Hi,
On Tue, Dec 10, 2024, at 19:29, Anton Khirnov wrote:
> Hi,
> Quoting Thomas Guillem via ffmpeg-devel (2024-11-29 11:44:21)
>> avcodec_get_hw_frames_parameters(), called by the user from get_format,
>> is allocating ctx->internal->hwaccel_priv_data. But the hardware
&g
Ping.
On Fri, Nov 29, 2024, at 11:44, Thomas Guillem via ffmpeg-devel wrote:
> avcodec_get_hw_frames_parameters(), called by the user from get_format,
> is allocating ctx->internal->hwaccel_priv_data. But the hardware
> decoding setup may fail on the user side and it may fallb
Sorry for the spam. "avcodec: uninit hwaccel in case of software decoder" is a
better version.
On Fri, Nov 29, 2024, at 09:37, Thomas Guillem via ffmpeg-devel wrote:
> This patch is an alternative to the patch "avcodec: add
> avcodec_reset_hw_frames_parameters():
>
>
avcodec_get_hw_frames_parameters(), called by the user from get_format,
is allocating ctx->internal->hwaccel_priv_data. But the hardware
decoding setup may fail on the user side and it may fallback to software
decoding. In that case, ctx->internal->hwaccel_priv_data is still
allocated but not used
Or you can check the alternative patch: "avcodec/pthread_frame: rework assert"
On Thu, Nov 28, 2024, at 16:57, Thomas Guillem via ffmpeg-devel wrote:
> usage example:
>
> AVBufferRef *hwframes_ref;
> int ret = avcodec_get_hw_frames_parameters
This patch is an alternative to the patch "avcodec: add
avcodec_reset_hw_frames_parameters():
On Fri, Nov 29, 2024, at 09:34, Thomas Guillem via ffmpeg-devel wrote:
> avcodec_get_hw_frames_parameters(), called by the user from get_format,
> is allocating ctx->internal->hwaccel_p
avcodec_get_hw_frames_parameters(), called by the user from get_format,
is allocating ctx->internal->hwaccel_priv_data. But the hardware
decoding setup may fail on the user side and it may fallback to software
decoding. In that case, ctx->internal->hwaccel_priv_data is still
allocated but not used
usage example:
AVBufferRef *hwframes_ref;
int ret = avcodec_get_hw_frames_parameters(ctx, hwdev_ref, hwfmt,
&hwframes_ref);
...
ret = av_hwframe_ctx_init(hwframes_ref);
if (ret < 0)
{
av_buffer_unref(&hwframes_ref);
Happy monthly anniversary to my not very interesting patch.
FYI, OpenSuse applied it for their ffmpeg build...
On Mon, Oct 14, 2024, at 10:54, Thomas Guillem via ffmpeg-devel wrote:
> Ping.
>
> On Mon, Oct 7, 2024, at 17:43, Thomas Guillem via ffmpeg-devel wrote:
>> Fixes the f
Thanks !
On Wed, Oct 30, 2024, at 14:34, Zhao Zhili wrote:
>> On Oct 30, 2024, at 21:19, Thomas Guillem via ffmpeg-devel
>> wrote:
>>
>> Happy monthly anniversary to my not very interesting patch.
>>
>> FYI, OpenSuse applied it for their ffmpeg b
Ping.
On Mon, Oct 7, 2024, at 17:43, Thomas Guillem via ffmpeg-devel wrote:
> Fixes the following assert:
>
> [7f1df83d17e0] vaapi generic error:
> avcodec_get_hw_frames_parameters failed: -22
> Assertion p_dst->hwaccel_threadsafe || (!dst->hwaccel &&
> !
Fixes the following assert:
[7f1df83d17e0] vaapi generic error: avcodec_get_hw_frames_parameters
failed: -22
Assertion p_dst->hwaccel_threadsafe || (!dst->hwaccel &&
!dst->internal->hwaccel_priv_data) failed at libavcodec/pthread_frame.c:349
Reproduced from VLC with VAAPI, when fallbacking
On Mon, Oct 7, 2024, at 17:12, Rémi Denis-Courmont wrote:
> Le 30 septembre 2024 17:27:46 GMT+09:00, Thomas Guillem via
> ffmpeg-devel a écrit :
>>Fixes the following assert:
>>
>>[7f1df83d17e0] vaapi generic error: avcodec_get_hw_frames_parameters
>>
Ping.
