Last minute fix for MTE, making sure the pages are
flagged as MTE before they are released.
* kvm-arm64/mmu/mte:
KVM: arm64: Set the MTE tag bit before releasing the page
Signed-off-by: Marc Zyngier <maz@kernel.org>
Setting a page flag without holding a reference to the page
is living dangerously. In the tag-writing path, we drop the
reference to the page by calling kvm_release_pfn_dirty(),
and only then set the PG_mte_tagged bit.
It would be safer to do it the other way round.
Fixes: f0376edb1d ("KVM: arm64: Add ioctl to fetch/store tags in a guest")
Cc: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Steven Price <steven.price@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/87k0mjidwb.wl-maz@kernel.org
KVM/arm64 support for MTE, courtesy of Steven Price.
It allows the guest to use memory tagging, and offers
a new userspace API to save/restore the tags.
* kvm-arm64/mmu/mte:
KVM: arm64: Document MTE capability and ioctl
KVM: arm64: Add ioctl to fetch/store tags in a guest
KVM: arm64: Expose KVM_ARM_CAP_MTE
KVM: arm64: Save/restore MTE registers
KVM: arm64: Introduce MTE VM feature
arm64: mte: Sync tags for pages where PTE is untagged
Signed-off-by: Marc Zyngier <maz@kernel.org>
A new capability (KVM_CAP_ARM_MTE) identifies that the kernel supports
granting a guest access to the tags, and provides a mechanism for the
VMM to enable it.
A new ioctl (KVM_ARM_MTE_COPY_TAGS) provides a simple way for a VMM to
access the tags of a guest without having to maintain a PROT_MTE mapping
in userspace. The above capability gates access to the ioctl.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Steven Price <steven.price@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210621111716.37157-7-steven.price@arm.com
The VMM may not wish to have it's own mapping of guest memory mapped
with PROT_MTE because this causes problems if the VMM has tag checking
enabled (the guest controls the tags in physical RAM and it's unlikely
the tags are correct for the VMM).
Instead add a new ioctl which allows the VMM to easily read/write the
tags from guest memory, allowing the VMM's mapping to be non-PROT_MTE
while the VMM can still read/write the tags for the purpose of
migration.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Steven Price <steven.price@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210621111716.37157-6-steven.price@arm.com
It's now safe for the VMM to enable MTE in a guest, so expose the
capability to user space.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Steven Price <steven.price@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210621111716.37157-5-steven.price@arm.com
Define the new system registers that MTE introduces and context switch
them. The MTE feature is still hidden from the ID register as it isn't
supported in a VM yet.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Steven Price <steven.price@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210621111716.37157-4-steven.price@arm.com
Add a new VM feature 'KVM_ARM_CAP_MTE' which enables memory tagging
for a VM. This will expose the feature to the guest and automatically
tag memory pages touched by the VM as PG_mte_tagged (and clear the tag
storage) to ensure that the guest cannot see stale tags, and so that
the tags are correctly saved/restored across swap.
Actually exposing the new capability to user space happens in a later
patch.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Steven Price <steven.price@arm.com>
[maz: move VM_SHARED sampling into the critical section]
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210621111716.37157-3-steven.price@arm.com
A KVM guest could store tags in a page even if the VMM hasn't mapped
the page with PROT_MTE. So when restoring pages from swap we will
need to check to see if there are any saved tags even if !pte_tagged().
However don't check pages for which pte_access_permitted() returns false
as these will not have been swapped out.
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Steven Price <steven.price@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210621111716.37157-2-steven.price@arm.com
Selftest updates from Andrew Jones, fixing the sysgreg list
expectations by dealing with multiple configurations, such
as with or without a PMU.
