va_trace_surface: dumpe surface content after vaEndPicture call

vaEndPicture will kick HW decode/encode, and at this time use CPU
to dump surface content can keep GPU/CPU in parallel.

Also for encode trace, don't need to wait current frame completions
but for decode, we must wait decode till it is completed

Signed-off-by: Austin Yuan <shengquan.yuan@intel.com>
(cherry picked from commit 0af365de4c03c12fc20183093a944f7513a1c7be)

1 files changed, 5 insertions, 8 deletions

diff --git a/va/va.c b/va/va.c
index e331f74..6210112 100644
--- a/va/va.c
+++ b/va/va.c
@@ -1069,19 +1069,16 @@ VAStatus vaEndPicture (
     VAContextID context
 )
 {
-  VAStatus va_status;
+  VAStatus va_status = VA_STATUS_SUCCESS;
   VADriverContextP ctx;
 
   CHECK_DISPLAY(dpy);
   ctx = CTX(dpy);
 
-  /* dump encode source surface */
-  VA_TRACE_SURFACE(va_TraceEndPicture, dpy, context, 0);
-  /* return directly if do dummy operation */
-  VA_FOOL_RETURN();
-  
-  va_status = ctx->vtable->vaEndPicture( ctx, context );
-  /* dump decode dest surface */
+  if (fool_codec == 0)
+      va_status = ctx->vtable->vaEndPicture( ctx, context );
+
+  /* dump surface content */
   VA_TRACE_SURFACE(va_TraceEndPicture, dpy, context, 1);
 
   return va_status;