh4 by flushing the cache on
get_user_pages_fast().  This is not necessary a good thing to do, but
get_user_pages() does this, so get_user_pages_fast() should too.

Please note the patch for mips arch that addresses the similar problem:
  https://kernel.googlesource.com/pub/scm/linux/kernel/git/ralf/linux/+/linux-3.4.50%5E!/#F0

They basically simply disable get_user_pages_fast() at all, using a
fall-back to get_user_pages().  But my fix is different, it adds an
explicit cache flushes.
Signed-off-by: NStas Sergeev <stsp@users.sourceforge.net>
Cc: Geert Uytterhoeven <geert+renesas@glider.be>
Cc: Kamal Dasu <kdasu.kdev@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

When RANDOMIZE_BASE (KASLR) is enabled; or the sum of all loaded
modules exceeds 512 MiB, then loading modules fails with a warning
(and hence a vmalloc allocation failure) because the PTEs for the
newly-allocated vmalloc address space are not zero.

  WARNING: CPU: 0 PID: 494 at linux/mm/vmalloc.c:128
           vmap_page_range_noflush+0x2a1/0x360()

This is caused by xen_setup_kernel_pagetables() copying
level2_kernel_pgt into level2_fixmap_pgt, overwriting many non-present
entries.

Without KASLR, the normal kernel image size only covers the first half
of level2_kernel_pgt and module space starts after that.

L4[511]->level3_kernel_pgt[510]->level2_kernel_pgt[  0..255]->kernel
                                                  [256..511]->module
                          [511]->level2_fixmap_pgt[  0..505]->module

This allows 512 MiB of of module vmalloc space to be used before
having to use the corrupted level2_fixmap_pgt entries.

With KASLR enabled, the kernel image uses the full PUD range of 1G and
module space starts in the level2_fixmap_pgt. So basically:

L4[511]->level3_kernel_pgt[510]->level2_kernel_pgt[0..511]->kernel
                          [511]->level2_fixmap_pgt[0..505]->module

And now no module vmalloc space can be used without using the corrupt
level2_fixmap_pgt entries.

Fix this by properly converting the level2_fixmap_pgt entries to MFNs,
and setting level1_fixmap_pgt as read-only.

A number of comments were also using the the wrong L3 offset for
level2_kernel_pgt.  These have been corrected.
Signed-off-by: NStefan Bader <stefan.bader@canonical.com>
Signed-off-by: NDavid Vrabel <david.vrabel@citrix.com>
Reviewed-by: NBoris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: stable@vger.kernel.org

Number of syscalls have to be updated too.
Signed-off-by: NMichal Simek <michal.simek@xilinx.com>

"Advance setup: menu is misleading that's why rename it.
Signed-off-by: NMichal Simek <michal.simek@xilinx.com>

mm/Kconfig is getting too big to be in root menu.
Move it to submenu.
Signed-off-by: NMichal Simek <michal.simek@xilinx.com>

When DYNAMIC_DEBUG enabled, pr_debug() depends on KBUILD_MODNAME which
also depends on the modules number in Makefile. The related information
in "scripts/Makefile.lib" line 94:

  # $(modname_flags) #defines KBUILD_MODNAME as the name of the module it will
  # end up in (or would, if it gets compiled in)
  # Note: Files that end up in two or more modules are compiled without the
  #       KBUILD_MODNAME definition. The reason is that any made-up name would
  #       differ in different configs.

For this case, 'radio-si470x-i2c.o' and 'radio-si470x-common.o' are in
one line, so cause compiling issue. And 'uaccess.h' is a common shared
header (not specially for drivers), so use pr_devel() instead of is OK.

