Merge branch 'arm/queue' into 'master'

arm/arm64: PMU and PSCI tests

See merge request kvm-unit-tests/kvm-unit-tests!39

arm/arm64: mmu: Rename mmu_get_pte() -> follow_pte()

The function get_pte() from mmu.c returns a pointer to the PTE
associated with the requested virtual address, mapping the virtual
address in the

arm/arm64: mmu: Rename mmu_get_pte() -> follow_pte()

The function get_pte() from mmu.c returns a pointer to the PTE
associated with the requested virtual address, mapping the virtual
address in the process if it's not already mapped.

mmu_get_pte() returns a pointer to the PTE if and only if the virtual is
mapped in pgtable, otherwise returns NULL. Rename it to follow_pte() to
avoid any confusion with get_pte(). follow_pte() also matches the name
of Linux kernel function with a similar purpose.

Also remove the mmu_enabled() check from the function, as the purpose of
the function is to get the mapping for the virtual address in the pgtable
supplied as the argument, not to translate the virtual address to a
physical address using the current translation; that's what
virt_to_phys() does.

Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Andrew Jones <andrew.jones@linux.dev>

show more ...

header guards: clean up some stragglers

Some headers had been missed during the initial header guard
standardization.

Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Laurent Vivier <l

header guards: clean up some stragglers

Some headers had been missed during the initial header guard
standardization.

Signed-off-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Laurent Vivier <lvivier@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Message-Id: <20210610135937.94375-2-cohuck@redhat.com>

show more ...

Merge branch 'arm/queue' into 'master'

arm/arm64: Fixes, improvements, and prep for target-efi

See merge request kvm-unit-tests/kvm-unit-tests!6

arm/arm64: Read system registers to get the state of the MMU

When we are in EL1 we can directly tell if the local cpu's MMU is on
by reading a system register (SCTRL/SCTRL_EL1). In EL0, we use the
r

arm/arm64: Read system registers to get the state of the MMU

When we are in EL1 we can directly tell if the local cpu's MMU is on
by reading a system register (SCTRL/SCTRL_EL1). In EL0, we use the
relevant cpumask. This way we don't have to rely on the cpu id in
thread_info when we are in setup executing in EL1.

In addition, this change:
* Removes mmu_disabled_cpu_count as it is no longer necessary and
assumed that calls to mmu_mark_enabled()/mmu_mark_disabled() were
serialized. This is currently true but a future change could
easily break that assumption.
* Changes mmu_disabled_mask to mmu_enabled_mask and inverts the
logic to track in a more intuitive way that all CPUs start with
the MMU OFF and at some point, we turn them ON.

Reviewed-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
[ Applied some code cleanups suggested by Alexandru Elisei ]
Signed-off-by: Andrew Jones <drjones@redhat.com>

show more ...

Merge branch 'arm-pull-11-11-2020' into 'master'

Arm pull 11.11.2020

See merge request kvm-unit-tests/kvm-unit-tests!2

arm: Add mmu_get_pte() to the MMU API

Add the mmu_get_pte() function that allows a test to get a pointer to
the PTE for a valid virtual address. Return NULL if the MMU is off.

Signed-off-by: Nikos

arm: Add mmu_get_pte() to the MMU API

Add the mmu_get_pte() function that allows a test to get a pointer to
the PTE for a valid virtual address. Return NULL if the MMU is off.

Signed-off-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Signed-off-by: Luc Maranget <Luc.Maranget@inria.fr>
Co-Developed-by: Nikos Nikoleris <nikos.nikoleris@arm.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Andrew Jones <drjones@redhat.com>

show more ...

Merge branch 'arm/queue' of https://github.com/rhdrjones/kvm-unit-tests into HEAD

lib: arm/arm64: Teach mmu_clear_user about block mappings

kvm-unit-tests uses block mappings, so let's expand the mmu_clear_user
function to handle those as well.

Now that the function knows about

lib: arm/arm64: Teach mmu_clear_user about block mappings

kvm-unit-tests uses block mappings, so let's expand the mmu_clear_user
function to handle those as well.

Now that the function knows about block mappings, we cannot simply
assume that if an address isn't mapped we can map it as a regular page.
Change the semantics of the function to fail quite loudly if the address
isn't mapped, and shift the burden on the caller to map the address as a
page or block mapping before calling mmu_clear_user.

Also make mmu_clear_user more flexible by adding a pgtable parameter,
instead of assuming that the change always applies to the current
translation tables.

Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Andrew Jones <drjones@redhat.com>

show more ...

lib: arm/arm64: Add function to clear the PTE_USER bit

The PTE_USER bit (AP[1]) in a page entry means that lower privilege levels
(EL0, on arm64, or PL0, on arm) can read and write from that memory

lib: arm/arm64: Add function to clear the PTE_USER bit

The PTE_USER bit (AP[1]) in a page entry means that lower privilege levels
(EL0, on arm64, or PL0, on arm) can read and write from that memory
location [1][2]. On arm64, it also implies PXN (Privileged execute-never)
when is set [3]. Add a function to clear the bit which we can use when we
want to execute code from that page or the prevent access from lower
exception levels.

Make it available to arm too, in case someone needs it at some point.

