On 30.10.20 15:01, Janosch Frank wrote:
> We can only have protected guest pages after a successful set secure
> parameters call as only then the UV allows imports and unpacks.
>
> By moving the test we can now also check for it in s390_reset_acc()
> and do an early return if it is 0.
>
> Signed-off-by: Janosch Frank <frankja@xxxxxxxxxxxxx>
Can we check this into devel to give it some test coverage?
> ---
> arch/s390/kvm/kvm-s390.c | 2 +-
> arch/s390/kvm/pv.c | 3 ++-
> arch/s390/mm/gmap.c | 2 ++
> 3 files changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
> index 6b74b92c1a58..08ea6c4735cd 100644
> --- a/arch/s390/kvm/kvm-s390.c
> +++ b/arch/s390/kvm/kvm-s390.c
> @@ -2312,7 +2312,7 @@ static int kvm_s390_handle_pv(struct kvm *kvm, struct kvm_pv_cmd *cmd)
> struct kvm_s390_pv_unp unp = {};
>
> r = -EINVAL;
> - if (!kvm_s390_pv_is_protected(kvm))
> + if (!kvm_s390_pv_is_protected(kvm) || !mm_is_protected(kvm->mm))
> break;
>
> r = -EFAULT;
> diff --git a/arch/s390/kvm/pv.c b/arch/s390/kvm/pv.c
> index eb99e2f95ebe..f5847f9dec7c 100644
> --- a/arch/s390/kvm/pv.c
> +++ b/arch/s390/kvm/pv.c
> @@ -208,7 +208,6 @@ int kvm_s390_pv_init_vm(struct kvm *kvm, u16 *rc, u16 *rrc)
> return -EIO;
> }
> kvm->arch.gmap->guest_handle = uvcb.guest_handle;
> - atomic_set(&kvm->mm->context.is_protected, 1);
> return 0;
> }
>
> @@ -228,6 +227,8 @@ int kvm_s390_pv_set_sec_parms(struct kvm *kvm, void *hdr, u64 length, u16 *rc,
> *rrc = uvcb.header.rrc;
> KVM_UV_EVENT(kvm, 3, "PROTVIRT VM SET PARMS: rc %x rrc %x",
> *rc, *rrc);
> + if (!cc)
> + atomic_set(&kvm->mm->context.is_protected, 1);
> return cc ? -EINVAL : 0;
> }
>
> diff --git a/arch/s390/mm/gmap.c b/arch/s390/mm/gmap.c
> index cfb0017f33a7..64795d034926 100644
> --- a/arch/s390/mm/gmap.c
> +++ b/arch/s390/mm/gmap.c
> @@ -2690,6 +2690,8 @@ static const struct mm_walk_ops reset_acc_walk_ops = {
> #include <linux/sched/mm.h>
> void s390_reset_acc(struct mm_struct *mm)
> {
> + if (!mm_is_protected(mm))
> + return;
> /*
> * we might be called during
> * reset: we walk the pages and clear
>
