drm/amdgpu: use doorbell mgr for kfd process doorbells

This patch:
- adds a doorbell object in kfd pdd structure.
- allocates doorbells for a process while creating its queue.
- frees the doorbells with pdd destroy.
- moves doorbell bitmap init function to kfd_doorbell.c

PS: This patch ensures that we don't break the existing KFD
    functionality, but now KFD userspace library should also
    create doorbell pages as AMDGPU GEM objects using libdrm
    functions in userspace. The reference code for the same
    is available with AMDGPU Usermode queue libdrm MR. Once
    this is done, we will not need to create process doorbells
    in kernel.

V2: - Do not use doorbell wrapper API, use amdgpu_bo_create_kernel
      instead (Alex).
    - Do not use custom doorbell structure, instead use separate
      variables for bo and doorbell_bitmap (Alex)
V3:
   - Do not allocate doorbell page with PDD, delay doorbell process
     page allocation until really needed (Felix)

Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian Koenig <christian.koenig@amd.com>
Cc: Felix Kuehling <Felix.Kuehling@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Felix Kuehling <Felilx.Kuehling@amd.com>
Signed-off-by: Shashank Sharma <shashank.sharma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

1 files changed, 75 insertions, 27 deletions

diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c b/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c
index dad81c14f6bc..216553307d3a 100644
--- a/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c
+++ b/drivers/gpu/drm/amd/amdkfd/kfd_doorbell.c
@@ -232,48 +232,96 @@ uint64_t kfd_get_number_elems(struct kfd_dev *kfd)
 
 }
 
+static int init_doorbell_bitmap(struct qcm_process_device *qpd,
+				struct kfd_dev *dev)
+{
+	unsigned int i;
+	int range_start = dev->shared_resources.non_cp_doorbells_start;
+	int range_end = dev->shared_resources.non_cp_doorbells_end;
+
+	if (!KFD_IS_SOC15(dev))
+		return 0;
+
+	/* Mask out doorbells reserved for SDMA, IH, and VCN on SOC15. */
+	pr_debug("reserved doorbell 0x%03x - 0x%03x\n", range_start, range_end);
+	pr_debug("reserved doorbell 0x%03x - 0x%03x\n",
+			range_start + KFD_QUEUE_DOORBELL_MIRROR_OFFSET,
+			range_end + KFD_QUEUE_DOORBELL_MIRROR_OFFSET);
+
+	for (i = 0; i < KFD_MAX_NUM_OF_QUEUES_PER_PROCESS / 2; i++) {
+		if (i >= range_start && i <= range_end) {
+			__set_bit(i, qpd->doorbell_bitmap);
+			__set_bit(i + KFD_QUEUE_DOORBELL_MIRROR_OFFSET,
+				  qpd->doorbell_bitmap);
+		}
+	}
+
+	return 0;
+}
+
 phys_addr_t kfd_get_process_doorbells(struct kfd_process_device *pdd)
 {
-	if (!pdd->doorbell_index) {
-		int r = kfd_alloc_process_doorbells(pdd->dev->kfd,
-						    &pdd->doorbell_index);
-		if (r < 0)
+	struct amdgpu_device *adev = pdd->dev->adev;
+	uint32_t first_db_index;
+
+	if (!pdd->qpd.proc_doorbells) {
+		if (kfd_alloc_process_doorbells(pdd->dev->kfd, pdd))
+			/* phys_addr_t 0 is error */
 			return 0;
 	}
 
-	return pdd->dev->kfd->doorbell_base +
-		pdd->doorbell_index * kfd_doorbell_process_slice(pdd->dev->kfd);
+	first_db_index = amdgpu_doorbell_index_on_bar(adev, pdd->qpd.proc_doorbells, 0);
+	return adev->doorbell.base + first_db_index * sizeof(uint32_t);
 }
 
-int kfd_alloc_process_doorbells(struct kfd_dev *kfd, unsigned int *doorbell_index)
+int kfd_alloc_process_doorbells(struct kfd_dev *kfd, struct kfd_process_device *pdd)
 {
-	int r = 0;
+	int r;
+	struct qcm_process_device *qpd = &pdd->qpd;
 
-	if (!kfd->shared_resources.enable_mes)
-		r = ida_simple_get(&kfd->doorbell_ida, 1,
-				   kfd->max_doorbell_slices, GFP_KERNEL);
-	else
-		r = amdgpu_mes_alloc_process_doorbells(
-				(struct amdgpu_device *)kfd->adev,
-				doorbell_index);
+	/* Allocate bitmap for dynamic doorbell allocation */
+	qpd->doorbell_bitmap = bitmap_zalloc(KFD_MAX_NUM_OF_QUEUES_PER_PROCESS,
+					     GFP_KERNEL);
+	if (!qpd->doorbell_bitmap) {
+		DRM_ERROR("Failed to allocate process doorbell bitmap\n");
+		return -ENOMEM;
+	}
 
-	if (r > 0)
-		*doorbell_index = r;
+	r = init_doorbell_bitmap(&pdd->qpd, kfd);
+	if (r) {
+		DRM_ERROR("Failed to initialize process doorbells\n");
+		r = -ENOMEM;
+		goto err;
+	}
 
-	if (r < 0)
-		pr_err("Failed to allocate process doorbells\n");
+	/* Allocate doorbells for this process */
+	r = amdgpu_bo_create_kernel(kfd->adev,
+				    kfd_doorbell_process_slice(kfd),
+				    PAGE_SIZE,
+				    AMDGPU_GEM_DOMAIN_DOORBELL,
+				    &qpd->proc_doorbells,
+				    NULL,
+				    NULL);
+	if (r) {
+		DRM_ERROR("Failed to allocate process doorbells\n");
+		goto err;
+	}
+	return 0;
 
+err:
+	bitmap_free(qpd->doorbell_bitmap);
+	qpd->doorbell_bitmap = NULL;
 	return r;
 }
 
-void kfd_free_process_doorbells(struct kfd_dev *kfd, unsigned int doorbell_index)
+void kfd_free_process_doorbells(struct kfd_dev *kfd, struct kfd_process_device *pdd)
 {
-	if (doorbell_index) {
-		if (!kfd->shared_resources.enable_mes)
-			ida_simple_remove(&kfd->doorbell_ida, doorbell_index);
-		else
-			amdgpu_mes_free_process_doorbells(
-					(struct amdgpu_device *)kfd->adev,
-					doorbell_index);
+	struct qcm_process_device *qpd = &pdd->qpd;
+
+	if (qpd->doorbell_bitmap) {
+		bitmap_free(qpd->doorbell_bitmap);
+		qpd->doorbell_bitmap = NULL;
 	}
+
+	amdgpu_bo_free_kernel(&qpd->proc_doorbells, NULL, NULL);
 }