* kvm-arm64/selftest/sysreg-list-fix:
KVM: arm64: Update MAINTAINERS to include selftests
KVM: arm64: selftests: get-reg-list: Split base and pmu registers
KVM: arm64: selftests: get-reg-list: Remove get-reg-list-sve
KVM: arm64: selftests: get-reg-list: Provide config selection option
KVM: arm64: selftests: get-reg-list: Prepare to run multiple configs at once
KVM: arm64: selftests: get-reg-list: Introduce vcpu configs
As the KVM/arm64 selftests are routed via the kvmarm tree,
add the relevant references to the MAINTAINERS file.
Suggested-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210622070732.zod7gaqhqo344vg6@gator
Since KVM commit 11663111cd ("KVM: arm64: Hide PMU registers from
userspace when not available") the get-reg-list* tests have been
failing with
...
... There are 74 missing registers.
The following lines are missing registers:
...
where the 74 missing registers are all PMU registers. This isn't a
bug in KVM that the selftest found, even though it's true that a
KVM userspace that wasn't setting the KVM_ARM_VCPU_PMU_V3 VCPU
flag, but still expecting the PMU registers to be in the reg-list,
would suddenly no longer have their expectations met. In that case,
the expectations were wrong, though, so that KVM userspace needs to
be fixed, and so does this selftest. The fix for this selftest is to
pull the PMU registers out of the base register sublist into their
own sublist and then create new, pmu-enabled vcpu configs which can
be tested.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210531103344.29325-6-drjones@redhat.com
Now that we can easily run the test for multiple vcpu configs, let's
merge get-reg-list and get-reg-list-sve into just get-reg-list. We
also add a final change to make it more possible to run multiple
tests, which is to fork the test, rather than directly run it. That
allows a test to fail, but subsequent tests can still run.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210531103344.29325-5-drjones@redhat.com
Add a new command line option that allows the user to select a specific
configuration, e.g. --config=sve will give the sve config. Also provide
help text and the --help/-h options.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210531103344.29325-4-drjones@redhat.com
We don't want to have to create a new binary for each vcpu config, so
prepare to run the test for multiple vcpu configs in a single binary.
We do this by factoring out the test from main() and then looping over
configs. When given '--list' we still never print more than a single
reg-list for a single vcpu config though, because it would be confusing
otherwise.
No functional change intended.
Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210531103344.29325-3-drjones@redhat.com
We already break register lists into sublists that get selected based
on vcpu config. However, since we only had two configs (vregs and sve),
we didn't structure the code very well to manage them. Restructure it
now to more cleanly handle register sublists that are dependent on the
vcpu config.
This patch has no intended functional change (except for the vcpu
config name now being prepended to all output).
Signed-off-by: Andrew Jones <drjones@redhat.com>
Reviewed-by: Ricardo Koller <ricarkol@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210531103344.29325-2-drjones@redhat.com
arm64 cache management function cleanup from Fuad Tabba,
shared with the arm64 tree.
* arm64/for-next/caches:
arm64: Rename arm64-internal cache maintenance functions
arm64: Fix cache maintenance function comments
arm64: sync_icache_aliases to take end parameter instead of size
arm64: __clean_dcache_area_pou to take end parameter instead of size
arm64: __clean_dcache_area_pop to take end parameter instead of size
arm64: __clean_dcache_area_poc to take end parameter instead of size
arm64: __flush_dcache_area to take end parameter instead of size
arm64: dcache_by_line_op to take end parameter instead of size
arm64: __inval_dcache_area to take end parameter instead of size
arm64: Fix comments to refer to correct function __flush_icache_range
arm64: Move documentation of dcache_by_line_op
arm64: assembler: remove user_alt
arm64: Downgrade flush_icache_range to invalidate
arm64: Do not enable uaccess for invalidate_icache_range
arm64: Do not enable uaccess for flush_icache_range
arm64: Apply errata to swsusp_arch_suspend_exit
arm64: assembler: add conditional cache fixups
arm64: assembler: replace `kaddr` with `addr`
Signed-off-by: Marc Zyngier <maz@kernel.org>
Fixes for the PMUv3 emulation of PMCR_EL0:
- Don't spuriously reset the cycle counter when resetting other counters
- Force PMCR_EL0 to become effective after having restored it
* kvm-arm64/pmu-fixes:
KVM: arm64: Restore PMU configuration on first run
KVM: arm64: Don't zero the cycle count register when PMCR_EL0.P is set
Restoring a guest with an active virtual PMU results in no perf
counters being instanciated on the host side. Not quite what
you'd expect from a restore.
