A test with the command below gives these errors:
arch/arm/boot/dts/rv1108-elgin-r1.dt.yaml:
bus_intmem@10080000: $nodename:0: 'bus_intmem@10080000'
does not match '^sram(@.*)?'
arch/arm/boot/dts/rv1108-evb.dt.yaml:
bus_intmem@10080000: $nodename:0: 'bus_intmem@10080000'
does not match '^sram(@.*)?'
Fix this error by adding sram to the bus_intmem nodename
in rv1108.dtsi.
make ARCH=arm dtbs_check
DT_SCHEMA_FILES=Documentation/devicetree/bindings/sram/sram.yaml
Signed-off-by: Johan Jonker <jbx6244@gmail.com>
Link: https://lore.kernel.org/r/20200228155354.27206-1-jbx6244@gmail.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
A test with the command below gives these errors:
arch/arm/boot/dts/rv1108-elgin-r1.dt.yaml: usb@30180000:
'g-use-dma' does not match any of the regexes: 'pinctrl-[0-9]+'
arch/arm/boot/dts/rv1108-evb.dt.yaml: usb@30180000:
'g-use-dma' does not match any of the regexes: 'pinctrl-[0-9]+'
arch/arm/boot/dts/rk3228-evb.dt.yaml: usb@30040000:
'g-use-dma' does not match any of the regexes: 'pinctrl-[0-9]+'
arch/arm/boot/dts/rk3229-evb.dt.yaml: usb@30040000:
'g-use-dma' does not match any of the regexes: 'pinctrl-[0-9]+'
arch/arm/boot/dts/rk3229-xms6.dt.yaml: usb@30040000:
'g-use-dma' does not match any of the regexes: 'pinctrl-[0-9]+'
'g-use-dma' is not a valid option in dwc2.yaml, so remove it
from all Rockchip dtsi files.
make ARCH=arm dtbs_check
DT_SCHEMA_FILES=Documentation/devicetree/bindings/usb/dwc2.yaml
g-use-dma was deprecated in november 2016, see
https://patchwork.kernel.org/patch/9420553/
Signed-off-by: Johan Jonker <jbx6244@gmail.com>
Link: https://lore.kernel.org/r/20200228113922.20266-1-jbx6244@gmail.com
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
This pull request contains Broadcom ARM-based SoCs Device Tree fixes for
5.6, please pull the following:
- Stefan adds missing Device Tree properties for the Raspberry Pi 3B and
4 LEDs to have proper default configuration
- Nicolas adds an alias for the PCIe root complex node that is looked up
by the Raspberry Pi firmware for patching in specific properties
* tag 'arm-soc/for-5.6/devicetree-fixes' of https://github.com/Broadcom/stblinux:
ARM: dts: bcm2711: Add pcie0 alias
ARM: dts: bcm283x: Add missing properties to the PWR LED
Link: https://lore.kernel.org/r/20200228181144.15148-2-f.fainelli@gmail.com
Signed-off-by: Olof Johansson <olof@lixom.net>
Few fixes for omaps for v5.6-rc cycle
This series of changes contains few code fixes for issues
recently discovered:
- A build fix for ARMv6 only configs when CONFIG_HAVE_ARM_SMCCC is
not set
- A fix for ti-sysc quirk handling for 1-wire hdq reset
And a handful of dts fixes that I had queued up and should
have already sent earlier instead of waiting for the code
fixes to get sorted out:
- Fix naming of vsys_3v3 regulator for dra7-evm
- Fix incorrect OPP node names for am437x-idk-evm
- Fix IPU1 mux clock parent source for dra7
- Add missing PWM property for dra7 timers 13 to 16
- Add missing dma-ranges for dra7 PCIe nodes
- Fix mmc3 max-frequency for dra76x
* tag 'omap-for-v5.6/fixes-rc3-signed' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap:
ARM: OMAP2+: Fix compile if CONFIG_HAVE_ARM_SMCCC is not set
arm: dts: dra76x: Fix mmc3 max-frequency
ARM: dts: dra7: Add "dma-ranges" property to PCIe RC DT nodes
bus: ti-sysc: Fix 1-wire reset quirk
ARM: dts: dra7-l4: mark timer13-16 as pwm capable
ARM: dts: dra7xx-clocks: Fixup IPU1 mux clock parent source
ARM: dts: am437x-idk-evm: Fix incorrect OPP node names
ARM: dts: dra7-evm: Rename evm_3v3 regulator to vsys_3v3
Link: https://lore.kernel.org/r/pull-1582903541-589933@atomide.com
Signed-off-by: Olof Johansson <olof@lixom.net>
G3D power domain in Exynos5420 SoC is supplied from PVDD_G3DS_1V0 and
PVDD_G3D_1V0. Besides the main GPU MALI module it also contains the power
domain control logic and clocks. Turning the power supplies off causes
the power domain to fail to operate properly if GPU drivers are loaded as
modules. GPU should use PVDD_G3D_1V0 supply mainly to control the DVFS.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
G3D power domain in Exynos5422 SoC is supplied from VDD_G3D. Besides the
main GPU MALI module it also contains the power domain control logic and
clocks. Turning the VDD_G3D power supply off causes the power domain to
fail to operate properly and breaks for example system suspend/resume.
