> -Original Message-
> From: ffmpeg-devel On Behalf Of
> James Almer
> Sent: Thursday, 30 September 2021 05:53
> To: ffmpeg-devel@ffmpeg.org
> Subject: Re: [FFmpeg-devel] [PATCH v5 3/7] libavformat/asfdec: Fix
> type of value_len
>
> On 9/29/2021 11:58 PM, Soft Works wrote:
> > The
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> James Almer
> Sent: Thursday, 30 September 2021 05:53
> To: ffmpeg-devel@ffmpeg.org
> Subject: Re: [FFmpeg-devel] [PATCH v5 3/7] libavformat/asfdec: Fix
> type of value_len
>
> On 9/29/2021 11:58 PM, Soft Works wrote:
> > The
On 9/29/2021 11:58 PM, Soft Works wrote:
The value_len is an uint32 not an int32 per spec. That
value must not be truncated, neither by casting to int, nor by any
conditional checks, because at the end of get_tag, this value is
needed to move forward in parsing. When the len value gets
modified,
Signed-off-by: softworkz
---
libavutil/frame.h | 13 +++--
1 file changed, 11 insertions(+), 2 deletions(-)
diff --git a/libavutil/frame.h b/libavutil/frame.h
index ff2540a20f..f07e690410 100644
--- a/libavutil/frame.h
+++ b/libavutil/frame.h
@@ -304,7 +304,8 @@ typedef struct AVFrame {
ping for review.
在 2021年9月27日 +0800 09:48，Sun Zhenliang ，写道：
> 在 2021年9月26日 +0800 20:40，myp...@gmail.com ，写道：
> > On Thu, Sep 16, 2021 at 11:24 AM sunzhenliang
> > wrote:
> > >
> > > Signed-off-by: sunzhenliang
> > > ---
> > > libavformat/amr.c | 22 ++
> > > 1 file changed,
Signed-off-by: softworkz
---
v5: Split into pieces as requested
libavformat/asfdec_f.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavformat/asfdec_f.c b/libavformat/asfdec_f.c
index c4bcf0904d..870d4bc7e6 100644
--- a/libavformat/asfdec_f.c
+++
Signed-off-by: softworkz
---
v5: Split into pieces as requested
libavformat/asfdec_f.c | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/libavformat/asfdec_f.c b/libavformat/asfdec_f.c
index 7ac4e460a7..c4bcf0904d 100644
--- a/libavformat/asfdec_f.c
+++
Signed-off-by: softworkz
---
v5: Split into pieces as requested
libavformat/asfdec_f.c | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/libavformat/asfdec_f.c b/libavformat/asfdec_f.c
index d35eec7082..7ac4e460a7 100644
--- a/libavformat/asfdec_f.c
+++
These three are closely related and can't be separated easily:
In get_tag, the code was adding 22 bytes (in order to allow
it to hold 64bit numbers as string) to the value len for creating
creating a buffer. This was unnecessarily imposing a
size-constraint on the value_len parameter.
The code
The value_len is an uint32 not an int32 per spec. That
value must not be truncated, neither by casting to int, nor by any
conditional checks, because at the end of get_tag, this value is
needed to move forward in parsing. When the len value gets
modified, the parsing may break.
Signed-off-by:
get_value had a return type of int, which means that reading
QWORDS (case 4) was broken due to truncation of the result from
avio_rl64().
Signed-off-by: softworkz
---
v5: Split into pieces as requested
libavformat/asfdec_f.c | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
The spec allows attachment sizes of up to UINT32_MAX while
we can handle only sizes up to INT32_MAX (in downstream
code).
The debug.assert in get_tag didn't really address this,
and truncating the value_len in calling methods cannot
be used because the length value is required in order to
lance.lmw...@gmail.com:
> From: Limin Wang
>
> It is simpler and more efficient compared to the current code.
>
> Signed-off-by: Limin Wang
> ---
> libavfilter/vf_addroi.c | 98
> +++--
> 1 file changed, 37 insertions(+), 61 deletions(-)
>
> diff
Hello,
>+pb_shuffle_low: times 4 db 1, 3, 5, 7, 9, 11, 13, 15, -1, -1, -1, -1, -1, -1,
>-1, -1
Why we times 4?
AVX2 provided instruction VPBROADCASTQ to load these constant into SIMD
register.
Moreover, the plane U/V also apply same algorithm to get improve.
Regards,
Min Chen
At 2021-09-30
With the accelerating by means of AVX512, the uyvytoyuv422 can be faster.
