diff options
| author | Julien Isorce <j.isorce@samsung.com> | 2016-10-19 15:37:04 +0100 |
|---|---|---|
| committer | Víctor Manuel Jáquez Leal <victorx.jaquez@intel.com> | 2017-02-02 20:33:36 +0100 |
| commit | 33af1fc5786a179a22ecb9c43021ddd24c382263 (patch) | |
| tree | 120531284baa45459d55606d0f4b39a14387fdf7 | |
| parent | 69a2406a20335ce045c4c67a9ff98e2c96d143c9 (diff) | |
vaapivideomemory: release proxy's data if downstream
The surface created for downstream is going to be filled by VAAPI
elements. So, the driver needs write access on that surface.
This patch releases the derived image held by the proxy, thus the
surface is unmarked as busy.
This is how it has to be done as discussed on libva mailing list.
https://bugzilla.gnome.org/show_bug.cgi?id=755072
| -rw-r--r-- | gst/vaapi/gstvaapivideomemory.c | 8 |
1 files changed, 8 insertions, 0 deletions
diff --git a/gst/vaapi/gstvaapivideomemory.c b/gst/vaapi/gstvaapivideomemory.c index ed1fdf75..ebb0065f 100644 --- a/gst/vaapi/gstvaapivideomemory.c +++ b/gst/vaapi/gstvaapivideomemory.c @@ -992,6 +992,14 @@ gst_vaapi_dmabuf_memory_new (GstAllocator * base_allocator, gst_mini_object_set_qdata (GST_MINI_OBJECT_CAST (mem), GST_VAAPI_BUFFER_PROXY_QUARK, dmabuf_proxy, (GDestroyNotify) gst_vaapi_buffer_proxy_unref); + + /* When a VA surface is going to be filled by a VAAPI element + * (decoder or VPP), it has _not_ be marked as busy in the driver. + * Releasing the surface's derived image, held by the buffer proxy, + * the surface will be unmarked as busy. */ + if (allocator->direction == GST_PAD_SRC) + gst_vaapi_buffer_proxy_release_data (dmabuf_proxy); + return mem; /* ERRORS */ |