Re: [PATCH V3 6/6] crypto/nx: Add P9 NX support for 842 compression engine

[Michael Neuling <mikey@neuling.org>]

> > > +static int nx842_powernv_crypto_init_vas(struct crypto_tfm *tfm)
> > > +{
> > > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0struct nx842_crypto_ctx *c=
tx =3D crypto_tfm_ctx(tfm);
> > > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0struct nx842_workmem *wmem=
;
> > > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0struct nx842_coproc *copro=
c;
> > > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0int ret;
> > > +
> > > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0ret =3D nx842_crypto_init(=
tfm, &nx842_powernv_driver);
> > > +
> > > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0if (ret)
> > > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=
=C2=A0=C2=A0=C2=A0=C2=A0return ret;
> > > +
> > > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0wmem =3D PTR_ALIGN((struct=
 nx842_workmem *)ctx->wmem,
> > > WORKMEM_ALIGN);
> > > +=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0=C2=A0coproc =3D per_cpu(coproc_=
inst, smp_processor_id());
> >=20
> > this is wrong.=C2=A0=C2=A0the crypto transform init function is not gua=
ranteed
> > to be called by the same processor that later uses it.=C2=A0=C2=A0Just =
because
> > that happens to be how zswap operates doesn't guarantee other crypto
> > users will do the same.
>=20
> Dan, Sorry missed this comment.=C2=A0
>=20
> Right, The actual crypto request can be executed on other processor than =
the
> CPU when the init is executed. The main goal is open send window on the N=
X
> engine which is on the same chip for the corresponding CPU. So we are OK =
if
> the request is scheduled on other CPU as long as it belongs to same chip.
> Otherwise in the worst case we will end up using remote NX.=C2=A0=C2=A0

You want the NX to be close to the requester CPU, but probably more importa=
ntly
you want the NX close to the memory it's going to be operating on.=C2=A0

Preferably they would all be on the same node.

Mikey=20