In order to fix this, force a writeback of PMCR_EL0 on the first
run of a vcpu (using a new request so that it happens once the
vcpu has been loaded). This will in turn create all the host-side
counters that were missing.
Reported-by: Jinank Jain <jinankj@amazon.de>
Tested-by: Jinank Jain <jinankj@amazon.de>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/87wnrbylxv.wl-maz@kernel.org
Link: https://lore.kernel.org/r/b53dfcf9bbc4db7f96154b1cd5188d72b9766358.camel@amazon.de
According to ARM DDI 0487G.a, page D13-3895, setting the PMCR_EL0.P bit to
1 has the following effect:
"Reset all event counters accessible in the current Exception level, not
including PMCCNTR_EL0, to zero."
Similar behaviour is described for AArch32 on page G8-7022. Make it so.
Fixes: c01d6a1802 ("KVM: arm64: pmu: Only handle supported event counters")
Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210618105139.83795-1-alexandru.elisei@arm.com
Cache maintenance updates from Yanan Wang, moving the CMOs
down into the page-table code. This ensures that we only issue
them when actually performing a mapping rather than upfront.
* kvm-arm64/mmu/stage2-cmos:
KVM: arm64: Move guest CMOs to the fault handlers
KVM: arm64: Tweak parameters of guest cache maintenance functions
KVM: arm64: Introduce mm_ops member for structure stage2_attr_data
KVM: arm64: Introduce two cache maintenance callbacks
We currently uniformly perform CMOs of D-cache and I-cache in function
user_mem_abort before calling the fault handlers. If we get concurrent
guest faults(e.g. translation faults, permission faults) or some really
unnecessary guest faults caused by BBM, CMOs for the first vcpu are
necessary while the others later are not.
By moving CMOs to the fault handlers, we can easily identify conditions
where they are really needed and avoid the unnecessary ones. As it's a
time consuming process to perform CMOs especially when flushing a block
range, so this solution reduces much load of kvm and improve efficiency
of the stage-2 page table code.
We can imagine two specific scenarios which will gain much benefit:
1) In a normal VM startup, this solution will improve the efficiency of
handling guest page faults incurred by vCPUs, when initially populating
stage-2 page tables.
2) After live migration, the heavy workload will be resumed on the
destination VM, however all the stage-2 page tables need to be rebuilt
at the moment. So this solution will ease the performance drop during
resuming stage.
Reviewed-by: Fuad Tabba <tabba@google.com>
Signed-off-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210617105824.31752-5-wangyanan55@huawei.com
Adjust the parameter "kvm_pfn_t pfn" of __clean_dcache_guest_page
and __invalidate_icache_guest_page to "void *va", which paves the
way for converting these two guest CMO functions into callbacks in
structure kvm_pgtable_mm_ops. No functional change.
Reviewed-by: Fuad Tabba <tabba@google.com>
Signed-off-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210617105824.31752-4-wangyanan55@huawei.com
Also add a mm_ops member for structure stage2_attr_data, since we
will move I-cache maintenance for guest stage-2 to the permission
path and as a result will need mm_ops for some callbacks.
Reviewed-by: Fuad Tabba <tabba@google.com>
Signed-off-by: Yanan Wang <wangyanan55@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210617105824.31752-3-wangyanan55@huawei.com
To prepare for performing CMOs for guest stage-2 in the fault handlers
in pgtable.c, here introduce two cache maintenance callbacks in struct
kvm_pgtable_mm_ops. We also adjust the comment alignment for the
existing part but make no real content change at all.
