On 8/2/2021 8:28 PM, Joerg Roedel wrote:
On Wed, Jul 28, 2021 at 10:52:20AM -0400, Tianyu Lan wrote:
+void hv_ghcb_msr_write(u64 msr, u64 value)
+{
+ union hv_ghcb *hv_ghcb;
+ void **ghcb_base;
+ unsigned long flags;
+
+ if (!ms_hyperv.ghcb_base)
+ return;
+
+ WARN_ON(in_nmi());
+
+ local_irq_save(flags);
+ ghcb_base = (void **)this_cpu_ptr(ms_hyperv.ghcb_base);
+ hv_ghcb = (union hv_ghcb *)*ghcb_base;
+ if (!hv_ghcb) {
+ local_irq_restore(flags);
+ return;
+ }
+
+ memset(hv_ghcb, 0x00, HV_HYP_PAGE_SIZE);
Do you really need to zero out the whole 4k? The validation bitmap
should be enough, there are no secrets on the page anyway.
Same in hv_ghcb_msr_read().
OK. Thanks for suggestion. I will have a try.
+enum es_result sev_es_ghcb_hv_call(struct ghcb *ghcb,
+ struct es_em_ctxt *ctxt,
+ u64 exit_code, u64 exit_info_1,
+ u64 exit_info_2)
{
enum es_result ret;
@@ -109,7 +109,16 @@ static enum es_result sev_es_ghcb_hv_call(struct ghcb *ghcb,
ghcb_set_sw_exit_info_1(ghcb, exit_info_1);
ghcb_set_sw_exit_info_2(ghcb, exit_info_2);
- sev_es_wr_ghcb_msr(__pa(ghcb));
+ /*
+ * Hyper-V runs paravisor with SEV. Ghcb page is allocated by
+ * paravisor and not needs to be updated in the Linux guest.
+ * Otherwise, the ghcb page's PA reported by paravisor is above
+ * VTOM. Hyper-V use this function with NULL for ctxt point and
+ * skip setting ghcb page in such case.
+ */
+ if (ctxt)
+ sev_es_wr_ghcb_msr(__pa(ghcb));
No, do not make this function work with ctxt==NULL. Instead, factor out
a helper function which contains what Hyper-V needs and use that in
sev_es_ghcb_hv_call() and Hyper-V code.
OK. Will update.
+union hv_ghcb {
+ struct ghcb ghcb;
+} __packed __aligned(PAGE_SIZE);
I am curious what this will end up being good for.
Hyper-V introduces a specific hypercall request in GHCB page and use
same union in the Linux Hyper-V code to read/write MSR and call the new
hypercall request.
