With VHE different exception levels are used between the host (EL2) and
guest (EL1) with a shared exception level for userpace (EL0). We can take
advantage of this and use the PMU's exception level filtering to avoid
enabling/disabling counters in the world-switch code. Instead we just
modify the counter type to include or exclude EL0 at vcpu_{load,put} time.
We also ensure that trapped PMU system register writes do not re-enable
EL0 when reconfiguring the backing perf events.
This approach completely avoids blackout windows seen with !VHE.
Suggested-by: Christoffer Dall <christoffer.dall@arm.com>
Signed-off-by: Andrew Murray <andrew.murray@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
The virt/arm core allocates a kvm_cpu_context_t percpu, at present this is
a typedef to kvm_cpu_context and is used to store host cpu context. The
kvm_cpu_context structure is also used elsewhere to hold vcpu context.
In order to use the percpu to hold additional future host information we
encapsulate kvm_cpu_context in a new structure and rename the typedef and
percpu to match.
Signed-off-by: Andrew Murray <andrew.murray@arm.com>
Reviewed-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
When pointer authentication is supported, a guest may wish to use it.
This patch adds the necessary KVM infrastructure for this to work, with
a semi-lazy context switch of the pointer auth state.
Pointer authentication feature is only enabled when VHE is built
in the kernel and present in the CPU implementation so only VHE code
paths are modified.
When we schedule a vcpu, we disable guest usage of pointer
authentication instructions and accesses to the keys. While these are
disabled, we avoid context-switching the keys. When we trap the guest
trying to use pointer authentication functionality, we change to eagerly
context-switching the keys, and enable the feature. The next time the
vcpu is scheduled out/in, we start again. However the host key save is
optimized and implemented inside ptrauth instruction/register access
trap.
Pointer authentication consists of address authentication and generic
authentication, and CPUs in a system might have varied support for
either. Where support for either feature is not uniform, it is hidden
from guests via ID register emulation, as a result of the cpufeature
framework in the host.
Unfortunately, address authentication and generic authentication cannot
be trapped separately, as the architecture provides a single EL2 trap
covering both. If we wish to expose one without the other, we cannot
prevent a (badly-written) guest from intermittently using a feature
which is not uniformly supported (when scheduled on a physical CPU which
supports the relevant feature). Hence, this patch expects both type of
authentication to be present in a cpu.
This switch of key is done from guest enter/exit assembly as preparation
for the upcoming in-kernel pointer authentication support. Hence, these
key switching routines are not implemented in C code as they may cause
pointer authentication key signing error in some situations.
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
[Only VHE, key switch in full assembly, vcpu_has_ptrauth checks
, save host key in ptrauth exception trap]
Signed-off-by: Amit Daniel Kachhap <amit.kachhap@arm.com>
Reviewed-by: Julien Thierry <julien.thierry@arm.com>
Cc: Christoffer Dall <christoffer.dall@arm.com>
Cc: kvmarm@lists.cs.columbia.edu
[maz: various fixups]
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Pull syscall numbering updates from Arnd Bergmann:
"arch: add pidfd and io_uring syscalls everywhere
This comes a bit late, but should be in 5.1 anyway: we want the newly
added system calls to be synchronized across all architectures in the
release.
I hope that in the future, any newly added system calls can be added
to all architectures at the same time, and tested there while they are
in linux-next, avoiding dependencies between the architecture
maintainer trees and the tree that contains the new system call"
* tag 'syscalls-5.1' of git://git.kernel.org/pub/scm/linux/kernel/git/arnd/asm-generic:
arch: add pidfd and io_uring syscalls everywhere
A few architectures use <asm/segment.h> internally, but nothing in
common code does. Remove all the empty or almost empty versions of it,
including the asm-generic one.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Enable following drivers for merged devices:
- Batteries with BQ27XXX chips for Minnie boards.
- Elan eKTH I2C touchscreen for Minnie boards.
- GPIO charger for all Veyron boards.
- Rockchip SARADC driver for all rk3288 boards.
- Rockchip eFUSE driver for all rk3288 boards.
- TPM security chip for all Veyron boards.
- ChromeOS EC userspace interface for all chromebooks boards.
- ChromeOS EC light and proximity sensors for some chromebooks.
Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The call to of_get_next_child returns a node pointer with refcount
incremented thus it must be explicitly decremented after the last
usage.
Detected by coccinelle with the following warnings:
./arch/arm/mach-rockchip/platsmp.c:250:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 241, but without a corresponding object release within this function.
./arch/arm/mach-rockchip/platsmp.c:260:2-8: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 241, but without a corresponding object release within this function.