Performance data(Less is better):
uyvytoyuv422_avx2 0.27309
uyvytoyuv422_avx5120.16229
Signed-off-by: Wu Jianhua
---
libswscale/x86/rgb2rgb.c | 6 ++
libswscale/x86/rgb_2_rgb.asm | 20
With the accelerating by means of AVX2, the uyvytoyuv422 can be faster
Performance data(Less is better):
uyvytoyuv422_sse20.50388
uyvytoyuv422_avx 0.46132
uyvytoyuv422_avx20.27309
Signed-off-by: Wu Jianhua
---
libswscale/x86/rgb2rgb.c | 6
Performance data(Less is better):
shuffle_bytes_avx2 0.94288
shuffle_bytes_avx5120.60049
Signed-off-by: Wu Jianhua
---
libswscale/x86/rgb2rgb.c | 13 +
libswscale/x86/rgb_2_rgb.asm | 8
2 files changed, 21 insertions(+)
diff --git
Performance data(Less is better):
shuffle_bytes_ssse3 3.64654
shuffle_bytes_avx20.94288
Signed-off-by: Wu Jianhua
---
libswscale/x86/rgb2rgb.c | 17 +++--
libswscale/x86/rgb_2_rgb.asm | 11 +++
2 files changed, 26 insertions(+), 2 deletions(-)
diff --git
On Thu, Sep 30, 2021 at 03:19:39AM +0200, Andreas Rheinhardt wrote:
> lance.lmw...@gmail.com:
> > From: Limin Wang
> >
> > Signed-off-by: Limin Wang
> > ---
> > libavfilter/vf_showinfo.c | 4 ++--
> > 1 file changed, 2 insertions(+), 2 deletions(-)
> >
> > diff --git
lance.lmw...@gmail.com:
> From: Limin Wang
>
> Signed-off-by: Limin Wang
> ---
> libavfilter/vf_showinfo.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/libavfilter/vf_showinfo.c b/libavfilter/vf_showinfo.c
> index 87c2bed..5d97ef6 100644
> ---
From: Limin Wang
Signed-off-by: Limin Wang
---
doc/protocols.texi | 10 ++
libavformat/libsrt.c | 7 +++
2 files changed, 17 insertions(+)
diff --git a/doc/protocols.texi b/doc/protocols.texi
index 726e5f1..9c246f8 100644
--- a/doc/protocols.texi
+++ b/doc/protocols.texi
@@
From: Limin Wang
Make all options string conversion consistent.
Signed-off-by: Limin Wang
---
libavformat/libsrt.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/libavformat/libsrt.c b/libavformat/libsrt.c
index 5aef0e7..c6308d1 100644
--- a/libavformat/libsrt.c
+++
From: Limin Wang
Signed-off-by: Limin Wang
---
libavformat/libsrt.c | 16
1 file changed, 8 insertions(+), 8 deletions(-)
diff --git a/libavformat/libsrt.c b/libavformat/libsrt.c
index a80f249..5aef0e7 100644
--- a/libavformat/libsrt.c
+++ b/libavformat/libsrt.c
@@ -397,10
From: Limin Wang
It is simpler and more efficient compared to the current code.
Signed-off-by: Limin Wang
---
libavfilter/vf_addroi.c | 98 +++--
1 file changed, 37 insertions(+), 61 deletions(-)
diff --git a/libavfilter/vf_addroi.c
From: Limin Wang
Signed-off-by: Limin Wang
---
libavfilter/vf_showinfo.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/libavfilter/vf_showinfo.c b/libavfilter/vf_showinfo.c
index 87c2bed..5d97ef6 100644
--- a/libavfilter/vf_showinfo.c
+++ b/libavfilter/vf_showinfo.c
Timo Rothenpieler:
> On 29.09.2021 19:31, Dennis Mungai wrote:
>> On Wed, 29 Sept 2021 at 20:22, Timo Rothenpieler
>> wrote:
>>
>>> On 29.09.2021 19:17, Dennis Mungai wrote:
A potential downside to this would be on QSV's side, which needs at
least
2 threads to prevent
Derek Buitenhuis:
> This is as per:
>* mp4ra: http://mp4ra.org/#/brands
>* Dolby Vision muxing spec (which is public):
>
>
> -Original Message-
> From: ffmpeg-devel On Behalf Of
> Andreas Rheinhardt
> Sent: Wednesday, 29 September 2021 10:28
> To: ffmpeg-devel@ffmpeg.org
> Subject: Re: [FFmpeg-devel] [PATCH v2 1/1] avfilter/frames: Ensure
> frames are writable when processing in-place
>
> Soft Works:
> >
>
From: Pierre-Anthony Lemieux
Signed-off-by: Pierre-Anthony Lemieux
---
Notes:
Modify the FFMPEG build system to add support for an IMF demuxer.
