On 8/10/21 6:33 AM, Borislav Petkov wrote:

On Wed, Jul 07, 2021 at 01:14:35PM -0500, Brijesh Singh wrote:

quoted

diff --git a/arch/x86/include/asm/sev-common.h b/arch/x86/include/asm/sev-common.h
index 23929a3010df..e75e29c05f59 100644
--- a/arch/x86/include/asm/sev-common.h
+++ b/arch/x86/include/asm/sev-common.h

@@ -63,9 +63,17 @@
  	(((((u64)reason_set) &  GHCB_MSR_TERM_REASON_SET_MASK) << GHCB_MSR_TERM_REASON_SET_POS) | \
  	((((u64)reason_val) & GHCB_MSR_TERM_REASON_MASK) << GHCB_MSR_TERM_REASON_POS))
  
+/* Error code from reason set 0 */

... Error codes...

Noted.

quoted

+#define SEV_TERM_SET_GEN		0
  #define GHCB_SEV_ES_GEN_REQ		0
  #define GHCB_SEV_ES_PROT_UNSUPPORTED	1
  
  #define GHCB_RESP_CODE(v)		((v) & GHCB_MSR_INFO_MASK)
  
+/* Linux specific reason codes (used with reason set 1) */

... Linux-specific ...

Noted.

quoted

+#define SEV_TERM_SET_LINUX		1

GHCB doc says:

"This document defines and owns reason code set 0x0"

Should it also say, reason code set 1 is allocated for Linux guest use?
I don't see why not...
 > Tom?

If Tom is okay with it then maybe in next version of the GHCB doc can 
add this text.