coresight: Make trace ID map spinlock local to the map

Reduce contention on the lock by replacing the global lock with one for
each map.

Signed-off-by: James Clark <james.clark@arm.com>
Reviewed-by: Mike Leach <mike.leach@linaro.org>
Signed-off-by: James Clark <james.clark@linaro.org>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Link: https://lore.kernel.org/r/20240722101202.26915-18-james.clark@linaro.org
(cherry picked from commit 988d40a)
Signed-off-by: Carol L Soto <csoto@nvidia.com>

drivers/hwtracing/coresight/coresight-core.c

@@ -1164,6 +1164,7 @@ struct coresight_device *coresight_register(struct coresight_desc *desc)
 
 	if (csdev->type == CORESIGHT_DEV_TYPE_SINK ||
 	    csdev->type == CORESIGHT_DEV_TYPE_LINKSINK) {
+		spin_lock_init(&csdev->perf_sink_id_map.lock);
 		csdev->perf_sink_id_map.cpu_map = alloc_percpu(atomic_t);
 		if (!csdev->perf_sink_id_map.cpu_map) {
 			kfree(csdev);

drivers/hwtracing/coresight/coresight-trace-id.c

@@ -15,12 +15,10 @@
 /* Default trace ID map. Used in sysfs mode and for system sources */
 static DEFINE_PER_CPU(atomic_t, id_map_default_cpu_ids) = ATOMIC_INIT(0);
 static struct coresight_trace_id_map id_map_default = {
-	.cpu_map = &id_map_default_cpu_ids
+	.cpu_map = &id_map_default_cpu_ids,
+	.lock = __SPIN_LOCK_UNLOCKED(id_map_default.lock)
 };
 
-/* lock to protect id_map and cpu data  */
-static DEFINE_SPINLOCK(id_map_lock);
-
 /* #define TRACE_ID_DEBUG 1 */
 #if defined(TRACE_ID_DEBUG) || defined(CONFIG_COMPILE_TEST)
 
@@ -123,11 +121,11 @@ static void coresight_trace_id_release_all(struct coresight_trace_id_map *id_map
 	unsigned long flags;
 	int cpu;
 
-	spin_lock_irqsave(&id_map_lock, flags);
+	spin_lock_irqsave(&id_map->lock, flags);
 	bitmap_zero(id_map->used_ids, CORESIGHT_TRACE_IDS_MAX);
 	for_each_possible_cpu(cpu)
 		atomic_set(per_cpu_ptr(id_map->cpu_map, cpu), 0);
-	spin_unlock_irqrestore(&id_map_lock, flags);
+	spin_unlock_irqrestore(&id_map->lock, flags);
 	DUMP_ID_MAP(id_map);
 }
 
@@ -136,7 +134,7 @@ static int _coresight_trace_id_get_cpu_id(int cpu, struct coresight_trace_id_map
 	unsigned long flags;
 	int id;
 
-	spin_lock_irqsave(&id_map_lock, flags);
+	spin_lock_irqsave(&id_map->lock, flags);
 
 	/* check for existing allocation for this CPU */
 	id = _coresight_trace_id_read_cpu_id(cpu, id_map);
@@ -163,7 +161,7 @@ static int _coresight_trace_id_get_cpu_id(int cpu, struct coresight_trace_id_map
 	atomic_set(per_cpu_ptr(id_map->cpu_map, cpu), id);
 
 get_cpu_id_out_unlock:
-	spin_unlock_irqrestore(&id_map_lock, flags);
+	spin_unlock_irqrestore(&id_map->lock, flags);
 
 	DUMP_ID_CPU(cpu, id);
 	DUMP_ID_MAP(id_map);
@@ -180,12 +178,12 @@ static void _coresight_trace_id_put_cpu_id(int cpu, struct coresight_trace_id_ma
 	if (!id)
 		return;
 
-	spin_lock_irqsave(&id_map_lock, flags);
+	spin_lock_irqsave(&id_map->lock, flags);
 
 	coresight_trace_id_free(id, id_map);
 	atomic_set(per_cpu_ptr(id_map->cpu_map, cpu), 0);
 
-	spin_unlock_irqrestore(&id_map_lock, flags);
+	spin_unlock_irqrestore(&id_map->lock, flags);
 	DUMP_ID_CPU(cpu, id);
 	DUMP_ID_MAP(id_map);
 }
@@ -195,10 +193,10 @@ static int coresight_trace_id_map_get_system_id(struct coresight_trace_id_map *i
 	unsigned long flags;
 	int id;
 
-	spin_lock_irqsave(&id_map_lock, flags);
+	spin_lock_irqsave(&id_map->lock, flags);
 	/* prefer odd IDs for system components to avoid legacy CPU IDS */
 	id = coresight_trace_id_alloc_new_id(id_map, 0, true);
-	spin_unlock_irqrestore(&id_map_lock, flags);
+	spin_unlock_irqrestore(&id_map->lock, flags);
 
 	DUMP_ID(id);
 	DUMP_ID_MAP(id_map);
@@ -209,9 +207,9 @@ static void coresight_trace_id_map_put_system_id(struct coresight_trace_id_map *
 {
 	unsigned long flags;
 
-	spin_lock_irqsave(&id_map_lock, flags);
+	spin_lock_irqsave(&id_map->lock, flags);
 	coresight_trace_id_free(id, id_map);
-	spin_unlock_irqrestore(&id_map_lock, flags);
+	spin_unlock_irqrestore(&id_map->lock, flags);
 
 	DUMP_ID(id);
 	DUMP_ID_MAP(id_map);

include/linux/coresight.h

@@ -232,6 +232,7 @@ struct coresight_trace_id_map {
 	DECLARE_BITMAP(used_ids, CORESIGHT_TRACE_IDS_MAX);
 	atomic_t __percpu *cpu_map;
 	atomic_t perf_cs_etm_session_active;
+	spinlock_t lock;
 };
 
 /**