From: James Almer <jamrial@gmail.com>
To: ffmpeg-devel@ffmpeg.org
Subject: Re: [FFmpeg-devel] [PATCH 2/5] avcodec/mjpegdec: Don't create unnecessary AVFrame reference
Date: Thu, 14 Apr 2022 13:39:46 -0300
Message-ID: <ead0259b-2694-12e5-cfa0-dad2a0d73175@gmail.com> (raw)
In-Reply-To: <AS8PR01MB79442B849DFE6AC6C1F5DE178FEF9@AS8PR01MB7944.eurprd01.prod.exchangelabs.com>
On 4/14/2022 12:57 PM, Andreas Rheinhardt wrote:
> MJPEG is an intra-codec, so it makes no sense to keep the reference.
> It is basically unused lateron anyway.
Wouldn't this mean you can remove the AV_GET_BUFFER_FLAG_REF flag from
the ff_get_buffer() call?
>
> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
> ---
> libavcodec/mjpegdec.c | 45 +++++++++++++++++++++----------------------
> 1 file changed, 22 insertions(+), 23 deletions(-)
>
> diff --git a/libavcodec/mjpegdec.c b/libavcodec/mjpegdec.c
> index 0e76bf4c26..939dedce33 100644
> --- a/libavcodec/mjpegdec.c
> +++ b/libavcodec/mjpegdec.c
> @@ -2589,8 +2589,7 @@ eoi_parser:
>
> av_freep(&s->hwaccel_picture_private);
> }
> - if ((ret = av_frame_ref(frame, s->picture_ptr)) < 0)
> - return ret;
> + av_frame_move_ref(frame, s->picture_ptr);
> s->got_picture = 0;
>
> frame->pkt_dts = s->pkt->dts;
> @@ -2675,9 +2674,9 @@ the_end:
> if (ret)
> return ret;
>
> - av_assert0(s->nb_components == av_pix_fmt_count_planes(s->picture_ptr->format));
> + av_assert0(s->nb_components == av_pix_fmt_count_planes(frame->format));
> for (p = 0; p<s->nb_components; p++) {
> - uint8_t *line = s->picture_ptr->data[p];
> + uint8_t *line = frame->data[p];
> int w = s->width;
> int h = s->height;
> if (!s->upscale_h[p])
> @@ -2737,7 +2736,7 @@ the_end:
> if (ret)
> return ret;
>
> - av_assert0(s->nb_components == av_pix_fmt_count_planes(s->picture_ptr->format));
> + av_assert0(s->nb_components == av_pix_fmt_count_planes(frame->format));
> for (p = 0; p < s->nb_components; p++) {
> uint8_t *dst;
> int w = s->width;
> @@ -2748,10 +2747,10 @@ the_end:
> w = AV_CEIL_RSHIFT(w, hshift);
> h = AV_CEIL_RSHIFT(h, vshift);
> }
> - dst = &((uint8_t *)s->picture_ptr->data[p])[(h - 1) * s->linesize[p]];
> + dst = &frame->data[p][(h - 1) * s->linesize[p]];
> for (i = h - 1; i; i--) {
> - uint8_t *src1 = &((uint8_t *)s->picture_ptr->data[p])[i * s->upscale_v[p] / (s->upscale_v[p] + 1) * s->linesize[p]];
> - uint8_t *src2 = &((uint8_t *)s->picture_ptr->data[p])[(i + 1) * s->upscale_v[p] / (s->upscale_v[p] + 1) * s->linesize[p]];
> + uint8_t *src1 = &frame->data[p][ i * s->upscale_v[p] / (s->upscale_v[p] + 1) * s->linesize[p]];
> + uint8_t *src2 = &frame->data[p][(i + 1) * s->upscale_v[p] / (s->upscale_v[p] + 1) * s->linesize[p]];
> if (s->upscale_v[p] != 2 && (src1 == src2 || i == h - 1)) {
> memcpy(dst, src1, w);
> } else {
> @@ -2768,36 +2767,36 @@ the_end:
> if (ret)
> return ret;
>
> - av_assert0(s->nb_components == av_pix_fmt_count_planes(s->picture_ptr->format));
> + av_assert0(s->nb_components == av_pix_fmt_count_planes(frame->format));
> for (index=0; index<s->nb_components; index++) {
> - uint8_t *dst = s->picture_ptr->data[index];