The related error with allmodconfig:

    CC [M]  drivers/media/radio/si470x/radio-si470x-i2c.o
    CC [M]  drivers/media/radio/si470x/radio-si470x-common.o
  In file included from include/linux/printk.h:257:0,
                   from include/linux/kernel.h:13,
                   from drivers/media/radio/si470x/radio-si470x.h:29,
                   from drivers/media/radio/si470x/radio-si470x-common.c:115:
  ./arch/microblaze/include/asm/uaccess.h: In function 'access_ok':
  include/linux/dynamic_debug.h:66:14: error: 'KBUILD_MODNAME' undeclared (first use in this function)
     .modname = KBUILD_MODNAME,   \
                ^
  include/linux/dynamic_debug.h:76:2: note: in expansion of macro 'DEFINE_DYNAMIC_DEBUG_METADATA'
    DEFINE_DYNAMIC_DEBUG_METADATA(descriptor, fmt);  \
    ^
  include/linux/printk.h:263:2: note: in expansion of macro 'dynamic_pr_debug'
    dynamic_pr_debug(fmt, ##__VA_ARGS__)
    ^
  ./arch/microblaze/include/asm/uaccess.h:101:3: note: in expansion of macro 'pr_debug'
     pr_debug("ACCESS fail: %s at 0x%08x (size 0x%x), seg 0x%08x\n",
     ^
Signed-off-by: NChen Gang <gang.chen.5i5j@gmail.com>
Signed-off-by: NMichal Simek <michal.simek@xilinx.com>

"entry.h" needs 'asmlinkage', and "asm/linkage.h" does not provide it.
So need include "linux/linkage.h" to use generic one instead of.

The related error (with allmodconfig under microblaze):

    CC [M]  drivers/net/ethernet/emulex/benet/be_main.o
  In file included from ./arch/microblaze/include/asm/processor.h:17:0,
                   from include/linux/prefetch.h:14,
                   from drivers/net/ethernet/emulex/benet/be_main.c:18:
  ./arch/microblaze/include/asm/entry.h:33:19: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'void'
   extern asmlinkage void do_notify_resume(struct pt_regs *regs, int in_syscall);
                     ^
Signed-off-by: NChen Gang <gang.chen.5i5j@gmail.com>
Signed-off-by: NMichal Simek <michal.simek@xilinx.com>

This patch wires up three new syscalls for powerpc. The three
new syscalls are seccomp, getrandom and memfd_create.
Signed-off-by: NPranith Kumar <bobby.prani@gmail.com>
Reviewed-by: NDavid Herrmann <dh.herrmann@gmail.com>

CONFIG_FHANDLE is a requirement for systemd and with the increasing
uptake of systemd within distros it makes sense for 64 bit defconfigs
to include it.
Signed-off-by: NCyril Bur <cyril.bur@au1.ibm.com>

As opal_message_init() uses machine_early_initcall(powernv, ), and
opal_hmi_handler_init() depends on that early initcall, so it also needs
use machine_* to check the machine_id.
Signed-off-by: NLi Zhong <zhong@linux.vnet.ibm.com>

ABIv2 kernels are failing to backtrace through the kernel. An example:

39.30%  readseek2_proce  [kernel.kallsyms]    [k] find_get_entry
            |
            --- find_get_entry
               __GI___libc_read

The problem is in valid_next_sp() where we check that the new stack
pointer is at least STACK_FRAME_OVERHEAD below the previous one.

ABIv1 has a minimum stack frame size of 112 bytes consisting of 48 bytes
and 64 bytes of parameter save area. ABIv2 changes that to 32 bytes
with no paramter save area.

STACK_FRAME_OVERHEAD is in theory the minimum stack frame size,
but we over 240 uses of it, some of which assume that it includes
space for the parameter area.