Reviewed-by: Fuad Tabba <tabba@google.com>
Signed-off-by: Yanan Wang <wangyanan55@huawei.com>
[maz: fixed up comments and renamed callbacks]
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210617105824.31752-2-wangyanan55@huawei.com
Covers fundamental tests for debug exceptions. The guest installs and
handle its debug exceptions itself, without KVM_SET_GUEST_DEBUG.
Signed-off-by: Ricardo Koller <ricarkol@google.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210611011020.3420067-7-ricarkol@google.com
Add the infrastructure needed to enable exception handling in aarch64
selftests. The exception handling defaults to an unhandled-exception
handler which aborts the test, just like x86. These handlers can be
overridden by calling vm_install_exception_handler(vector) or
vm_install_sync_handler(vector, ec). The unhandled exception reporting
from the guest is done using the ucall type introduced in a previous
commit, UCALL_UNHANDLED.
The exception handling code is inspired on kvm-unit-tests.
Signed-off-by: Ricardo Koller <ricarkol@google.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210611011020.3420067-6-ricarkol@google.com
Move GUEST_ASSERT_EQ to a common header, kvm_util.h, for other
architectures and tests to use. Also modify __GUEST_ASSERT so it can be
reused to implement GUEST_ASSERT_EQ.
Signed-off-by: Ricardo Koller <ricarkol@google.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210611011020.3420067-5-ricarkol@google.com
x86, the only arch implementing exception handling, reports unhandled
vectors using port IO at a specific port number. This replicates what
ucall already does.
Introduce a new ucall type, UCALL_UNHANDLED, for guests to report
unhandled exceptions. Then replace the x86 unhandled vector exception
reporting to use it instead of port IO. This new ucall type will be
used in the next commits by arm64 to report unhandled vectors as well.
Tested: Forcing a page fault in the ./x86_64/xapic_ipi_test
halter_guest_code() shows this:
$ ./x86_64/xapic_ipi_test
...
Unexpected vectored event in guest (vector:0xe)
Signed-off-by: Ricardo Koller <ricarkol@google.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210611011020.3420067-4-ricarkol@google.com
The guest in sync_regs_test does raw ucalls by directly accessing the
ucall IO port. It makes these ucalls without setting %rdi to a `struct
ucall`, which is what a ucall uses to pass messages. The issue is that
if the host did a get_ucall (the receiver side), it would try to access
the `struct ucall` at %rdi=0 which would lead to an error ("No mapping
for vm virtual address, gva: 0x0").
This issue is currently benign as there is no get_ucall in
sync_regs_test; however, that will change in the next commit as it
changes the unhandled exception reporting mechanism to use ucalls. In
that case, every vcpu_run is followed by a get_ucall to check if the
guest is trying to report an unhandled exception.
Fix this in advance by setting %rdi to a UCALL_NONE struct ucall for the
sync_regs_test guest.
Tested with gcc-[8,9,10], and clang-[9,11].
Signed-off-by: Ricardo Koller <ricarkol@google.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210611011020.3420067-3-ricarkol@google.com
Rename the vm_handle_exception function to a name that indicates more
clearly that it installs something: vm_install_exception_handler.
Reported-by: kernel test robot <oliver.sang@intel.com>
Suggested-by: Marc Zyngier <maz@kernel.org>
Suggested-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Ricardo Koller <ricarkol@google.com>
Reviewed-by: Andrew Jones <drjones@redhat.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210611011020.3420067-2-ricarkol@google.com
Host stage-2 optimisations from Quentin Perret
* kvm-arm64/mmu/reduce-vmemmap-overhead:
KVM: arm64: Use less bits for hyp_page refcount
KVM: arm64: Use less bits for hyp_page order
KVM: arm64: Remove hyp_pool pointer from struct hyp_page
KVM: arm64: Unify MMIO and mem host stage-2 pools
KVM: arm64: Remove list_head from hyp_page
KVM: arm64: Use refcount at hyp to check page availability
KVM: arm64: Move hyp_pool locking out of refcount helpers
The hyp_page refcount is currently encoded on 4 bytes even though we
never need to count that many objects in a page. Make it 2 bytes to save
some space in the vmemmap.
