Re: [Xen-devel] [PATCH v2] x86/shadow: Correct guest behaviour when creating PTEs above maxphysaddr

[Tim Deegan <tim@xxxxxxx>]

At 16:04 +0000 on 14 Feb (1487088291), Andrew Cooper wrote:
> Hmm ok.  With the other bugfix of not dropping the first line, this hunk
> is now simply:
> 
> @@ -504,7 +505,7 @@ void recalculate_cpuid_policy(struct domain *d)
>  
>      p->extd.maxphysaddr = min(p->extd.maxphysaddr, max->extd.maxphysaddr);
>      p->extd.maxphysaddr = min_t(uint8_t, p->extd.maxphysaddr,
> -                                d->arch.paging.gfn_bits + PAGE_SHIFT);
> +                                paging_max_paddr_bits(d));
>      p->extd.maxphysaddr = max_t(uint8_t, p->extd.maxphysaddr,
>                                  (p->basic.pae || p->basic.pse36) ? 36 :
> 32);

That looks good to me.  Reviewed-by: Tim Deegan <tim@xxxxxxx>

> >> @@ -360,6 +361,21 @@ void paging_dump_vcpu_info(struct vcpu *v);
> >>  int paging_set_allocation(struct domain *d, unsigned int pages,
> >>                            bool *preempted);
> >>  
> >> +/* Maxphysaddr supportable by the paging infrastructure. */
> >> +static inline unsigned int paging_max_paddr_bits(const struct domain *d)
> >> +{
> >> +    unsigned int bits = paging_mode_hap(d) ? hap_paddr_bits : paddr_bits;
> >> +
> >> +    if ( !IS_ENABLED(BIGMEM) && paging_mode_shadow(d) &&
> >> +         (!is_pv_domain(d) || opt_allow_superpage) )
> >> +    {
> >> +        /* Shadowed superpages store GFNs in 32-bit page_info fields. */
> >> +        bits = min(bits, 32U + PAGE_SHIFT);
> >> +    }
> >> +
> >> +    return bits;
> >> +}
> > Does this really need to be an inline function? With the overall goal
> > of not including every header everywhere, I particularly dislike the
> > need to include xen/kconfig.h here for things to build.
> 
> It is not on a critically path, but it seems wasteful to force something
> this small to be out of line.

It could be a macro, too.  FWIW I agree with you that a static inline
is best.

Cheers,

Tim.

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel