On Wed, 1 Feb 2017 21:36:04 +
Mark Thompson wrote:
> This avoids freeing and reallocating lots of surfaces when it isn't
> necessary do so (for example, when seeking in the same stream).
> Since all of the hwaccel data is cleared on reinitialisation, this
> adds a new
On Wed, 1 Feb 2017 21:32:21 +
Mark Thompson wrote:
> For use by codec implementations which can allocate frames internally.
> ---
> Now with explicit prohibition on the user reading either field.
>
>
> doc/APIchanges | 3 +++
> libavcodec/avcodec.h | 23
Adds a new driver quirk to indicate no support at all for surface
attributes.
Based on a patch by wm4 .
---
libavutil/hwcontext_vaapi.c | 79 ++---
libavutil/hwcontext_vaapi.h | 7
2 files changed, 52 insertions(+), 34
This avoids freeing and reallocating lots of surfaces when it isn't
necessary do so (for example, when seeking in the same stream).
Since all of the hwaccel data is cleared on reinitialisation, this
adds a new hw_frames_ctx member to AVCodecInternal to store the
reused frames context. Memory use
---
Now just writes the hw_device_ctx field (no unrefs).
avconv_vaapi.c | 128 ++---
1 file changed, 13 insertions(+), 115 deletions(-)
diff --git a/avconv_vaapi.c b/avconv_vaapi.c
index 584b8b4df..13e8a8a71 100644
--- a/avconv_vaapi.c
+++
In this case, the user only supplies a device and the frame context
is allocated internally by lavc.
---
Conforming to the rules in the previous patch. Since the user isn't allowed to
look at the AVCodecContext fields we can continue to store the references
there, which also does the right
For use by codec implementations which can allocate frames internally.
---
Now with explicit prohibition on the user reading either field.
doc/APIchanges | 3 +++
libavcodec/avcodec.h | 23 ++-
libavcodec/decode.c | 1 +
libavcodec/utils.c | 1 +
On Wed, Feb 1, 2017 at 6:37 PM, Diego Biurrun wrote:
> On Wed, Feb 01, 2017 at 01:12:47PM +0100, Vittorio Giovara wrote:
>> On Wed, Feb 1, 2017 at 12:23 PM, Diego Biurrun wrote:
>> > On Wed, Feb 01, 2017 at 10:51:15AM +0100, Vittorio Giovara wrote:
>> >> On
On Wed, Feb 01, 2017 at 01:12:47PM +0100, Vittorio Giovara wrote:
> On Wed, Feb 1, 2017 at 12:23 PM, Diego Biurrun wrote:
> > On Wed, Feb 01, 2017 at 10:51:15AM +0100, Vittorio Giovara wrote:
> >> On Tue, Jan 24, 2017 at 6:12 PM, Diego Biurrun wrote:
> >> >
On Wed, Feb 1, 2017 at 12:23 PM, Diego Biurrun wrote:
> On Wed, Feb 01, 2017 at 10:51:15AM +0100, Vittorio Giovara wrote:
>> On Tue, Jan 24, 2017 at 6:12 PM, Diego Biurrun wrote:
>> > --- a/configure
>> > +++ b/configure
>> > @@ -4165,9 +4165,7 @@ EOF
>> >
>>
Something got mangled in the email.
The concept of the patch below seems good, check for every block
candidate once you find the magic nibble.
Could you please resend the two squashed in one?
Below some more comments.
On 27/01/2017 20:06, Devesh Singh wrote:
> From
On 24/01/2017 18:12, Diego Biurrun wrote:
> ---
> configure | 18 --
> 1 file changed, 4 insertions(+), 14 deletions(-)
Not against it
___
libav-devel mailing list
libav-devel@libav.org
On Wed, Feb 01, 2017 at 10:51:15AM +0100, Vittorio Giovara wrote:
> On Tue, Jan 24, 2017 at 6:12 PM, Diego Biurrun wrote:
> > --- a/configure
> > +++ b/configure
> > @@ -4165,9 +4165,7 @@ EOF
> >
> > -check_cc < > -void foo(void) { __asm__ volatile ("" ::); }
> > -EOF
> >
On Wed, Feb 1, 2017 at 11:34 AM, Anton Khirnov wrote:
> Quoting Vittorio Giovara (2017-02-01 10:44:11)
>> On Wed, Feb 1, 2017 at 10:25 AM, Anton Khirnov wrote:
>> > Video decoders always consume full packets.
