Merge tag 'x86_cpu_for_v5.18_rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull x86 cpu feature updates from Borislav Petkov:

 - Merge the AMD and Intel PPIN code into a shared one by both vendors.
   Add the PPIN number to sysfs so that sockets can be identified when
   replacement is needed

 - Minor fixes and cleanups

* tag 'x86_cpu_for_v5.18_rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/cpu: Clear SME feature flag when not in use
  x86/cpufeatures: Put the AMX macros in the word 18 block
  topology/sysfs: Add PPIN in sysfs under cpu topology
  topology/sysfs: Add format parameter to macro defining "show" functions for proc
  x86/cpu: Read/save PPIN MSR during initialization
  x86/cpu: X86_FEATURE_INTEL_PPIN finally has a CPUID bit
  x86/cpu: Merge Intel and AMD ppin_init() functions
  x86/CPU/AMD: Use default_groups in kobj_type

Documentation/ABI/stable/sysfs-devices-system-cpu

@@ -86,6 +86,10 @@ What:           /sys/devices/system/cpu/cpuX/topology/die_cpus
 Description:    internal kernel map of CPUs within the same die.
 Values:         hexadecimal bitmask.
 
+What:           /sys/devices/system/cpu/cpuX/topology/ppin
+Description:    per-socket protected processor inventory number
+Values:         hexadecimal.
+
 What:           /sys/devices/system/cpu/cpuX/topology/die_cpus_list
 Description:    human-readable list of CPUs within the same die.
                 The format is like 0-3, 8-11, 14,17.

Documentation/ABI/testing/sysfs-devices-system-cpu

@@ -73,6 +73,7 @@ What:		/sys/devices/system/cpu/cpuX/topology/core_id
 		/sys/devices/system/cpu/cpuX/topology/physical_package_id
 		/sys/devices/system/cpu/cpuX/topology/thread_siblings
 		/sys/devices/system/cpu/cpuX/topology/thread_siblings_list
+		/sys/devices/system/cpu/cpuX/topology/ppin
 Date:		December 2008
 Contact:	Linux kernel mailing list <linux-kernel@vger.kernel.org>
 Description:	CPU topology files that describe a logical CPU's relationship
@@ -103,6 +104,11 @@ Description:	CPU topology files that describe a logical CPU's relationship
 		thread_siblings_list: human-readable list of cpuX's hardware
 		threads within the same core as cpuX
 
+		ppin: human-readable Protected Processor Identification
+		Number of the socket the cpu# belongs to. There should be
+		one per physical_package_id. File is readable only to
+		admin.
+
 		See Documentation/admin-guide/cputopology.rst for more information.
 
 

arch/x86/include/asm/cpufeatures.h

@@ -299,9 +299,6 @@
 /* Intel-defined CPU features, CPUID level 0x00000007:1 (EAX), word 12 */
 #define X86_FEATURE_AVX_VNNI		(12*32+ 4) /* AVX VNNI instructions */
 #define X86_FEATURE_AVX512_BF16		(12*32+ 5) /* AVX512 BFLOAT16 instructions */
-#define X86_FEATURE_AMX_BF16		(18*32+22) /* AMX bf16 Support */
-#define X86_FEATURE_AMX_TILE		(18*32+24) /* AMX tile Support */
-#define X86_FEATURE_AMX_INT8		(18*32+25) /* AMX int8 Support */
 
 /* AMD-defined CPU features, CPUID level 0x80000008 (EBX), word 13 */
 #define X86_FEATURE_CLZERO		(13*32+ 0) /* CLZERO instruction */
@@ -390,7 +387,10 @@
 #define X86_FEATURE_TSXLDTRK		(18*32+16) /* TSX Suspend Load Address Tracking */
 #define X86_FEATURE_PCONFIG		(18*32+18) /* Intel PCONFIG */
 #define X86_FEATURE_ARCH_LBR		(18*32+19) /* Intel ARCH LBR */
+#define X86_FEATURE_AMX_BF16		(18*32+22) /* AMX bf16 Support */
 #define X86_FEATURE_AVX512_FP16		(18*32+23) /* AVX512 FP16 */
+#define X86_FEATURE_AMX_TILE		(18*32+24) /* AMX tile Support */
+#define X86_FEATURE_AMX_INT8		(18*32+25) /* AMX int8 Support */
 #define X86_FEATURE_SPEC_CTRL		(18*32+26) /* "" Speculation Control (IBRS + IBPB) */
 #define X86_FEATURE_INTEL_STIBP		(18*32+27) /* "" Single Thread Indirect Branch Predictors */
 #define X86_FEATURE_FLUSH_L1D		(18*32+28) /* Flush L1D cache */