We need to work through all our stack defines and rationalise them
but let's fix perf now by creating STACK_FRAME_MIN_SIZE and using
in valid_next_sp(). This fixes the issue:

30.64%  readseek2_proce  [kernel.kallsyms]    [k] find_get_entry
            |
            --- find_get_entry
               pagecache_get_page
               generic_file_read_iter
               new_sync_read
               vfs_read
               sys_read
               syscall_exit
               __GI___libc_read

Cc: stable@vger.kernel.org # 3.16+
Reported-by: NAneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: NAnton Blanchard <anton@samba.org>

Values acquired from Open Firmware are in 32-bit big endian format
and need to be handled on little endian architectures.  This patch
ensures values are in cpu endian when hotplugging memory.
Signed-off-by: NThomas Falcon <tlfalcon@linux.vnet.ibm.com>

The atmel,clk-divisors property is taking 4 divisors, if less are
provided, the clock registration will fail.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

Actually register clocks from device tree when using the common clock
framework.
Signed-off-by: NAlexandre Belloni <alexandre.belloni@free-electrons.com>
Acked-by: NBoris Brezillon <boris.brezillon@free-electrons.com>
[nicolas.ferre@atmel.com: add at91 to function name]
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

The at91sam9g20 SOC uses its own pllb implementation which is different
from the one inherited from at91sam9260 SOC.
Signed-off-by: NGaël PORTAY <gael.portay@gmail.com>
Acked-by: NBoris Brezillon <boris.brezillon@free-electrons.com>
Signed-off-by: NNicolas Ferre <nicolas.ferre@atmel.com>

DRA7 evm REV G and later boards uses a vtt regulator for DDR3
termination and this is controlled by gpio7_11. This gpio is
configured in boot loader. gpio7_11, which is only available only on
Pad A22, in previous boards, is connected only to an unused pad on
expansion connector EXP_P3 and is safe to be muxed as GPIO on all
DRA7-evm versions (without a need to spin off another dts file).

Since gpio7_11 is used to control VTT and should not be reset or kept
in idle state during boot up else VTT will be disconnected and DDR
gets corrupted. So, as part of this change, mark gpio7 as no-reset and
no-idle on init.
Signed-off-by: NLokesh Vutla <lokeshvutla@ti.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

While auditing the various pin ctrl configurations using the following
command:
grep PIN_ arch/arm/boot/dts/dra7-evm.dts|(while read line;
do
	v=`echo "$line" | sed -e "s/\s\s*/|/g" | cut -d '|' -f1 |
		cut -d 'x' -f2|tr [a-z] [A-Z]`;
	HEX=`echo "obase=16;ibase=16;4A003400+$v"| bc`;
	echo "$HEX ===> $line";
done)
against DRA75x/74x NDA TRM revision S(SPRUHI2S August 2014),
documentation errors were found for spi1 pinctrl. Fix the same.

Fixes: 6e58b8f1 ("ARM: dts: DRA7: Add the dts files for dra7 SoC and dra7-evm board")
Signed-off-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

Both QSPI and GPMC-NAND share the same Pin (A8) from the SoC for Chip Select
functionality. So both can't be enabled simultaneously.

Disable QSPI node to prevent the pin conflict as well as
be similar to 3.12 release.

CC: Sourav Poddar <sourav.poddar@ti.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Reviewed-by: NPekon Gupta <pekon@pek-sem.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

For NAND read & write wait pin monitoring must be kept disabled as the
wait pin is only used to indicate NAND device ready status and not to
extend each read/write cycle.

So don't print a warning if wait pin is specified while read/write
monitoring is not in the device tree.

Sanity check wait pin number irrespective if read/write monitoring is
set or not.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Reviewed-by: NPekon Gupta <pekon@pek-sem.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

NAND uses wait pin only to indicate device readiness after
a block/page operation. It is not use to extend individual
read/write cycle and so read/write wait pin monitoring must
be disabled for NAND.

Add gpmc wait pin information as the NAND uses wait pin 0
for device ready indication.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Reviewed-by: NPekon Gupta <pekon@pek-sem.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

NAND uses wait pin only to indicate device readiness after
a block/page operation. It is not use to extend individual
read/write cycle and so read/write wait pin monitoring must
be disabled for NAND.

