x86/mce: Add support for physical address valid bit

Starting with Zen6, AMD's Scalable MCA systems will incorporate two new bits in
MCA_STATUS and MCA_CONFIG MSRs. These bits will indicate if a valid System
Physical Address (SPA) is present in MCA_ADDR.

PhysAddrValidSupported bit (MCA_CONFIG[11]) serves as the architectural
indicator and states if PhysAddrV bit (MCA_STATUS[54]) is Reserved or if it
indicates validity of SPA in MCA_ADDR.

PhysAddrV bit (MCA_STATUS[54]) advertises if MCA_ADDR contains valid SPA or if
it is implementation specific.

Use and prefer MCA_STATUS[PhysAddrV] when checking for a usable address.

Signed-off-by: Avadhut Naik <avadhut.naik@amd.com>
Signed-off-by: Borislav Petkov (AMD) <bp@alien8.de>
Link: https://patch.msgid.link/20251118191731.181269-1-avadhut.naik@amd.com

Author: Avadhut Naik

arch/x86/include/asm/mce.h

@@ -48,6 +48,7 @@
 
 /* AMD-specific bits */
 #define MCI_STATUS_TCC		BIT_ULL(55)  /* Task context corrupt */
+#define MCI_STATUS_PADDRV	BIT_ULL(54)  /* Valid System Physical Address */
 #define MCI_STATUS_SYNDV	BIT_ULL(53)  /* synd reg. valid */
 #define MCI_STATUS_DEFERRED	BIT_ULL(44)  /* uncorrected error, deferred exception */
 #define MCI_STATUS_POISON	BIT_ULL(43)  /* access poisonous data */
@@ -62,6 +63,7 @@
  */
 #define MCI_CONFIG_MCAX		0x1
 #define MCI_CONFIG_FRUTEXT	BIT_ULL(9)
+#define MCI_CONFIG_PADDRV	BIT_ULL(11)
 #define MCI_IPID_MCATYPE	0xFFFF0000
 #define MCI_IPID_HWID		0xFFF
 

arch/x86/kernel/cpu/mce/amd.c

@@ -87,6 +87,8 @@ struct smca_bank {
 	const struct smca_hwid *hwid;
 	u32 id;			/* Value of MCA_IPID[InstanceId]. */
 	u8 sysfs_id;		/* Value used for sysfs name. */
+	u64 paddrv	:1,	/* Physical Address Valid bit in MCA_CONFIG */
+	    __reserved	:63;
 };
 
 static DEFINE_PER_CPU_READ_MOSTLY(struct smca_bank[MAX_NR_BANKS], smca_banks);
@@ -327,6 +329,9 @@ static void smca_configure(unsigned int bank, unsigned int cpu)
 
 		this_cpu_ptr(mce_banks_array)[bank].lsb_in_status = !!(low & BIT(8));
 
+		if (low & MCI_CONFIG_PADDRV)
+			this_cpu_ptr(smca_banks)[bank].paddrv = 1;
+
 		wrmsr(smca_config, low, high);
 	}
 
@@ -790,9 +795,9 @@ bool amd_mce_is_memory_error(struct mce *m)
 }
 
 /*
- * AMD systems do not have an explicit indicator that the value in MCA_ADDR is
- * a system physical address. Therefore, individual cases need to be detected.
- * Future cases and checks will be added as needed.
+ * Some AMD systems have an explicit indicator that the value in MCA_ADDR is a
+ * system physical address. Individual cases though, need to be detected for
+ * other systems. Future cases will be added as needed.
  *
  * 1) General case
  *	a) Assume address is not usable.
@@ -806,6 +811,8 @@ bool amd_mce_is_memory_error(struct mce *m)
  *	a) Reported in legacy bank 4 with extended error code (XEC) 8.
  *	b) MCA_STATUS[43] is *not* defined as poison in legacy bank 4. Therefore,
  *	   this bit should not be checked.
+ * 4) MCI_STATUS_PADDRVAL is set
+ *	a) Will provide a valid system physical address.
  *
  * NOTE: SMCA UMC memory errors fall into case #1.
  */
@@ -819,6 +826,9 @@ bool amd_mce_usable_address(struct mce *m)
 			return false;
 	}
 
+	if (this_cpu_ptr(smca_banks)[m->bank].paddrv)
+		return m->status & MCI_STATUS_PADDRV;
+
 	/* Check poison bit for all other bank types. */
 	if (m->status & MCI_STATUS_POISON)
 		return true;