arch/x86/include/asm/processor.h

@@ -119,6 +119,8 @@ struct cpuinfo_x86 {
 	int			x86_cache_mbm_width_offset;
 	int			x86_power;
 	unsigned long		loops_per_jiffy;
+	/* protected processor identification number */
+	u64			ppin;
 	/* cpuid returned max cores value: */
 	u16			x86_max_cores;
 	u16			apicid;

arch/x86/include/asm/topology.h

@@ -110,6 +110,7 @@ extern const struct cpumask *cpu_clustergroup_mask(int cpu);
 #define topology_logical_die_id(cpu)		(cpu_data(cpu).logical_die_id)
 #define topology_die_id(cpu)			(cpu_data(cpu).cpu_die_id)
 #define topology_core_id(cpu)			(cpu_data(cpu).cpu_core_id)
+#define topology_ppin(cpu)			(cpu_data(cpu).ppin)
 
 extern unsigned int __max_die_per_package;
 

arch/x86/kernel/cpu/amd.c

@@ -394,35 +394,6 @@ static void amd_detect_cmp(struct cpuinfo_x86 *c)
 	per_cpu(cpu_llc_id, cpu) = c->cpu_die_id = c->phys_proc_id;
 }
 
-static void amd_detect_ppin(struct cpuinfo_x86 *c)
-{
-	unsigned long long val;
-
-	if (!cpu_has(c, X86_FEATURE_AMD_PPIN))
-		return;
-
-	/* When PPIN is defined in CPUID, still need to check PPIN_CTL MSR */
-	if (rdmsrl_safe(MSR_AMD_PPIN_CTL, &val))
-		goto clear_ppin;
-
-	/* PPIN is locked in disabled mode, clear feature bit */
-	if ((val & 3UL) == 1UL)
-		goto clear_ppin;
-
-	/* If PPIN is disabled, try to enable it */
-	if (!(val & 2UL)) {
-		wrmsrl_safe(MSR_AMD_PPIN_CTL,  val | 2UL);
-		rdmsrl_safe(MSR_AMD_PPIN_CTL, &val);
-	}
-
-	/* If PPIN_EN bit is 1, return from here; otherwise fall through */
-	if (val & 2UL)
-		return;
-
-clear_ppin:
-	clear_cpu_cap(c, X86_FEATURE_AMD_PPIN);
-}
-
 u32 amd_get_nodes_per_socket(void)
 {
 	return nodes_per_socket;
@@ -585,6 +556,8 @@ static void early_detect_mem_encrypt(struct cpuinfo_x86 *c)
 	 *	      the SME physical address space reduction value.
 	 *	      If BIOS has not enabled SME then don't advertise the
 	 *	      SME feature (set in scattered.c).
+	 *	      If the kernel has not enabled SME via any means then
+	 *	      don't advertise the SME feature.
 	 *   For SEV: If BIOS has not enabled SEV then don't advertise the
 	 *            SEV and SEV_ES feature (set in scattered.c).
 	 *
@@ -607,6 +580,9 @@ static void early_detect_mem_encrypt(struct cpuinfo_x86 *c)
 		if (IS_ENABLED(CONFIG_X86_32))
 			goto clear_all;
 
+		if (!sme_me_mask)
+			setup_clear_cpu_cap(X86_FEATURE_SME);
+
 		rdmsrl(MSR_K7_HWCR, msr);
 		if (!(msr & MSR_K7_HWCR_SMMLOCK))
 			goto clear_sev;
@@ -947,7 +923,6 @@ static void init_amd(struct cpuinfo_x86 *c)
 	amd_detect_cmp(c);
 	amd_get_topology(c);
 	srat_detect_node(c);
-	amd_detect_ppin(c);
 
 	init_amd_cacheinfo(c);
 

arch/x86/kernel/cpu/common.c

@@ -88,6 +88,83 @@ EXPORT_SYMBOL_GPL(get_llc_id);
 /* L2 cache ID of each logical CPU */
 DEFINE_PER_CPU_READ_MOSTLY(u16, cpu_l2c_id) = BAD_APICID;
 