> - int w = s->picture_ptr->width;
> - int h = s->picture_ptr->height;
> + uint8_t *dst = frame->data[index];
> + int w = frame->width;
> + int h = frame->height;
> if(index && index<3){
> w = AV_CEIL_RSHIFT(w, hshift);
> h = AV_CEIL_RSHIFT(h, vshift);
> }
> if(dst){
> - uint8_t *dst2 = dst + s->picture_ptr->linesize[index]*(h-1);
> + uint8_t *dst2 = dst + frame->linesize[index]*(h-1);
> for (i=0; i<h/2; i++) {
> for (j=0; j<w; j++)
> FFSWAP(int, dst[j], dst2[j]);
> - dst += s->picture_ptr->linesize[index];
> - dst2 -= s->picture_ptr->linesize[index];
> + dst += frame->linesize[index];
> + dst2 -= frame->linesize[index];
> }
> }
> }
> }
> if (s->adobe_transform == 0 && s->avctx->pix_fmt == AV_PIX_FMT_GBRAP) {
> - int w = s->picture_ptr->width;
> - int h = s->picture_ptr->height;
> + int w = frame->width;
> + int h = frame->height;
> av_assert0(s->nb_components == 4);
> for (i=0; i<h; i++) {
> int j;
> uint8_t *dst[4];
> for (index=0; index<4; index++) {
> - dst[index] = s->picture_ptr->data[index]
> - + s->picture_ptr->linesize[index]*i;
> + dst[index] = frame->data[index]
> + + frame->linesize[index]*i;
> }
> for (j=0; j<w; j++) {
> int k = dst[3][j];
> @@ -2812,15 +2811,15 @@ the_end:
> }
> }
> if (s->adobe_transform == 2 && s->avctx->pix_fmt == AV_PIX_FMT_YUVA444P) {
> - int w = s->picture_ptr->width;
> - int h = s->picture_ptr->height;
> + int w = frame->width;
> + int h = frame->height;
> av_assert0(s->nb_components == 4);
> for (i=0; i<h; i++) {
> int j;
> uint8_t *dst[4];
> for (index=0; index<4; index++) {
> - dst[index] = s->picture_ptr->data[index]
> - + s->picture_ptr->linesize[index]*i;
> + dst[index] = frame->data[index]
> + + frame->linesize[index]*i;
> }
> for (j=0; j<w; j++) {
> int k = dst[3][j];
_______________________________________________
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".
next prev parent reply other threads:[~2022-04-14 16:39 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-04-14 15:56 [FFmpeg-devel] [PATCH 1/5] avcodec/mjpegdec: Always reset got_picture at the beginnig of decoding Andreas Rheinhardt
2022-04-14 15:57 ` [FFmpeg-devel] [PATCH 2/5] avcodec/mjpegdec: Don't create unnecessary AVFrame reference Andreas Rheinhardt
2022-04-14 16:39 ` James Almer [this message]
2022-04-14 16:43 ` Andreas Rheinhardt
2022-04-14 16:46 ` James Almer
2022-04-14 16:51 ` Andreas Rheinhardt
2022-04-14 15:57 ` [FFmpeg-devel] [PATCH 3/5] avcodec/mjpegdec: Avoid copying data when flipping image Andreas Rheinhardt
2022-04-17 12:57 ` Michael Niedermayer
2022-04-14 15:57 ` [FFmpeg-devel] [PATCH 4/5] avcodec/mjpegbdec: Don't create unnecessary AVFrame reference Andreas Rheinhardt
2022-04-14 15:57 ` [FFmpeg-devel] [PATCH 5/5] avcodec/mjpegbdec: Don't use GetBit-API for byte-aligned reads Andreas Rheinhardt
2022-04-15 22:15 ` [FFmpeg-devel] [PATCH 1/5] avcodec/mjpegdec: Always reset got_picture at the beginnig of decoding Michael Niedermayer
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=ead0259b-2694-12e5-cfa0-dad2a0d73175@gmail.com \
--to=jamrial@gmail.com \
--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