shinys000114/linux
1
0
Fork 0
mirror of https://github.com/hardkernel/linux.git synced 2026-03-25 12:00:22 +09:00
Code Issues Packages Projects Releases Wiki Activity

ANDROID: KVM: arm64: Do not allow memslot modifications once a PVM has run

Browse Source 
Currently trying to move or delete a memslot results in a warning
and a failure. Userspace shouldn't be able to trigger kernel
warnings.

The cause is that in protected mode, stage-2 is managed by hyp.
Modifying a memslot flushes the shadow memslot, which tries to
unmap any stage-2 mapped pages.

Bug: 226890762
Signed-off-by: Fuad Tabba <tabba@google.com>
Change-Id: Icc6a0aada76e8492285cd5509bad1ee57700af7c
This commit is contained in:
Fuad Tabba
2022-04-07 16:12:41 +01:00
parent 8be6e93244
commit 0a227f89cf
1 changed files with 11 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

11
arch/arm64/kvm/mmu.c
View File

@@ -1857,6 +1857,13 @@ int kvm_arch_prepare_memory_region(struct kvm *kvm,
change != KVM_MR_FLAGS_ONLY)
return 0;
/* In protected mode, cannot modify memslots once a VM has run. */
if (is_protected_kvm_enabled() &&
(change == KVM_MR_DELETE || change == KVM_MR_MOVE) &&
kvm->arch.pkvm.shadow_handle) {
return -EPERM;
}
/*
* Prevent userspace from creating a memory region outside of the IPA
* space addressable by the KVM guest IPA space.
@@ -1926,6 +1933,10 @@ void kvm_arch_flush_shadow_memslot(struct kvm *kvm,
gpa_t gpa = slot->base_gfn << PAGE_SHIFT;
phys_addr_t size = slot->npages << PAGE_SHIFT;
/* Stage-2 is managed by hyp in protected mode. */
if (is_protected_kvm_enabled())
return;
spin_lock(&kvm->mmu_lock);
unmap_stage2_range(&kvm->arch.mmu, gpa, size);
spin_unlock(&kvm->mmu_lock);