+static struct ppin_info {
+	int	feature;
+	int	msr_ppin_ctl;
+	int	msr_ppin;
+} ppin_info[] = {
+	[X86_VENDOR_INTEL] = {
+		.feature = X86_FEATURE_INTEL_PPIN,
+		.msr_ppin_ctl = MSR_PPIN_CTL,
+		.msr_ppin = MSR_PPIN
+	},
+	[X86_VENDOR_AMD] = {
+		.feature = X86_FEATURE_AMD_PPIN,
+		.msr_ppin_ctl = MSR_AMD_PPIN_CTL,
+		.msr_ppin = MSR_AMD_PPIN
+	},
+};
+
+static const struct x86_cpu_id ppin_cpuids[] = {
+	X86_MATCH_FEATURE(X86_FEATURE_AMD_PPIN, &ppin_info[X86_VENDOR_AMD]),
+	X86_MATCH_FEATURE(X86_FEATURE_INTEL_PPIN, &ppin_info[X86_VENDOR_INTEL]),
+
+	/* Legacy models without CPUID enumeration */
+	X86_MATCH_INTEL_FAM6_MODEL(IVYBRIDGE_X, &ppin_info[X86_VENDOR_INTEL]),
+	X86_MATCH_INTEL_FAM6_MODEL(HASWELL_X, &ppin_info[X86_VENDOR_INTEL]),
+	X86_MATCH_INTEL_FAM6_MODEL(BROADWELL_D, &ppin_info[X86_VENDOR_INTEL]),
+	X86_MATCH_INTEL_FAM6_MODEL(BROADWELL_X, &ppin_info[X86_VENDOR_INTEL]),
+	X86_MATCH_INTEL_FAM6_MODEL(SKYLAKE_X, &ppin_info[X86_VENDOR_INTEL]),
+	X86_MATCH_INTEL_FAM6_MODEL(ICELAKE_X, &ppin_info[X86_VENDOR_INTEL]),
+	X86_MATCH_INTEL_FAM6_MODEL(ICELAKE_D, &ppin_info[X86_VENDOR_INTEL]),
+	X86_MATCH_INTEL_FAM6_MODEL(SAPPHIRERAPIDS_X, &ppin_info[X86_VENDOR_INTEL]),
+	X86_MATCH_INTEL_FAM6_MODEL(XEON_PHI_KNL, &ppin_info[X86_VENDOR_INTEL]),
+	X86_MATCH_INTEL_FAM6_MODEL(XEON_PHI_KNM, &ppin_info[X86_VENDOR_INTEL]),
+
+	{}
+};
+
+static void ppin_init(struct cpuinfo_x86 *c)
+{
+	const struct x86_cpu_id *id;
+	unsigned long long val;
+	struct ppin_info *info;
+
+	id = x86_match_cpu(ppin_cpuids);
+	if (!id)
+		return;
+
+	/*
+	 * Testing the presence of the MSR is not enough. Need to check
+	 * that the PPIN_CTL allows reading of the PPIN.
+	 */
+	info = (struct ppin_info *)id->driver_data;
+
+	if (rdmsrl_safe(info->msr_ppin_ctl, &val))
+		goto clear_ppin;
+
+	if ((val & 3UL) == 1UL) {
+		/* PPIN locked in disabled mode */
+		goto clear_ppin;
+	}
+
+	/* If PPIN is disabled, try to enable */
+	if (!(val & 2UL)) {
+		wrmsrl_safe(info->msr_ppin_ctl,  val | 2UL);
+		rdmsrl_safe(info->msr_ppin_ctl, &val);
+	}
+
+	/* Is the enable bit set? */
+	if (val & 2UL) {
+		c->ppin = __rdmsr(info->msr_ppin);
+		set_cpu_cap(c, info->feature);
+		return;
+	}
+
+clear_ppin:
+	clear_cpu_cap(c, info->feature);
+}
+
 /* correctly size the local cpu masks */
 void __init setup_cpu_local_masks(void)
 {
@@ -1655,6 +1732,8 @@ static void identify_cpu(struct cpuinfo_x86 *c)
 			c->x86_capability[i] |= boot_cpu_data.x86_capability[i];
 	}
 
+	ppin_init(c);
+
 	/* Init Machine Check Exception if available. */
 	mcheck_cpu_init(c);
 

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

@@ -993,6 +993,7 @@ static struct attribute *default_attrs[] = {
 	NULL,	/* possibly interrupt_enable if supported, see below */
 	NULL,
 };
+ATTRIBUTE_GROUPS(default);
 