MAINTAINERS | 1 +
configure| 3 ++-
doc/demuxers.texi| 6 ++
libavformat/Makefile | 2 ++
From: Pierre-Anthony Lemieux
Signed-off-by: Pierre-Anthony Lemieux
---
Notes:
Tests for the public API of the IMF demuxer.
libavformat/tests/imf.c | 476
1 file changed, 476 insertions(+)
create mode 100644 libavformat/tests/imf.c
diff --git
From: Pierre-Anthony Lemieux
Signed-off-by: Pierre-Anthony Lemieux
---
Notes:
Implements the IMF demuxer.
libavformat/imfdec.c | 646 +++
1 file changed, 646 insertions(+)
create mode 100644 libavformat/imfdec.c
diff --git a/libavformat/imfdec.c
From: Pierre-Anthony Lemieux
Signed-off-by: Pierre-Anthony Lemieux
---
Notes:
Implements IMF Composition Playlist (CPL) parsing.
libavformat/imf_cpl.c | 652 ++
1 file changed, 652 insertions(+)
create mode 100644 libavformat/imf_cpl.c
diff --git
From: Pierre-Anthony Lemieux
Signed-off-by: Pierre-Anthony Lemieux
---
Notes:
The IMF demuxer accepts as input an IMF CPL. The assets referenced by the
CPL can be
contained in multiple deliveries, each defined by an ASSETMAP file:
ffmpeg -assetmaps ,,... -i
If
On Tue, Sep 28, 2021 at 5:31 PM Derek Buitenhuis
wrote:
>
> This is as per:
>* mp4ra: http://mp4ra.org/#/brands
>* Dolby Vision muxing spec (which is public):
>
>
On Wed, Sep 15, 2021 at 10:09:49PM +0200, Andreas Rheinhardt wrote:
> Michael Niedermayer:
> > Fixes: signed integer overflow: 559334865 * 4 cannot be represented in type
> > 'int'
> > Fixes:
> > 37929/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_APE_fuzzer-6751932295806976
> >
> >
Fixes: left shift of negative value -1
Fixes:
39223/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_H264_fuzzer-5498831521841152
Found-by: continuous fuzzing process
https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer
---
libavcodec/h264_slice.c
Fixes: memleak
Fixes:
39166/clusterfuzz-testcase-minimized-ffmpeg_dem_IVR_fuzzer-5153276690038784
Found-by: continuous fuzzing process
https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
Signed-off-by: Michael Niedermayer
---
libavformat/rmdec.c | 4
1 file changed, 4
Fixes: runtime error: signed integer overflow: 727298502 * 3 cannot be
represented in type 'int'
Fixes:
39172/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_APE_fuzzer-638602483033702
Found-by: continuous fuzzing process
https://github.com/google/oss-fuzz/tree/master/projects/ffmpeg
On Wed, Sep 29, 2021 at 7:52 PM Arif Driessen wrote:
> > Better add separate section, and than link to it with each filter docs,
> instead copy paste lines 100 times.
>
> If we do add a separate section, I'm not sure where to put it. Perhaps
> under "@chapter See Also"?
>
> I thought about
> Better add separate section, and than link to it with each filter docs,
instead copy paste lines 100 times.
If we do add a separate section, I'm not sure where to put it. Perhaps
under "@chapter See Also"?
I thought about something like a "@chapter Common Filter Options", but this
makes it
On Wed, 29 Sept 2021 at 20:37, Timo Rothenpieler
wrote:
> On 29.09.2021 19:31, Dennis Mungai wrote:
> > On Wed, 29 Sept 2021 at 20:22, Timo Rothenpieler
> > wrote:
> >
> >> On 29.09.2021 19:17, Dennis Mungai wrote:
> >>> A potential downside to this would be on QSV's side, which needs at
>
On 29.09.2021 19:31, Dennis Mungai wrote:
On Wed, 29 Sept 2021 at 20:22, Timo Rothenpieler
wrote:
On 29.09.2021 19:17, Dennis Mungai wrote:
A potential downside to this would be on QSV's side, which needs at least
2 threads to prevent dead-locking, see
On Wed, 29 Sept 2021 at 20:22, Timo Rothenpieler
wrote:
> On 29.09.2021 19:17, Dennis Mungai wrote:
> > A potential downside to this would be on QSV's side, which needs at least
> > 2 threads to prevent dead-locking, see
> >
>
On Tue, Sep 28, 2021 at 6:38 PM chen wrote:
> Hello,
>
>
> Excuse me, how about FMADD on AVX2 platform?
>
>
> For example
> +mulps m7, m7, m14
> +addps m0, m0, m7
>
> ==>
>
>
> fmadd231ps m0,m7,m14
>
>
Interesting, does having AVX2 guarantee having FMA instructions?
