Re: [PATCH v2 05/11] KVM: s390: Add optional storage key checking to MEMOP IOCTL

[Janis Schoetterl-Glausch]

On Wed, 2022-02-09 at 10:08 +0100, Christian Borntraeger wrote:
> 
> Am 09.02.22 um 09:49 schrieb Janis Schoetterl-Glausch:
> > On 2/9/22 08:34, Christian Borntraeger wrote:
> > > Am 07.02.22 um 17:59 schrieb Janis Schoetterl-Glausch:
> > > > User space needs a mechanism to perform key checked accesses when
> > > > emulating instructions.
> > > > 
> > > > The key can be passed as an additional argument.
> > > > Having an additional argument is flexible, as user space can
> > > > pass the guest PSW's key, in order to make an access the same way the
> > > > CPU would, or pass another key if necessary.
> > > > 
> > > > Signed-off-by: Janis Schoetterl-Glausch <scgl@xxxxxxxxxxxxx>
> > > > Acked-by: Janosch Frank <frankja@xxxxxxxxxxxxx>
> > > > Reviewed-by: Claudio Imbrenda <imbrenda@xxxxxxxxxxxxx>
> > > > ---
> > > >    arch/s390/kvm/kvm-s390.c | 49 +++++++++++++++++++++++++++++++---------
> > > >    include/uapi/linux/kvm.h |  8 +++++--
> > > >    2 files changed, 44 insertions(+), 13 deletions(-)
> > > > 
> > > > diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
> > > > index cf347e1a4f17..71e61fb3f0d9 100644
> > > > --- a/arch/s390/kvm/kvm-s390.c
> > > > +++ b/arch/s390/kvm/kvm-s390.c
> > > > @@ -32,6 +32,7 @@
> > > >    #include <linux/sched/signal.h>
> > > >    #include <linux/string.h>
> > > >    #include <linux/pgtable.h>
> > > > +#include <linux/bitfield.h>
> > > >      #include <asm/asm-offsets.h>
> > > >    #include <asm/lowcore.h>
> > > > @@ -2359,6 +2360,11 @@ static int kvm_s390_handle_pv(struct kvm *kvm, struct kvm_pv_cmd *cmd)
> > > >        return r;
> > > >    }
> > > >    +static bool access_key_invalid(u8 access_key)
> > > > +{
> > > > +    return access_key > 0xf;
> > > > +}
> > > > +
> > > >    long kvm_arch_vm_ioctl(struct file *filp,
> > > >                   unsigned int ioctl, unsigned long arg)
> > > >    {
> > > > @@ -4687,34 +4693,54 @@ static long kvm_s390_guest_mem_op(struct kvm_vcpu *vcpu,
> > > >                      struct kvm_s390_mem_op *mop)
> > > >    {
> > > >        void __user *uaddr = (void __user *)mop->buf;
> > > > +    u8 access_key = 0, ar = 0;
> > > >        void *tmpbuf = NULL;
> > > > +    bool check_reserved;
> > > >        int r = 0;
> > > >        const u64 supported_flags = KVM_S390_MEMOP_F_INJECT_EXCEPTION
> > > > -                    | KVM_S390_MEMOP_F_CHECK_ONLY;
> > > > +                    | KVM_S390_MEMOP_F_CHECK_ONLY
> > > > +                    | KVM_S390_MEMOP_F_SKEY_PROTECTION;
> > > >    -    if (mop->flags & ~supported_flags || mop->ar >= NUM_ACRS || !mop->size)
> > > > +    if (mop->flags & ~supported_flags || !mop->size)
> > > >            return -EINVAL;
> > > > -
> > > >        if (mop->size > MEM_OP_MAX_SIZE)
> > > >            return -E2BIG;
> > > > -
> > > >        if (kvm_s390_pv_cpu_is_protected(vcpu))
> > > >            return -EINVAL;
> > > > -
> > > >        if (!(mop->flags & KVM_S390_MEMOP_F_CHECK_ONLY)) {
> > > >            tmpbuf = vmalloc(mop->size);
> > > >            if (!tmpbuf)
> > > >                return -ENOMEM;
> > > >        }
> > > > +    ar = mop->ar;
> > > > +    mop->ar = 0;
> > > 
> > > Why this assignment to 0?
> > 
> > It's so the check of reserved below works like that, they're all part of the anonymous union.
> 
> Ah, I see. This is ugly :-)

Yes :)
> 
> > > > +    if (ar >= NUM_ACRS)
> > > > +        return -EINVAL;
> > > > +    if (mop->flags & KVM_S390_MEMOP_F_SKEY_PROTECTION) {
> > > > +        access_key = mop->key;
> > > > +        mop->key = 0;
> > > 
> > > and this? I think we can leave mop unchanged.
> > > 
> > > In fact, why do we add the ar and access_key variable?
> > > This breaks the check from above (if (mop->flags & ~supported_flags || mop->ar >= NUM_ACRS || !mop->size))  into two checks
> > > and it will create a memleak for tmpbuf.
> > 
> > I can move the allocation down, goto out or get rid of the reserved check and keep everything as before.
> > First is simpler, but second makes handling that case more explicit and might help in the future.
> 
> Maybe add a reserved_02 field in the anon struct and check this for being zero and get rid of the local variables?

I think that would require us adding new fields in the struct by putting them in a union with reserved_02 and so on,
which could get rather messy.

Maybe a comment is good enought?
> 
> > Patch 6 has the same issue in the vm ioctl handler.
> > > Simply use mop->key and mop->ar below and get rid of the local variables.
> > > The structure has no concurrency and gcc will handle that just as the local variable.
> > > 
> > > Other than that this looks good.
> > 
> > [...]
> >