January 2026 - ffmpeg-cvslog - lists.ffmpeg.org

[ffmpeg] avutil/hwcontext_d3d11va: added resource and heap flags to DeviceContext (branch master)
by Dmitrii Ovchinnikov 26 Jan '26

26 Jan '26

This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg. The following commit(s) were added to refs/heads/master by this push: new 714152dd9a avutil/hwcontext_d3d11va: added resource and heap flags to DeviceContext 714152dd9a is described below commit 714152dd9a0ea260a25346ca276f06096ab4b7ad Author: Dmitrii Ovchinnikov <ovchinnikov.dmitrii(a)gmail.com> AuthorDate: Tue Oct 7 16:12:42 2025 +0200 Commit: Dmitrii Ovchinnikov <ovchinnikov.dmitrii(a)gmail.com> CommitDate: Mon Jan 26 10:39:10 2026 +0000 avutil/hwcontext_d3d11va: added resource and heap flags to DeviceContext --- doc/APIchanges | 3 +++ libavutil/hwcontext_d3d11va.c | 15 +++++++++++++++ libavutil/hwcontext_d3d11va.h | 14 ++++++++++++++ libavutil/version.h | 2 +- 4 files changed, 33 insertions(+), 1 deletion(-) diff --git a/doc/APIchanges b/doc/APIchanges index e68376fc3a..ee3167bb14 100644 --- a/doc/APIchanges +++ b/doc/APIchanges @@ -2,6 +2,9 @@ The last version increases of all libraries were on 2025-03-28 API changes, most recent first: +2026-01-xx - xxxxxxxxxx - lavu 60.24.100 - hwcontext_d3d11va.h + Add BindFlags and MiscFlags to AVD3D11VADeviceContext + 2026-01-04 - xxxxxxxxxx - lavu 60.23.100 - cpu.h Add AV_CPU_FLAG_ARM_CRC. diff --git a/libavutil/hwcontext_d3d11va.c b/libavutil/hwcontext_d3d11va.c index 83414ad0d4..418fe9dfda 100644 --- a/libavutil/hwcontext_d3d11va.c +++ b/libavutil/hwcontext_d3d11va.c @@ -290,6 +290,9 @@ static int d3d11va_frames_init(AVHWFramesContext *ctx) return AVERROR(EINVAL); } + hwctx->BindFlags |= device_hwctx->BindFlags; + hwctx->MiscFlags |= device_hwctx->MiscFlags; + ctx->initial_pool_size = FFMIN(ctx->initial_pool_size, MAX_ARRAY_SIZE); texDesc = (D3D11_TEXTURE2D_DESC){ @@ -710,6 +713,18 @@ static int d3d11va_device_create(AVHWDeviceContext *ctx, const char *device, } #endif + if (av_dict_get(opts, "SHADER", NULL, 0)) + device_hwctx->BindFlags |= D3D11_BIND_SHADER_RESOURCE; + + if (av_dict_get(opts, "UAV", NULL, 0)) + device_hwctx->BindFlags |= D3D11_BIND_UNORDERED_ACCESS; + + if (av_dict_get(opts, "RTV", NULL, 0)) + device_hwctx->BindFlags |= D3D11_BIND_RENDER_TARGET; + + if (av_dict_get(opts, "SHARED", NULL, 0)) + device_hwctx->MiscFlags |= D3D11_RESOURCE_MISC_SHARED; + return 0; } diff --git a/libavutil/hwcontext_d3d11va.h b/libavutil/hwcontext_d3d11va.h index 77d2d72f1b..b87b1e8fa2 100644 --- a/libavutil/hwcontext_d3d11va.h +++ b/libavutil/hwcontext_d3d11va.h @@ -94,6 +94,20 @@ typedef struct AVD3D11VADeviceContext { void (*lock)(void *lock_ctx); void (*unlock)(void *lock_ctx); void *lock_ctx; + /** + * D3D11_TEXTURE2D_DESC.BindFlags to be applied to D3D11 resources allocated + * for frames using this device context. + * + * It applies globally to all AVD3D11VAFramesContext allocated from this device context. + */ + UINT BindFlags; + /** + * D3D11_TEXTURE2D_DESC.MiscFlags to be applied to D3D11 resources allocated + * for frames using this device context. + * + * It applies globally to all AVD3D11VAFramesContext allocated from this device context. + */ + UINT MiscFlags; } AVD3D11VADeviceContext; /** diff --git a/libavutil/version.h b/libavutil/version.h index 9ed38790c7..9d308ed059 100644 --- a/libavutil/version.h +++ b/libavutil/version.h @@ -79,7 +79,7 @@ */ #define LIBAVUTIL_VERSION_MAJOR 60 -#define LIBAVUTIL_VERSION_MINOR 23 +#define LIBAVUTIL_VERSION_MINOR 24 #define LIBAVUTIL_VERSION_MICRO 100 #define LIBAVUTIL_VERSION_INT AV_VERSION_INT(LIBAVUTIL_VERSION_MAJOR, \