./arch/arm/mach-rockchip/platsmp.c:263:1-7: ERROR: missing of_node_put; acquired a node pointer with refcount incremented on line 241, but without a corresponding object release within this function.
Signed-off-by: Wen Yang <wen.yang99@zte.com.cn>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Heiko Stuebner <heiko@sntech.de>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-rockchip@lists.infradead.org
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
The EFI stub is entered with the caches and MMU enabled by the
firmware, and once the stub is ready to hand over to the decompressor,
we clean and disable the caches.
The cache clean routines use CP15 barrier instructions, which can be
disabled via SCTLR. Normally, when using the provided cache handling
routines to enable the caches and MMU, this bit is enabled as well.
However, but since we entered the stub with the caches already enabled,
this routine is not executed before we call the cache clean routines,
resulting in undefined instruction exceptions if the firmware never
enabled this bit.
So set the bit explicitly in the EFI entry code, but do so in a way that
guarantees that the resulting code can still run on v6 cores as well
(which are guaranteed to have CP15 barriers enabled)
Cc: <stable@vger.kernel.org> # v4.9+
Acked-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
When CONFIG_ARM_MPU is not defined, the base address of v7M SCB register
is not initialized with correct value. This prevents enabling I/D caches
when the L1 cache poilcy is applied in kernel.
Fixes: 3c24121039 ("ARM: 8756/1: NOMMU: Postpone MPU activation till __after_proc_init")
Signed-off-by: Tigran Tadevosyan <tigran.tadevosyan@arm.com>
Signed-off-by: Vladimir Murzin <vladimir.murzin@arm.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Naresh Kamboju recently reported that the function-graph tracer crashes
on ARM. The function-graph tracer assumes that the kernel is built with
frame pointers.
We explicitly disabled the function-graph tracer when building Thumb2,
since the Thumb2 ABI doesn't have frame pointers.
We recently changed the way the unwinder method was selected, which
seems to have made it more likely that we can end up with the function-
graph tracer enabled but without the kernel built with frame pointers.
Fix up the function graph tracer dependencies so the option is not
available when we have no possibility of having frame pointers, and
adjust the dependencies on the unwinder option to hide the non-frame
pointer unwinder options if the function-graph tracer is enabled.
Reviewed-by: Masami Hiramatsu <mhiramat@kernel.org>
Tested-by: Masami Hiramatsu <mhiramat@kernel.org>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
We use $(LD) to link vmlinux, modules, decompressors, etc.
VDSO is the only exceptional case where $(CC) is used as the linker
driver, but I do not know why we need to do so. VDSO uses a special
linker script, and does not link standard libraries at all.
I changed the Makefile to use $(LD) rather than $(CC). I confirmed
the same vdso.so.raw was still produced.
Users will be able to use their favorite linker (e.g. lld instead of
of bfd) by passing LD= from the command line.
My plan is to rewrite all VDSO Makefiles to use $(LD), then delete
cc-ldoption.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Remove the -mno-warn-deprecated assembler flag to make sure the GNU
assembler warns in case non-unified syntax is used.
Signed-off-by: Stefan Agner <stefan@agner.ch>
Acked-by: Nicolas Pitre <nico@linaro.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Currently LLVM's integrated assembler does not recognize .w form
of the pld instructions (LLVM Bug 40972 [0]):
./arch/arm/include/asm/processor.h:133:5: error: invalid instruction
"pldw.wt%a0 n"
^
<inline asm>:2:1: note: instantiated into assembly here
pldw.w [r0]
^
1 error generated.
The W macro for generating wide instructions when targeting Thumb-2
is not strictly required for the preload data instructions (pld, pldw)
since they are only available as wide instructions. The GNU assembler
works with or without the .w appended when compiling an Thumb-2 kernel.
Drop the macro to work around LLVM Bug 40972 issue.
[0] https://bugs.llvm.org/show_bug.cgi?id=40972
Signed-off-by: Stefan Agner <stefan@agner.ch>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Convert the conditional infix to a postfix to make sure this inline
assembly is unified syntax. Since gcc assumes non-unified syntax
when emitting ARM instructions, make sure to define the syntax as
unified.
This allows to use LLVM's integrated assembler.
Additionally, for GCC ".syntax unified" for inline assembly.
When compiling non-Thumb2 GCC always emits a ".syntax divided"
at the beginning of the inline assembly which makes the
assembler fail. Since GCC 5 there is the -masm-syntax-unified
GCC option which make GCC assume unified syntax asm and hence
emits ".syntax unified" even in ARM mode. However, the option
is broken since GCC version 6 (see GCC PR88648 [1]). Work
around by adding ".syntax unified" as part of the inline
assembly.
