From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from ffbox0-bg.ffmpeg.org (ffbox0-bg.ffmpeg.org [79.124.17.100]) by master.gitmailbox.com (Postfix) with ESMTPS id CD68F4F0F1 for ; Sun, 15 Jun 2025 17:02:30 +0000 (UTC) Received: from [127.0.1.1] (localhost [127.0.0.1]) by ffbox0-bg.ffmpeg.org (Postfix) with ESMTP id 66F2668D9AA; Sun, 15 Jun 2025 19:58:26 +0300 (EEST) Received: from mail-pl1-f173.google.com (mail-pl1-f173.google.com [209.85.214.173]) by ffbox0-bg.ffmpeg.org (Postfix) with ESMTPS id A2F7168D972 for ; Sun, 15 Jun 2025 19:58:24 +0300 (EEST) Received: by mail-pl1-f173.google.com with SMTP id d9443c01a7336-234bfe37cccso46717455ad.0 for ; Sun, 15 Jun 2025 09:58:24 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20230601; t=1750006703; x=1750611503; darn=ffmpeg.org; h=cc:to:mime-version:content-transfer-encoding:fcc:subject:date :references:in-reply-to:message-id:from:from:to:cc:subject:date :message-id:reply-to; bh=8grzezaY5DbfI39doXCy++/qhhg93MuMIfsq94Fs1fQ=; b=kLD+y8JKub2CW9kcQq0KKko2O9v/juSoaiThbC29cUvu5XF6xP6241ZwiVIi/FXO6W cbWOXsyoglwgCj3avbdF0X5lXK64/FV9BBKaOy8T2sKs5l651YdWvdvjtt6mlD0TEy8o NECQ2i/xHWX9ggVkCCECoqIKFvkkG91RjKkfPXCMcqQDVSurzAL8nALjGJsNUCDcFX9L UwzKJKKm+Q3NtLxtGDZl2mdrlZmDxZU5CGT21jSOVBqxk2hRB9dY5JGitqryR0OFNtwN GdYiqWOxsy9OMbUma3wVA1MxAk/doBI4ZXfnkx596lufub8PQxycHoXIGmXXCYfNLp4L CV6A== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1750006703; x=1750611503; h=cc:to:mime-version:content-transfer-encoding:fcc:subject:date :references:in-reply-to:message-id:from:x-gm-message-state:from:to :cc:subject:date:message-id:reply-to; bh=8grzezaY5DbfI39doXCy++/qhhg93MuMIfsq94Fs1fQ=; b=sbe5BNeeV9MvXEEYIpR66oUuj2TiZ7usDpqWGArejCofzUiI5Drh5HIKfy6v6wshIi JbtLnF6ozqRwV61MH95f5fp48+pIchOCE7XOWaOZQu3UfWE4fJqVYfiQtQNNU8u1WvrD BZhC7Z0ygQyA8TeV3MjGh8zk+26gXYAr7MDxIovi61X4haW6Cu3FHo08mY2SFHXn63UF lwIOeIMXTX51L6aC1m9hSaXxCn6s+Ek338Vlh1XhJsghpI7KYUPD951He0wxrK3T5Z9+ MvB/VeHeksFXSZp0rzMOx+p1DVc+8sDq+nFFNKBwbRsjD7hK3290mxYfykhtS4fAKQAP SFyQ== X-Gm-Message-State: AOJu0YxxFgNQ8FUu4pjZmatojfbaHCbn8wVS/uZrEG6H4l1CKWzjE7Hu LmBcCeK9MRaPmfu1+UOxefKnH6xTD5d6Jufof7XD1kq7ojVcXfV2BLxfI4S9iQ== X-Gm-Gg: ASbGncuyQrbNW5XX7maMxhHNs24w0jTKBNw65bPghRcg2dXuCjIcZ6miEYl86vbXY6W XEIijjuFbKeF4nQtZ0VttJYjAPaoy0rG/PkcQ+2YubB/6MGR0418xJca02NvvQrbmM/MhfvI9I5 u1AO+4FTJ/9mwimQsIk5/sS++hF0aIi/HHYev/fWZuCS/r1EFQWUhGzFBntSY3NvVPeXe4XrnFB jENF4evTVqXFf2ObGXvVi05miuoxXZdCi56HJAKy7eXv0CBVVdYBI0DZrqC95YMJGDd6rRL8bIf LjqbIICjhuB/LmO1ohMHzf0l7M9FHi10a9hVjUBVu4dEm5aaAcmgCw4BMm0LaeF5cOK1ktYGUfE 7ID/XBw4CFkY7xPy4 X-Google-Smtp-Source: AGHT+IHQ4vy3IAgBodNPuFzAfNW1/0c8YqLJ+lyFJcR7OvnrL2YHzJ6ZvU4KedBbqr+c5sqYc2WyyA== X-Received: by 2002:a17:902:da81:b0:234:a063:e2c1 with SMTP id d9443c01a7336-2366b17b667mr91432935ad.42.1750006702763; Sun, 15 Jun 2025 09:58:22 -0700 (PDT) Received: from [127.0.0.1] (master.gitmailbox.com. [34.83.118.50]) by smtp.gmail.com with ESMTPSA id d9443c01a7336-2365d8a4d82sm46404945ad.72.2025.06.15.09.58.22 (version=TLS1_2 cipher=ECDHE-ECDSA-AES128-GCM-SHA256 bits=128/128); Sun, 15 Jun 2025 09:58:22 -0700 (PDT) From: Andreas Rheinhardt X-Google-Original-From: Andreas Rheinhardt Message-Id: <12acaf174811b9615db9cfe9201e3f11081f3229.1750006480.git.ffmpegagent@gmail.com> In-Reply-To: References: Date: Sun, 15 Jun 2025 16:54:29 +0000 Fcc: Sent MIME-Version: 1.0 To: ffmpeg-devel@ffmpeg.org Subject: [FFmpeg-devel] [PATCH 38/47] avcodec/mpegvideo: Allocate ac_val jointly X-BeenThere: ffmpeg-devel@ffmpeg.