Merge drm/drm-next into drm-misc-next

Let's start the 5.19 development cycle.

Signed-off-by: Maxime Ripard <maxime@cerno.tech>

1 files changed, 16 insertions, 8 deletions

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
index ee59b6928b11..e85e347eb670 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c
@@ -32,6 +32,7 @@
 
 #include <drm/amdgpu_drm.h>
 #include <drm/drm_syncobj.h>
+#include "amdgpu_cs.h"
 #include "amdgpu.h"
 #include "amdgpu_trace.h"
 #include "amdgpu_gmc.h"
@@ -127,8 +128,6 @@ static int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, union drm_amdgpu_cs
 		goto free_chunk;
 	}
 
-	mutex_lock(&p->ctx->lock);
-
 	/* skip guilty context job */
 	if (atomic_read(&p->ctx->guilty) == 1) {
 		ret = -ECANCELED;
@@ -341,7 +340,7 @@ static void amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev,
 	if (free_vram >= 128 * 1024 * 1024 || free_vram >= total_vram / 8) {
 		s64 min_us;
 
-		/* Be more aggresive on dGPUs. Try to fill a portion of free
+		/* Be more aggressive on dGPUs. Try to fill a portion of free
 		 * VRAM now.
 		 */
 		if (!(adev->flags & AMD_IS_APU))
@@ -585,6 +584,16 @@ static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p,
 		}
 	}
 
+	/* Move fence waiting after getting reservation lock of
+	 * PD root. Then there is no need on a ctx mutex lock.
+	 */
+	r = amdgpu_ctx_wait_prev_fence(p->ctx, p->entity);
+	if (unlikely(r != 0)) {
+		if (r != -ERESTARTSYS)
+			DRM_ERROR("amdgpu_ctx_wait_prev_fence failed.\n");
+		goto error_validate;
+	}
+
 	amdgpu_cs_get_threshold_for_moves(p->adev, &p->bytes_moved_threshold,
 					  &p->bytes_moved_vis_threshold);
 	p->bytes_moved = 0;
@@ -700,7 +709,6 @@ static void amdgpu_cs_parser_fini(struct amdgpu_cs_parser *parser, int error,
 	dma_fence_put(parser->fence);
 
 	if (parser->ctx) {
-		mutex_unlock(&parser->ctx->lock);
 		amdgpu_ctx_put(parser->ctx);
 	}
 	if (parser->bo_list)
@@ -775,12 +783,12 @@ static int amdgpu_cs_vm_handling(struct amdgpu_cs_parser *p)
 				memcpy(ib->ptr, kptr, chunk_ib->ib_bytes);
 				amdgpu_bo_kunmap(aobj);
 
-				r = amdgpu_ring_parse_cs(ring, p, j);
+				r = amdgpu_ring_parse_cs(ring, p, p->job, ib);
 				if (r)
 					return r;
 			} else {
 				ib->ptr = (uint32_t *)kptr;
-				r = amdgpu_ring_patch_cs_in_place(ring, p, j);
+				r = amdgpu_ring_patch_cs_in_place(ring, p, p->job, ib);
 				amdgpu_bo_kunmap(aobj);
 				if (r)
 					return r;
@@ -944,7 +952,7 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev,
 	if (parser->job->uf_addr && ring->funcs->no_user_fence)
 		return -EINVAL;
 
-	return amdgpu_ctx_wait_prev_fence(parser->ctx, parser->entity);
+	return 0;
 }
 
 static int amdgpu_cs_process_fence_dep(struct amdgpu_cs_parser *p,
@@ -1365,7 +1373,6 @@ int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp)
 		goto out;
 
 	r = amdgpu_cs_submit(&parser, cs);
-
 out:
 	amdgpu_cs_parser_fini(&parser, r, reserved_buffers);
 
@@ -1511,6 +1518,7 @@ int amdgpu_cs_fence_to_handle_ioctl(struct drm_device *dev, void *data,
 		return 0;
 
 	default:
+		dma_fence_put(fence);
 		return -EINVAL;
 	}
 }