It appears that while we are now able to properly hide PMU
registers from the guest when a PMU isn't available (either
because none has been configured, the host doesn't have
the PMU support compiled in, or that the HW doesn't have
one at all), we are still exposing more than we should to
userspace.

Introduce a visibility callback gating all the PMU registers,
which covers both usrespace and guest.

Signed-off-by: Marc Zyngier <maz@kernel.org>

Commit d82755b2 ("KVM: arm64: Kill off CONFIG_KVM_ARM_HOST") deletes
CONFIG_KVM_ARM_HOST option, it should use CONFIG_KVM instead.

Just remove CONFIG_KVM_ARM_HOST here.

Fixes: d82755b2 ("KVM: arm64: Kill off CONFIG_KVM_ARM_HOST")
Signed-off-by: Shannon Zhao <shannon.zhao@linux.alibaba.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/1609760324-92271-1-git-send-email-shannon.zhao@linux.alibaba.com

KVM_ARM_PMU only existed for the benefit of 32bit ARM hosts,
and makes no sense now that we are 64bit only. Get rid of it.

Signed-off-by: Marc Zyngier <maz@kernel.org>

Although not a problem right now, it flared up while working
on some other aspects of the code-base. Remove the useless
semicolon.

Signed-off-by: Marc Zyngier <maz@kernel.org>

The __init annotations on hyp_cpu_pm_{init,exit} are obviously incorrect,
and the build system shouts at you if you enable DEBUG_SECTION_MISMATCH.

Nothing really bad happens as we never execute that code outside of the
init context, but we can't label the callers as __int either, as kvm_init
isn't __init itself. Oh well.

Signed-off-by: Marc Zyngier <maz@kernel.org>
Reviewed-by: Nathan Chancellor <natechancellor@gmail.com>
Link: https://lore.kernel.org/r/20201223120854.255347-1-maz@kernel.org

dist->ready setting is pointlessly spread across the two vgic
backends, while it could be consolidated in kvm_vgic_map_resources().

Move it there, and slightly simplify the flows in both backends.

Suggested-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>

KVM_ARM_VCPU_INIT ioctl calls kvm_reset_vcpu(), which in turn resets the
PMU with a call to kvm_pmu_vcpu_reset(). The function zeroes the PMU
chained counters bitmap and stops all the counters with a perf event
attached. Because it is called before the VCPU has had the chance to run,
no perf events are in use and none are released.

kvm_arm_pmu_v3_enable(), called by kvm_vcpu_first_run_init() only if the
VCPU has been initialized, also resets the PMU. kvm_pmu_vcpu_reset() in
this case does the exact same thing as the previous call, so remove it.

Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201201150157.223625-6-alexandru.elisei@arm.com

vgic_v3_map_resources() returns -EBUSY if the VGIC isn't initialized,
update the comment to kvm_vgic_map_resources() to match what the function
does.

Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201201150157.223625-5-alexandru.elisei@arm.com

kvm_vgic_map_resources() is called when a VCPU if first run and it maps all
the VGIC MMIO regions. To prevent double-initialization, the VGIC uses the
ready variable to keep track of the state of resources and the global KVM
mutex to protect against concurrent accesses. After the lock is taken, the
variable is checked again in case another VCPU took the lock between the
current VCPU reading ready equals false and taking the lock.

The double-checked lock pattern is spread across four different functions:
in kvm_vcpu_first_run_init(), in kvm_vgic_map_resource() and in
vgic_{v2,v3}_map_resources(), which makes it hard to reason about and
introduces minor code duplication. Consolidate the checks in
kvm_vgic_map_resources(), where the lock is taken.

No functional change intended.

Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201201150157.223625-4-alexandru.elisei@arm.com

kvm_timer_enable() is called in kvm_vcpu_first_run_init() after
kvm_vgic_map_resources() if the VGIC wasn't ready. kvm_vgic_map_resources()
is the only place where kvm->arch.vgic.ready is set to true.

For a v2 VGIC, kvm_vgic_map_resources() will attempt to initialize the VGIC
and set the initialized flag.