>> > ---
>> > libavcodec/h264dec.c | 20
On Wed, Feb 1, 2017 at 11:52 AM, Anton Khirnov wrote:
> Currently it incorrectly compares bits with bytes.
>
> Also, move the check right before where it's relevant, so that the
> correct number of remaining bits is used.
>
> CC: libav-sta...@libav.org
> ---
>
Currently it incorrectly compares bits with bytes.
Also, move the check right before where it's relevant, so that the
correct number of remaining bits is used.
CC: libav-sta...@libav.org
---
libavcodec/svq3.c | 9 -
1 file changed, 4 insertions(+), 5 deletions(-)
diff --git
Quoting Vittorio Giovara (2017-02-01 10:47:21)
> On Wed, Feb 1, 2017 at 10:15 AM, Anton Khirnov wrote:
> > diff --git a/libavcodec/h264dec.c b/libavcodec/h264dec.c
> > index b27aa54..771520b 100644
> > --- a/libavcodec/h264dec.c
> > +++ b/libavcodec/h264dec.c
> > @@ -383,14
Quoting Vittorio Giovara (2017-02-01 10:44:11)
> On Wed, Feb 1, 2017 at 10:25 AM, Anton Khirnov wrote:
> > Video decoders always consume full packets.
> > ---
> > libavcodec/h264dec.c | 20 +++-
> > 1 file changed, 3 insertions(+), 17 deletions(-)
>
> Do you
On Tue, Jan 24, 2017 at 6:12 PM, Diego Biurrun wrote:
> ---
> @@ -4452,11 +4448,8 @@ check_code cc arm_neon.h "int16x8_t test =
> vdupq_n_s16(0)" && enable intrinsics_
>
> check_ldflags -Wl,--as-needed
>
> -if check_func dlopen; then
> -ldl=
> -elif check_func dlopen -ldl;
On Tue, Jan 24, 2017 at 6:12 PM, Diego Biurrun wrote:
> Weak dependencies on external libraries do not obviate having to
> explicitly enable these libraries, so the weak dependency does not
> simplify the configure command line nor have any real effect.
> ---
> configure | 1 -
On Tue, Jan 24, 2017 at 6:12 PM, Diego Biurrun wrote:
> ---
> configure | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/configure b/configure
> index 90bc8b2..9ff6ad0 100755
> --- a/configure
> +++ b/configure
> @@ -2471,6 +2471,7 @@
On Tue, Jan 24, 2017 at 6:12 PM, Diego Biurrun wrote:
> Simplifies checking for conditions in external library headers and
> aborting if said conditions are not met.
> ---
> configure | 17 +++--
> 1 file changed, 11 insertions(+), 6 deletions(-)
probably ok
--
On Tue, Jan 24, 2017 at 6:12 PM, Diego Biurrun wrote:
> Simplifies checking for external library headers and aborting if
> the external library support was requested, but is not available.
