Re: [PATCH] selftests/kvm: touch all pages of args on each memstress iteration

[Vipin Sharma]

On Wed, Apr 12, 2023 at 1:09 PM Paolo Bonzini <pbonzini@xxxxxxxxxx> wrote:
>
> Access the same memory addresses on each iteration of the memstress
> guest code.  This ensures that the state of KVM's page tables
> is the same after every iteration, including the pages that host the
> guest page tables for args and vcpu_args.
>

AMD and eptad=0 reasoning you gave in
https://lore.kernel.org/lkml/199f404d-c08e-3895-6ce3-36b21514f487@xxxxxxxxxx
should also be included here.

"It also reproduces on Intel with pml=0 and eptad=0; the reason is due
to the different semantics of dirty bits for page-table pages on AMD
and Intel.  Both AMD and eptad=0 Intel treat those as writes, therefore
more pages are dropped before the repopulation phase when dirty logging
is disabled.

The "missing" page had been included in the population phase because it
hosts the page tables for vcpu_args, but repopulation does not need it."

> This difference is visible on the dirty_log_page_splitting_test
> on AMD machines.
>
> Signed-off-by: Paolo Bonzini <pbonzini@xxxxxxxxxx>
> ---
>  tools/testing/selftests/kvm/lib/memstress.c | 3 +++
>  1 file changed, 3 insertions(+)
>
> diff --git a/tools/testing/selftests/kvm/lib/memstress.c b/tools/testing/selftests/kvm/lib/memstress.c
> index 3632956c6bcf..df457452d146 100644
> --- a/tools/testing/selftests/kvm/lib/memstress.c
> +++ b/tools/testing/selftests/kvm/lib/memstress.c
> @@ -65,6 +65,9 @@ void memstress_guest_code(uint32_t vcpu_idx)
>         GUEST_ASSERT(vcpu_args->vcpu_idx == vcpu_idx);
>
>         while (true) {
> +               for (i = 0; i < sizeof(memstress_args); i += args->guest_page_size)
> +                       (void) *((volatile char *)args + i);
> +
>                 for (i = 0; i < pages; i++) {
>                         if (args->random_access)
>                                 page = guest_random_u32(&rand_state) % pages;
> --
> 2.39.1
>

Apart from the commit log.
Reviewed-by: Vipin Sharma <vipinsh@xxxxxxxxxx>