Adds a sunxi-cedrus-drv-video libVA backend

This VA backend uses v4l2's Frame API proposal to interface with the "sunxi-cedrus" video driver on Allwinner SoC. Only a few parts of the code are really dependent on sunxi-cedrus and this VA backend could be reused for other v4l drivers using the Frame API.
2016-08-05 12:23:15 +02:00
commit e263c9542c
28 changed files with 2805 additions and 0 deletions
@@ -0,0 +1,167 @@
+/*
+ * Copyright (c) 2016 Florent Revest, <florent.revest@free-electrons.com>
+ *               2007 Intel Corporation. All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL PRECISION INSIGHT AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ */
+
+#include "sunxi_cedrus_drv_video.h"
+#include "context.h"
+#include "va_config.h"
+#include "surface.h"
+
+#include <stdlib.h>
+#include <string.h>
+
+#include <assert.h>
+
+#include <sys/ioctl.h>
+
+#include <linux/videodev2.h>
+
+/*
+ * A Context is a global data structure used for rendering a video of a certain
+ * format. When a context is created, input buffers are created and v4l's output
+ * (which is the compressed data input queue, since capture is the real output)
+ * format is set.
+ */
+
+VAStatus sunxi_cedrus_CreateContext(VADriverContextP ctx, VAConfigID config_id,
+		int picture_width, int picture_height, int flag,
+		VASurfaceID *render_targets, int num_render_targets,
+		VAContextID *context)
+{
+	INIT_DRIVER_DATA
+	VAStatus vaStatus = VA_STATUS_SUCCESS;
+	object_config_p obj_config;
+	int i;
+	struct v4l2_create_buffers create_bufs;
+	struct v4l2_format fmt;
+	enum v4l2_buf_type type;
+
+	obj_config = CONFIG(config_id);
+	if (NULL == obj_config)
+	{
+		vaStatus = VA_STATUS_ERROR_INVALID_CONFIG;
+		return vaStatus;
+	}
+
+	int contextID = object_heap_allocate(&driver_data->context_heap);
+	object_context_p obj_context = CONTEXT(contextID);
+	if (NULL == obj_context)
+	{
+		vaStatus = VA_STATUS_ERROR_ALLOCATION_FAILED;
+		return vaStatus;
+	}
+
+	obj_context->context_id  = contextID;
+	*context = contextID;
+	obj_context->current_render_target = -1;
+	obj_context->config_id = config_id;
+	obj_context->picture_width = picture_width;
+	obj_context->picture_height = picture_height;
+	obj_context->num_render_targets = num_render_targets;
+	obj_context->render_targets = (VASurfaceID *) malloc(num_render_targets * sizeof(VASurfaceID));
+	obj_context->num_rendered_surfaces = 0;
+
+	if (obj_context->render_targets == NULL)
+	{
+		vaStatus = VA_STATUS_ERROR_ALLOCATION_FAILED;
+		return vaStatus;
+	}
+
+	for(i = 0; i < num_render_targets; i++)
+	{
+		if (NULL == SURFACE(render_targets[i]))
+		{
+			vaStatus = VA_STATUS_ERROR_INVALID_SURFACE;
+			break;
+		}
+		obj_context->render_targets[i] = render_targets[i];
+	}
+	obj_context->flags = flag;
+
+	/* Error recovery */
+	if (VA_STATUS_SUCCESS != vaStatus)
+	{
+		obj_context->context_id = -1;
+		obj_context->config_id = -1;
+		free(obj_context->render_targets);
+		obj_context->render_targets = NULL;
+		obj_context->num_render_targets = 0;
+		obj_context->flags = 0;
+		object_heap_free(&driver_data->context_heap, (object_base_p) obj_context);
+	}
+
+	memset(&(fmt), 0, sizeof(fmt));
+	fmt.type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE;
+	fmt.fmt.pix_mp.width = picture_width;
+	fmt.fmt.pix_mp.height = picture_height;
+	fmt.fmt.pix_mp.plane_fmt[0].sizeimage = INPUT_BUFFER_MAX_SIZE;
+	switch(obj_config->profile) {
+		default:
+			vaStatus = VA_STATUS_ERROR_UNSUPPORTED_PROFILE;
+			return vaStatus;
+			break;
+	}
+	fmt.fmt.pix_mp.field = V4L2_FIELD_ANY;
+	fmt.fmt.pix_mp.num_planes = 1;
+	assert(ioctl(driver_data->mem2mem_fd, VIDIOC_S_FMT, &fmt)==0);
+
+	memset (&create_bufs, 0, sizeof (struct v4l2_create_buffers));
+	create_bufs.count = INPUT_BUFFERS_NB;
+	create_bufs.memory = V4L2_MEMORY_MMAP;
+	create_bufs.format.type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE;
+	assert(ioctl(driver_data->mem2mem_fd, VIDIOC_G_FMT, &create_bufs.format)==0);
+	assert(ioctl(driver_data->mem2mem_fd, VIDIOC_CREATE_BUFS, &create_bufs)==0);
+
+	type = V4L2_BUF_TYPE_VIDEO_OUTPUT_MPLANE;
+	assert(ioctl(driver_data->mem2mem_fd, VIDIOC_STREAMON, &type)==0);
+
+	type = V4L2_BUF_TYPE_VIDEO_CAPTURE_MPLANE;
+	assert(ioctl(driver_data->mem2mem_fd, VIDIOC_STREAMON, &type)==0);
+
+	return vaStatus;
+}
+
+VAStatus sunxi_cedrus_DestroyContext(VADriverContextP ctx, VAContextID context)
+{
+	INIT_DRIVER_DATA
+	object_context_p obj_context = CONTEXT(context);
+	assert(obj_context);
+
+	obj_context->context_id = -1;
+	obj_context->config_id = -1;
+	obj_context->picture_width = 0;
+	obj_context->picture_height = 0;
+	if (obj_context->render_targets)
+		free(obj_context->render_targets);
+	obj_context->render_targets = NULL;
+	obj_context->num_render_targets = 0;
+	obj_context->flags = 0;
+
+	obj_context->current_render_target = -1;
+
+	object_heap_free(&driver_data->context_heap, (object_base_p) obj_context);
+
+	return VA_STATUS_SUCCESS;
+}
+