1 0

[ffmpeg] ffmpeg_demux: fix readrate to respect initial burst (branch master)
by Gyan Doshi 26 Jan '26

26 Jan '26

This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg. The following commit(s) were added to refs/heads/master by this push: new fe0813d6e2 ffmpeg_demux: fix readrate to respect initial burst fe0813d6e2 is described below commit fe0813d6e23e07deac804bfcdabfebf1fd2f6b0b Author: Gyan Doshi <ffmpeg(a)gyani.pro> AuthorDate: Tue Jan 20 17:32:00 2026 +0530 Commit: Gyan Doshi <ffmpeg(a)gyani.pro> CommitDate: Mon Jan 26 15:37:22 2026 +0530 ffmpeg_demux: fix readrate to respect initial burst Fixes #21510 --- fftools/ffmpeg_demux.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/fftools/ffmpeg_demux.c b/fftools/ffmpeg_demux.c index 193967fce1..701f9ddc34 100644 --- a/fftools/ffmpeg_demux.c +++ b/fftools/ffmpeg_demux.c @@ -525,6 +525,9 @@ static void readrate_sleep(Demuxer *d) pts = av_rescale(ds->dts, 1000000, AV_TIME_BASE); now = av_gettime_relative(); wc_elapsed = now - d->wallclock_start; + + if (pts <= stream_ts_offset + initial_burst) continue; + max_pts = stream_ts_offset + initial_burst + (int64_t)(wc_elapsed * d->readrate); lag = FFMAX(max_pts - pts, 0); if ( (!ds->lag && lag > 0.3 * AV_TIME_BASE) || ( lag > ds->lag + 0.3 * AV_TIME_BASE) ) {

1 0

[ffmpeg] swscale/utils: zero init filter memory as before (branch master)
by Kacper Michajłow 26 Jan '26

26 Jan '26

This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg. The following commit(s) were added to refs/heads/master by this push: new 10db62d205 swscale/utils: zero init filter memory as before 10db62d205 is described below commit 10db62d20527d7f17d1cce8e1c22c86c33be7350 Author: Kacper Michajłow <kasper93(a)gmail.com> AuthorDate: Sun Jan 25 16:57:09 2026 +0100 Commit: Niklas Haas <ffmpeg(a)haasn.dev> CommitDate: Mon Jan 26 09:45:16 2026 +0000 swscale/utils: zero init filter memory as before Commit 795bb37a39e163cbb5b6d897ebb0ca90ea7b449a removed zeroing of those buffers, without mention, which introduces corrupted output. Fixes: 795bb37a39e163cbb5b6d897ebb0ca90ea7b449a Fixes: https://github.com/mpv-player/mpv/issues/17317 --- libswscale/utils.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/libswscale/utils.c b/libswscale/utils.c index 24942fe6da..195af846d4 100644 --- a/libswscale/utils.c +++ b/libswscale/utils.c @@ -388,7 +388,7 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, if (dstFilter) filter2Size += dstFilter->length - 1; av_assert0(filter2Size > 0); - filter2 = av_malloc_array(dstW, filter2Size * sizeof(*filter2)); + filter2 = av_calloc(dstW, filter2Size * sizeof(*filter2)); if (!filter2) goto nomem; for (i = 0; i < dstW; i++) { @@ -557,7 +557,7 @@ static av_cold int initFilter(int16_t **outFilter, int32_t **filterPos, // Note the +1 is for the MMX scaler which reads over the end /* align at 16 for AltiVec (needed by hScale_altivec_real) */ - *outFilter = av_malloc_array(dstW + 3, *outFilterSize * sizeof(**outFilter)); + *outFilter = av_calloc(dstW + 3, *outFilterSize * sizeof(**outFilter)); if (!*outFilter) goto nomem;