On Mon, Sep 30, 2024, at 13:44, Thomas Guillem via ffmpeg-devel wrote:
> cf. https://trac.ffmpeg.org/ticket/11013
>
>
> On Mon, Sep 30, 2024, at 10:27, Thomas Guillem via ffmpeg-devel wrote:
>> Fixes the following assert:
>>
>> [7f1
cf. https://trac.ffmpeg.org/ticket/11013
On Mon, Sep 30, 2024, at 10:27, Thomas Guillem via ffmpeg-devel wrote:
> Fixes the following assert:
>
> [7f1df83d17e0] vaapi generic error:
> avcodec_get_hw_frames_parameters failed: -22
> Assertion p_dst->hwaccel_threadsafe
Fixes the following assert:
[7f1df83d17e0] vaapi generic error: avcodec_get_hw_frames_parameters
failed: -22
Assertion p_dst->hwaccel_threadsafe || (!dst->hwaccel &&
!dst->internal->hwaccel_priv_data) failed at libavcodec/pthread_frame.c:349
Reproduced from VLC with VAAPI, when fallbacking
On Tue, Sep 12, 2023, at 15:11, Andreas Rheinhardt wrote:
> Thomas Guillem via ffmpeg-devel:
>> Same than the previous commit but with FF_DECODE_ERROR_DECODE_SLICES
>>
>> Fix the following data-race:
>>
>> WARNING: ThreadSanitizer: data race (pid=55935)
>&
Set the FF_DECODE_ERROR_CONCEALMENT_ACTIVE flags on the AVFrane before
outputing it. Store in in the H264Picture in the meantime, where it
won't be read/write by other threads.
Fix the following data-race:
WARNING: ThreadSanitizer: data race (pid=55134)
Write of size 4 at 0x7b507f78 by thre
Same than the previous commit but with FF_DECODE_ERROR_DECODE_SLICES
Fix the following data-race:
WARNING: ThreadSanitizer: data race (pid=55935)
Write of size 4 at 0x7b509378 by thread T1 (mutexes: write M608):
#0 decode_nal_units src/libavcodec/h264dec.c:742 (ffmpeg+0xb19dd6)
#1 h
This will allow to fix data-races when ff_er_frame_end() is called after
ff_thread_finish_setup()
---
libavcodec/error_resilience.c | 12 ++--
libavcodec/error_resilience.h | 2 +-
libavcodec/h263dec.c | 6 --
libavcodec/h264dec.c | 3 ++-
libavcodec/mpeg12dec.c
Hello,
Sorry, but I'm not allowed to share these cameras' footage.
I was able to test them, and it runs fine on last ffmpeg.
On Thu, Feb 17, 2022, at 15:32, Michael Niedermayer wrote:
> On Thu, Feb 17, 2022 at 11:01:38AM +0100, Thomas Guillem wrote:
>> ---
>> libavf
---
libavformat/riff.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/libavformat/riff.c b/libavformat/riff.c
index 27a9706510..23e6306b66 100644
--- a/libavformat/riff.c
+++ b/libavformat/riff.c
@@ -32,6 +32,7 @@
* files use it as well.
*/
const AVCodecTag ff_codec_bmp_tags[] = {
+
On Fri, Dec 3, 2021, at 09:52, Anton Khirnov wrote:
> Quoting Thomas Guillem (2021-12-02 20:44:33)
>>
>> On Thu, Dec 2, 2021, at 15:05, Andreas Rheinhardt wrote:
>> > 1. hw_device_ctx is in the public AVCodecContext, not in the private
>> > AVCodecInternal.
>
On Thu, Dec 2, 2021, at 15:05, Andreas Rheinhardt wrote:
> 1. hw_device_ctx is in the public AVCodecContext, not in the private
> AVCodecInternal.
> 2. ctx->internal is intended to be allocated for all the internal
> AVCodecContexts; the check is only there to ensure that we don't crash
> if an a
On Thu, Dec 2, 2021, at 14:38, Andreas Rheinhardt wrote:
> Thomas Guillem:
>> Reproduced when using the VAAPI va module on VLC 4.0. No leaks when
>> setting thread count to 1.
>> ---
>> libavcodec/pthread_frame.c | 1 +
>> 1 file changed, 1 insertion(
Reproduced when using the VAAPI va module on VLC 4.0. No leaks when
setting thread count to 1.
---
libavcodec/pthread_frame.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/libavcodec/pthread_frame.c b/libavcodec/pthread_frame.c
index 73b1b7d7d9..4c578aa44a 100644
--- a/libavcodec/pthread_fra
On Thu, Jul 19, 2018, at 18:27, Aman Gupta wrote:
> On Fri, Jul 6, 2018 at 10:55 PM Thomas Guillem wrote:
>
> >
> >
> > On Thu, Jul 5, 2018, at 20:49, Aman Gupta wrote:
> > > On Wed, Jul 4, 2018 at 3:05 AM Thomas Guillem wrote:
> > >
> > >
n?
>
> Has a radar been filed with Apple about this?
This issue doesn't reproduce on the iOS 12 beta (with the same device), so I
guess they are aware or it has been fixed.