This patch also gets rid of the below warning when NAND is
accessed for the first time.

omap_l3_noc 44000000.ocp: L3 application error: target 13 mod:1 (unclearable)
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Reviewed-by: NPekon Gupta <pekon@pek-sem.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

am437x-gp-evm uses a NAND chip with page size 4096 bytes
and spare area of 225 bytes per page.

For such a setup it is preferrable to use BCH16 ECC scheme over
BCH8. This also makes it compatible with ROM code ECC scheme so
we can boot with NAND after flashing from kernel.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Reviewed-by: NPekon Gupta <pekon@pek-sem.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

am43x-epos-evm uses a NAND chip with page size 4096 bytes
and spare area of 225 bytes per page.

For such a setup it is preferrable to use BCH16 ECC scheme over
BCH8. This also makes it compatible with ROM code ECC scheme so
we can boot with NAND after flashing from kernel.
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Reviewed-by: NPekon Gupta <pekon@pek-sem.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

Size should be 64KiB instead of 92KiB.
Signed-off-by: NFelipe Balbi <balbi@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

On the GP EVM, the ambient light sensor is limited to 100KHz on the
I2C bus.

So use 100kHz for I2C on the GP EVM due to this limitation on the
ambient light sensor.
Reported-by: NAparna Balasubramanian <aparnab@ti.com>
Signed-off-by: NNishanth Menon <nm@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

The 8th NAND partition should be named "NAND.u-boot-env.backup1"
instead of "NAND.u-boot-env". This is to be consistent with other
TI boards as well as u-boot.

CC: Pekon Gupta <pekon@pek-sem.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NSekhar Nori <nsekhar@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

The I2C3 pins are taken from pads E21 (GPIO6_14) and
F20 (GPIO6_15). Use the right pinmux register and mode.

Also set the I2C3 bus frequency to a safer 400KHz than
3.4Mhz.

CC: Peter Ujfalusi <peter.ujfalusi@ti.com>
Signed-off-by: NRoger Quadros <rogerq@ti.com>
Signed-off-by: NTony Lindgren <tony@atomide.com>

The arm64 interrupt migration code on cpu offline calls
irqchip.irq_set_affinity() with the argument force=true. Originally
this argument had no effect because it was not used by any interrupt
chip driver and there was no semantics defined.

This changed with commit 01f8fa4f ("genirq: Allow forcing cpu
affinity of interrupts") which made the force argument useful to route
interrupts to not yet online cpus without checking the target cpu
against the cpu online mask. The following commit ffde1de6
("irqchip: gic: Support forced affinity setting") implemented this for
the GIC interrupt controller.

As a consequence the cpu offline irq migration fails if CPU0 is
offlined, because CPU0 is still set in the affinity mask and the
validation against cpu online mask is skipped to the force argument
being true. The following first_cpu(mask) selection always selects
CPU0 as the target.

Commit 601c9421("arm64: use cpu_online_mask when using forced
irq_set_affinity") intended to fix the above mentioned issue but
introduced another issue where affinity can be migrated to a wrong
CPU due to unconditional copy of cpu_online_mask.

As with for arm, solve the issue by calling irq_set_affinity() with
force=false from the CPU offline irq migration code so the GIC driver
validates the affinity mask against CPU online mask and therefore
removes CPU0 from the possible target candidates. Also revert the
changes done in the commit 601c9421 as it's no longer needed.

Tested on Juno platform.

Fixes: 601c9421("arm64: use cpu_online_mask when using forced
	irq_set_affinity")
Signed-off-by: NSudeep Holla <sudeep.holla@arm.com>
Acked-by: NMark Rutland <mark.rutland@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: <stable@vger.kernel.org> # 3.10.x
Signed-off-by: NWill Deacon <will.deacon@arm.com>

Structure name and variable name were erroneously interchanged
Signed-off-by: NNoam Camus <noamc@ezchip.com>
Acked-by: NVineet Gupta <vgupta@synopsys.com>
[ Also removed pointless cast from "void *".  - Linus ]
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

fc95ca72 introduces a memset in
kvmppc_alloc_hpt since the general CMA doesn't clear the memory it
allocates.