As overflows are more likely to happen as well, make sure to catch those
with a BUG in the increment function.
Signed-off-by: Quentin Perret <qperret@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210608114518.748712-8-qperret@google.com
The hyp_page order is currently encoded on 4 bytes even though it is
guaranteed to be smaller than this. Make it 2 bytes to reduce the hyp
vmemmap overhead.
Signed-off-by: Quentin Perret <qperret@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210608114518.748712-7-qperret@google.com
Each struct hyp_page currently contains a pointer to a hyp_pool struct
where the page should be freed if its refcount reaches 0. However, this
information can always be inferred from the context in the EL2 code, so
drop the pointer to save a few bytes in the vmemmap.
Signed-off-by: Quentin Perret <qperret@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210608114518.748712-6-qperret@google.com
We currently maintain two separate memory pools for the host stage-2,
one for pages used in the page-table when mapping memory regions, and
the other to map MMIO regions. The former is large enough to map all of
memory with page granularity and the latter can cover an arbitrary
portion of IPA space, but allows to 'recycle' pages.
However, this split makes accounting difficult to manage as pages at
intermediate levels of the page-table may be used to map both memory and
MMIO regions. Simplify the scheme by merging both pools into one. This
means we can now hit the -ENOMEM case in the memory abort path, but
we're still guaranteed forward-progress in the worst case by unmapping
MMIO regions. On the plus side this also means we can usually map a lot
more MMIO space at once if memory ranges happen to be mapped with block
mappings.
Signed-off-by: Quentin Perret <qperret@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210608114518.748712-5-qperret@google.com
The list_head member of struct hyp_page is only needed when the page is
attached to a free-list, which by definition implies the page is free.
As such, nothing prevents us from using the page itself to store the
list_head, hence reducing the size of the vmemmap.
Signed-off-by: Quentin Perret <qperret@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210608114518.748712-4-qperret@google.com
The hyp buddy allocator currently checks the struct hyp_page list node
to see if a page is available for allocation or not when trying to
coalesce memory. Now that decrementing the refcount and attaching to
the buddy tree is done in the same critical section, we can rely on the
refcount of the buddy page to be in sync, which allows to replace the
list node check by a refcount check. This will ease removing the list
node from struct hyp_page later on.
Signed-off-by: Quentin Perret <qperret@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210608114518.748712-3-qperret@google.com
The hyp_page refcount helpers currently rely on the hyp_pool lock for
serialization. However, this means the refcounts can't be changed from
the buddy allocator core as it already holds the lock, which means pages
have to go through odd transient states.
For example, when a page is freed, its refcount is set to 0, and the
lock is transiently released before the page can be attached to a free
list in the buddy tree. This is currently harmless as the allocator
checks the list node of each page to see if it is available for
allocation or not, but it means the page refcount can't be trusted to
represent the state of the page even if the pool lock is held.
In order to fix this, remove the pool locking from the refcount helpers,
and move all the logic to the buddy allocator. This will simplify the
removal of the list node from struct hyp_page in a later patch.
Signed-off-by: Quentin Perret <qperret@google.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210608114518.748712-2-qperret@google.com
MMIO block mapping support from Keqian Zhu, allowing larger
(and lazy) mappings for devices assigned to guests.
* kvm-arm64/mmu/MMIO-block-mapping:
KVM: arm64: Try stage2 block mapping for host device MMIO
KVM: arm64: Remove the creation time's mapping of MMIO regions
Rework the KVM GIC and timer to cope with lesser HW such as
the Apple M1 SoC.
