Re: [PATCH 04/10] s390/mm: force swiotlb for protected virtualization

[Claudio Imbrenda <imbrenda@linux.ibm.com>]

On Thu, 9 May 2019 14:05:20 -0400
"Jason J. Herne" <jjherne@linux.ibm.com> wrote:

[...]

> > +#define sme_me_mask    0ULL
> > +
> > +static inline bool sme_active(void) { return false; }
> > +extern bool sev_active(void);
> > +  
> 
> I noticed this patch always returns false for sme_active. Is it safe
> to assume that whatever fixups are required on x86 to deal with sme
> do not apply to s390?

yes, on x86 sev_active returns false if SEV is enabled. SME is for
host memory encryption. from arch/x86/mm/mem_encrypt.c:

bool sme_active(void)
{
        return sme_me_mask && !sev_enabled;
}

and it makes sense because you can't have both SME and SEV enabled on
the same kernel, because either you're running on bare metal (and then
you can have SME) __or__ you are running as a guest (and then you can
have SEV). The key difference is that DMA operations don't need
bounce buffers with SME, but they do with SEV.

I hope this clarifies your doubts :)

[...]