From: Lynne <dev@lynne.ee> To: FFmpeg development discussions and patches <ffmpeg-devel@ffmpeg.org> Subject: Re: [FFmpeg-devel] [PATCH] avfft: avoid overreads with RDFT API users Date: Fri, 9 Feb 2024 20:30:38 +0100 (CET) Message-ID: <NqELQbP--3-9@lynne.ee> (raw) In-Reply-To: <ZcZ6jjFW7JKaaMTk@kusa> Feb 9, 2024, 20:19 by matthieu.bouron@gmail.com: > On Fri, Feb 09, 2024 at 06:27:50PM +0100, Lynne wrote: > >> The new API requires an extra array member at the very end, >> which old API users did not do. >> >> This disables in-place RDFT transforms and instead >> does the transform out of place by copying once, there shouldn't >> be a significant loss of speed as our in-place FFT requires a reorder >> which is likely more expensive in the majority of cases to do. >> >> Patch attached. >> >> From 763f2e8941dc3dd4282ad42574bd8d451a256a53 Mon Sep 17 00:00:00 2001 >> From: Lynne <dev@lynne.ee> >> Date: Fri, 9 Feb 2024 18:17:54 +0100 >> Subject: [PATCH] avfft: avoid overreads with RDFT API users >> >> The new API requires an extra array member at the very end, >> which old API users did not do. >> >> This disables in-place RDFT transforms and instead >> does the transform out of place by copying once, there shouldn't >> be a significant loss of speed as our in-place FFT requires a reorder >> which is likely more expensive in the majority of cases to do. >> --- >> libavcodec/avfft.c | 31 +++++++++++++++++++++++++------ >> 1 file changed, 25 insertions(+), 6 deletions(-) >> >> diff --git a/libavcodec/avfft.c b/libavcodec/avfft.c >> index 999b5ed79a..485f216427 100644 >> --- a/libavcodec/avfft.c >> +++ b/libavcodec/avfft.c >> @@ -152,7 +152,7 @@ RDFTContext *av_rdft_init(int nbits, enum RDFTransformType trans) >> return NULL; >> >> ret = av_tx_init(&s->ctx, &s->fn, AV_TX_FLOAT_RDFT, trans == IDFT_C2R, >> - 1 << nbits, &scale, AV_TX_INPLACE); >> + 1 << nbits, &scale, 0x0); >> if (ret < 0) { >> av_free(s); >> return NULL; >> @@ -162,17 +162,35 @@ RDFTContext *av_rdft_init(int nbits, enum RDFTransformType trans) >> s->len = 1 << nbits; >> s->inv = trans == IDFT_C2R; >> >> + s->tmp = av_malloc((s->len + 2)*sizeof(float)); >> + if (!s->tmp) { >> + av_tx_uninit(&s->ctx); >> + av_free(s); >> + return NULL; >> + } >> + >> return (RDFTContext *)s; >> } >> >> void av_rdft_calc(RDFTContext *s, FFTSample *data) >> { >> AVTXWrapper *w = (AVTXWrapper *)s; >> - if (w->inv) >> - FFSWAP(float, data[1], data[w->len]); >> - w->fn(w->ctx, data, (void *)data, w->stride); >> - if (!w->inv) >> - FFSWAP(float, data[1], data[w->len]); >> + float *src = w->inv ? w->tmp : (float *)data; >> + float *dst = w->inv ? (float *)data : w->tmp; >> + >> + if (w->inv) { >> + memcpy(src, data, w->len*sizeof(float)); >> + >> + src[w->len] = src[1]; >> + src[1] = 0.0f; >> + } >> + >> + w->fn(w->ctx, dst, (void *)src, w->stride); >> + >> + if (!w->inv) { >> + dst[1] = dst[w->len]; >> + memcpy(data, dst, w->len*sizeof(float)); >> + } >> } >> >> av_cold void av_rdft_end(RDFTContext *s) >> @@ -180,6 +198,7 @@ av_cold void av_rdft_end(RDFTContext *s) >> if (s) { >> AVTXWrapper *w = (AVTXWrapper *)s; >> av_tx_uninit(&w->ctx); >> + av_free(&w->tmp); >> > > av_free(w->tmp); > Yup, discussed on IRC. >> av_free(w); >> } >> } >> -- >> 2.43.0.381.gb435a96ce8 >> > > Hi, > > I tested the patch with the av_free() adjustment and it fixed the crash / > memory corruption I was encountering on macOS. > Thanks, will push soon. _______________________________________________ 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".
prev parent reply other threads:[~2024-02-09 19:30 UTC|newest] Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top 2024-02-09 17:27 Lynne 2024-02-09 19:18 ` Matthieu Bouron 2024-02-09 19:30 ` Lynne [this message]
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=NqELQbP--3-9@lynne.ee \ --to=dev@lynne.ee \ --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