diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index 4de1147119515e078ab9bcfc5ea7acc58b1f936f..6e1fea473a668dbaae89c9d55d93c40202bd477e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -81,6 +81,7 @@ extern int amdgpu_vm_size;
 extern int amdgpu_vm_block_size;
 extern int amdgpu_enable_scheduler;
 extern int amdgpu_sched_jobs;
+extern int amdgpu_sched_hw_submission;
 
 #define AMDGPU_WAIT_IDLE_TIMEOUT_IN_MS	        3000
 #define AMDGPU_MAX_USEC_TIMEOUT			100000	/* 100 ms */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
index 319de441e9076aa2710bdb78bb2033dffef7d364..8bb5610c2d25aa7b402381f05df55c43af5df011 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c
@@ -77,6 +77,7 @@ int amdgpu_vm_block_size = -1;
 int amdgpu_exp_hw_support = 0;
 int amdgpu_enable_scheduler = 0;
 int amdgpu_sched_jobs = 16;
+int amdgpu_sched_hw_submission = 2;
 
 MODULE_PARM_DESC(vramlimit, "Restrict VRAM for testing, in megabytes");
 module_param_named(vramlimit, amdgpu_vram_limit, int, 0600);
@@ -147,6 +148,9 @@ module_param_named(enable_scheduler, amdgpu_enable_scheduler, int, 0444);
 MODULE_PARM_DESC(sched_jobs, "the max number of jobs supported in the sw queue (default 16)");
 module_param_named(sched_jobs, amdgpu_sched_jobs, int, 0444);
 
+MODULE_PARM_DESC(sched_hw_submission, "the max number of HW submissions (default 2)");
+module_param_named(sched_hw_submission, amdgpu_sched_hw_submission, int, 0444);
+
 static struct pci_device_id pciidlist[] = {
 #ifdef CONFIG_DRM_AMDGPU_CIK
 	/* Kaveri */
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
index b0e15b574606d6bdd9293caff250f633ce6c6d19..67d9277d616847d27a5aa815e2aaf9826a571fbf 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
@@ -926,7 +926,8 @@ void amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring)
 	if (amdgpu_enable_scheduler) {
 		ring->scheduler = amd_sched_create((void *)ring->adev,
 						   &amdgpu_sched_ops,
-						   ring->idx, 5, 0);
+						   ring->idx, 5, 0,
+						   amdgpu_sched_hw_submission);
 		if (!ring->scheduler)
 			DRM_ERROR("Failed to create scheduler on ring %d.\n",
 				  ring->idx);
diff --git a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c
index 87993e06ba37dd3b4828df277c95ccd4795c4040..042da7ddf8db0bdb213ab6f4b3724c5646e461ab 100644
--- a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c
+++ b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c
@@ -468,7 +468,8 @@ struct amd_gpu_scheduler *amd_sched_create(void *device,
 					   struct amd_sched_backend_ops *ops,
 					   unsigned ring,
 					   unsigned granularity,
-					   unsigned preemption)
+					   unsigned preemption,
+					   unsigned hw_submission)
 {
 	struct amd_gpu_scheduler *sched;
 	char name[20] = "gpu_sched[0]";
@@ -495,7 +496,7 @@ struct amd_gpu_scheduler *amd_sched_create(void *device,
 
 	init_waitqueue_head(&sched->wait_queue);
 	if(kfifo_alloc(&sched->active_hw_rq,
-		       AMD_MAX_ACTIVE_HW_SUBMISSION * sizeof(void *),
+		       hw_submission * sizeof(void *),
 		       GFP_KERNEL)) {
 		kfree(sched);
 		return NULL;
diff --git a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h
index 52577a88b05457891453968b6352bc9eeadc80c4..7f6bc26d2d29db21ae4e2688de4fa802c4f15eb4 100644
--- a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h
+++ b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.h
@@ -26,8 +26,6 @@
 
 #include <linux/kfifo.h>
 
-#define AMD_MAX_ACTIVE_HW_SUBMISSION		2
-
 #define AMD_KERNEL_CONTEXT_ID			0
 #define AMD_KERNEL_PROCESS_ID			0
 
@@ -127,7 +125,8 @@ struct amd_gpu_scheduler *amd_sched_create(void *device,
 				struct amd_sched_backend_ops *ops,
 				uint32_t ring,
 				uint32_t granularity,
-				uint32_t preemption);
+				uint32_t preemption,
+				uint32_t hw_submission);
 
 int amd_sched_destroy(struct amd_gpu_scheduler *sched);