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: FFmpeg development discussions and patches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Reply-To: FFmpeg development discussions and patches Cc: Andreas Rheinhardt Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Errors-To: ffmpeg-devel-bounces@ffmpeg.org Sender: "ffmpeg-devel" Archived-At: List-Archive: List-Post: From: Andreas Rheinhardt They are currently allocated separately per slice; allocating them jointly allows to avoid saving them in ff_update_duplicate_context(). The way it is done also avoids allocating ac_val for encoders that don't need it (e.g. H.263 or H.263+ with AIC). This entailed moving setting nb_slices to ff_mpv_init_context_frame() which is called from ff_mpv_common_frame_size_change(). The resultant nb_slices will always be one when called from a decoder using ff_mpv_common_frame_size_change(). Signed-off-by: Andreas Rheinhardt --- libavcodec/mpegvideo.c | 60 ++++++++++++++++++-------------------- libavcodec/mpegvideo_enc.c | 36 +++++++++++++++-------- 2 files changed, 52 insertions(+), 44 deletions(-) diff --git a/libavcodec/mpegvideo.c b/libavcodec/mpegvideo.c index e528935f68..ae0c847aa4 100644 --- a/libavcodec/mpegvideo.c +++ b/libavcodec/mpegvideo.c @@ -120,18 +120,6 @@ static av_cold int init_duplicate_context(MpegEncContext *s) return AVERROR(ENOMEM); s->block = s->blocks[0]; - if (s->out_format == FMT_H263) { - int mb_height = s->msmpeg4_version == MSMP4_VC1 ? - FFALIGN(s->mb_height, 2) : s->mb_height; - int y_size = s->b8_stride * (2 * mb_height + 1); - int c_size = s->mb_stride * (mb_height + 1); - int yc_size = y_size + 2 * c_size; - /* ac values */ - if (!FF_ALLOCZ_TYPED_ARRAY(s->ac_val_base, yc_size)) - return AVERROR(ENOMEM); - s->ac_val = s->ac_val_base + s->b8_stride + 1; - } - return 0; } @@ -171,7 +159,6 @@ static av_cold void free_duplicate_context(MpegEncContext *s) s->sc.linesize = 0; av_freep(&s->blocks); - av_freep(&s->ac_val_base); s->block = NULL; } @@ -192,7 +179,6 @@ static void backup_duplicate_context(MpegEncContext *bak, MpegEncContext *src) COPY(block); COPY(start_mb_y); COPY(end_mb_y); - COPY(ac_val_base); COPY(ac_val); #undef COPY } @@ -245,15 +231,34 @@ static av_cold void free_buffer_pools(BufferPoolContext *pools) av_cold int ff_mpv_init_context_frame(MpegEncContext *s) { + int nb_slices = (HAVE_THREADS && + s->avctx->active_thread_type & FF_THREAD_SLICE) ? + s->avctx->thread_count : 1; BufferPoolContext *const pools = &s->buffer_pools; int y_size, c_size, yc_size, i, mb_array_size, mv_table_size, x, y; int mb_height; + if (s->encoding && s->avctx->slices) + nb_slices = s->avctx->slices; + if (s->codec_id == AV_CODEC_ID_MPEG2VIDEO && !s->progressive_sequence) s->mb_height = (s->height + 31) / 32 * 2; else s->mb_height = (s->height + 15) / 16; + if (nb_slices > MAX_THREADS || (nb_slices > s->mb_height && s->mb_height)) { + int max_slices; + if (s->mb_height) + max_slices = FFMIN(MAX_THREADS, s->mb_height); + else + max_slices = MAX_THREADS; + av_log(s->avctx, AV_LOG_WARNING, "too many threads/slices (%d)," + " reducing to %d\n", nb_slices, max_slices); + nb_slices = max_slices; + } + + s->slice_context_count = nb_slices; + /* VC-1 can change from being progressive to interlaced on a per-frame * basis. We therefore allocate certain buffers so big that they work * in both instances. */ @@ -332,6 +337,14 @@ av_cold int ff_mpv_init_context_frame(MpegEncContext *s) } if (s->h263_pred || s->h263_aic || !s->encoding) { + size_t allslice_yc_size = yc_size * (s->encoding ? nb_slices : 1); + if (s->out_format == FMT_H263) { + /* ac values */ + if (!FF_ALLOCZ_TYPED_ARRAY(s->ac_val_base, allslice_yc_size)) + return AVERROR(ENOMEM); + s->ac_val = s->ac_val_base + s->b8_stride + 1; + } + /* dc values */ // MN: we need these for error resilience of intra-frames // Allocating them unconditionally for decoders also means @@ -381,14 +394,8 @@ av_cold int ff_mpv_init_context_frame(MpegEncContext *s) */ av_cold int ff_mpv_common_init(MpegEncContext *s) { - int nb_slices = (HAVE_THREADS && - s->avctx->active_thread_type & FF_THREAD_SLICE) ? - s->avctx->thread_count : 1; int ret; - if (s->encoding && s->avctx->slices) - nb_slices = s->avctx->slices; - if (s->avctx->pix_fmt == AV_PIX_FMT_NONE) { av_log(s->avctx, AV_LOG_ERROR, "decoding to AV_PIX_FMT_NONE is not supported.\n"); @@ -411,20 +418,8 @@ av_cold int ff_mpv_common_init(MpegEncContext *s) if ((ret = ff_mpv_init_context_frame(s))) goto fail; - if (nb_slices > MAX_THREADS || (nb_slices > s->mb_height && s->mb_height)) { - int max_slices; - if (s->mb_height) - max_slices = FFMIN(MAX_THREADS, s->mb_height); - else - max_slices = MAX_THREADS; - av_log(s->avctx, AV_LOG_WARNING, "too many threads/slices (%d)," - " reducing to %d\n", nb_slices, max_slices); - nb_slices = max_slices; - } - s->context_initialized = 1; s->thread_context[0] = s; - s->slice_context_count = nb_slices; // if (s->width && s->height) { if (!s->encoding) { @@ -450,6 +445,7 @@ av_cold void ff_mpv_free_context_frame(MpegEncContext *s) for (int j = 0; j < 2; j++) s->p_field_mv_table[i][j] = NULL; + av_freep(&s->ac_val_base); av_freep(&s->dc_val_base); av_freep(&s->coded_block_base); av_freep(&s->mbintra_table); diff --git a/libavcodec/mpegvideo_enc.c b/libavcodec/mpegvideo_enc.c index 870cf7f5a7..3bec6b57cf 100644 --- a/libavcodec/mpegvideo_enc.c +++ b/libavcodec/mpegvideo_enc.c @@ -518,24 +518,36 @@ static av_cold int init_slice_buffers(MPVMainEncContext *const m) static_assert(DCT_ERROR_SIZE * MAX_THREADS + ALIGN - 1 <= SIZE_MAX, "Need checks for potential overflow."); unsigned nb_slices = s->c.slice_context_count; + char *dct_error = NULL; - if (!m->noise_reduction) - return 0; + if (m->noise_reduction) { + if (!FF_ALLOCZ_TYPED_ARRAY(s->dct_offset, 2)) + return AVERROR(ENOMEM); + dct_error = av_mallocz(ALIGN - 1 + nb_slices * DCT_ERROR_SIZE); + if (!dct_error) + return AVERROR(ENOMEM); + m->dct_error_sum_base = dct_error; + dct_error += FFALIGN((uintptr_t)dct_error, ALIGN) - (uintptr_t)dct_error; + } - if (!FF_ALLOCZ_TYPED_ARRAY(s->dct_offset, 2)) - return AVERROR(ENOMEM); - char *dct_error = av_mallocz(ALIGN - 1 + nb_slices * DCT_ERROR_SIZE); - if (!dct_error) - return AVERROR(ENOMEM); - m->dct_error_sum_base = dct_error; - dct_error += FFALIGN((uintptr_t)dct_error, ALIGN) - (uintptr_t)dct_error; + const int y_size = s->c.b8_stride * (2 * s->c.mb_height + 1); + const int c_size = s->c.mb_stride * (s->c.mb_height + 1); + const int yc_size = y_size + 2 * c_size; + ptrdiff_t offset = 0; for (unsigned i = 0; i < nb_slices; ++i) { MPVEncContext *const s2 = s->c.enc_contexts[i]; - s2->dct_offset = s->dct_offset; - s2->dct_error_sum = (void*)dct_error; - dct_error += DCT_ERROR_SIZE; + if (dct_error) { + s2->dct_offset = s->dct_offset; + s2->dct_error_sum = (void*)dct_error; + dct_error += DCT_ERROR_SIZE; + } + + if (s2->c.ac_val) { + s2->c.ac_val += offset; + offset += yc_size; + } } return 0; } -- ffmpeg-codebot _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-request@ffmpeg.org with subject "unsubscribe".