>
> On Wed, Jul 4, 2018 at 1:44 PM Thomas Guillem wrote:
>
> > On Wed, Jul 4, 2018, at 0
On Thu, Jul 5, 2018, at 20:49, Aman Gupta wrote:
> On Wed, Jul 4, 2018 at 3:05 AM Thomas Guillem wrote:
>
> > On macOS, a zero rc_max_rate cause an error from
> > VTSessionSetProperty(kVTCompressionPropertyKey_DataRateLimits).
> >
> > on iOS (depending on devic
On Wed, Jul 4, 2018, at 09:05, Thomas Guillem wrote:
> On macOS, a zero rc_max_rate cause an error from
> VTSessionSetProperty(kVTCompressionPropertyKey_DataRateLimits).
>
> on iOS (depending on device/version), a zero rc_max_rate cause invalid
> arguments from the vtenc_output_cal
On Tue, Jul 3, 2018, at 20:53, Carl Eugen Hoyos wrote:
> 2018-07-03 17:05 GMT+02:00, Thomas Guillem :
> > On macOS, a zero rc_max_rate cause an error from
> > VTSessionSetProperty(kVTCompressionPropertyKey_DataRateLimits).
> >
> > on iOS (depending on device/version
---
libavcodec/videotoolboxenc.c | 72 ++--
1 file changed, 36 insertions(+), 36 deletions(-)
diff --git a/libavcodec/videotoolboxenc.c b/libavcodec/videotoolboxenc.c
index aa9aae7e05..050e5cefee 100644
--- a/libavcodec/videotoolboxenc.c
+++ b/libavcodec/videotoolb
On macOS, a zero rc_max_rate cause an error from
VTSessionSetProperty(kVTCompressionPropertyKey_DataRateLimits).
on iOS (depending on device/version), a zero rc_max_rate cause invalid
arguments from the vtenc_output_callback after few frames and then a crash
within the VideoToolbox library.
---
l
On macOS, a zero rc_max_rate cause an error from
VTSessionSetProperty(kVTCompressionPropertyKey_DataRateLimits).
on iOS (depending on device/version), a zero rc_max_rate cause invalid
arguments from the vtenc_output_callback after few frames and then a crash
within the VideoToolbox library.
---
l
Ping?
On Thu, Jun 14, 2018, at 17:48, Thomas Guillem wrote:
> Cf. comment. Restart the VT session when the APP goes from foreground to
> background and vice versa.
> ---
> libavcodec/videotoolboxenc.c | 23 +--
> 1 file changed, 21 insertions(+), 2 deletio
Cf. comment. Restart the VT session when the APP goes from foreground to
background and vice versa.
---
libavcodec/videotoolboxenc.c | 23 +--
1 file changed, 21 insertions(+), 2 deletions(-)
diff --git a/libavcodec/videotoolboxenc.c b/libavcodec/videotoolboxenc.c
index 106005
Cf. comment. Restart the VT session when the APP goes from foreground to
background and vice versa.
---
libavcodec/videotoolboxenc.c | 25 +++--
1 file changed, 23 insertions(+), 2 deletions(-)
diff --git a/libavcodec/videotoolboxenc.c b/libavcodec/videotoolboxenc.c
index 1060
Split vtenc_init() into vtenc_init() (VTEncContext initialization) and
vtenc_configure_encoder() (creates the vt session).
This commit will allow to restart the vt session while encoding.
---
libavcodec/videotoolboxenc.c | 48 +---
1 file changed, 28 insertions(+),
The leak could happen when the vtenc_create_encoder() function failed.
---
libavcodec/videotoolboxenc.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/libavcodec/videotoolboxenc.c b/libavcodec/videotoolboxenc.c
index 7796a685c2..9f2a71b15d 100644
--- a/libavcodec/videotoo
On Fri, Dec 8, 2017, at 17:15, Mark Thompson wrote:
> From: Thomas Guillem
>
> This fixes the use of old ad-hoc methods which are still supported by
> some
> hwaccels which also support newer methods (DXVA2, VAAPI, VDPAU,
> videotoolbox) - without the method being visible h
---
libavcodec/hwaccel.h | 8
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/libavcodec/hwaccel.h b/libavcodec/hwaccel.h
index 16ee822920..292a87f6b5 100644
--- a/libavcodec/hwaccel.h
+++ b/libavcodec/hwaccel.h
@@ -75,15 +75,15 @@ typedef struct AVCodecHWConfigInternal {
Add a boolean to specify if an API can work with
AV_CODEC_HW_CONFIG_METHOD_AD_HOC.
---
libavcodec/hwaccel.h | 17 +
1 file changed, 9 insertions(+), 8 deletions(-)
diff --git a/libavcodec/hwaccel.h b/libavcodec/hwaccel.h
index ae55527c2f..16ee822920 100644
--- a/libavcodec/hwaccel
42 matches
Mail list logo