1 0

[ffmpeg] branch master updated (6e65d1c945 -> bf4d5037b4)
by ffmpeg-git 25 Jan '26

25 Jan '26

This is an automated email from the git hooks/post-receive script. Git pushed a change to branch master in repository ffmpeg. from 6e65d1c945 avcodec/motion_est: Fix left shifts of negative numbers new 982244818b avcodec/x86/h264_deblock: Remove unused macros new d53ff30215 tests/checkasm/h264dsp: Don't test loop filter strength 0 new a22149ab3d avcodec/x86/h264_deblock: Remove always-false branches new 899475326b avcodec/x86/h264_deblock: Simplify splatting new b1140d3c98 avcodec/x86/h264_deblock: Remove obsolete macro parameters new dbdf514c17 avcodec/x86/h264_deblock_10bit: Remove custom stack allocation code new aab0946eae avcodec/x86/h264_deblock_10bit: Remove mmxext functions new eaaf45fd79 avcodec/x86/h264_deblock_10bit: Simplify r0+4*r1 new 9882973935 avcodec/x86/h264_deblock: Avoid reloading constant new 8428a412bc avcodec/x86/h264_deblock: Avoid MMX in deblock_h_luma_8 new db66e057eb avcodec/x86/h264_deblock: Avoid reload new 489aaf4e1c avcodec/x86/h264_deblock: Don't sign-extend stride new bf4d5037b4 avcodec/h264dsp: Remove redundant h264 from H264DSPCtx member names The 13 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "add" were already present in the repository and have only been added to this reference. Summary of changes: libavcodec/aarch64/h264dsp_init_aarch64.c | 72 ++++---- libavcodec/arm/h264dsp_init_arm.c | 38 ++-- libavcodec/h264_loopfilter.c | 30 ++-- libavcodec/h264_mb.c | 54 +++--- libavcodec/h264_mb_template.c | 28 +-- libavcodec/h264dsp.c | 78 ++++----- libavcodec/h264dsp.h | 114 ++++++------ libavcodec/loongarch/h264dsp_init_loongarch.c | 86 +++++----- libavcodec/mips/h264dsp_init_mips.c | 108 ++++++------ libavcodec/ppc/h264dsp.c | 30 ++-- libavcodec/riscv/h264dsp_init.c | 88 +++++----- libavcodec/svq3.c | 8 +- libavcodec/x86/h264_deblock.asm | 238 ++++++++++---------------- libavcodec/x86/h264_deblock_10bit.asm | 132 ++------------ libavcodec/x86/h264dsp_init.c | 193 ++++++++++----------- tests/checkasm/h264dsp.c | 92 +++++----- 16 files changed, 613 insertions(+), 776 deletions(-)

2 13

[ffmpeg] avcodec/motion_est: Fix left shifts of negative numbers (branch master)
by Andreas Rheinhardt 25 Jan '26

