On 2/10/23 15:47, Sean Christopherson wrote:
> diff --git a/arch/x86/kernel/apic/io_apic.c b/arch/x86/kernel/apic/io_apic.c
> index a868b76cd3d4..08f65ed439d9 100644
> --- a/arch/x86/kernel/apic/io_apic.c
> +++ b/arch/x86/kernel/apic/io_apic.c
> @@ -2682,11 +2682,16 @@ static void io_apic_set_fixmap(enum fixed_addresses idx, phys_addr_t phys)
> {
> pgprot_t flags = FIXMAP_PAGE_NOCACHE;
>
> - /*
> - * Ensure fixmaps for IOAPIC MMIO respect memory encryption pgprot
> - * bits, just like normal ioremap():
> - */
> - flags = pgprot_decrypted(flags);
> + if (cc_platform_has(CC_ATTR_GUEST_MEM_ENCRYPT)) {
> + /*
> + * Ensure fixmaps for IOAPIC MMIO respect memory encryption pgprot
> + * bits, just like normal ioremap():
> + */
> + if (x86_platform.hyper.is_private_mmio(phys))
> + flags = pgprot_encrypted(flags);
> + else
> + flags = pgprot_decrypted(flags);
> + }
I don't completely hate this. Thinking to the future, I'd hope that
future platforms will include information about which physical addresses
are shared or private. This might even vary per device, but this
interface would still work.
I _think_ it would be nicer to wrap both checks up in a helper where the
comments can be more detailed, like:
if (cc_private_mmio(phys))
flags = pgprot_encrypted(flags);
else
flags = pgprot_decrypted(flags);
but I honestly don't feel that strongly about it.
It does seem a bit odd that there's a new CC_ATTR_GUEST_MEM_ENCRYPT
check wrapping this whole thing. I guess the trip through
pgprot_decrypted() is harmless on normal platforms, though.
> diff --git a/arch/x86/mm/ioremap.c b/arch/x86/mm/ioremap.c
> index 6453fbaedb08..0baec766b921 100644
> --- a/arch/x86/mm/ioremap.c
> +++ b/arch/x86/mm/ioremap.c
> @@ -116,6 +116,9 @@ static void __ioremap_check_other(resource_size_t addr, struct ioremap_desc *des
> if (!cc_platform_has(CC_ATTR_GUEST_MEM_ENCRYPT))
> return;
>
> + if (x86_platform.hyper.is_private_mmio(addr))
> + desc->flags |= IORES_MAP_ENCRYPTED;
> +
> if (!IS_ENABLED(CONFIG_EFI))
> return;
>