GPU should use VDD_G3D supply mainly to control the DVFS.
Fixes: 1a5a85c564 ("ARM: dts: exynos: Add Mali/GPU node on Exynos5420 and enable it on Odroid XU3/4")
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
The following commit:
c7225494b ("efi/arm: Work around missing cache maintenance in decompressor handover")
modified the EFI handover code written in assembler to work around the
missing cache maintenance of the piece of code that is executed after the
MMU and caches are turned off.
Due to the fact that this sequence incorporates a subroutine call, cleaning
that code from the cache is not a matter of simply passing the start and end of
the currently running subroutine into cache_clean_flush(), which is why
instead, the code jumps across into the cleaned copy of the image.
However, this assumes that this copy is executable, and this means we
expect EFI_LOADER_DATA regions to be executable as well, which is not
a reasonable assumption to make, even if this is true for most UEFI
implementations today.
So change this back, and add a cache_clean_flush() call to cover the
remaining code in the subroutine, and any code it may execute in the
context of cache_off().
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: linux-efi@vger.kernel.org
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: David Hildenbrand <david@redhat.com>
Cc: Heinrich Schuchardt <xypron.glpk@gmx.de>
Cc: Tom Lendacky <thomas.lendacky@amd.com>
Link: https://lore.kernel.org/r/20200228121408.9075-5-ardb@kernel.org
Commit 73f9de0c7f ("ARM: dts: uniphier: Add generic compatible string
for I2C EEPROM") did not touch this node.
Add the compatible string prefixed "atmel," so that this matches to the
OF table.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Follow the standard nodename pattern
"^(cache-controller|cpu)(@[0-9a-f,]+)*$" defined in
schemas/cache-controller.yaml of dt-schema.
Otherwise, after the dt-binding is converted to json-schema,
'make ARCH=arm dtbs_check' will show warnings like this:
l2-cache@500c0000: $nodename:0: 'l2-cache@500c0000' does not match '^(cache-controller|cpu)(@[0-9a-f,]+)*$'
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Follow the standard nodename pattern "^nand-controller(@.*)?" defined
in Documentation/devicetree/bindings/mtd/nand-controller.yaml
Otherwise, after the dt-binding is converted to json-schema,
'make ARCH=arm dtbs_check' will show warnings like this:
nand@68000000: $nodename:0: 'nand@68000000' does not match '^nand-controller(@.*)?'
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Follow the standard nodename pattern "^interrupt-controller(@[0-9a-f,]+)*$"
defined in schemas/interrupt-controller.yaml of dt-schema.
Otherwise, after the dt-binding is converted to json-schema,
make ARCH=arm dtbs_check' will show warnings like this:
aidet@5fc20000: $nodename:0: 'aidet@5fc20000' does not match '^interrupt-controller(@[0-9a-f,]+)*$'
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Follow the standard nodename pattern "^mmc(@.*)?$" defined in
Documentation/devicetree/bindings/mmc/mmc-controller.yaml
Otherwise, after the dt-binding is converted to json-schema,
'make ARCH=arm dtbs_check' will show warnings like this:
sdhc@5a000000: $nodename:0: 'sdhc@5a000000' does not match '^mmc(@.*)?$'
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Some bcm2711 revisions have different DMA constraints on the their PCIE
bus. The lower common denominator, being able to access the lower 3GB of
memory, is the default setting for now. Newer SoC revisions are able to
access the whole memory space.
Raspberry Pi 4's firmware is aware of this limitation and will correct
the PCIE's dma-ranges property if a pcie0 alias is available. So add
it.