25 Jan '26

This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg. The following commit(s) were added to refs/heads/master by this push: new 6e65d1c945 avcodec/motion_est: Fix left shifts of negative numbers 6e65d1c945 is described below commit 6e65d1c94533feefc7237bd75fc26d4b89a72f46 Author: Andreas Rheinhardt <andreas.rheinhardt(a)outlook.com> AuthorDate: Wed Jan 21 00:28:35 2026 +0100 Commit: Andreas Rheinhardt <andreas.rheinhardt(a)outlook.com> CommitDate: Sun Jan 25 22:46:39 2026 +0100 avcodec/motion_est: Fix left shifts of negative numbers Fixes ticket #21486. Reviewed-by: James Almer <jamrial(a)gmail.com> Signed-off-by: Andreas Rheinhardt <andreas.rheinhardt(a)outlook.com> --- libavcodec/motion_est.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/libavcodec/motion_est.c b/libavcodec/motion_est.c index 33166f2a90..8ca6f16a7c 100644 --- a/libavcodec/motion_est.c +++ b/libavcodec/motion_est.c @@ -1082,11 +1082,9 @@ int ff_pre_estimate_p_frame_motion(MPVEncContext *const s, get_limits(s, 16*mb_x, 16*mb_y, 0); c->skip=0; - P_LEFT[0] = s->p_mv_table[xy + 1][0]; + P_LEFT[0] = FFMAX(s->p_mv_table[xy + 1][0], c->xmin * (1 << shift)); P_LEFT[1] = s->p_mv_table[xy + 1][1]; - if(P_LEFT[0] < (c->xmin<<shift)) P_LEFT[0] = (c->xmin<<shift); - /* special case for first line */ if (s->c.first_slice_line) { c->pred_x= P_LEFT[0];

1 0

[ffmpeg] branch master updated (1e1dde8798 -> 8966101fa6)
by ffmpeg-git 25 Jan '26

25 Jan '26

This is an automated email from the git hooks/post-receive script. Git pushed a change to branch master in repository ffmpeg. from 1e1dde8798 avcodec/libx265: map ffmpeg log level to x265 log level new 0886e50c6b lavc/hevc: add aarch64 neon for 8-bit dequant new 24f296c7a1 lavc/hevc: optimize dequant for shift=0 case (identity transform) new ce89d974c8 lavc/hevc: add aarch64 neon for 10-bit dequant new 8966101fa6 lavc/hevc: add aarch64 neon for 12-bit dequant The 4 revisions listed above as "new" are entirely new to this repository and will be described in separate emails. The revisions listed as "add" were already present in the repository and have only been added to this reference. Summary of changes: libavcodec/aarch64/Makefile | 1 + libavcodec/aarch64/hevcdsp_dequant_neon.S | 371 ++++++++++++++++++++++++++++++ libavcodec/aarch64/hevcdsp_init_aarch64.c | 52 +++++ libavcodec/hevc/dsp_template.c | 23 +- tests/checkasm/Makefile | 2 +- tests/checkasm/checkasm.c | 1 + tests/checkasm/checkasm.h | 1 + tests/checkasm/hevc_dequant.c | 76 ++++++ tests/fate/checkasm.mak | 1 + 9 files changed, 526 insertions(+), 2 deletions(-) create mode 100644 libavcodec/aarch64/hevcdsp_dequant_neon.S create mode 100644 tests/checkasm/hevc_dequant.c

2 4

[ffmpeg] avcodec/libx265: map ffmpeg log level to x265 log level (branch master)
by Zhao Zhili 25 Jan '26

25 Jan '26

This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg. The following commit(s) were added to refs/heads/master by this push: new 1e1dde8798 avcodec/libx265: map ffmpeg log level to x265 log level 1e1dde8798 is described below commit 1e1dde87986de78ce493534e6389f549cbb54d17 Author: Zhao Zhili <quinkblack(a)foxmail.com> AuthorDate: Wed Jan 14 21:28:00 2026 +0800 Commit: Zhao Zhili <quinkblack(a)foxmail.com> CommitDate: Sun Jan 25 13:09:30 2026 +0800 avcodec/libx265: map ffmpeg log level to x265 log level Previously x265 encoder used its default log level regardless of FFmpeg's log level setting. Note the log level can be overwritten by x265-params. Fix #21462 Signed-off-by: Zhao Zhili <zhilizhao(a)tencent.com> --- libavcodec/libx265.c | 19 +++++++++++++++++++ 1 file changed, 19 insertions(+) diff --git a/libavcodec/libx265.c b/libavcodec/libx265.c index 89cd7d5dbe..7d488c0f4b 100644 --- a/libavcodec/libx265.c +++ b/libavcodec/libx265.c @@ -251,6 +251,24 @@ static int handle_side_data(AVCodecContext *avctx, const x265_api *api, return 0; } +static int get_x265_log_level(AVCodecContext *avctx) +{ + int level = av_log_get_level() + avctx->log_level_offset; + + if (level <= AV_LOG_QUIET) + return X265_LOG_NONE; + if (level <= AV_LOG_ERROR) + return X265_LOG_ERROR; + if (level <= AV_LOG_WARNING) + return X265_LOG_WARNING; + if (level <= AV_LOG_INFO) + return X265_LOG_INFO; + if (level <= AV_LOG_DEBUG) + return X265_LOG_DEBUG; + + return X265_LOG_FULL; +} + static av_cold int libx265_encode_init(AVCodecContext *avctx) { libx265Context *ctx = avctx->priv_data; @@ -286,6 +304,7 @@ static av_cold int libx265_encode_init(AVCodecContext *avctx) return AVERROR(EINVAL); } + ctx->params->logLevel = get_x265_log_level(avctx); ctx->params->frameNumThreads = avctx->thread_count; if (avctx->framerate.num > 0 && avctx->framerate.den > 0) { ctx->params->fpsNum = avctx->framerate.num;