[1] ARM DDI 0406C.d, Table B3-6
[2] ARM DDI 0487E.a, table D5-28
[3] ARM DDI 0487E.a, table D5-33

Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Andrew Jones <drjones@redhat.com>

show more ...

arm: enable vmalloc

Unlike x86, ARM always enables virtual memory so it always switches
early from phys_alloc to vmalloc. The changes mostly involve using
alloc_page directly whenever physical addr

arm: enable vmalloc

Unlike x86, ARM always enables virtual memory so it always switches
early from phys_alloc to vmalloc. The changes mostly involve using
alloc_page directly whenever physical addresses are needed, and of course
implementing the architecture-dependent callbacks that vmalloc needs.

The 32-bit root page table doesn't need a full page, but it is simpler
to just allocate one. With split files for 32-bit and 64-bit, we could
just place mmu_idmap in .bss.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

arm: use better-matching types for virtual and physical addresses

Make virtual addresses use uintptr_t and physical addresses
use phys_addr_t.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

arm: make pgtable.h private to mmu.c

The pgd/pmd/pud definitions are not needed outside the architecture-
dependent MMU functions. In fact, it probably would be simpler to
have completely different

arm: make pgtable.h private to mmu.c

The pgd/pmd/pud definitions are not needed outside the architecture-
dependent MMU functions. In fact, it probably would be simpler to
have completely different files for 32-bit and 64-bit ARM. However,
as a first and possibly less controversial step, hide them inside
that file.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

arm/arm64: fix mmu_enabled

We recently modified how mmu_enabled works in order to speed
spinlocks up (see c33efcf3 and b141dbac). Unfortunately c33efcf3
was a bit hasty in removing mmu_set_enabled.

arm/arm64: fix mmu_enabled

We recently modified how mmu_enabled works in order to speed
spinlocks up (see c33efcf3 and b141dbac). Unfortunately c33efcf3
was a bit hasty in removing mmu_set_enabled. I had forgotten
one of the reasons I introduced it was because secondaries
don't go through mmu_enable(), they go straight to asm_mmu_enable
from secondary_entry. This patch brings it back, albeit renamed
to mmu_mark_enabled to match mmu_mark_disabled.

The patch also moves mmu_mark_disabled from psci code to smp code,
as psci code shouldn't need to care about mmu stuff, and because
we now have a nice balance in the smp code - we mark the mmu disabled
for the secondary before we kick it, and then mark it enabled on the
init.

Finally, a bit unrelated, I've added a lengthy comment to
__mmu_enabled in order to warn about what can't be called from there.
We need to be cautious there because mmu_enabled() gets scattered
around easily, being in basic functions, and thus there's risk of
recursion.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Reported-by: Levente Kurusa <lkurusa@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

arm/arm64: speed up spinlocks and atomic ops

spinlock torture tests made it clear that checking mmu_enabled()
every time we call spin_lock is a bad idea. As most tests will
want the MMU enabled the

arm/arm64: speed up spinlocks and atomic ops

spinlock torture tests made it clear that checking mmu_enabled()
every time we call spin_lock is a bad idea. As most tests will
want the MMU enabled the entire time, then we can inline a light
weight "nobody disabled the mmu" check, and bail out early.

Adding a light weight inlined check vs. a compile-time flag
suggested by Paolo.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

arm/arm64: drop mmu_set_enabled

The mmu is enabled automatically for all cpus, they must disable it
themselves if they don't want it on. Switch from managing a cpumask
of enabled cpus to one of disa

arm/arm64: drop mmu_set_enabled

The mmu is enabled automatically for all cpus, they must disable it
themselves if they don't want it on. Switch from managing a cpumask
of enabled cpus to one of disabled cpus. This allows us to remove
the mmu_set_enabled call from secondary_cinit, and the function all
together.

Signed-off-by: Andrew Jones <drjones@redhat.com>
[Place CPUs in MMU disabled state at startup. - Paolo]
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

arm/arm64: Introduce mmu_disable

Allow unit test cpus to disable the MMU. Why not? We want the
test framework to be as flexible as possible. Callers will have
to deal with the cache coherency fallou

arm/arm64: Introduce mmu_disable

Allow unit test cpus to disable the MMU. Why not? We want the
test framework to be as flexible as possible. Callers will have
to deal with the cache coherency fallout... Cache flush support
is still forthcoming to the framework though.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...

arm/arm64: make mmu_on per cpu

We introduced a variable called mmu_on to the mmu implementation
because unit tests may want to run with the mmu off, yet still
call into common code that could lead t

arm/arm64: make mmu_on per cpu

We introduced a variable called mmu_on to the mmu implementation
because unit tests may want to run with the mmu off, yet still
call into common code that could lead to {Load,Store}-Exclusive
instructions - which is illegal. So, the mmu_on variable was added
and made query-able (through mmu_enabled()) in order to guard those
paths. But, mmu_on is really a per cpu concept, so for smp we need
to change it. As it's just a bool, we can easily make it per cpu by
changing it into a cpumask. We rename it more appropriately too.

Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>

show more ...

arm: prepare mmu code for arm64

* don't assume 1G PGDIR_SIZE or L1_CACHE_BYTES pgd alignment
* use page level descriptors for non-I/O memory
* apply new pgd/pud/pmd/pte methods
* split mmu.h to shar

arm: prepare mmu code for arm64

* don't assume 1G PGDIR_SIZE or L1_CACHE_BYTES pgd alignment
* use page level descriptors for non-I/O memory
* apply new pgd/pud/pmd/pte methods
* split mmu.h to share function declarations
* use more generic flag names in mmu.c

Signed-off-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

show more ...