LoongArch: Add SMT (Simultaneous Multi-Threading) support

Loongson-3A6000 has SMT (Simultaneous Multi-Threading) support, each
physical core has two logical cores (threads). This patch add SMT probe
and scheduler support via ACPI PPTT.

If SCHED_SMT enabled, Loongson-3A6000 is treated as 4 cores, 8 threads;
If SCHED_SMT disabled, Loongson-3A6000 is treated as 8 cores, 8 threads.

Remove smp_num_siblings to support HMP (Heterogeneous Multi-Processing).

Signed-off-by: Liupu Wang <wangliupu@loongson.cn>
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>

3 files changed, 43 insertions, 14 deletions

diff --git a/arch/loongarch/kernel/acpi.c b/arch/loongarch/kernel/acpi.c
index 98f431157e4c..9450e09073eb 100644
--- a/arch/loongarch/kernel/acpi.c
+++ b/arch/loongarch/kernel/acpi.c
@@ -33,6 +33,8 @@ u64 acpi_saved_sp;
 
 #define PREFIX			"ACPI: "
 
+struct acpi_madt_core_pic acpi_core_pic[NR_CPUS];
+
 void __init __iomem * __acpi_map_table(unsigned long phys, unsigned long size)
 {
 
@@ -99,6 +101,7 @@ acpi_parse_processor(union acpi_subtable_headers *header, const unsigned long en
 
 	acpi_table_print_madt_entry(&header->common);
 #ifdef CONFIG_SMP
+	acpi_core_pic[processor->core_id] = *processor;
 	set_processor_mask(processor->core_id, processor->flags);
 #endif
 
@@ -140,6 +143,35 @@ static void __init acpi_process_madt(void)
 	loongson_sysconf.nr_cpus = num_processors;
 }
 
+int pptt_enabled;
+
+int __init parse_acpi_topology(void)
+{
+	int cpu, topology_id;
+
+	for_each_possible_cpu(cpu) {
+		topology_id = find_acpi_cpu_topology(cpu, 0);
+		if (topology_id < 0) {
+			pr_warn("Invalid BIOS PPTT\n");
+			return -ENOENT;
+		}
+
+		if (acpi_pptt_cpu_is_thread(cpu) <= 0)
+			cpu_data[cpu].core = topology_id;
+		else {
+			topology_id = find_acpi_cpu_topology(cpu, 1);
+			if (topology_id < 0)
+				return -ENOENT;
+
+			cpu_data[cpu].core = topology_id;
+		}
+	}
+
+	pptt_enabled = 1;
+
+	return 0;
+}
+
 #ifndef CONFIG_SUSPEND
 int (*acpi_suspend_lowlevel)(void);
 #else
diff --git a/arch/loongarch/kernel/proc.c b/arch/loongarch/kernel/proc.c
index 0d82907b5404..d4b270630bb5 100644
--- a/arch/loongarch/kernel/proc.c
+++ b/arch/loongarch/kernel/proc.c
@@ -49,6 +49,7 @@ static int show_cpuinfo(struct seq_file *m, void *v)
 	seq_printf(m, "processor\t\t: %ld\n", n);
 	seq_printf(m, "package\t\t\t: %d\n", cpu_data[n].package);
 	seq_printf(m, "core\t\t\t: %d\n", cpu_data[n].core);
+	seq_printf(m, "global_id\t\t: %d\n", cpu_data[n].global_id);
 	seq_printf(m, "CPU Family\t\t: %s\n", __cpu_family[n]);
 	seq_printf(m, "Model Name\t\t: %s\n", __cpu_full_name[n]);
 	seq_printf(m, "CPU Revision\t\t: 0x%02x\n", version);
diff --git a/arch/loongarch/kernel/smp.c b/arch/loongarch/kernel/smp.c
index a858a468f746..255967ff8c36 100644
--- a/arch/loongarch/kernel/smp.c
+++ b/arch/loongarch/kernel/smp.c
@@ -8,6 +8,7 @@
  * Copyright (C) 2000, 2001 Silicon Graphics, Inc.
  * Copyright (C) 2000, 2001, 2003 Broadcom Corporation
  */
+#include <linux/acpi.h>
 #include <linux/cpu.h>
 #include <linux/cpumask.h>
 #include <linux/init.h>
@@ -37,10 +38,6 @@ EXPORT_SYMBOL(__cpu_number_map);
 int __cpu_logical_map[NR_CPUS];		/* Map logical to physical */
 EXPORT_SYMBOL(__cpu_logical_map);
 
-/* Number of threads (siblings) per CPU core */
-int smp_num_siblings = 1;
-EXPORT_SYMBOL(smp_num_siblings);
-
 /* Representing the threads (siblings) of each logical CPU */
 cpumask_t cpu_sibling_map[NR_CPUS] __read_mostly;
 EXPORT_SYMBOL(cpu_sibling_map);
@@ -229,9 +226,12 @@ void __init loongson_prepare_cpus(unsigned int max_cpus)
 {
 	int i = 0;
 
+	parse_acpi_topology();
+
 	for (i = 0; i < loongson_sysconf.nr_cpus; i++) {
 		set_cpu_present(i, true);
 		csr_mail_send(0, __cpu_logical_map[i], 0);
+		cpu_data[i].global_id = __cpu_logical_map[i];
 	}
 
 	per_cpu(cpu_state, smp_processor_id()) = CPU_ONLINE;
@@ -272,10 +272,10 @@ void loongson_init_secondary(void)
 	numa_add_cpu(cpu);
 #endif
 	per_cpu(cpu_state, cpu) = CPU_ONLINE;
-	cpu_data[cpu].core =
-		     cpu_logical_map(cpu) % loongson_sysconf.cores_per_package;
 	cpu_data[cpu].package =
 		     cpu_logical_map(cpu) / loongson_sysconf.cores_per_package;
+	cpu_data[cpu].core = pptt_enabled ? cpu_data[cpu].core :
+		     cpu_logical_map(cpu) % loongson_sysconf.cores_per_package;
 }
 
 void loongson_smp_finish(void)
@@ -381,14 +381,10 @@ static inline void set_cpu_sibling_map(int cpu)
 
 	cpumask_set_cpu(cpu, &cpu_sibling_setup_map);
 
-	if (smp_num_siblings <= 1)
-		cpumask_set_cpu(cpu, &cpu_sibling_map[cpu]);
-	else {
-		for_each_cpu(i, &cpu_sibling_setup_map) {
-			if (cpus_are_siblings(cpu, i)) {
-				cpumask_set_cpu(i, &cpu_sibling_map[cpu]);
-				cpumask_set_cpu(cpu, &cpu_sibling_map[i]);
-			}
+	for_each_cpu(i, &cpu_sibling_setup_map) {
+		if (cpus_are_siblings(cpu, i)) {
+			cpumask_set_cpu(i, &cpu_sibling_map[cpu]);
+			cpumask_set_cpu(cpu, &cpu_sibling_map[i]);
 		}
 	}
 }