>
> Regards,
> Min
On Wed, Sep 29, 2021 at 12:54:02PM -0300, James Almer wrote:
> On 9/29/2021 12:31 PM, Michael Niedermayer wrote:
> > On Tue, Sep 28, 2021 at 10:19:48PM -0300, James Almer wrote:
> > > On 9/28/2021 7:28 PM, Michael Niedermayer wrote:
> > > > Fixes: Timeout
> > > > Fixes:
> > > >
On 29.09.2021 19:17, Dennis Mungai wrote:
A potential downside to this would be on QSV's side, which needs at least
2 threads to prevent dead-locking, see
https://github.com/Intel-Media-SDK/MediaSDK/blob/master/_studio/mfx_lib/scheduler/linux/src/mfx_scheduler_core_ischeduler.cpp#L90-L93
Is
On Wed, 29 Sept 2021 at 20:02, Timo Rothenpieler
wrote:
> ---
> fftools/ffmpeg.c | 16 ++--
> 1 file changed, 10 insertions(+), 6 deletions(-)
>
> Not sure if there is any downside to this.
> Threading for hwaccel does not make a whole lot of sense, and at least
> in case of nvdec
---
fftools/ffmpeg.c | 16 ++--
1 file changed, 10 insertions(+), 6 deletions(-)
Not sure if there is any downside to this.
Threading for hwaccel does not make a whole lot of sense, and at least
in case of nvdec wastes a lot of VRAM for no performance gain, and
specially on high core
will apply 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-requ...@ffmpeg.org with subject "unsubscribe".
On 9/29/2021 12:31 PM, Michael Niedermayer wrote:
On Tue, Sep 28, 2021 at 10:19:48PM -0300, James Almer wrote:
On 9/28/2021 7:28 PM, Michael Niedermayer wrote:
Fixes: Timeout
Fixes:
39089/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_MSNSIREN_fuzzer-6677219854909440
Found-by: continuous
On Tue, Sep 28, 2021 at 10:19:48PM -0300, James Almer wrote:
> On 9/28/2021 7:28 PM, Michael Niedermayer wrote:
> > Fixes: Timeout
> > Fixes:
> > 39089/clusterfuzz-testcase-minimized-ffmpeg_AV_CODEC_ID_MSNSIREN_fuzzer-6677219854909440
> >
> > Found-by: continuous fuzzing process
> >
tis 2021-09-28 klockan 16:41 +0200 skrev Andreas Rheinhardt:
> ff_encode_preinit() already checked the pixel format via
> AVCodec.pix_fmts.
>
> Signed-off-by: Andreas Rheinhardt
> ---
> libavcodec/zmbvenc.c | 3 ---
> 1 file changed, 3 deletions(-)
>
> diff --git a/libavcodec/zmbvenc.c
On Wed, Sep 29, 2021 at 12:51 PM Arif Driessen wrote:
> Here's your patch for `afftfilt` win_func. (copied from the surround
> filter, but checked against the code to ensure it's correct). If you like
> it, I'll do it for all the win_funcs.
>
Better add separate section, and than link to it
Here's your patch for `afftfilt` win_func. (copied from the surround
filter, but checked against the code to ensure it's correct). If you like
it, I'll do it for all the win_funcs.
---
doc/filters.texi | 27 ++-
1 file changed, 26 insertions(+), 1 deletion(-)
diff --git
Soft Works:
>
>
>> -Original Message-
>> From: ffmpeg-devel On Behalf Of
>> James Almer
>> Sent: Tuesday, 28 September 2021 22:08
>> To: ffmpeg-devel@ffmpeg.org
>> Subject: Re: [FFmpeg-devel] [PATCH v2 1/1] avfilter/frames: Ensure
>> frames are writable when processing in-place
>>
>> On
On Tue, 2021-09-28 at 16:53 -0300, James Almer wrote:
> On 9/23/2021 3:28 AM, Haihao Xiang wrote:
> > In oneVPL, MFXLoad() and MFXCreateSession() are required to create a
> > workable mfx session[1]
> >
> > Add AccelerationMode config filter for D3D9/D3D11 session (galinart)
> >
> > The default
On Tue, 2021-09-28 at 16:48 -0300, James Almer wrote:
> On 9/23/2021 3:28 AM, Haihao Xiang wrote:
> > The following Cflags has been added to libmfx.pc, so mfx/ prefix is no
> > longer needed when including mfx headers in FFmpeg.
> > Cflags: -I${includedir} -I${includedir}/mfx
> >
> > Some old
57 matches
Mail list logo
