Re: [PATCH v7 05/12] arm/sve: save/restore SVE context switch

[Julien Grall <julien@xxxxxxx>]

Hi Luca,

On 25/05/2023 11:01, Luca Fancellu wrote:
> > On 25 May 2023, at 10:09, Julien Grall <julien@xxxxxxx> wrote:
> > > diff --git a/xen/arch/arm/include/asm/arm64/vfp.h 
> > > b/xen/arch/arm/include/asm/arm64/vfp.h
> > > index e6e8c363bc16..4aa371e85d26 100644
> > > --- a/xen/arch/arm/include/asm/arm64/vfp.h
> > > +++ b/xen/arch/arm/include/asm/arm64/vfp.h
> > > @@ -6,7 +6,19 @@
> > >     struct vfp_state
> > >   {
> > > +    /*
> > > +     * When SVE is enabled for the guest, fpregs memory will be used to
> > > +     * save/restore P0-P15 registers, otherwise it will be used for the V0-V31
> > > +     * registers.
> > > +     */
> > >       uint64_t fpregs[64] __vfp_aligned;
> > > +    /*
> > > +     * When SVE is enabled for the guest, sve_zreg_ctx_end points to memory
> > > +     * where Z0-Z31 registers and FFR can be saved/restored, it points at the
> > > +     * end of the Z0-Z31 space and at the beginning of the FFR space, it's done
> > > +     * like that to ease the save/restore assembly operations.
> > > +     */
> > > +    uint64_t *sve_zreg_ctx_end;
> > Sorry I only noticed now. But shouldn't this be protected with #ifdef 
> > CONFIG_SVE? Same...
> >
> > >       register_t fpcr;
> > >       register_t fpexc32_el2;
> > >       register_t fpsr;
> > > diff --git a/xen/arch/arm/include/asm/domain.h 
> > > b/xen/arch/arm/include/asm/domain.h
> > > index 331da0f3bcc3..814652d92568 100644
> > > --- a/xen/arch/arm/include/asm/domain.h
> > > +++ b/xen/arch/arm/include/asm/domain.h
> > > @@ -195,6 +195,8 @@ struct arch_vcpu
> > >       register_t tpidrro_el0;
> > >         /* HYP configuration */
> > > +    register_t zcr_el1;
> > > +    register_t zcr_el2;
> > ... here.
> Sure I can protect them. It was done on purpose before to avoid ifdefs but I 
> think saving space
> is better here and also there won’t be any use of them when the config is off.
I wasn't thinking about saving space. I was more thinking about catching 
any (mis)use of the fields in common code. With the #ifdef, the 
compilation would fail.
Cheers,

--
Julien Grall