sub, vo_opengl: use packed sub-bitmaps directly if available

The previous few commits changed sd_lavc.c's output to packed RGB sub-
images. In particular, this means all sub-bitmaps are part of a larger,
single bitmap. Change the vo_opengl OSD code such that it can make use
of this, and upload the pre-packed image, instead of packing and copying
them again.

This complicates the upload code a bit (4 code paths due to messy PBO
handling). The plan is to make sub-bitmaps always packed, but some more
work is required to reach this point. The plan is to pack libass images
as well. Since this implies a copy, this will make it easy to refcount
the result.

(This is all targeted towards vo_opengl. Other VOs, vo_xv, vo_x11, and
vo_wayland in particular, will become less efficient. Although at least
vo_vdpau and vo_direct3d could be switched to the new method as well.)

3 files changed, 30 insertions, 1 deletions

diff --git a/sub/img_convert.c b/sub/img_convert.c
index 2015e49ca6..348811e773 100644
--- a/sub/img_convert.c
+++ b/sub/img_convert.c
@@ -80,6 +80,7 @@ bool osd_scale_rgba(struct osd_conv_cache *c, struct sub_bitmaps *imgs)
 
     talloc_free(c->parts);
     imgs->parts = c->parts = talloc_array(c, struct sub_bitmap, src.num_parts);
+    imgs->packed = NULL;
 
     // Note: we scale all parts, since most likely all need scaling anyway, and
     //       to get a proper copy of all data in the imgs list.
@@ -158,6 +159,7 @@ bool osd_conv_ass_to_rgba(struct osd_conv_cache *c, struct sub_bitmaps *imgs)
     imgs->format = SUBBITMAP_RGBA;
     imgs->parts = c->part;
     imgs->num_parts = num_bb;
+    imgs->packed = NULL;
 
     size_t newsize = 0;
     for (int n = 0; n < num_bb; n++) {
diff --git a/sub/osd.h b/sub/osd.h
index b50b72c587..7cfc695873 100644
--- a/sub/osd.h
+++ b/sub/osd.h
@@ -42,6 +42,11 @@ struct sub_bitmap {
     int x, y;
     int dw, dh;
 
+    // If the containing struct sub_bitmaps has the packed field set, then this
+    // is the position within the source. (Strictly speaking this is redundant
+    // with the bitmap pointer.)
+    int src_x, src_y;
+
     struct {
         uint32_t color;
     } libass;
@@ -60,6 +65,17 @@ struct sub_bitmaps {
     struct sub_bitmap *parts;
     int num_parts;
 
+    // Packed representation of the bitmap data. If non-NULL, then the
+    // parts[].bitmap pointer points into the image data here (and stride will
+    // correspond to packed->stride[0]).
+    //  SUBBITMAP_RGBA: IMGFMT_BGRA (exact match)
+    // Other formats have this set to NULL.
+    struct mp_image *packed;
+
+    // Bounding box for the packed image. All parts will be within the bounding
+    // box. (The origin of the box is at (0,0).)
+    int packed_w, packed_h;
+
     int change_id;  // Incremented on each change
 };
 
diff --git a/sub/sd_lavc.c b/sub/sd_lavc.c
index cc72efc54d..b7420d33c6 100644
--- a/sub/sd_lavc.c
+++ b/sub/sd_lavc.c
@@ -44,6 +44,7 @@ struct sub {
     struct sub_bitmap *inbitmaps;
     int count;
     struct mp_image *data;
+    int bound_w, bound_h;
     int src_w, src_h;
     double pts;
     double endpts;
@@ -254,7 +255,10 @@ static void read_sub_bitmaps(struct sd *sd, struct sub *sub)
     struct pos bb[2];
     packer_get_bb(priv->packer, bb);
 
-    if (!sub->data || sub->data->w < bb[1].x || sub->data->h < bb[1].y) {
+    sub->bound_w = bb[1].x;
+    sub->bound_h = bb[1].y;
+
+    if (!sub->data || sub->data->w < sub->bound_w || sub->data->h < sub->bound_h) {
         talloc_free(sub->data);
         sub->data = mp_image_alloc(IMGFMT_BGRA, priv->packer->w, priv->packer->h);
         if (!sub->data) {
@@ -279,6 +283,8 @@ static void read_sub_bitmaps(struct sd *sd, struct sub *sub)
         b->h = r->h;
         b->x = r->x;
         b->y = r->y;
+        b->src_x = pos.x;
+        b->src_y = pos.y;
         b->stride = sub->data->stride[0];
         b->bitmap = sub->data->planes[0] + pos.y * b->stride + pos.x * 4;
 
@@ -307,6 +313,8 @@ static void read_sub_bitmaps(struct sd *sd, struct sub *sub)
         }
 
         b->bitmap = (char*)b->bitmap - extend * b->stride - extend * 4;
+        b->src_x -= extend;
+        b->src_y -= extend;
         b->x -= extend;
         b->y -= extend;
         b->w += extend * 2;
@@ -443,6 +451,9 @@ static void get_bitmaps(struct sd *sd, struct mp_osd_res d, double pts,
     if (priv->displayed_id != current->id)
         res->change_id++;
     priv->displayed_id = current->id;
+    res->packed = current->data;
+    res->packed_w = current->bound_w;
+    res->packed_h = current->bound_h;
     res->format = SUBBITMAP_RGBA;
 
     double video_par = 0;