drm/amdkfd: allow users to target recommended SDMA engines

Certain GPUs have better copy performance over xGMI on specific
SDMA engines depending on the source and destination GPU.
Allow users to create SDMA queues on these recommended engines.
Close to 2x overall performance has been observed with this
optimization.

Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

1 files changed, 37 insertions, 1 deletions

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
index fdc76c24b2e7..f0bfeb35246f 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c
@@ -1532,6 +1532,41 @@ static int allocate_sdma_queue(struct device_queue_manager *dqm,
 			q->sdma_id % kfd_get_num_xgmi_sdma_engines(dqm->dev);
 		q->properties.sdma_queue_id = q->sdma_id /
 			kfd_get_num_xgmi_sdma_engines(dqm->dev);
+	} else if (q->properties.type == KFD_QUEUE_TYPE_SDMA_BY_ENG_ID) {
+		int i, num_queues, num_engines, eng_offset = 0, start_engine;
+		bool free_bit_found = false, is_xgmi = false;
+
+		if (q->properties.sdma_engine_id < kfd_get_num_sdma_engines(dqm->dev)) {
+			num_queues = get_num_sdma_queues(dqm);
+			num_engines = kfd_get_num_sdma_engines(dqm->dev);
+			q->properties.type = KFD_QUEUE_TYPE_SDMA;
+		} else {
+			num_queues = get_num_xgmi_sdma_queues(dqm);
+			num_engines = kfd_get_num_xgmi_sdma_engines(dqm->dev);
+			eng_offset = kfd_get_num_sdma_engines(dqm->dev);
+			q->properties.type = KFD_QUEUE_TYPE_SDMA_XGMI;
+			is_xgmi = true;
+		}
+
+		/* Scan available bit based on target engine ID. */
+		start_engine = q->properties.sdma_engine_id - eng_offset;
+		for (i = start_engine; i < num_queues; i += num_engines) {
+
+			if (!test_bit(i, is_xgmi ? dqm->xgmi_sdma_bitmap : dqm->sdma_bitmap))
+				continue;
+
+			clear_bit(i, is_xgmi ? dqm->xgmi_sdma_bitmap : dqm->sdma_bitmap);
+			q->sdma_id = i;
+			q->properties.sdma_queue_id = q->sdma_id / num_engines;
+			free_bit_found = true;
+			break;
+		}
+
+		if (!free_bit_found) {
+			dev_err(dev, "No more SDMA queue to allocate for target ID %i\n",
+				q->properties.sdma_engine_id);
+			return -ENOMEM;
+		}
 	}
 
 	pr_debug("SDMA engine id: %d\n", q->properties.sdma_engine_id);
@@ -1784,7 +1819,8 @@ static int create_queue_cpsch(struct device_queue_manager *dqm, struct queue *q,
 	}
 
 	if (q->properties.type == KFD_QUEUE_TYPE_SDMA ||
-		q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI) {
+		q->properties.type == KFD_QUEUE_TYPE_SDMA_XGMI ||
+		q->properties.type == KFD_QUEUE_TYPE_SDMA_BY_ENG_ID) {
 		dqm_lock(dqm);
 		retval = allocate_sdma_queue(dqm, q, qd ? &qd->sdma_id : NULL);
 		dqm_unlock(dqm);