On 12/11/2018 10:05, Jan Beulich wrote:
>>>> On 09.11.18 at 16:26, <andrew.coop...@citrix.com> wrote:
>> The shadow code doesn't support 1G superpages, and will hand #PF[RSVD] back
>> to
>> guests.
> So this change is then instead of trying to make shadow code cope?
At the moment, that is an unknown quantity of work. Ideally yes we
should get 1G superpage support working in the shadow code, but it
definitely isn't top of the priority list atm.
>
>> --- a/xen/arch/x86/pv/dom0_build.c
>> +++ b/xen/arch/x86/pv/dom0_build.c
>> @@ -148,7 +148,14 @@ static __init void setup_pv_physmap(struct domain *d,
>> unsigned long pgtbl_pfn,
>> pl3e += l3_table_offset(vphysmap_start);
>> if ( !l3e_get_intpte(*pl3e) )
>> {
>> - if ( cpu_has_page1gb &&
>> + /*
>> + * 1G superpages aren't supported by the shadow code. Avoid
>> using
>> + * them we are liable to need to start shadowing dom0. This
> I think there's a word missing here - "when" perhaps?
Oops yes.
>
>> + * assumes that there are no circumstances where we will
>> activate
>> + * logdirty mode on dom0.
>> + */
>> + if ( (!IS_ENABLED(CONFIG_SHADOW_PAGING) ||
>> + !d->arch.pv.check_l1tf) && cpu_has_page1gb &&
>> !(vphysmap_start & ((1UL << L3_PAGETABLE_SHIFT) - 1)) &&
>> vphysmap_end >= vphysmap_start + (1UL <<
>> L3_PAGETABLE_SHIFT) &&
>> (page = alloc_domheap_pages(d,
> Reviewed-by: Jan Beulich <jbeul...@suse.com>
Thanks,
~Andrew
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