For a v3 VGIC, kvm_vgic_map_resources() will return an error code if the
VGIC isn't already initialized.

The end result is that if we've reached kvm_timer_enable(), the VGIC is
initialzed and ready and vgic_initialized() will always be true, so remove
this check.

Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
Reviewed-by: Eric Auger <eric.auger@redhat.com>
[maz: added comment about vgic initialisation, as suggested by Eric]
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201201150157.223625-3-alexandru.elisei@arm.com

The API documentation states that general error codes are not detailed, but
errors with specific meanings are. On arm64, KVM_RUN can return error
numbers with a different meaning than what is described by POSIX or the C99
standard (as taken from man 3 errno).

Absent from the newly documented error codes is ERANGE which can be
returned when making a change to the EL2 stage 1 tables if the address is
larger than the largest supported input address. Assuming no bugs in the
implementation, that is not possible because the input addresses which are
mapped are the result of applying the macro kern_hyp_va() on kernel virtual
addresses.

CC: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201201150157.223625-2-alexandru.elisei@arm.com

Although there is nothing wrong with the current host PSCI relay
implementation, we can clean it up and remove some of the helpers
that do not improve the overall readability of the legacy PSCI 0.1
handling.

Opportunity is taken to turn the bitmap into a set of booleans,
and creative use of preprocessor macros make init and check
more concise/readable.

Suggested-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>

Move function for skipping host instruction in the host trap handler to
a header file containing analogical helpers for guests.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201208142452.87237-7-dbrazdil@google.com

Minor cleanup removing unused includes.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201208142452.87237-6-dbrazdil@google.com

Small cleanup moving declarations of hyp-exported variables to
kvm_host.h and using macros to avoid having to refer to them with
kvm_nvhe_sym() in host.

No functional change intended.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201208142452.87237-5-dbrazdil@google.com

Computing the hyp VA layout is redundant when the kernel runs in EL2 and
hyp shares its VA mappings. Make calling kvm_compute_layout()
conditional on not just CONFIG_KVM but also !is_kernel_in_hyp_mode().

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201208142452.87237-4-dbrazdil@google.com

init_hyp_physvirt_offset() computes PA from a kernel VA. Conversion to
kernel linear-map is required first but the code used kvm_ksym_ref() for
this purpose. Under VHE that is a NOP and resulted in a runtime warning.
Replace kvm_ksym_ref with lm_alias.

Reported-by: Qian Cai <qcai@redhat.com>
Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201208142452.87237-3-dbrazdil@google.com

PSCI driver exposes a struct containing the PSCI v0.1 function IDs
configured in the DT. However, the struct does not convey the
information whether these were set from DT or contain the default value
zero. This could be a problem for PSCI proxy in KVM protected mode.

Extend config passed to KVM with a bit mask with individual bits set
depending on whether the corresponding function pointer in psci_ops is
set, eg. set bit for PSCI_CPU_SUSPEND if psci_ops.cpu_suspend != NULL.

Previously config was split into multiple global variables. Put
everything into a single struct for convenience.

Reported-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201208142452.87237-2-dbrazdil@google.com

We reset the guest's view of PMCR_EL0 unconditionally, based on
the host's view of this register. It is however legal for an
implementation not to provide any PMU, resulting in an UNDEF.

The obvious fix is to skip the reset of this shadow register
when no PMU is available, sidestepping the issue entirely.
If no PMU is available, the guest is not able to request
a virtual PMU anyway, so not doing nothing is the right thing
to do!

It is unlikely that this bug can hit any HW implementation
though, as they all provide a PMU. It has been found using nested
virt with the host KVM not implementing the PMU itself.

Fixes: ab946834 ("arm64: KVM: Add access handler for PMCR register")
Reviewed-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201210083059.1277162-1-maz@kernel.org

Signed-off-by: Marc Zyngier <maz@kernel.org>

Conflict resolution gone astray results in the kernel not booting
on VHE-capable HW when VHE support is disabled. Thankfully spotted
by David.

Reported-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>

Signed-off-by: Marc Zyngier <maz@kernel.org>

With protected nVHE hyp code interception host's PSCI SMCs, the host
starts seeing new CPUs boot in EL1 instead of EL2. The kernel logic
that keeps track of the boot mode needs to be adjusted.

Add a static key enabled if KVM protected mode initialization is
successful.

When the key is enabled, is_hyp_mode_available continues to report
`true` because its users either treat it as a check whether KVM will be
/ was initialized, or whether stub HVCs can be made (eg. hibernate).

is_hyp_mode_mismatched is changed to report `false` when the key is
enabled. That's because all cores' modes matched at the point of KVM
init and KVM will not allow cores not present at init to boot. That
said, the function is never used after KVM is initialized.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-27-dbrazdil@google.com

While protected KVM is installed, start trapping all host SMCs.
For now these are simply forwarded to EL3, except PSCI
CPU_ON/CPU_SUSPEND/SYSTEM_SUSPEND which are intercepted and the
hypervisor installed on newly booted cores.

Create new constant HCR_HOST_NVHE_PROTECTED_FLAGS with the new set of HCR
flags to use while the nVHE vector is installed when the kernel was
booted with the protected flag enabled. Switch back to the default HCR
flags when switching back to the stub vector.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-26-dbrazdil@google.com

KVM by default keeps the stub vector installed and installs the nVHE
vector only briefly for init and later on demand. Change this policy
to install the vector at init and then never uninstall it if the kernel
was given the protected KVM command line parameter.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-25-dbrazdil@google.com

Add a handler of SYSTEM_SUSPEND host PSCI SMCs. The semantics are
equivalent to CPU_SUSPEND, typically called on the last online CPU.
Reuse the same entry point and boot args struct as CPU_SUSPEND.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-24-dbrazdil@google.com

Add a handler of CPU_SUSPEND host PSCI SMCs. The SMC can either enter
a sleep state indistinguishable from a WFI or a deeper sleep state that
behaves like a CPU_OFF+CPU_ON except that the core is still considered
online while asleep.

The handler saves r0,pc of the host and makes the same call to EL3 with
the hyp CPU entry point. It either returns back to the handler and then
back to the host, or wakes up into the entry point and initializes EL2
state before dropping back to EL1. No EL2 state needs to be
saved/restored for this purpose.

CPU_ON and CPU_SUSPEND are both implemented using struct psci_boot_args
to store the state upon powerup, with each CPU having separate structs
for CPU_ON and CPU_SUSPEND so that CPU_SUSPEND can operate locklessly
and so that a CPU_ON call targeting a CPU cannot interfere with
a concurrent CPU_SUSPEND call on that CPU.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-23-dbrazdil@google.com

Add a handler of the CPU_ON PSCI call from host. When invoked, it looks
up the logical CPU ID corresponding to the provided MPIDR and populates
the state struct of the target CPU with the provided x0, pc. It then
calls CPU_ON itself, with an entry point in hyp that initializes EL2
state before returning ERET to the provided PC in EL1.

There is a simple atomic lock around the boot args struct. If it is
already locked, CPU_ON will return PENDING_ON error code.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-22-dbrazdil@google.com

All nVHE hyp code is currently executed as handlers of host's HVCs. This
will change as nVHE starts intercepting host's PSCI CPU_ON SMCs. The
newly booted CPU will need to initialize EL2 state and then enter the
host. Add __host_enter function that branches into the existing
host state-restoring code after the trap handler would have returned.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-21-dbrazdil@google.com

In preparation for adding a CPU entry point in nVHE hyp code, extract
most of __do_hyp_init hypervisor initialization code into a common
helper function. This will be invoked by the entry point to install KVM
on the newly booted CPU.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-20-dbrazdil@google.com

Forward the following PSCI SMCs issued by host to EL3 as they do not
require the hypervisor's intervention. This assumes that EL3 correctly
implements the PSCI specification.

Only function IDs implemented in Linux are included.

Where both 32-bit and 64-bit variants exist, it is assumed that the host
will always use the 64-bit variant.

