Re: [PATCH v1 1/1] KVM: s390: Fix gmap_helper_zap_one_page() again

From: Marc Hartmayer
Date: Mon Nov 10 2025 - 10:43:51 EST


On Thu, Nov 06, 2025 at 04:25 PM +0100, Claudio Imbrenda <imbrenda@xxxxxxxxxxxxx> wrote:
> A few checks were missing in gmap_helper_zap_one_page(), which can lead
> to memory corruption in the guest under specific circumstances.
>
> Add the missing checks.
>
> Fixes: 5deafa27d9ae ("KVM: s390: Fix to clear PTE when discarding a swapped page")
> Reported-by: Marc Hartmayer <mhartmay@xxxxxxxxxxxxx>
> Signed-off-by: Claudio Imbrenda <imbrenda@xxxxxxxxxxxxx>
> ---
> arch/s390/mm/gmap_helpers.c | 9 +++++++--
> 1 file changed, 7 insertions(+), 2 deletions(-)
>
> diff --git a/arch/s390/mm/gmap_helpers.c b/arch/s390/mm/gmap_helpers.c
> index d4c3c36855e2..38a2d82cd88a 100644
> --- a/arch/s390/mm/gmap_helpers.c
> +++ b/arch/s390/mm/gmap_helpers.c
> @@ -47,6 +47,7 @@ static void ptep_zap_swap_entry(struct mm_struct *mm, swp_entry_t entry)
> void gmap_helper_zap_one_page(struct mm_struct *mm, unsigned long vmaddr)
> {
> struct vm_area_struct *vma;
> + unsigned long pgstev;
> spinlock_t *ptl;
> pgste_t pgste;
> pte_t *ptep;
> @@ -65,9 +66,13 @@ void gmap_helper_zap_one_page(struct mm_struct *mm, unsigned long vmaddr)
> if (pte_swap(*ptep)) {
> preempt_disable();
> pgste = pgste_get_lock(ptep);
> + pgstev = pgste_val(pgste);
>
> - ptep_zap_swap_entry(mm, pte_to_swp_entry(*ptep));
> - pte_clear(mm, vmaddr, ptep);
> + if ((pgstev & _PGSTE_GPS_USAGE_MASK) == _PGSTE_GPS_USAGE_UNUSED ||
> + (pgstev & _PGSTE_GPS_ZERO)) {
> + ptep_zap_swap_entry(mm, pte_to_swp_entry(*ptep));
> + pte_clear(mm, vmaddr, ptep);
> + }
>
> pgste_set_unlock(ptep, pgste);
> preempt_enable();
> --
> 2.51.1

Thanks for the fix.

Tested-by: Marc Hartmayer <mhartmay@xxxxxxxxxxxxx>