Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
 help / color / mirror / Atom feed
From: Marton Balint <cus@passwd.hu>
To: ffmpeg-devel@ffmpeg.org
Cc: Marton Balint <cus@passwd.hu>
Subject: [FFmpeg-devel] [PATCH 11/12] fftools/ffmpeg_filter: always reap all available frames before requesting new ones
Date: Tue, 24 Jun 2025 21:23:09 +0200
Message-ID: <20250624192318.7430-11-cus@passwd.hu> (raw)
In-Reply-To: <20250624192318.7430-1-cus@passwd.hu>

alfilter_graph_request_oldest() might return EAGAIN and produce a frame on not
the oldest sink.

Fixes ticket #11597.
Fixes excessive frame buffering in #10959.
Fixes excessive frame buffering in #11366.

Signed-off-by: Marton Balint <cus@passwd.hu>
---
 fftools/ffmpeg_filter.c | 31 ++++++++++++++++---------------
 1 file changed, 16 insertions(+), 15 deletions(-)

diff --git a/fftools/ffmpeg_filter.c b/fftools/ffmpeg_filter.c
index a0dc4c745e..f6e496158c 100644
--- a/fftools/ffmpeg_filter.c
+++ b/fftools/ffmpeg_filter.c
@@ -2669,6 +2669,22 @@ static int read_frames(FilterGraph *fg, FilterGraphThread *fgt,
     while (fgp->nb_outputs_done < fg->nb_outputs) {
         int ret;
 
+        /* Reap all buffers present in the buffer sinks */
+        for (int i = 0; i < fg->nb_outputs; i++) {
+            OutputFilterPriv *ofp = ofp_from_ofilter(fg->outputs[i]);
+
+            ret = 0;
+            while (!ret) {
+                ret = fg_output_step(ofp, fgt, frame);
+                if (ret < 0)
+                    return ret;
+            }
+        }
+
+        // return after one iteration, so that scheduler can rate-control us
+        if (did_step && fgp->have_sources)
+            return 0;
+
         ret = avfilter_graph_request_oldest(fgt->graph);
         if (ret == AVERROR(EAGAIN)) {
             fgt->next_in = choose_input(fg, fgt);
@@ -2684,21 +2700,6 @@ static int read_frames(FilterGraph *fg, FilterGraphThread *fgt,
         }
         fgt->next_in = fg->nb_inputs;
 
-        // return after one iteration, so that scheduler can rate-control us
-        if (did_step && fgp->have_sources)
-            return 0;
-
-        /* Reap all buffers present in the buffer sinks */
-        for (int i = 0; i < fg->nb_outputs; i++) {
-            OutputFilterPriv *ofp = ofp_from_ofilter(fg->outputs[i]);
-
-            ret = 0;
-            while (!ret) {
-                ret = fg_output_step(ofp, fgt, frame);
-                if (ret < 0)
-                    return ret;
-            }
-        }
         did_step = 1;
     }
 
-- 
2.43.0

_______________________________________________
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".

  parent reply	other threads:[~2025-06-24 19:25 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-06-24 19:22 [FFmpeg-devel] [PATCH 01/12] avfilter: factorize requesting an input frame from multi output filters Marton Balint
2025-06-24 19:23 ` [FFmpeg-devel] [PATCH 02/12] avfilter/filters: simplify FF_FILTER_FORWARD_WANTED_ANY Marton Balint
2025-06-24 19:23 ` [FFmpeg-devel] [PATCH 03/12] avfilter/avfilter: fix forwarding EOF for simple API filters in filter_activate_default Marton Balint
2025-06-24 19:23 ` [FFmpeg-devel] [PATCH 04/12] fate/filter-video: add ffprobe test for dual output select filter Marton Balint
2025-06-24 19:23 ` [FFmpeg-devel] [PATCH 05/12] avfilter/avfilter: always forward request frame in filter_activate_default Marton Balint
2025-06-24 19:23 ` [FFmpeg-devel] [PATCH 06/12] avfilter/avfilter: make filter_activate_default request frames on behalf of sinks Marton Balint
2025-06-24 19:23 ` [FFmpeg-devel] [PATCH 07/12] tests/fate/filter-audio: add anullsink test Marton Balint
2025-06-24 19:23 ` [FFmpeg-devel] [PATCH 08/12] avfilter: signal an empty buffersrc with an explicit activate error code Marton Balint
2025-06-24 19:23 ` [FFmpeg-devel] [PATCH 09/12] avfilter/buffersink: keep requesting frames if one activation of the graph does not provide one Marton Balint
2025-06-24 19:23 ` [FFmpeg-devel] [PATCH 10/12] avfilter/f_select: port to activate Marton Balint
2025-06-24 19:23 ` Marton Balint [this message]
2025-06-24 19:23 ` [FFmpeg-devel] [PATCH 12/12] avfilter/ffmpeg_filter: rate control all filter graphs Marton Balint

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20250624192318.7430-11-cus@passwd.hu \
    --to=cus@passwd.hu \
    --cc=ffmpeg-devel@ffmpeg.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Git Inbox Mirror of the ffmpeg-devel mailing list - see https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://master.gitmailbox.com/ffmpegdev/0 ffmpegdev/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 ffmpegdev ffmpegdev/ https://master.gitmailbox.com/ffmpegdev \
		ffmpegdev@gitmailbox.com
	public-inbox-index ffmpegdev

Example config snippet for mirrors.


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git