On 6/5/2023 8:09 PM, Vitaly Kuznetsov wrote:
Tianyu Lan <ltykernel@xxxxxxxxx> writes:
int hv_call_deposit_pages(int node, u64 partition_id, u32 num_pages);
diff --git a/arch/x86/kernel/cpu/mshyperv.c b/arch/x86/kernel/cpu/mshyperv.c
index c7969e806c64..9186453251f7 100644
--- a/arch/x86/kernel/cpu/mshyperv.c
+++ b/arch/x86/kernel/cpu/mshyperv.c
@@ -402,8 +402,12 @@ static void __init ms_hyperv_init_platform(void)
pr_info("Hyper-V: Isolation Config: Group A 0x%x, Group B 0x%x\n",
ms_hyperv.isolation_config_a, ms_hyperv.isolation_config_b);
- if (hv_get_isolation_type() == HV_ISOLATION_TYPE_SNP)
+
+ if (cc_platform_has(CC_ATTR_GUEST_SEV_SNP)) {
+ static_branch_enable(&isolation_type_en_snp);
+ } else if (hv_get_isolation_type() == HV_ISOLATION_TYPE_SNP) {
static_branch_enable(&isolation_type_snp);
Nitpick: In case 'isolation_type_snp' and 'isolation_type_en_snp' are
mutually exclusive, I'd suggest we rename the former: it is quite
un-intuitive that for an enlightened SNP guest '&isolation_type_snp' is
NOT enabled. E.g. we can use
'isol_type_snp_paravisor'
and
'isol_type_snp_enlightened'
(I also don't like 'isolation_type_en_snp' name as 'en' normally stands
for 'enabled')
Hi Vitaly:
I will do such rename the function in the following patchset and this
will not affect SEV-SNP function.