* kvm-arm64/m1:
irqchip/apple-aic: Advertise some level of vGICv3 compatibility
KVM: arm64: timer: Add support for SW-based deactivation
KVM: arm64: timer: Refactor IRQ configuration
KVM: arm64: vgic: Implement SW-driven deactivation
KVM: arm64: vgic: move irq->get_input_level into an ops structure
KVM: arm64: vgic: Let an interrupt controller advertise lack of HW deactivation
KVM: arm64: vgic: Be tolerant to the lack of maintenance interrupt masking
KVM: arm64: Handle physical FIQ as an IRQ while running a guest
irqchip/gic: Split vGIC probing information from the GIC code
The MMIO region of a device maybe huge (GB level), try to use
block mapping in stage2 to speedup both map and unmap.
Compared to normal memory mapping, we should consider two more
points when try block mapping for MMIO region:
1. For normal memory mapping, the PA(host physical address) and
HVA have same alignment within PUD_SIZE or PMD_SIZE when we use
the HVA to request hugepage, so we don't need to consider PA
alignment when verifing block mapping. But for device memory
mapping, the PA and HVA may have different alignment.
2. For normal memory mapping, we are sure hugepage size properly
fit into vma, so we don't check whether the mapping size exceeds
the boundary of vma. But for device memory mapping, we should pay
attention to this.
This adds get_vma_page_shift() to get page shift for both normal
memory and device MMIO region, and check these two points when
selecting block mapping size for MMIO region.
Signed-off-by: Keqian Zhu <zhukeqian1@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210507110322.23348-3-zhukeqian1@huawei.com
The MMIO regions may be unmapped for many reasons and can be remapped
by stage2 fault path. Map MMIO regions at creation time becomes a
minor optimization and makes these two mapping path hard to sync.
Remove the mapping code while keep the useful sanity check.
Signed-off-by: Keqian Zhu <zhukeqian1@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210507110322.23348-2-zhukeqian1@huawei.com
The CPUs in the Apple M1 SoC partially implement a virtual GICv3
CPU interface, although one that is incapable of HW deactivation
of interrupts, nor masking the maintenance interrupt.
Advertise the support to KVM.
Signed-off-by: Marc Zyngier <maz@kernel.org>
In order to deal with the lack of active state, we need to use
the mask/unmask primitives (after all, the active state is just an
additional mask on top of the normal one).
To avoid adding a bunch of ugly conditionals in the timer and vgic
code, let's use a timer-specific irqdomain to deal with the state
conversion. Yes, this is an unexpected use of irqdomains, but
there is no reason not to be just as creative as the designers
of the HW...
This involves overloading the vcpu_affinity, set_irqchip_state
and eoi callbacks so that the rest of the KVM code can continue
ignoring the oddities of the underlying platform.
Signed-off-by: Marc Zyngier <maz@kernel.org>
As we are about to add some more things to the timer IRQ
configuration, move this code out of the main timer init code
into its own set of functions.
No functional changes.
Signed-off-by: Marc Zyngier <maz@kernel.org>
In order to deal with these systems that do not offer HW-based
deactivation of interrupts, let implement a SW-based approach:
- When the irq is queued into a LR, treat it as a pure virtual
interrupt and set the EOI flag in the LR.
- When the interrupt state is read back from the LR, force a
deactivation when the state is invalid (neither active nor
pending)
Interrupts requiring such treatment get the VGIC_SW_RESAMPLE flag.
Signed-off-by: Marc Zyngier <maz@kernel.org>
We already have the option to attach a callback to an interrupt
to retrieve its pending state. As we are planning to expand this
facility, move this callback into its own data structure.
This will limit the size of individual interrupts as the ops
structures can be shared across multiple interrupts.
Signed-off-by: Marc Zyngier <maz@kernel.org>
The vGIC, as architected by ARM, allows a virtual interrupt to
trigger the deactivation of a physical interrupt. This allows
the following interrupt to be delivered without requiring an exit.
However, some implementations have choosen not to implement this,
meaning that we will need some unsavoury workarounds to deal with this.
On detecting such a case, taint the kernel and spit a nastygram.
We'll deal with this in later patches.
Signed-off-by: Marc Zyngier <maz@kernel.org>