[0] https://gcc.gnu.org/onlinedocs/gcc/ARM-Options.html#index-masm-syntax-unified
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88648
Signed-off-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
Unified assembly syntax requires conditionals to be postfixes.
TUSER() currently only takes a single argument which then gets
appended t (with translation) on every instruction.
This fixes a build error when using LLVM's integrated assembler:
In file included from kernel/futex.c:72:
./arch/arm/include/asm/futex.h:116:3: error: invalid instruction, did you mean: strt?
"2: " TUSER(streq) " %3, [%4]n"
^
<inline asm>:5:4: note: instantiated into assembly here
2: streqt r2, [r4]
^~~~~~
Additionally, for GCC ".syntax unified" for inline assembly.
When compiling non-Thumb2 GCC always emits a ".syntax divided"
at the beginning of the inline assembly which makes the
assembler fail. Since GCC 5 there is the -masm-syntax-unified
GCC option which make GCC assume unified syntax asm and hence
emits ".syntax unified" even in ARM mode. However, the option
is broken since GCC version 6 (see GCC PR88648 [1]). Work
around by adding ".syntax unified" as part of the inline
assembly.
[0] https://gcc.gnu.org/onlinedocs/gcc/ARM-Options.html#index-masm-syntax-unified
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88648
Signed-off-by: Stefan Agner <stefan@agner.ch>
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk>
The AHB queue manager and Network Processing Engines are
present on all IXP4xx SoCs, so we add them to the overarching
device tree include.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Instead of using hardcoded base address implicitly
obtained through <linux/io.h>, pass the physical base
for the QMGR block as a memory resource and remap
it in the driver.
Also pass the two IRQs as resources and obtain them
in the driver.
Use devm_* accessors and simplify the error path in the
process. Drop memory region request as this is done by
the devm_ioremap* functions.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Instead of using hardcoded base addresses implicitly
obtained through <linux/io.h>, pass the physical base
for the three NPE blocks as memory resources and remap
these in the driver.
Drop the memory request region business, this will
anyways be done by devm_* remapping functions.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Instead of registering everything related to the QMGR
unconditionally in the module_init() call (which will
never work with multiplatform) create a platform device
and probe the QMGR like any other device.
Put the device second in the list of devices added for
the platform so it is there when the dependent network
and crypto drivers probe later on.
This probe() path will not be taken unconditionally on
device tree boots, so remove the DT guard.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Instead of registering everything related to the NPE
unconditionally in the module_init() call (which will
never work with multiplatform) create a platform device
and probe the NPE like any other device.
Put the device first in the list of devices added for
the platform so it is there when the dependent network
and crypto drivers probe later on.
This probe() path will not be taken unconditionally on
device tree boots, so remove the DT guard.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
This moves the IXP4xx Queue Manager and Network Processing
Engine headers out of the <mack/*> include path as that is
incompatible with multiplatform.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
The Network Processing Engine and Queue Manager are
versatile firmware components used by several IXP4xx
drivers.
Drivers are relying on getting access to these components
using <mach/*> headers which does not work with
multiplatform. We need to find a better place for the
drivers to live.
Let's first move them to drivers/soc and the start to
refactor a bit by passing resources and moving headers.
This patch introduce static IRQ assignments but that
will be fixed by later patches in this series.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
This adds a device tree for the IXP4xx-based Linksys
NSLU2 and Gateworks GW2358 which encompass the Gateworks
Cambria family.
These will be the first IXP4xx device tree platforms.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
This adds a minimal support for booting IXP4xx systems
from device tree.
We have to add hacks to the QMGR, NPE and notably also
ethernet and watchdog drivers so that they don't crash
the platform: these drivers are unconditionally starting
to grab regions of statically remapped IO space with no
concern of the device model or other platforms.
We will go in and properly fix these drivers as we go
along but for now this hack gets us to a place where we
can start working on proper device tree support for these
platforms.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Pull in v5.1-rc6 to resolve two conflicts. One is in BFQ, in just a
comment, and is trivial. The other one is a conflict due to a later fix
in the bio multi-page work, and needs a bit more care.
* tag 'v5.1-rc6': (770 commits)
Linux 5.1-rc6
block: make sure that bvec length can't be overflow
block: kill all_q_node in request_queue
x86/cpu/intel: Lower the "ENERGY_PERF_BIAS: Set to normal" message's log priority
coredump: fix race condition between mmget_not_zero()/get_task_mm() and core dumping
mm/kmemleak.c: fix unused-function warning
init: initialize jump labels before command line option parsing
kernel/watchdog_hld.c: hard lockup message should end with a newline
kcov: improve CONFIG_ARCH_HAS_KCOV help text
mm: fix inactive list balancing between NUMA nodes and cgroups
mm/hotplug: treat CMA pages as unmovable
proc: fixup proc-pid-vm test
proc: fix map_files test on F29
mm/vmstat.c: fix /proc/vmstat format for CONFIG_DEBUG_TLBFLUSH=y CONFIG_SMP=n
mm/memory_hotplug: do not unlock after failing to take the device_hotplug_lock
mm: swapoff: shmem_unuse() stop eviction without igrab()
mm: swapoff: take notice of completion sooner
mm: swapoff: remove too limiting SWAP_UNUSE_MAX_TRIES
mm: swapoff: shmem_find_swap_entries() filter out other types
slab: store tagged freelist for off-slab slabmgmt
...