Fixes: d5c8dc0d4c ("ARM: dts: bcm2711: Enable PCIe controller")
Signed-off-by: Nicolas Saenz Julienne <nsaenzjulienne@suse.de>
Reviewed-by: Phil Elwell <phil@raspberrypi.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
This adds the missing properties to the PWR LED for the RPi 3 & 4 boards,
which are already set for the other boards. Without them we will lose
the LED state after suspend.
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Tested-by: Peter Robinson <pbrobinson@gmail.com>
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Renesas Fixes for v5.6
- Restore R-Car M3-W support,
- Drop deprecated compatible value to ease DT binding conversion to
json-schema.
* tag 'renesas-fixes-for-v5.6-tag1' of git://git.kernel.org/pub/scm/linux/kernel/git/geert/renesas-devel:
ARM: dts: r8a7779: Remove deprecated "renesas, rcar-sata" compatible value
arm64: defconfig: Replace ARCH_R8A7796 by ARCH_R8A77960
Link: https://lore.kernel.org/r/20200226105236.18368-1-geert+renesas@glider.be
Signed-off-by: Olof Johansson <olof@lixom.net>
The OMAP4 DSP hwmod is used in DT, but the DT node is not
probing any real driver. The DSP device-tree node shall be
updated as per the new OMAP remoteproc bindings with the
underneath MMU device relying on the ti-sysc infrastructure.
Drop the legacy hwmod data for the DSP device along with the
custom ti,hwmods property. They have to be dropped together
since the early platform data init code is based on the custom
ti,hwmods property.
Signed-off-by: Suman Anna <s-anna@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Add idle_states table for CPU on am437x. Currently just add C1 state
which gates the MPU clock domain.
Signed-off-by: Dave Gerlach <d-gerlach@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Add idle_states table for CPU on am335x. Currently just add C1 state
which gates the MPU clock domain.
Signed-off-by: Dave Gerlach <d-gerlach@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
There is a second PWM unit available in the PL I/O block.
Add a node and pinmux definition for it.
Signed-off-by: Mans Rullgard <mans@mansr.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Update the v7 cache_clean_flush routine to take into account the
memory range passed in r0/r1, and perform cache maintenance by
virtual address on this range instead of set/way maintenance, which
is inappropriate for the purpose of maintaining the cached state of
memory contents.
Since this removes any use of the stack in the implementation of
cache_clean_flush(), we can also drop some code that manages the
value of the stack pointer before calling it.
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
In preparation for turning the decompressor's cache clean/flush
operations into proper by-VA maintenance for v7 cores, pass the
start and end addresses of the regions that need cache maintenance
into cache_clean_flush in registers r0 and r1.
Currently, all implementations of cache_clean_flush ignore these
values, so no functional change is expected as a result of this
patch.
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Before adding another reference to the inflated image size, factor
out the slightly complicated way of loading the unaligned little-endian
constant from the end of the compressed data.
Tested-by: Tony Lindgren <tony@atomide.com>
Tested-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
The last 8MB of physical memory on Artik520 family boards is reserved for
secure firmware. Adjust the total amount of the memory defined in
exynos3250-artik5.dtsi to match the memory available for the Linux kernel.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
dra76x is not affected by i887 which requires mmc3 node to be limited to
a max frequency of 64 MHz. Fix this by overwriting the correct value in
the the dra76 specific dtsi.
Fixes: 895bd4b3e5 ("ARM: dts: Add support for dra76-evm")
Cc: stable@vger.kernel.org
Signed-off-by: Faiz Abbas <faiz_abbas@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
'dma-ranges' in a PCI bridge node does correctly set dma masks for PCI
devices not described in the DT. Certain DRA7 platforms (e.g., DRA76)
has RAM above 32-bit boundary (accessible with LPAE config) though the
PCIe bridge will be able to access only 32-bits. Add 'dma-ranges'
property in PCIe RC DT nodes to indicate the host bridge can access
only 32 bits.
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Pull EFI updates for v5.7 from Ard Biesheuvel:
This time, the set of changes for the EFI subsystem is much larger than
usual. The main reasons are:
- Get things cleaned up before EFI support for RISC-V arrives, which will
increase the size of the validation matrix, and therefore the threshold to
making drastic changes,
- After years of defunct maintainership, the GRUB project has finally started
to consider changes from the distros regarding UEFI boot, some of which are
highly specific to the way x86 does UEFI secure boot and measured boot,
based on knowledge of both shim internals and the layout of bootparams and
the x86 setup header. Having this maintenance burden on other architectures
(which don't need shim in the first place) is hard to justify, so instead,
we are introducing a generic Linux/UEFI boot protocol.
