| Message ID | 1472603168-3871-2-git-send-email-qyot27@gmail.com |
|---|---|
| State | Rejected |
| Delegated to: | Stephen Hutchinson |
| Headers | show |
2016-08-31 2:26 GMT+02:00 Stephen Hutchinson <qyot27@gmail.com>: > -//AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer > + AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer Is this related to alpha offsets? Carl Eugen
On 8/31/2016 9:37 AM, Carl Eugen Hoyos wrote: > 2016-08-31 2:26 GMT+02:00 Stephen Hutchinson <qyot27@gmail.com>: >> -//AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer >> + AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer > > Is this related to alpha offsets? > It's part of the changes to the header upstream. The commit message probably should have been 'minor catch-up, inc. alpha offsets'.
On 8/30/2016 8:26 PM, Stephen Hutchinson wrote: > --- > compat/avisynth/avisynth_c.h | 8 ++++++-- > 1 file changed, 6 insertions(+), 2 deletions(-) > > diff --git a/compat/avisynth/avisynth_c.h b/compat/avisynth/avisynth_c.h > index 605b92a..2f84dd1 100644 > --- a/compat/avisynth/avisynth_c.h > +++ b/compat/avisynth/avisynth_c.h > @@ -533,7 +533,11 @@ typedef struct AVS_VideoFrame { > volatile long refcount; > AVS_VideoFrameBuffer * vfb; > int offset, pitch, row_size, height, offsetU, offsetV, pitchUV; // U&V offsets are from top of picture. > - int row_sizeUV, heightUV; > + int row_sizeUV, heightUV; // for Planar RGB offsetU, offsetV is for the 2nd and 3rd Plane. > + // for Planar RGB pitchUV and row_sizeUV = 0, because when no VideoInfo (MakeWriteable) > + // the decision on existance of UV is checked by zero pitch > + // AVS+ extension, avisynth.h: class does not break plugins if appended here > + int offsetA, pitchA, row_sizeA; // 4th alpha plane support, pitch and row_size is 0 is none > } AVS_VideoFrame; > > // Access functions for AVS_VideoFrame > @@ -753,7 +757,7 @@ enum { > AVS_CPUF_SSSE3 = 0x200, // Core 2 > AVS_CPUF_SSE4 = 0x400, // Penryn, Wolfdale, Yorkfield > AVS_CPUF_SSE4_1 = 0x400, > -//AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer > + AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer > AVS_CPUF_SSE4_2 = 0x1000, // Nehalem > //AVS_CPUF_AVX2 = 0x2000, // Haswell > //AVS_CPUF_AVX512 = 0x4000, // Knights Landing > Consider this patch revoked. The changes don't affect FFmpeg's handling of the discrete alpha formats, so my being cautious and sending this patch wasn't necessary.
2016-08-31 16:16 GMT+02:00 Stephen Hutchinson <qyot27@gmail.com>: > On 8/31/2016 9:37 AM, Carl Eugen Hoyos wrote: >> >> 2016-08-31 2:26 GMT+02:00 Stephen Hutchinson <qyot27@gmail.com>: >>> >>> -//AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer >>> + AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer >> >> >> Is this related to alpha offsets? >> > > It's part of the changes to the header upstream. Then "sync with upstream header" would be less confusing imo. Sorry for missing your message, Carl Eugen
On 9/27/2016 3:12 AM, Carl Eugen Hoyos wrote: > 2016-08-31 16:16 GMT+02:00 Stephen Hutchinson <qyot27@gmail.com>: >> On 8/31/2016 9:37 AM, Carl Eugen Hoyos wrote: >>> >>> 2016-08-31 2:26 GMT+02:00 Stephen Hutchinson <qyot27@gmail.com>: >>>> >>>> -//AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer >>>> + AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer >>> >>> >>> Is this related to alpha offsets? >>> >> >> It's part of the changes to the header upstream. > > Then "sync with upstream header" would be less confusing imo. > > Sorry for missing your message, Carl Eugen > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > It's not necessary to sync with upstream yet, so the header update patch was revoked. Only the Planar RGB fix patch is actually necessary.