> ---
> configure | 15 +++
> 1 file changed, 11 insertions(+), 4
On Tue, Jan 24, 2017 at 6:12 PM, Diego Biurrun wrote:
> ---
> configure | 4 +---
> 1 file changed, 1 insertion(+), 3 deletions(-)
>
> diff --git a/configure b/configure
> index f7436c5..c543c05 100755
> --- a/configure
> +++ b/configure
> @@ -4165,9 +4165,7 @@ EOF
> sym=$($nm
On Tue, Jan 24, 2017 at 6:12 PM, Diego Biurrun wrote:
> ---
> configure | 27 +--
> 1 file changed, 13 insertions(+), 14 deletions(-)
>
> diff --git a/configure b/configure
> index 2c8b57e..f7436c5 100755
> --- a/configure
> +++ b/configure
> @@ -4097,26
On Tue, Jan 24, 2017 at 6:12 PM, Diego Biurrun wrote:
> ---
> configure | 14 --
> 1 file changed, 4 insertions(+), 10 deletions(-)
>
> diff --git a/configure b/configure
> index 61d1807..2c8b57e 100755
> --- a/configure
> +++ b/configure
> @@ -3746,18 +3746,12 @@
On Wed, Feb 1, 2017 at 10:15 AM, Anton Khirnov wrote:
> diff --git a/libavcodec/h264dec.c b/libavcodec/h264dec.c
> index b27aa54..771520b 100644
> --- a/libavcodec/h264dec.c
> +++ b/libavcodec/h264dec.c
> @@ -383,14 +383,16 @@ static av_cold int h264_decode_init(AVCodecContext
On 01/02/2017 10:25, Anton Khirnov wrote:
> Video decoders always consume full packets.
> ---
> libavcodec/h264dec.c | 20 +++-
> 1 file changed, 3 insertions(+), 17 deletions(-)
>
Possibly OK.
___
libav-devel mailing list
On Wed, Feb 1, 2017 at 10:25 AM, Anton Khirnov wrote:
> Video decoders always consume full packets.
> ---
> libavcodec/h264dec.c | 20 +++-
> 1 file changed, 3 insertions(+), 17 deletions(-)
Do you mean that get_consumed_bytes() was always equal to buf_size?
On 01/02/2017 10:39, Anton Khirnov wrote:
> This is no longer done automatically for filters marked as
> hwframe-aware.
> ---
> libavfilter/vf_scale_npp.c | 8 +---
> 1 file changed, 5 insertions(+), 3 deletions(-)
>
> diff --git a/libavfilter/vf_scale_npp.c b/libavfilter/vf_scale_npp.c
>
On 01/02/2017 10:15, Anton Khirnov wrote:
> This is extremely fragile against reordering and hides what is actually
> being copied. Copy all the fields manually instead.
> ---
> libavcodec/pthread_frame.c | 6 +++---
> 1 file changed, 3 insertions(+), 3 deletions(-)
Sure.
On 01/02/2017 10:15, Anton Khirnov wrote:
> This is fragile and dangerous, allocate separate private data for each
> per-thread context.
> ---
> libavcodec/pthread_frame.c | 19 ---
> 1 file changed, 16 insertions(+), 3 deletions(-)
>
Seems fine.
This is no longer done automatically for filters marked as
hwframe-aware.
---
libavfilter/vf_scale_npp.c | 8 +---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git a/libavfilter/vf_scale_npp.c b/libavfilter/vf_scale_npp.c
index 0e636a9..be1f81f 100644
---
On 01/02/2017 10:15, Anton Khirnov wrote:
> ---
> libavcodec/decode.c | 6 --
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/libavcodec/decode.c b/libavcodec/decode.c
> index d5210d0..44cab68 100644
> --- a/libavcodec/decode.c
> +++ b/libavcodec/decode.c
> @@ -923,8
On Wed, Feb 1, 2017 at 10:25 AM, Anton Khirnov wrote:
> ---
> libavcodec/h264dec.c | 5 +
> 1 file changed, 1 insertion(+), 4 deletions(-)
>
> diff --git a/libavcodec/h264dec.c b/libavcodec/h264dec.c
> index b92795d..6ce0287 100644
> --- a/libavcodec/h264dec.c
> +++
On Wed, Feb 1, 2017 at 10:25 AM, Anton Khirnov wrote:
> It is never modified anymore.
> ---
> libavcodec/h264_ps.h| 3 +--
> libavcodec/h264_slice.c | 2 +-
> 2 files changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/libavcodec/h264_ps.h b/libavcodec/h264_ps.h
>
On 01/02/2017 10:15, Anton Khirnov wrote:
> This is a constant codec property, so a capability flag is more appropriate.
