libavcodec/vvc/ctu.c

@@ -1444,20 +1444,25 @@ static void merge_data_block(VVCLocalContext *lc)
     }
 }
 
-static void merge_data_ibc(VVCLocalContext *lc)
+static int merge_data_ibc(VVCLocalContext *lc)
 {
     const VVCFrameContext* fc = lc->fc;
     const VVCSPS* sps         = fc->ps.sps;
     MotionInfo *mi            = &lc->cu->pu.mi;
     int merge_idx             = 0;
+    int ret;
 
     mi->pred_flag = PF_IBC;
 
     if (sps->max_num_ibc_merge_cand > 1)
         merge_idx = ff_vvc_merge_idx(lc);
 
-    ff_vvc_luma_mv_merge_ibc(lc, merge_idx, &mi->mv[L0][0]);
+    ret = ff_vvc_luma_mv_merge_ibc(lc, merge_idx, &mi->mv[L0][0]);
+    if (ret)
+        return ret;
     ff_vvc_store_mv(lc, mi);
+
+    return 0;
 }
 
 static int hls_merge_data(VVCLocalContext *lc)
@@ -1466,11 +1471,14 @@ static int hls_merge_data(VVCLocalContext *lc)
     const VVCPH  *ph            = &fc->ps.ph;
     const CodingUnit *cu        = lc->cu;
     PredictionUnit *pu          = &lc->cu->pu;
+    int ret;
 
     pu->merge_gpm_flag = 0;
     pu->mi.num_sb_x = pu->mi.num_sb_y = 1;
     if (cu->pred_mode == MODE_IBC) {
-        merge_data_ibc(lc);
+        ret = merge_data_ibc(lc);
+        if (ret)
+            return ret;
     } else {
         if (ph->max_num_subblock_merge_cand > 0 && cu->cb_width >= 8 && cu->cb_height >= 8)
             pu->merge_subblock_flag = ff_vvc_merge_subblock_flag(lc);
@@ -1596,6 +1604,7 @@ static int mvp_data_ibc(VVCLocalContext *lc)
     int mvp_l0_flag           = 0;
     int amvr_shift            = 4;
     Mv *mv                    = &mi->mv[L0][0];
+    int ret;
 
     mi->pred_flag = PF_IBC;
     mi->num_sb_x  = 1;
@@ -1607,7 +1616,9 @@ static int mvp_data_ibc(VVCLocalContext *lc)
     if (sps->r->sps_amvr_enabled_flag && (mv->x || mv->y))
         amvr_shift = ff_vvc_amvr_shift(lc, pu->inter_affine_flag, cu->pred_mode, 1);
 
-    ff_vvc_mvp_ibc(lc, mvp_l0_flag, amvr_shift, mv);
+    ret = ff_vvc_mvp_ibc(lc, mvp_l0_flag, amvr_shift, mv);
+    if (ret)
+        return ret;
     ff_vvc_store_mv(lc, mi);
 
     return 0;

libavcodec/vvc/dec.c

@@ -1022,7 +1022,7 @@ static av_cold int vvc_decode_init(AVCodecContext *avctx)
     static AVOnce init_static_once = AV_ONCE_INIT;
     const int cpu_count            = av_cpu_count();
     const int delayed              = FFMIN(cpu_count, VVC_MAX_DELAYED_FRAMES);
-    const int thread_count         = avctx->thread_count ? avctx->thread_count : delayed;
+    int thread_count               = avctx->thread_count ? avctx->thread_count : delayed;
     int ret;
 
     s->avctx = avctx;
@@ -1049,6 +1049,8 @@ static av_cold int vvc_decode_init(AVCodecContext *avctx)
             return ret;
     }
 
+    if (thread_count == 1)
+        thread_count = 0;
     s->executor = ff_vvc_executor_alloc(s, thread_count);
     if (!s->executor)
         return AVERROR(ENOMEM);

libavcodec/vvc/mvs.c

@@ -1695,17 +1695,34 @@ static void ibc_merge_candidates(VVCLocalContext *lc, const int merge_idx, Mv *m
     memset(mv, 0, sizeof(*mv));
 }
 
-void ff_vvc_mvp_ibc(VVCLocalContext *lc, const int mvp_l0_flag, const int amvr_shift, Mv *mv)
+static int ibc_check_mv(VVCLocalContext *lc, Mv *mv)
+{
+    const VVCFrameContext *fc = lc->fc;
+    const VVCSPS *sps         = lc->fc->ps.sps;
+    const CodingUnit *cu      = lc->cu;
+    const Mv *bv              = &cu->pu.mi.mv[L0][0];
+
+    if (sps->ctb_size_y < ((cu->y0 + (bv->y >> 4)) & (sps->ctb_size_y - 1)) + cu->cb_height) {
+        av_log(fc->log_ctx, AV_LOG_ERROR, "IBC region spans multiple CTBs.\n");
+        return AVERROR_INVALIDDATA;
+    }
+
+    return 0;
+}
+
+int ff_vvc_mvp_ibc(VVCLocalContext *lc, const int mvp_l0_flag, const int amvr_shift, Mv *mv)
 {
     LOCAL_ALIGNED_8(Mv, mvp, [1]);
 
     ibc_merge_candidates(lc, mvp_l0_flag, mvp);
     ibc_add_mvp(mv, mvp, amvr_shift);
+    return ibc_check_mv(lc, mv);
 }
 
