From: Will Deacon <will@xxxxxxxxxx> If a guest performs cache maintenance on a read-only memslot, we should inform userspace rather than skip the instruction altogether. Signed-off-by: Will Deacon <will@xxxxxxxxxx> Signed-off-by: Marc Zyngier <maz@xxxxxxxxxx> Cc: Marc Zyngier <maz@xxxxxxxxxx> Cc: Quentin Perret <qperret@xxxxxxxxxx> Link: https://lore.kernel.org/r/20200729102821.23392-4-will@xxxxxxxxxx --- arch/arm64/kvm/mmu.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/arch/arm64/kvm/mmu.c b/arch/arm64/kvm/mmu.c index 14c6a9df5c9f..85b0ec9dd9ef 100644 --- a/arch/arm64/kvm/mmu.c +++ b/arch/arm64/kvm/mmu.c @@ -2121,7 +2121,7 @@ int kvm_handle_guest_abort(struct kvm_vcpu *vcpu, struct kvm_run *run) * So let's assume that the guest is just being * cautious, and skip the instruction. */ - if (kvm_vcpu_dabt_is_cm(vcpu)) { + if (kvm_is_error_hva(hva) && kvm_vcpu_dabt_is_cm(vcpu)) { kvm_skip_instr(vcpu, kvm_vcpu_trap_il_is32bit(vcpu)); ret = 1; goto out_unlock; -- 2.27.0