On Tue, Sep 27, 2016 at 12:48:18PM -0400, Stephen Hutchinson wrote: > On 9/27/2016 3:12 AM, Carl Eugen Hoyos wrote: > >2016-08-31 16:16 GMT+02:00 Stephen Hutchinson <qyot27@gmail.com>: > >>On 8/31/2016 9:37 AM, Carl Eugen Hoyos wrote: > >>> > >>>2016-08-31 2:26 GMT+02:00 Stephen Hutchinson <qyot27@gmail.com>: > >>>> > >>>>-//AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer > >>>>+ AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer > >>> > >>> > >>>Is this related to alpha offsets? > >>> > >> > >>It's part of the changes to the header upstream. > > > >Then "sync with upstream header" would be less confusing imo. > > > >Sorry for missing your message, Carl Eugen > >_______________________________________________ > >ffmpeg-devel mailing list > >ffmpeg-devel@ffmpeg.org > >http://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > > > It's not necessary to sync with upstream yet, so > the header update patch was revoked. Only the > Planar RGB fix patch is actually necessary. can you update the status for the patch(es) on https://patchwork.ffmpeg.org/project/ffmpeg/list/?submitter=49 ? so developers know what needs a review, needs to be applied and what is on hold/revovked, ... thx [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB You can kill me, but you cannot change the truth.
On 9/27/2016 4:11 PM, Michael Niedermayer wrote: > can you update the status for the patch(es) on > https://patchwork.ffmpeg.org/project/ffmpeg/list/?submitter=49 > ? > > so developers know what needs a review, needs to be applied and what > is on hold/revovked, ... > > thx > Done. I think I did it right, hopefully.
On Tue, Sep 27, 2016 at 10:07:20PM -0400, Stephen Hutchinson wrote: > On 9/27/2016 4:11 PM, Michael Niedermayer wrote: > >can you update the status for the patch(es) on > >https://patchwork.ffmpeg.org/project/ffmpeg/list/?submitter=49 > >? > > > >so developers know what needs a review, needs to be applied and what > >is on hold/revovked, ... > > > >thx > > > > Done. I think I did it right, hopefully. thx [...]
On 9/27/2016 10:07 PM, Stephen Hutchinson wrote: > On 9/27/2016 4:11 PM, Michael Niedermayer wrote: >> can you update the status for the patch(es) on >> https://patchwork.ffmpeg.org/project/ffmpeg/list/?submitter=49 >> ? >> >> so developers know what needs a review, needs to be applied and what >> is on hold/revovked, ... >> >> thx >> > > Done. I think I did it right, hopefully. Since there haven't been any responses, I've changed the status on the Planar RGB fix patch on Patchwork to Accepted (which I assume is what says it needs to be applied). This should definitely go in before 3.2 is split off.
On 10/17/2016 9:48 PM, Stephen Hutchinson wrote: > On 9/27/2016 10:07 PM, Stephen Hutchinson wrote: >> On 9/27/2016 4:11 PM, Michael Niedermayer wrote: >>> can you update the status for the patch(es) on >>> https://patchwork.ffmpeg.org/project/ffmpeg/list/?submitter=49 >>> ? >>> >>> so developers know what needs a review, needs to be applied and what >>> is on hold/revovked, ... >>> >>> thx >>> >> >> Done. I think I did it right, hopefully. > > Since there haven't been any responses, I've changed the status on > the Planar RGB fix patch on Patchwork to Accepted (which I assume is > what says it needs to be applied). This should definitely go in before > 3.2 is split off. Ok, after seeing the patch disappear from the lists requiring action, apparently 'Accepted' was wrong. The meaning of Patchwork's states as it relates to submitted patches needs to be documented somewhere, because Googling it wasn't bringing it up when I looked.
diff --git a/compat/avisynth/avisynth_c.h b/compat/avisynth/avisynth_c.h index 605b92a..2f84dd1 100644 --- a/compat/avisynth/avisynth_c.h +++ b/compat/avisynth/avisynth_c.h @@ -533,7 +533,11 @@ typedef struct AVS_VideoFrame { volatile long refcount; AVS_VideoFrameBuffer * vfb; int offset, pitch, row_size, height, offsetU, offsetV, pitchUV; // U&V offsets are from top of picture. - int row_sizeUV, heightUV; + int row_sizeUV, heightUV; // for Planar RGB offsetU, offsetV is for the 2nd and 3rd Plane. + // for Planar RGB pitchUV and row_sizeUV = 0, because when no VideoInfo (MakeWriteable) + // the decision on existance of UV is checked by zero pitch + // AVS+ extension, avisynth.h: class does not break plugins if appended here + int offsetA, pitchA, row_sizeA; // 4th alpha plane support, pitch and row_size is 0 is none } AVS_VideoFrame; // Access functions for AVS_VideoFrame @@ -753,7 +757,7 @@ enum { AVS_CPUF_SSSE3 = 0x200, // Core 2 AVS_CPUF_SSE4 = 0x400, // Penryn, Wolfdale, Yorkfield AVS_CPUF_SSE4_1 = 0x400, -//AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer + AVS_CPUF_AVX = 0x800, // Sandy Bridge, Bulldozer AVS_CPUF_SSE4_2 = 0x1000, // Nehalem //AVS_CPUF_AVX2 = 0x2000, // Haswell //AVS_CPUF_AVX512 = 0x4000, // Knights Landing