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Running reboot command on the TQMa7 board would just hang infinite
at the end of the system shutdown process.
Handling of i.MX7 errata e10574:
Watchdog: A watchdog timeout or software trigger will not reset the SOC.
Moved pinctrl from common mba7 to common tqma7 dtsi as it improves
readability of errata handling. Most integrators of this SoM will
likely use the development board as inspiration for handling this
SoC issue.
Signed-off-by: Bruno Thomsen <bruno.thomsen@gmail.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
Kirkwood has always had the ability to retrieve the local-mac-address
from the hardware (usually this was configured by the bootloader). This
is particularly useful when dealing with a legacy non-DT aware
bootloader.
The "error" message just indicated that the board used an old bootloader
and in many cases users can't do anything about this. The message
probably should have been pr_info() to inform the user that the kernel
has been helpful but rather than than let's remove it entirely to make
the kernel less noisy.
Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Signed-off-by: Gregory CLEMENT <gregory.clement@bootlin.com>
The first interrupt is for the regular watchdog timeout. Normally the
RSTOUT line will trigger a reset before this interrupt fires but on
systems with a non-standard reset it may still trigger.
The second interrupt is for a timer1 which is used as a pre-timeout for
the watchdog.
Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Signed-off-by: Gregory CLEMENT <gregory.clement@bootlin.com>
Each eTSEC MAC has its own TBI (SGMII) PCS and private MDIO bus.
But due to a DTS oversight, both SGMII-compatible MACs of the LS1021 SoC
are pointing towards the same internal PCS. Therefore nobody is
controlling the internal PCS of eTSEC0.
Upon initial ndo_open, the SGMII link is ok by virtue of U-boot
initialization. But upon an ifdown/ifup sequence, the code path from
ndo_open -> init_phy -> gfar_configure_serdes does not get executed for
the PCS of eTSEC0 (and is executed twice for MAC eTSEC1). So the SGMII
link remains down for eTSEC0. On the LS1021A-TWR board, to signal this
failure condition, the PHY driver keeps printing
'803x_aneg_done: SGMII link is not ok'.
Also, it changes compatible of mdio0 to "fsl,etsec2-mdio" to match
mdio1 device.
Fixes: 055223d4d2 ("ARM: dts: ls1021a: Enable the eTSEC ports on QDS and TWR")
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Reviewed-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Acked-by: Li Yang <leoyang.li@nxp.com>
Signed-off-by: Shawn Guo <shawnguo@kernel.org>
All 4 SPI controllers on NXP LPC32xx SoC support SPI slaves discerning them
by one cell address value, set it as default to avoid duplication in board
device tree files.
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
NXP LPC3220 and LPC3230 SoCs do NOT contain a MAC controller, so,
since for now there is just one dtsi file for all variants of
NXP LPC32xx SoCs, it is reasonable to disable the controller
by default and enable it in device tree files of particular boards.
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
The I2S controllers found on NXP LPC32xx SoCs are not yet in
use by any boards supported in upstream, disable the controllers
by default.
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
This is a non-functional change, all inconsistent hexadecimal values
found in the file are now fixed.
Taking a chance to interfere into some non-functional change I add
my copyright notice for work done during the last few years.
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
Both controllers are described in lpc32xx.dtsi and there is no any
specific platform data added in the platform file.
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
While the majority of platform data was moved to device tree description
the list of included header files remained untouched, the change cleans
it up to an irreducible and observable subset.
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
While the UDA1380 is described in some lpc3250 device trees, there is
currently no real user of that codec. Anyway, if the codec needs a clock,
it should take it explicitly.
lpc3250_machine_init is called for all the lpc32xx machines and some are
using test1_clk (for example to strobe an HW watchdog). Overwriting
TEST_CLK_SEL prevents booting those platforms.
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Tested-by: Gregory CLEMENT <gregory.clement@bootlin.com>
Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
This localizes the <mach/irqs.h> header to the mach-ixp4xx
directory, removes NR_IRQS and switches IXP4xx over to using
SPARSE_IRQ.
This is a prerequisite for DT support.
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