1 0

[ffmpeg] libavfilter/showcwt: fix OOB write for DU/RL position init (branch master)
by Ruikai Peng 24 Jan '26

24 Jan '26

This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg. The following commit(s) were added to refs/heads/master by this push: new 905a432403 libavfilter/showcwt: fix OOB write for DU/RL position init 905a432403 is described below commit 905a4324030e859a95bbb9901998c55128b7150d Author: Ruikai Peng <ruikai(a)pwno.io> AuthorDate: Fri Jan 16 22:32:35 2026 -0500 Commit: michaelni <michael(a)niedermayer.cc> CommitDate: Sat Jan 24 22:05:34 2026 +0000 libavfilter/showcwt: fix OOB write for DU/RL position init In config_output() for direction=DU/RL, the position is initialized to s->sono_size, which equals h or w when bar=0. That position is later used as an in-bounds pixel coordinate without clamping in draw(), causing writes past the end of the output planes. Repro: ffmpeg -f lavfi -i sine=frequency=1000:sample_rate=44100 \ -filter_complex "[0:a]showcwt=s=640x512:bar=0:direction=du[v]" \ -map "[v]" -frames:v 1 -f null - AddressSanitizer: heap-buffer-overflow ... WRITE of size 1 Initialize and wrap the DU/RL position to sono_size - 1 (or 0 when empty), preventing out-of-bounds row/column writes when bar=0 while preserving existing slide behavior. --- libavfilter/avf_showcwt.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/libavfilter/avf_showcwt.c b/libavfilter/avf_showcwt.c index 5cb2e9f9fa..839bdc7983 100644 --- a/libavfilter/avf_showcwt.c +++ b/libavfilter/avf_showcwt.c @@ -1012,7 +1012,7 @@ static int config_output(AVFilterLink *outlink) break; case DIRECTION_RL: case DIRECTION_DU: - s->pos = s->sono_size; + s->pos = FFMAX(s->sono_size - 1, 0); break; } @@ -1087,7 +1087,7 @@ static int output_frame(AVFilterContext *ctx) case DIRECTION_RL: s->pos--; if (s->pos < 0) { - s->pos = s->sono_size; + s->pos = FFMAX(s->sono_size - 1, 0); s->new_frame = 1; } break; @@ -1101,7 +1101,7 @@ static int output_frame(AVFilterContext *ctx) case DIRECTION_DU: s->pos--; if (s->pos < 0) { - s->pos = s->sono_size; + s->pos = FFMAX(s->sono_size - 1, 0); s->new_frame = 1; } break; @@ -1115,7 +1115,7 @@ static int output_frame(AVFilterContext *ctx) break; case DIRECTION_RL: case DIRECTION_DU: - s->pos = s->sono_size; + s->pos = FFMAX(s->sono_size - 1, 0); break; } break;

1 0

[ffmpeg] avformat/udp: don't bind sockets to 0.0.0.0 by default for unicast UDP inputs (branch master)
by Rost Kurylo 24 Jan '26