 * SMCs that only return information about the system
   * PSCI_VERSION        - PSCI version implemented by EL3
   * PSCI_FEATURES       - optional features supported by EL3
   * AFFINITY_INFO       - power state of core/cluster
   * MIGRATE_INFO_TYPE   - whether Trusted OS can be migrated
   * MIGRATE_INFO_UP_CPU - resident core of Trusted OS
 * operations which do not affect the hypervisor
   * MIGRATE             - migrate Trusted OS to a different core
   * SET_SUSPEND_MODE    - toggle OS-initiated mode
 * system shutdown/reset
   * SYSTEM_OFF
   * SYSTEM_RESET
   * SYSTEM_RESET2

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-19-dbrazdil@google.com

Add a host-initialized constant to KVM nVHE hyp code for converting
between EL2 linear map virtual addresses and physical addresses.
Also add `__hyp_pa` macro that performs the conversion.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-18-dbrazdil@google.com

Add a handler of PSCI SMCs in nVHE hyp code. The handler is initialized
with the version used by the host's PSCI driver and the function IDs it
was configured with. If the SMC function ID matches one of the
configured PSCI calls (for v0.1) or falls into the PSCI function ID
range (for v0.2+), the SMC is handled by the PSCI handler. For now, all
SMCs return PSCI_RET_NOT_SUPPORTED.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-17-dbrazdil@google.com

Add handler of host SMCs in KVM nVHE trap handler. Forward all SMCs to
EL3 and propagate the result back to EL1. This is done in preparation
for validating host SMCs in KVM protected mode.

The implementation assumes that firmware uses SMCCC v1.2 or older. That
means x0-x17 can be used both for arguments and results, other GPRs are
preserved.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-16-dbrazdil@google.com

When KVM starts validating host's PSCI requests, it will need to map
MPIDR back to the CPU ID. To this end, copy cpu_logical_map into nVHE
hyp memory when KVM is initialized.

Only copy the information for CPUs that are online at the point of KVM
initialization so that KVM rejects CPUs whose features were not checked
against the finalized capabilities.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-15-dbrazdil@google.com

When compiling with __KVM_NVHE_HYPERVISOR__, redefine per_cpu_offset()
to __hyp_per_cpu_offset() which looks up the base of the nVHE per-CPU
region of the given cpu and computes its offset from the
.hyp.data..percpu section.

This enables use of per_cpu_ptr() helpers in nVHE hyp code. Until now
only this_cpu_ptr() was supported by setting TPIDR_EL2.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-14-dbrazdil@google.com

Add rules for renaming the .data..ro_after_init ELF section in KVM nVHE
object files to .hyp.data..ro_after_init, linking it into the kernel
and mapping it in hyp at runtime.

The section is RW to the host, then mapped RO in hyp. The expectation is
that the host populates the variables in the section and they are never
changed by hyp afterwards.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-13-dbrazdil@google.com

MAIR_EL2 and TCR_EL2 are currently initialized from their _EL1 values.
This will not work once KVM starts intercepting PSCI ON/SUSPEND SMCs
and initializing EL2 state before EL1 state.

Obtain the EL1 values during KVM init and store them in the init params
struct. The struct will stay in memory and can be used when booting new
cores.

Take the opportunity to move copying the T0SZ value from idmap_t0sz in
KVM init rather than in .hyp.idmap.text. This avoids the need for the
idmap_t0sz symbol alias.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-12-dbrazdil@google.com

Once we start initializing KVM on newly booted cores before the rest of
the kernel, parameters to __do_hyp_init will need to be provided by EL2
rather than EL1. At that point it will not be possible to pass its three
arguments directly because PSCI_CPU_ON only supports one context
argument.

Refactor __do_hyp_init to accept its parameters in a struct. This
prepares the code for KVM booting cores as well as removes any limits on
the number of __do_hyp_init arguments.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-11-dbrazdil@google.com

KVM precomputes the hyp VA of __kvm_hyp_host_vector, essentially a
constant (minus ASLR), before passing it to __kvm_hyp_init.
Now that we have alternatives for converting kimg VA to hyp VA, replace
this with computing the constant inside __kvm_hyp_init, thus removing
the need for an argument.

Signed-off-by: David Brazdil <dbrazdil@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20201202184122.26046-10-dbrazdil@google.com