Summary of changes:
- Boot time GDT handling changes (Arvind)
- Simplify handling of EFI properties table on arm64
- Generic EFI stub cleanups, to improve command line handling, file I/O,
memory allocation, etc.
- Introduce a generic initrd loading method based on calling back into
the firmware, instead of relying on the x86 EFI handover protocol or
device tree.
- Introduce a mixed mode boot method that does not rely on the x86 EFI
handover protocol either, and could potentially be adopted by other
architectures (if another one ever surfaces where one execution mode
is a superset of another)
- Clean up the contents of struct efi, and move out everything that
doesn't need to be stored there.
- Incorporate support for UEFI spec v2.8A changes that permit firmware
implementations to return EFI_UNSUPPORTED from UEFI runtime services at
OS runtime, and expose a mask of which ones are supported or unsupported
via a configuration table.
- Various documentation updates and minor code cleanups (Heinrich)
- Partial fix for the lack of by-VA cache maintenance in the decompressor
on 32-bit ARM. Note that these patches were deliberately put at the
beginning so they can be used as a stable branch that will be shared with
a PR containing the complete fix, which I will send to the ARM tree.
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This enables passive cooling by down-regulating CPU voltage
and frequency.
For the trip points, I used values from the BSP code directly.
The critical trip point value is 30°C above the maximum recommended
ambient temperature (70°C) for the SoC from the datasheet, so there's
some headroom even at such a high ambient temperature.
Signed-off-by: Ondrej Jirman <megous@megous.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
This enables passive cooling by down-regulating CPU voltage
and frequency.
For trip points, I used a slightly lowered values from the BSP
code. The critical temperature of 110°C from BSP code seemed
like a lot, so I rounded it off to 100°C.
The critical trip point value is 30°C above the maximum recommended
ambient temperature (70°C) for the SoC from the datasheet, so there's
some headroom even at such a high ambient temperature.
Signed-off-by: Ondrej Jirman <megous@megous.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
i.MX fixes for 5.6:
- Build v7_cpu_resume() unconditionally to fix system hang in case that
suspend is disabled but cpuidle support is enabled.
- Drop unexisting Ethernet PHY device from imx8qxp-mek board.
- Fix SRAM compatible strings on imx6dl-colibri-eval-v3 board.
- Fix imx-scu driver to make sure that all messages words are written
sequentially.
- A series from Leonard Crestez to fix i.MX SC API users, having all
messages aligned on 4 bytes.
- Fix eMMC supply for phycore-som board.
- Drop bogus frequency setting from imx7-colibri SD/MMC device, so that
HS200 mode starts working and delivers better performance.
- Fix opp-supported-hw for i.MX7D to get consumer and industrial parts
work with correct frequency settings.
- Restore MDIO compatible to the correct one for LS1021A SoC.
* tag 'imx-fixes-5.6' of git://git.kernel.org/pub/scm/linux/kernel/git/shawnguo/linux:
soc: imx-scu: Align imx sc msg structs to 4
firmware: imx: Align imx_sc_msg_req_cpu_start to 4
firmware: imx: scu-pd: Align imx sc msg structs to 4
firmware: imx: misc: Align imx sc msg structs to 4
firmware: imx: scu: Ensure sequential TX
ARM: dts: imx7-colibri: Fix frequency for sd/mmc
arm64: dts: imx8qxp-mek: Remove unexisting Ethernet PHY
ARM: dts: imx6dl-colibri-eval-v3: fix sram compatible properties
ARM: dts: ls1021a: Restore MDIO compatible to gianfar
ARM: dts: imx7d: fix opp-supported-hw
ARM: imx: build v7_cpu_resume() unconditionally
ARM: dts: imx6: phycore-som: fix emmc supply
Link: https://lore.kernel.org/r/20200224120334.GH27688@dragon
Signed-off-by: Olof Johansson <olof@lixom.net>
Fix LCD backlight issue for droid4
There was a bit of an integration glitch with the LED backlight series.
The LED related parts got merged into v5.6-rc1, but the actual backlight
driver got left out.
This caused an issue on at least droid4 where the LCD backlight can not
yet be enabled automatically. And the LCD backlight can no longer be
enabled manually either via sysfs.