24 Jan '26

This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg. The following commit(s) were added to refs/heads/master by this push: new 4c9953592f avformat/udp: don't bind sockets to 0.0.0.0 by default for unicast UDP inputs 4c9953592f is described below commit 4c9953592ff55feb917a13f0d3fe14e3f9451584 Author: Rost Kurylo <rost.kurylo(a)netint.ca> AuthorDate: Tue Jan 13 11:17:57 2026 -0800 Commit: Marton Balint <cus(a)passwd.hu> CommitDate: Sat Jan 24 20:26:24 2026 +0000 avformat/udp: don't bind sockets to 0.0.0.0 by default for unicast UDP inputs --- libavformat/udp.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/libavformat/udp.c b/libavformat/udp.c index 61e80c86c5..0874d18397 100644 --- a/libavformat/udp.c +++ b/libavformat/udp.c @@ -700,6 +700,7 @@ static int udp_open(URLContext *h, const char *uri, int flags) struct sockaddr_storage my_addr; socklen_t len; int ret; + const char *bind_addr = NULL; h->is_streamed = 1; @@ -759,7 +760,12 @@ static int udp_open(URLContext *h, const char *uri, int flags) if ((s->is_multicast || s->local_port < 0) && (h->flags & AVIO_FLAG_READ)) s->local_port = port; - udp_fd = udp_socket_create(h, &my_addr, &len, s->localaddr); + if (!s->is_multicast && (s->localaddr == NULL || s->localaddr[0] == '\0') && h->flags == AVIO_FLAG_READ) + bind_addr = hostname; + else + bind_addr = s->localaddr; + + udp_fd = udp_socket_create(h, &my_addr, &len, bind_addr); if (udp_fd < 0) { ret = AVERROR(EIO); goto fail;

1 0

[ffmpeg] lavc/j2kdec: Do not ignore colour association for packed formats (branch master)
by Carl Eugen Hoyos 24 Jan '26

24 Jan '26

This is an automated email from the git hooks/post-receive script. Git pushed a commit to branch master in repository ffmpeg. The following commit(s) were added to refs/heads/master by this push: new aab0c23cb8 lavc/j2kdec: Do not ignore colour association for packed formats aab0c23cb8 is described below commit aab0c23cb848b9db10aae9849505221ba7988776 Author: Carl Eugen Hoyos <ceffmpeg(a)gmail.com> AuthorDate: Sun Jan 4 23:50:53 2026 +0300 Commit: Marton Balint <cus(a)passwd.hu> CommitDate: Sat Jan 24 20:25:05 2026 +0000 lavc/j2kdec: Do not ignore colour association for packed formats Fixes ticket #9468. Signed-off-by: Carl Eugen Hoyos <ceffmpeg(a)gmail.com --- libavcodec/jpeg2000dec.c | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/libavcodec/jpeg2000dec.c b/libavcodec/jpeg2000dec.c index 276c0a6e43..e068ddaa82 100644 --- a/libavcodec/jpeg2000dec.c +++ b/libavcodec/jpeg2000dec.c @@ -2348,9 +2348,12 @@ static inline int tile_codeblocks(const Jpeg2000DecoderContext *s, Jpeg2000Tile int h = tile->comp[compno].coord[1][1] - \ ff_jpeg2000_ceildiv(s->image_offset_y, s->cdy[compno]); \ int plane = 0; \ + ptrdiff_t dstoffset = 0; \ \ if (planar) \ plane = s->cdef[compno] ? s->cdef[compno]-1 : (s->ncomponents-1); \ + else \ + dstoffset = s->cdef[compno] ? s->cdef[compno] - 1 : compno; \ \ y = tile->comp[compno].coord[1][0] - \ ff_jpeg2000_ceildiv(s->image_offset_y, s->cdy[compno]); \ @@ -2360,7 +2363,7 @@ static inline int tile_codeblocks(const Jpeg2000DecoderContext *s, Jpeg2000Tile \ x = tile->comp[compno].coord[0][0] - \ ff_jpeg2000_ceildiv(s->image_offset_x, s->cdx[compno]); \ - dst = line + x * pixelsize + compno*!planar; \ + dst = line + x * pixelsize + dstoffset; \ \ if (codsty->transform == FF_DWT97) { \ for (; x < w; x++) { \

1 0