> ---
> doc/multithreading.txt | 5 +++--
> libavcodec/h264dec.c | 5 ++---
> libavcodec/hevcdec.c | 5 ++---
> libavcodec/internal.h | 20
---
libavcodec/h264dec.c | 5 +
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/libavcodec/h264dec.c b/libavcodec/h264dec.c
index b92795d..6ce0287 100644
--- a/libavcodec/h264dec.c
+++ b/libavcodec/h264dec.c
@@ -699,10 +699,7 @@ static int h264_decode_frame(AVCodecContext *avctx,
Video decoders always consume full packets.
---
libavcodec/h264dec.c | 20 +++-
1 file changed, 3 insertions(+), 17 deletions(-)
diff --git a/libavcodec/h264dec.c b/libavcodec/h264dec.c
index 6ce0287..a68a9ad 100644
--- a/libavcodec/h264dec.c
+++ b/libavcodec/h264dec.c
@@ -628,19
Current code is excessively complicated and hard to read. Use the same
logic as in the HEVC decoder, augmented with detection of insufficient
delay.
---
libavcodec/h264_picture.c | 2 +-
libavcodec/h264_refs.c| 18 +
libavcodec/h264_slice.c | 185
It is never modified anymore.
---
libavcodec/h264_ps.h| 3 +--
libavcodec/h264_slice.c | 2 +-
2 files changed, 2 insertions(+), 3 deletions(-)
diff --git a/libavcodec/h264_ps.h b/libavcodec/h264_ps.h
index 9a32d93..1b482f3 100644
--- a/libavcodec/h264_ps.h
+++ b/libavcodec/h264_ps.h
@@
---
libavcodec/decode.c | 6 --
1 file changed, 4 insertions(+), 2 deletions(-)
diff --git a/libavcodec/decode.c b/libavcodec/decode.c
index d5210d0..44cab68 100644
--- a/libavcodec/decode.c
+++ b/libavcodec/decode.c
@@ -923,8 +923,10 @@ static int update_frame_pool(AVCodecContext *avctx,
Use the AVFrame API to properly allocate and free frames for delayed
release.
---
libavcodec/pthread_frame.c | 40 +---
1 file changed, 25 insertions(+), 15 deletions(-)
diff --git a/libavcodec/pthread_frame.c b/libavcodec/pthread_frame.c
index
This is extremely fragile against reordering and hides what is actually
being copied. Copy all the fields manually instead.
---
libavcodec/pthread_frame.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/libavcodec/pthread_frame.c b/libavcodec/pthread_frame.c
index
This is fragile and dangerous, allocate separate private data for each
per-thread context.
---
libavcodec/pthread_frame.c | 19 ---
1 file changed, 16 insertions(+), 3 deletions(-)
diff --git a/libavcodec/pthread_frame.c b/libavcodec/pthread_frame.c
index d4a3406..12d5481 100644
Currently the frame pool used by the default get_buffer2()
implementation is a single struct, allocated when opening the decoder.
A pointer to it is simply copied to each frame thread and we assume that
no thread attempts to modify it at an unexpected time. This is rather
fragile and potentially
This is a constant codec property, so a capability flag is more appropriate.
---
doc/multithreading.txt | 5 +++--
libavcodec/h264dec.c | 5 ++---
libavcodec/hevcdec.c | 5 ++---
libavcodec/internal.h | 20 +---
libavcodec/mimic.c | 3 +--
The current design, where
- proper init is called for the first per-thread context
- first thread's private data is copied into private data for all the
other threads
- a "fixup" function is called for all the other threads to e.g.
allocate dynamically allocated data
is very fragile and hard
On 01/02/2017 01:05, Vittorio Giovara wrote:
> IMO the final executable should still stay in the build directory.
+1
___
libav-devel mailing list
libav-devel@libav.org
https://lists.libav.org/mailman/listinfo/libav-devel
From: Alexandra Hájková
---
Changed suspicious
-AV_WL32(>gb_slice.buffer[1], header ^ s->watermark_key);
+AV_WL32(>slice_buf[1], header ^ s->watermark_key);
to
-AV_WL32(>gb_slice.buffer[1], header ^ s->watermark_key);
+
50 matches
Mail list logo