On Wed, Nov 17, 2021 at 1:20 AM Hou Wenlong
<houwenlong93@xxxxxxxxxxxxxxxxx> wrote:
>
> Since tlb flush has been done for legacy MMU before
> kvm_tdp_mmu_zap_collapsible_sptes(), so the parameter flush
> should be false for kvm_tdp_mmu_zap_collapsible_sptes().
>
> Fixes: e2209710ccc5d ("KVM: x86/mmu: Skip rmap operations if rmaps not allocated")
> Signed-off-by: Hou Wenlong <houwenlong93@xxxxxxxxxxxxxxxxx>
Haha, I'm glad we're thinking along similar lines. I just sent a patch
yesterday to remove the flush parameter from that function entirely:
https://lore.kernel.org/lkml/20211115234603.2908381-2-bgardon@xxxxxxxxxx/
I'll CC you on that patch.
> ---
> arch/x86/kvm/mmu/mmu.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/arch/x86/kvm/mmu/mmu.c b/arch/x86/kvm/mmu/mmu.c
> index d57319e596a9..4b2be04e9862 100644
> --- a/arch/x86/kvm/mmu/mmu.c
> +++ b/arch/x86/kvm/mmu/mmu.c
> @@ -5853,7 +5853,7 @@ static bool kvm_mmu_zap_collapsible_spte(struct kvm *kvm,
> void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
> const struct kvm_memory_slot *slot)
> {
> - bool flush = false;
> + bool flush;
>
> if (kvm_memslots_have_rmaps(kvm)) {
> write_lock(&kvm->mmu_lock);
> @@ -5870,7 +5870,7 @@ void kvm_mmu_zap_collapsible_sptes(struct kvm *kvm,
>
> if (is_tdp_mmu_enabled(kvm)) {
> read_lock(&kvm->mmu_lock);
> - flush = kvm_tdp_mmu_zap_collapsible_sptes(kvm, slot, flush);
> + flush = kvm_tdp_mmu_zap_collapsible_sptes(kvm, slot, false);
> if (flush)
> kvm_arch_flush_remote_tlbs_memslot(kvm, slot);
> read_unlock(&kvm->mmu_lock);
> --
> 2.31.1
>
