From 6dabe0f74a967fd6f4af7cab3da65b08f06cb088 Mon Sep 17 00:00:00 2001
From: Joakim Plate <elupus@ecce.se>
Date: Tue, 5 Jul 2011 00:18:32 +0200
Subject: [PATCH] Added ability to enable workaround for dxva2 decoding using
 older ATI cards

The workaround need to be enabled per pci id which can not
be detected inside ffmpeg. So this adds a flag that enabled
the alternate behavior.
---
 libavcodec/dxva2.h      |  2 ++
 libavcodec/dxva2_h264.c | 17 +++++++++++++++--
 2 files changed, 17 insertions(+), 2 deletions(-)

diff --git a/libavcodec/dxva2.h b/libavcodec/dxva2.h
index 5c5fe21e2f..6eb494b9fe 100644
--- a/libavcodec/dxva2.h
+++ b/libavcodec/dxva2.h
@@ -27,6 +27,8 @@
 
 #include <dxva2api.h>
 
+#define FF_DXVA2_WORKAROUND_SCALING_LIST_ZIGZAG 1 ///< Work around for DXVA2 and old UVD/UVD+ ATI video cards
+
 /**
  * This structure is used to provides the necessary configurations and data
  * to the DXVA2 FFmpeg HWAccel implementation.
diff --git a/libavcodec/dxva2_h264.c b/libavcodec/dxva2_h264.c
index bc80e982fb..21157bfe1e 100644
--- a/libavcodec/dxva2_h264.c
+++ b/libavcodec/dxva2_h264.c
@@ -113,6 +113,9 @@ static void fill_picture_parameters(struct dxva_context *ctx, const H264Context
 
     pp->bit_depth_luma_minus8         = h->sps.bit_depth_luma - 8;
     pp->bit_depth_chroma_minus8       = h->sps.bit_depth_chroma - 8;
+    if (ctx->workaround & FF_DXVA2_WORKAROUND_SCALING_LIST_ZIGZAG)
+        pp->Reserved16Bits            = 0;
+    else
     pp->Reserved16Bits                = 3; /* FIXME is there a way to detect the right mode ? */
     pp->StatusReportFeedbackNumber    = 1 + ctx->report_id++;
     pp->CurrFieldOrderCnt[0] = 0;
@@ -150,10 +153,19 @@ static void fill_picture_parameters(struct dxva_context *ctx, const H264Context
     //pp->SliceGroupMap[810];               /* XXX not implemented by FFmpeg */
 }
 
-static void fill_scaling_lists(const H264Context *h, DXVA_Qmatrix_H264 *qm)
+static void fill_scaling_lists(struct dxva_context *ctx, const H264Context *h, DXVA_Qmatrix_H264 *qm)
 {
     unsigned i, j;
     memset(qm, 0, sizeof(*qm));
+    if (ctx->workaround & FF_DXVA2_WORKAROUND_SCALING_LIST_ZIGZAG) {
+        for (i = 0; i < 6; i++)
+            for (j = 0; j < 16; j++)
+                qm->bScalingLists4x4[i][j] = h->pps.scaling_matrix4[i][j];
+
+        for (i = 0; i < 2; i++)
+            for (j = 0; j < 64; j++)
+                qm->bScalingLists8x8[i][j] = h->pps.scaling_matrix8[i][j];
+    } else {
     for (i = 0; i < 6; i++)
         for (j = 0; j < 16; j++)
             qm->bScalingLists4x4[i][j] = h->pps.scaling_matrix4[i][zigzag_scan[j]];
@@ -161,6 +173,7 @@ static void fill_scaling_lists(const H264Context *h, DXVA_Qmatrix_H264 *qm)
     for (i = 0; i < 2; i++)
         for (j = 0; j < 64; j++)
             qm->bScalingLists8x8[i][j] = h->pps.scaling_matrix8[i][ff_zigzag_direct[j]];
+    }
 }
 
 static int is_slice_short(struct dxva_context *ctx)
@@ -370,7 +383,7 @@ static int start_frame(AVCodecContext *avctx,
     fill_picture_parameters(ctx, h, &ctx_pic->pp);
 
     /* Fill up DXVA_Qmatrix_H264 */
-    fill_scaling_lists(h, &ctx_pic->qm);
+    fill_scaling_lists(ctx, h, &ctx_pic->qm);
 
     ctx_pic->slice_count    = 0;
     ctx_pic->bitstream_size = 0;