The integration glitch happened because some pending comments from me.
There was some confusion on which device tree property we should use for
the default brightness property.
After discussing how to fix this on the mailing lists, we came to the
conclusion that it's best to fix this issue properly by adding the
missing driver.
The other solutions would mean backpedaling and try to come up with some
temporary solution that really does not solve the issue for users. The
patch for led_bl.c has been around for quite a while and tested by many
users and was assumed to be merged as part of the LED backlight series.
For the brightness property to use, we ended up using the more common
"default-brightness-level" rather than "default-brightness" used by
some backlight drivers.
* tag 'omap-for-v5.6/droid4-lcd-fix-signed' of git://git.kernel.org/pub/scm/linux/kernel/git/tmlind/linux-omap:
ARM: dts: droid4: Configure LED backlight for lm3532
backlight: add led-backlight driver
Link: https://lore.kernel.org/r/pull-1582303901-96693@atomide.com
Signed-off-by: Olof Johansson <olof@lixom.net>
We are missing AESS for omap5. Looks like it's similar to what we have
for omap4, and this gets ti-sysc interconnect target module driver to
detect it properly.
Note that we currently have no child device driver available for it.
Cc: H. Nikolaus Schaller <hns@goldelico.com>
Cc: Matthijs van Duin <matthijsvanduin@gmail.com>
Cc: Peter Ujfalusi <peter.ujfalusi@ti.com>
Cc: Tero Kristo <t-kristo@ti.com>
Tested-by: Nikolaus Schaller <hns@goldelico.com>
Signed-off-by: Tony Lindgren <tony@atomide.com>
USB-ID signal has a pullup on the schematic, but in reality it's not
pulled up, so add a GPIO pullup. And we also need a usb0_vbus_power-supply
for VBUS detection.
This fixes OTG mode detection and charging issues on TBS A711 tablet.
The issues came from ID pin reading 0, causing host mode to be enabled,
when it should not be, leading to DRVVBUS being enabled, which disabled
the charger.
Fixes: f2f221c781 ("ARM: dts: sun8i: a711: Enable USB OTG")
Signed-off-by: Ondrej Jirman <megous@megous.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Lowering the voltage solves the quick image degradation over time
(minutes), that was probably caused by overheating.
Signed-off-by: Ondrej Jirman <megous@megous.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Property dr_mode = "otg" is the default in sun8i-a83t.dtsi
Signed-off-by: Ondrej Jirman <megous@megous.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
What works:
- Serial console
- mmc0, mmc2 (both microSD card slots on the board)
- All buttons (gpio and lradc based)
- Power LED
- PMIC
- RTC
- USB OTG/gadgets mode
- Realtek USB WiFi
- Display backlight
- eInk display SPI NOR flash memory
Signed-off-by: Ondrej Jirman <megous@megous.com>
Signed-off-by: Maxime Ripard <maxime@cerno.tech>
Node name should be generic, use "ocotp-ctrl" instead of "ocotp"
for all i.MX6 SoCs.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Node name should be generic, use "watchdog" instead of "wdog" for
wdog nodes.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Now that we have added new ways to load the initrd or the mixed mode
kernel, we will also need a way to tell the loader about this. Add
symbolic constants for the PE/COFF major/minor version numbers (which
fortunately have always been 0x0 for all architectures), so that we
can bump them later to document the capabilities of the stub.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
The fixed regulators defined for Arndale5250 boards have no control lines,
so mark them as 'always-on' to better describe the hardware and also kill
the strange messages like 'MAIN_DC: disabling' after boot.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
According to the schematic, both eMMC and SDMMC use dedicated fixed
regulators connected directly to the DC5V and MAIN_DC rails. Remove the
GPX1-1 line assigned to the MMC regulator, because such control
connection doesn't exist. Also change its name to VDD_MMC to avoid
conflict with LDO18 output of S5M8767 PMIC.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Krzysztof Kozlowski <krzk@kernel.org>
Expose efi_entry() as the PE/COFF entrypoint directly, instead of
jumping into a wrapper that fiddles with stack buffers and other
stuff that the compiler is much better at. The only reason this
code exists is to obtain a pointer to the base of the image, but
we can get the same value from the loaded_image protocol, which
we already need for other reasons anyway.
Update the return type as well, to make it consistent with what
is required for a PE/COFF executable entrypoint.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