However, the size argument passed to memset is computed from a signed value
and its signed bit is extended by the cast the compiler is doing. This lead
to extremely large size value when dealing with order value >= 31, and
almost all the memory following the allocated space is cleaned. As a
consequence, the system is panicing and may even fail spawning the kdump
kernel.

This fix makes use of an unsigned value for the memset's size argument to
avoid sign extension. Among this fix, another shift operation which may
lead to signed extended value too is also fixed.

Cc: Alexey Kardashevskiy <aik@ozlabs.ru>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Alexander Graf <agraf@suse.de>
Cc: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Signed-off-by: NLaurent Dufour <ldufour@linux.vnet.ibm.com>
Signed-off-by: NPaolo Bonzini <pbonzini@redhat.com>

The quite-recently-added analog-tv-connector bindings say that the
compatible string for composite video connector is
"composite-connector". That string is also used in the omap3-n900.dts
file. However, the connector driver uses "composite-video-connector", so
this has never worked.

While changing the driver's compatible string to "composite-connector"
would be safer, as published DT bindings should not be changed, I'd
rather fix the bindings in this case for two reasons:

* composite-connector is a bit too generic name, as it doesn't even hint
  at video.
* it's clear that this has never worked, which means no one has used
  those bindings, which should make it safe to change this.
Signed-off-by: NTomi Valkeinen <tomi.valkeinen@ti.com>
Acked-by: NLaurent Pinchart <laurent.pinchart@ideasonboard.com>

Analogous to commit 8858d88a
that fixed commit 70b41abc
"ARM: ux500: move MSP pin control to the device tree"
accidentally activated MSP2, giving rise to a boot scroll
scream as the kernel attempts to probe a driver for it and
fails to obtain DMA channel 14.

For some reason I forgot to fix this on the Snowball. Fix
this up by marking the node disabled again.

Cc: Lee Jones <lee.jones@linaro.org>
Signed-off-by: NLinus Walleij <linus.walleij@linaro.org>
Tested-by: NKevin Hilman <khilman@linaro.org>
Signed-off-by: NKevin Hilman <khilman@linaro.org>

commit 0944fe3f ("s390/mm: implement software referenced bits")
triggered another paging/storage key corruption. There is an
unhandled invalid->valid pte change where we have to set the real
storage key from the pgste.
When doing paging a guest page might be swapcache or swap and when
faulted in it might be read-only and due to a parallel scan old.
An do_wp_page will make it writeable and young. Due to software
reference tracking this page was invalid and now becomes valid.
Signed-off-by: NChristian Borntraeger <borntraeger@de.ibm.com>
Acked-by: NMartin Schwidefsky <schwidefsky@de.ibm.com>
Cc: stable@vger.kernel.org # v3.12+

Since 3.12 or more precisely  commit 0944fe3f ("s390/mm:
implement software referenced bits") guest storage keys get
corrupted during paging. This commit added another valid->invalid
translation for page tables - namely ptep_test_and_clear_young.
We have to transfer the storage key into the pgste in that case.
Signed-off-by: NChristian Borntraeger <borntraeger@de.ibm.com>
Acked-by: NMartin Schwidefsky <schwidefsky@de.ibm.com>
Cc: stable@vger.kernel.org # v3.12+

It turns out that vendors are relying on the format of /proc/cpuinfo,
and we've even spotted out-of-tree hacks attempting to make it look
identical to the format used by arch/arm/. That means we can't afford to
churn this interface in mainline, so revert the recent reformatting of
the file for arm64 pending discussions on the list to find out what
people actually want.

This reverts commit d7a49086.
Acked-by: NMark Rutland <mark.rutland@arm.com>
Signed-off-by: NWill Deacon <will.deacon@arm.com>