Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:3 as permitted sender) client-ip=2620:137:e000::3:3;
Message-ID: <44148ab315f28a6d77627675cbde26977418c5df.camel@linux.ibm.com>
Subject: Re: [PATCH 4/4] KVM: s390: Minor refactor of base/ext facility lists
From:   Nina Schoetterl-Glausch <nsg@linux.ibm.com>
To:     Claudio Imbrenda <imbrenda@linux.ibm.com>
Cc:     Christian Borntraeger <borntraeger@linux.ibm.com>,
        Janosch Frank <frankja@linux.ibm.com>,
        Heiko Carstens <hca@linux.ibm.com>,
        Alexander Gordeev <agordeev@linux.ibm.com>,
        Vasily Gorbik <gor@linux.ibm.com>,
        David Hildenbrand <david@redhat.com>,
        linux-kernel@vger.kernel.org, kvm@vger.kernel.org,
        linux-s390@vger.kernel.org,
        Cornelia Huck <cornelia.huck@de.ibm.com>,
        Sven Schnelle <svens@linux.ibm.com>,
        Michael Mueller <mimu@linux.vnet.ibm.com>,
        David Hildenbrand <dahi@linux.vnet.ibm.com>
Date:   Mon, 06 Nov 2023 12:38:55 +0100
In-Reply-To: <20231103193254.7deef2e5@p-imbrenda>
References: <20231103173008.630217-1-nsg@linux.ibm.com>
         <20231103173008.630217-5-nsg@linux.ibm.com>
         <20231103193254.7deef2e5@p-imbrenda>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.48.4 (3.48.4-1.fc38) 
MIME-Version: 1.0
Precedence: bulk

On Fri, 2023-11-03 at 19:32 +0100, Claudio Imbrenda wrote:
> On Fri,  3 Nov 2023 18:30:08 +0100
> Nina Schoetterl-Glausch <nsg@linux.ibm.com> wrote:
>=20
> > Directly use the size of the arrays instead of going through the
> > indirection of kvm_s390_fac_size().
> > Don't use magic number for the number of entries in the non hypervisor
> > managed facility bit mask list.
> > Make the constraint of that number on kvm_s390_fac_base obvious.
> > Get rid of implicit double anding of stfle_fac_list.
> >=20
> > Signed-off-by: Nina Schoetterl-Glausch <nsg@linux.ibm.com>
> > ---
> >=20
> >=20
> > I found it confusing before and think it's nicer this way but
> > it might be needless churn.
>=20
> some things are probably overkill

I mostly wanted to get rid of kvm_s390_fac_size() since it's meaning
wasn't all that clear IMO. It's not the size of the host's facility list,
it's not the size of the guest's facility list (same as host right now,
but could be different in the future) it's the size of the arrays.

But like I said, none of it is necessary and while I'm reasonably sure
I didn't break anything, you never know :).

[...]

> >  arch/s390/kvm/kvm-s390.c | 44 +++++++++++++++++-----------------------
> >  1 file changed, 19 insertions(+), 25 deletions(-)
> >=20
> > diff --git a/arch/s390/kvm/kvm-s390.c b/arch/s390/kvm/kvm-s390.c
> > index b3f17e014cab..e00ab2f38c89 100644
> > --- a/arch/s390/kvm/kvm-s390.c
> > +++ b/arch/s390/kvm/kvm-s390.c

[...]

> >  /*
> >   * Extended feature mask. Consists of the defines in FACILITIES_KVM_CP=
UMODEL
> >   * and defines the facilities that can be enabled via a cpu model.
> >   */
> > -static unsigned long kvm_s390_fac_ext[SIZE_INTERNAL] =3D { FACILITIES_=
KVM_CPUMODEL };
> > -
> > -static unsigned long kvm_s390_fac_size(void)
> > -{
> > -	BUILD_BUG_ON(SIZE_INTERNAL > S390_ARCH_FAC_MASK_SIZE_U64);
> > -	BUILD_BUG_ON(SIZE_INTERNAL > S390_ARCH_FAC_LIST_SIZE_U64);
> > -	BUILD_BUG_ON(SIZE_INTERNAL * sizeof(unsigned long) >
> > -		sizeof(stfle_fac_list));
> > -
> > -	return SIZE_INTERNAL;
> > -}
> > +static const unsigned long kvm_s390_fac_ext[] =3D { FACILITIES_KVM_CPU=
MODEL };
>=20
> this was sized to [SIZE_INTERNAL], now it doesn't have a fixed size. is
> this intentional?

Yes, it's as big as it needs to be, that way it cannot be too small, so one
less thing to consider.

[...]
> >  /* available cpu features supported by kvm */
> >  static DECLARE_BITMAP(kvm_s390_available_cpu_feat, KVM_S390_VM_CPU_FEA=
T_NR_BITS);
> > @@ -3341,13 +3333,16 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned =
long type)
> >  	kvm->arch.sie_page2->kvm =3D kvm;
> >  	kvm->arch.model.fac_list =3D kvm->arch.sie_page2->fac_list;
> > =20
> > -	for (i =3D 0; i < kvm_s390_fac_size(); i++) {
> > +	for (i =3D 0; i < ARRAY_SIZE(kvm_s390_fac_base); i++) {
> >  		kvm->arch.model.fac_mask[i] =3D stfle_fac_list[i] &
> > -					      (kvm_s390_fac_base[i] |
> > -					       kvm_s390_fac_ext[i]);
> > +					      kvm_s390_fac_base[i];
> >  		kvm->arch.model.fac_list[i] =3D stfle_fac_list[i] &
> >  					      kvm_s390_fac_base[i];
> >  	}
> > +	for (i =3D 0; i < ARRAY_SIZE(kvm_s390_fac_ext); i++) {
> > +		kvm->arch.model.fac_mask[i] |=3D stfle_fac_list[i] &
> > +					       kvm_s390_fac_ext[i];
> > +	}
>=20
> I like it better when it's all in one place, instead of having two loops

Hmm, it's the result of the arrays being different lengths now.

[...]

> > -	for (i =3D 0; i < 16; i++)
> > -		kvm_s390_fac_base[i] |=3D
> > -			stfle_fac_list[i] & nonhyp_mask(i);
> > +	for (i =3D 0; i < HMFAI_DWORDS; i++)
> > +		kvm_s390_fac_base[i] |=3D nonhyp_mask(i);
>=20
> where did the stfle_fac_list[i] go?

I deleted it. That's what I meant by "Get rid of implicit double
anding of stfle_fac_list".
Besides it being redundant I didn't like it conceptually.
kvm_s390_fac_base specifies the facilities we support, regardless
if they're installed in the configuration. The hypervisor managed
ones are unconditionally declared via FACILITIES_KVM and we can add
the non hypervisor managed ones unconditionally, too.

> >  	r =3D __kvm_s390_init();
> >  	if (r)
>=20