 #define to_block(k)	container_of(k, struct threshold_block, kobj)
 #define to_attr(a)	container_of(a, struct threshold_attr, attr)
@@ -1029,7 +1030,7 @@ static void threshold_block_release(struct kobject *kobj);
 
 static struct kobj_type threshold_ktype = {
 	.sysfs_ops		= &threshold_ops,
-	.default_attrs		= default_attrs,
+	.default_groups		= default_groups,
 	.release		= threshold_block_release,
 };
 
@@ -1101,10 +1102,10 @@ static int allocate_threshold_blocks(unsigned int cpu, struct threshold_bank *tb
 	b->threshold_limit	= THRESHOLD_MAX;
 
 	if (b->interrupt_capable) {
-		threshold_ktype.default_attrs[2] = &interrupt_enable.attr;
+		default_attrs[2] = &interrupt_enable.attr;
 		b->interrupt_enable = 1;
 	} else {
-		threshold_ktype.default_attrs[2] = NULL;
+		default_attrs[2] = NULL;
 	}
 
 	INIT_LIST_HEAD(&b->miscj);

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

@@ -138,12 +138,7 @@ void mce_setup(struct mce *m)
 	m->socketid = cpu_data(m->extcpu).phys_proc_id;
 	m->apicid = cpu_data(m->extcpu).initial_apicid;
 	m->mcgcap = __rdmsr(MSR_IA32_MCG_CAP);
-
-	if (this_cpu_has(X86_FEATURE_INTEL_PPIN))
-		m->ppin = __rdmsr(MSR_PPIN);
-	else if (this_cpu_has(X86_FEATURE_AMD_PPIN))
-		m->ppin = __rdmsr(MSR_AMD_PPIN);
-
+	m->ppin = cpu_data(m->extcpu).ppin;
 	m->microcode = boot_cpu_data.microcode;
 }
 

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

@@ -470,47 +470,6 @@ void intel_clear_lmce(void)
 	wrmsrl(MSR_IA32_MCG_EXT_CTL, val);
 }
 
-static void intel_ppin_init(struct cpuinfo_x86 *c)
-{
-	unsigned long long val;
-
-	/*
-	 * Even if testing the presence of the MSR would be enough, we don't
-	 * want to risk the situation where other models reuse this MSR for
-	 * other purposes.
-	 */
-	switch (c->x86_model) {
-	case INTEL_FAM6_IVYBRIDGE_X:
-	case INTEL_FAM6_HASWELL_X:
-	case INTEL_FAM6_BROADWELL_D:
-	case INTEL_FAM6_BROADWELL_X:
-	case INTEL_FAM6_SKYLAKE_X:
-	case INTEL_FAM6_ICELAKE_X:
-	case INTEL_FAM6_ICELAKE_D:
-	case INTEL_FAM6_SAPPHIRERAPIDS_X:
-	case INTEL_FAM6_XEON_PHI_KNL:
-	case INTEL_FAM6_XEON_PHI_KNM:
-
-		if (rdmsrl_safe(MSR_PPIN_CTL, &val))
-			return;
-
-		if ((val & 3UL) == 1UL) {
-			/* PPIN locked in disabled mode */
-			return;
-		}
-
-		/* If PPIN is disabled, try to enable */
-		if (!(val & 2UL)) {
-			wrmsrl_safe(MSR_PPIN_CTL,  val | 2UL);
-			rdmsrl_safe(MSR_PPIN_CTL, &val);
-		}
-
-		/* Is the enable bit set? */
-		if (val & 2UL)
-			set_cpu_cap(c, X86_FEATURE_INTEL_PPIN);
-	}
-}
-
 /*
  * Enable additional error logs from the integrated
  * memory controller on processors that support this.
@@ -535,7 +494,6 @@ void mce_intel_feature_init(struct cpuinfo_x86 *c)
 {
 	intel_init_cmci();
 	intel_init_lmce();
-	intel_ppin_init(c);
 	intel_imc_init(c);
 }
 

arch/x86/kernel/cpu/scattered.c

@@ -26,6 +26,7 @@ struct cpuid_bit {
 static const struct cpuid_bit cpuid_bits[] = {
 	{ X86_FEATURE_APERFMPERF,       CPUID_ECX,  0, 0x00000006, 0 },
 	{ X86_FEATURE_EPB,		CPUID_ECX,  3, 0x00000006, 0 },
+	{ X86_FEATURE_INTEL_PPIN,	CPUID_EBX,  0, 0x00000007, 1 },
 	{ X86_FEATURE_CQM_LLC,		CPUID_EDX,  1, 0x0000000f, 0 },
 	{ X86_FEATURE_CQM_OCCUP_LLC,	CPUID_EDX,  0, 0x0000000f, 1 },
 	{ X86_FEATURE_CQM_MBM_TOTAL,	CPUID_EDX,  1, 0x0000000f, 1 },

