RGBCube/serenity
1
Fork 0
mirror of https://github.com/RGBCube/serenity synced 2025-07-26 12:07:45 +00:00
Code Issues Activity Actions

LibVideo/VP9: Add FIXME about implementation of tiled decoding

Browse source
This commit is contained in:
Zaggy1024 2022-11-28 01:39:42 -06:00 • committed by Andreas Kling
parent 2f043a0bd4
commit 9df72080a1
1 changed files with 4 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

4
Userland/Libraries/LibVideo/VP9/Parser.cpp
View file

@ -852,6 +852,10 @@ DecoderErrorOr<void> Parser::decode_tiles(FrameContext& frame_context)
clear_above_context(frame_context);
NonZeroTokens above_non_zero_tokens = DECODER_TRY_ALLOC(create_non_zero_tokens(blocks_to_sub_blocks(frame_context.columns()), frame_context.color_config.subsampling_x));
// FIXME: To implement tiled decoding, we'll need to pre-parse the tile positions and sizes into a 2D vector of ReadonlyBytes,
// then run through each column of tiles in top to bottom order afterward. Each column can be sent to a worker thread
// for execution. Each worker thread will want to create a set of above contexts sized to its tile width, then provide
// those to each tile as it decodes them.
for (auto tile_row = 0; tile_row < tile_rows; tile_row++) {
for (auto tile_col = 0; tile_col < tile_cols; tile_col++) {
auto last_tile = (tile_row == tile_rows - 1) && (tile_col == tile_cols - 1);