[FFmpeg-devel] [PATCH] lavc: Add device context field to AVCodecContext

[Mark Thompson]

For use by codec implementations which can allocate frames internally.
---
On 01/02/17 14:03, Timo Rothenpieler wrote:
> Am 17.01.2017 um 23:28 schrieb Mark Thompson:
>> For use by codec implementations which require a device to operation
>> but will allocate frames internally.
>> ---
>> The following patches make use of it for qsv.  Is this definition appropriate to solve the issues with cuvid device setup as well?
>>
>>
>>  doc/APIchanges       |  3 +++
>>  libavcodec/avcodec.h | 38 ++++++++++++++++++++++++++++++++++++--
>>  libavcodec/utils.c   |  1 +
>>  libavcodec/version.h |  2 +-
>>  4 files changed, 41 insertions(+), 3 deletions(-)
> 
> ping about this.
> I'd really like to use it for cuvid and get rid of some hackery.

This got more discussion in libav, see below for a patch expressing the currently-proposed semantics.

(See also <https://lists.libav.org/pipermail/libav-devel/2017-February/082084.html> for a full implementation for VAAPI, I haven't ported it here yet.)

Thanks,

- Mark


 libavcodec/avcodec.h | 22 +++++++++++++++++++++-
 1 file changed, 21 insertions(+), 1 deletion(-)

diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
index 1e681e989b..cd83e6b27f 100644
--- a/libavcodec/avcodec.h
+++ b/libavcodec/avcodec.h
@@ -3533,7 +3533,8 @@ typedef struct AVCodecContext {
     /**
      * A reference to the AVHWFramesContext describing the input (for encoding)
      * or output (decoding) frames. The reference is set by the caller and
-     * afterwards owned (and freed) by libavcodec.
+     * afterwards owned (and freed) by libavcodec - it should never be read by
+     * the caller after being set.
      *
      * - decoding: This field should be set by the caller from the get_format()
      *             callback. The previous reference (if any) will always be
@@ -3583,6 +3584,25 @@ typedef struct AVCodecContext {
      */
     int64_t max_pixels;
 
+    /**
+     * A reference to the AVHWDeviceContext describing the device which will
+     * be used by a hardware encoder/decoder.  The reference is set by the
+     * caller and afterwards owned (and freed) by libavcodec - it should never
+     * be read by the caller after being set.
+     *
+     * This should only be used if either the codec device does not require
+     * hardware frames or any that are used are allocated internally by
+     * libavcodec.  If the user wishes to supply any of the frames used as
+     * encoder input or decoder output then hw_frames_ctx should be used
+     * instead.
+     *
+     * - decoding: This field should be set by the caller from the get_format()
+     *             callback.  The previous reference (if any) will always be
+     *             unreffed by libavcodec before the get_format() call.
+     *
+     * - encoding: This field should be set before avcodec_open2() is called.
+     */
+    AVBufferRef *hw_device_ctx;
 } AVCodecContext;
 
 AVRational av_codec_get_pkt_timebase         (const AVCodecContext *avctx);
-- 
2.11.0