arch/x86/kernel/process.c

@@ -765,8 +765,11 @@ void stop_this_cpu(void *dummy)
 	 * without the encryption bit, they don't race each other when flushed
 	 * and potentially end up with the wrong entry being committed to
 	 * memory.
+	 *
+	 * Test the CPUID bit directly because the machine might've cleared
+	 * X86_FEATURE_SME due to cmdline options.
 	 */
-	if (boot_cpu_has(X86_FEATURE_SME))
+	if (cpuid_eax(0x8000001f) & BIT(0))
 		native_wbinvd();
 	for (;;) {
 		/*

drivers/base/topology.c

@@ -14,11 +14,11 @@
 #include <linux/hardirq.h>
 #include <linux/topology.h>
 
-#define define_id_show_func(name)					\
+#define define_id_show_func(name, fmt)					\
 static ssize_t name##_show(struct device *dev,				\
 			   struct device_attribute *attr, char *buf)	\
 {									\
-	return sysfs_emit(buf, "%d\n", topology_##name(dev->id));	\
+	return sysfs_emit(buf, fmt "\n", topology_##name(dev->id));	\
 }
 
 #define define_siblings_read_func(name, mask)					\
@@ -42,22 +42,25 @@ static ssize_t name##_list_read(struct file *file, struct kobject *kobj,	\
 					off, count);				\
 }
 
-define_id_show_func(physical_package_id);
+define_id_show_func(physical_package_id, "%d");
 static DEVICE_ATTR_RO(physical_package_id);
 
 #ifdef TOPOLOGY_DIE_SYSFS
-define_id_show_func(die_id);
+define_id_show_func(die_id, "%d");
 static DEVICE_ATTR_RO(die_id);
 #endif
 
 #ifdef TOPOLOGY_CLUSTER_SYSFS
-define_id_show_func(cluster_id);
+define_id_show_func(cluster_id, "%d");
 static DEVICE_ATTR_RO(cluster_id);
 #endif
 
-define_id_show_func(core_id);
+define_id_show_func(core_id, "%d");
 static DEVICE_ATTR_RO(core_id);
 
+define_id_show_func(ppin, "0x%llx");
+static DEVICE_ATTR_ADMIN_RO(ppin);
+
 define_siblings_read_func(thread_siblings, sibling_cpumask);
 static BIN_ATTR_RO(thread_siblings, 0);
 static BIN_ATTR_RO(thread_siblings_list, 0);
@@ -87,7 +90,7 @@ static BIN_ATTR_RO(package_cpus, 0);
 static BIN_ATTR_RO(package_cpus_list, 0);
 
 #ifdef TOPOLOGY_BOOK_SYSFS
-define_id_show_func(book_id);
+define_id_show_func(book_id, "%d");
 static DEVICE_ATTR_RO(book_id);
 define_siblings_read_func(book_siblings, book_cpumask);
 static BIN_ATTR_RO(book_siblings, 0);
@@ -95,7 +98,7 @@ static BIN_ATTR_RO(book_siblings_list, 0);
 #endif
 
 #ifdef TOPOLOGY_DRAWER_SYSFS
-define_id_show_func(drawer_id);
+define_id_show_func(drawer_id, "%d");
 static DEVICE_ATTR_RO(drawer_id);
 define_siblings_read_func(drawer_siblings, drawer_cpumask);
 static BIN_ATTR_RO(drawer_siblings, 0);
@@ -145,6 +148,7 @@ static struct attribute *default_attrs[] = {
 #ifdef TOPOLOGY_DRAWER_SYSFS
 	&dev_attr_drawer_id.attr,
 #endif
+	&dev_attr_ppin.attr,
 	NULL
 };
 

include/linux/topology.h

@@ -211,6 +211,9 @@ static inline int cpu_to_mem(int cpu)
 #ifndef topology_drawer_id
 #define topology_drawer_id(cpu)			((void)(cpu), -1)
 #endif
+#ifndef topology_ppin
+#define topology_ppin(cpu)			((void)(cpu), 0ull)
+#endif
 #ifndef topology_sibling_cpumask
 #define topology_sibling_cpumask(cpu)		cpumask_of(cpu)
 #endif