-void ff_vvc_luma_mv_merge_ibc(VVCLocalContext *lc, const int merge_idx, Mv *mv)
+int ff_vvc_luma_mv_merge_ibc(VVCLocalContext *lc, const int merge_idx, Mv *mv)
 {
     ibc_merge_candidates(lc, merge_idx, mv);
+    return ibc_check_mv(lc, mv);
 }
 
 static int affine_mvp_constructed_cp(NeighbourContext *ctx,

libavcodec/vvc/mvs.h

@@ -30,9 +30,9 @@ void ff_vvc_clip_mv(Mv *mv);
 void ff_vvc_mv_scale(Mv *dst, const Mv *src, int td, int tb);
 void ff_vvc_luma_mv_merge_mode(VVCLocalContext *lc, int merge_idx, int ciip_flag, MvField *mv);
 void ff_vvc_luma_mv_merge_gpm(VVCLocalContext *lc, const int merge_gpm_idx[2], MvField *mv);
-void ff_vvc_luma_mv_merge_ibc(VVCLocalContext *lc, int merge_idx, Mv *mv);
+int ff_vvc_luma_mv_merge_ibc(VVCLocalContext *lc, int merge_idx, Mv *mv);
 void ff_vvc_mvp(VVCLocalContext *lc, const int *mvp_lx_flag, const int amvr_shift,  MotionInfo *mi);
-void ff_vvc_mvp_ibc(VVCLocalContext *lc, int mvp_l0_flag, int amvr_shift, Mv *mv);
+int ff_vvc_mvp_ibc(VVCLocalContext *lc, int mvp_l0_flag, int amvr_shift, Mv *mv);
 void ff_vvc_sb_mv_merge_mode(VVCLocalContext *lc, int merge_subblock_idx, PredictionUnit *pu);
 void ff_vvc_affine_mvp(VVCLocalContext *lc, const int *mvp_lx_flag, const int amvr_shift, MotionInfo* mi);
 void ff_vvc_store_sb_mvs(const VVCLocalContext *lc, PredictionUnit *pu);

libavcodec/vvc/thread.c

@@ -443,8 +443,11 @@ static int run_inter(VVCContext *s, VVCLocalContext *lc, VVCTask *t)
 {
     VVCFrameContext *fc = lc->fc;
     const CTU *ctu      = fc->tab.ctus + t->rs;
+    int ret;
 
-    ff_vvc_predict_inter(lc, t->rs);
+    ret = ff_vvc_predict_inter(lc, t->rs);
+    if (ret < 0)
+        return ret;
 
     if (ctu->has_dmvr)
         report_frame_progress(fc, t->ry, VVC_PROGRESS_MV);
@@ -454,9 +457,7 @@ static int run_inter(VVCContext *s, VVCLocalContext *lc, VVCTask *t)
 
 static int run_recon(VVCContext *s, VVCLocalContext *lc, VVCTask *t)
 {
-    ff_vvc_reconstruct(lc, t->rs, t->rx, t->ry);
-
-    return 0;
+    return ff_vvc_reconstruct(lc, t->rs, t->rx, t->ry);
 }
 
 static int run_lmcs(VVCContext *s, VVCLocalContext *lc, VVCTask *t)

libavutil/executor.c

@@ -82,9 +82,11 @@ static int run_one_task(AVExecutor *e, void *lc)
         /* nothing */;
     if (*prev) {
         AVTask *t = remove_task(prev, *prev);
-        ff_mutex_unlock(&e->lock);
+        if (e->thread_count > 0)
+            ff_mutex_unlock(&e->lock);
         cb->run(t, lc, cb->user_data);
-        ff_mutex_lock(&e->lock);
+        if (e->thread_count > 0)
+            ff_mutex_lock(&e->lock);
         return 1;
     }
     return 0;
@@ -146,14 +148,17 @@ AVExecutor* av_executor_alloc(const AVTaskCallbacks *cb, int thread_count)
         return NULL;
     e->cb = *cb;
 
-    e->local_contexts = av_calloc(thread_count, e->cb.local_context_size);
+    e->local_contexts = av_calloc(FFMAX(thread_count, 1), e->cb.local_context_size);
     if (!e->local_contexts)
         goto free_executor;
 
-    e->threads = av_calloc(thread_count, sizeof(*e->threads));
+    e->threads = av_calloc(FFMAX(thread_count, 1), sizeof(*e->threads));
     if (!e->threads)
         goto free_executor;
 
+    if (!thread_count)
+        return e;
+
     has_lock = !ff_mutex_init(&e->lock, NULL);
     has_cond = !ff_cond_init(&e->cond, NULL);
 
@@ -175,9 +180,12 @@ free_executor:
 
 void av_executor_free(AVExecutor **executor)
 {
+    int thread_count;
+
     if (!executor || !*executor)
         return;
-    executor_free(*executor, 1, 1);
+    thread_count = (*executor)->thread_count;
+    executor_free(*executor, thread_count, thread_count);
     *executor = NULL;
 }
 
@@ -195,9 +203,9 @@ void av_executor_execute(AVExecutor *e, AVTask *t)
     ff_cond_signal(&e->cond);
     ff_mutex_unlock(&e->lock);
 
-#if !HAVE_THREADS
-    // We are running in a single-threaded environment, so we must handle all tasks ourselves
-    while (run_one_task(e, e->local_contexts))
-        /* nothing */;
-#endif
+    if (!e->thread_count || !HAVE_THREADS) {
+        // We are running in a single-threaded environment, so we must handle all tasks ourselves
+        while (run_one_task(e, e->local_contexts))
+            /* nothing */;
+    }
 }

libavutil/executor.h

@@ -46,7 +46,7 @@ typedef struct AVTaskCallbacks {
 /**
  * Alloc executor
  * @param callbacks callback structure for executor
- * @param thread_count worker thread number
+ * @param thread_count worker thread number, 0 for run on caller's thread directly
  * @return return the executor
  */
 AVExecutor* av_executor_alloc(const AVTaskCallbacks *callbacks, int thread_count);