On Sun, Apr 28, 2013 at 04:38:48PM -0400, Derek Buitenhuis wrote:
> On 2013-04-27 8:20 PM, Diego Biurrun wrote:
> > --- a/libavcodec/x86/Makefile
> > +++ b/libavcodec/x86/Makefile
> > @@ -10,7 +10,8 @@ OBJS-$(CONFIG_H264CHROMA)              += 
> > x86/h264chroma_init.o
> >  OBJS-$(CONFIG_H264DSP)                 += x86/h264dsp_init.o
> >  OBJS-$(CONFIG_H264PRED)                += x86/h264_intrapred_init.o
> >  OBJS-$(CONFIG_H264QPEL)                += x86/h264_qpel.o
> > -OBJS-$(CONFIG_HPELDSP)                 += x86/hpeldsp_init.o
> > +OBJS-$(CONFIG_HPELDSP)                 += x86/fpel_mmx.o                \
> > +                                          x86/hpeldsp_init.o
> >  OBJS-$(CONFIG_LPC)                     += x86/lpc.o
> >  OBJS-$(CONFIG_MLP_DECODER)             += x86/mlpdsp.o
> >  OBJS-$(CONFIG_MPEGAUDIODSP)            += x86/mpegaudiodec.o
> > @@ -33,6 +34,7 @@ OBJS-$(CONFIG_XMM_CLOBBER_TEST)        += x86/w64xmmtest.o
> >  
> >  MMX-OBJS-$(CONFIG_DSPUTIL)             += x86/dsputil_mmx.o             \
> >                                            x86/fdct.o                    \
> > +                                          x86/fpel_mmx.o                \
> >                                            x86/idct_mmx_xvid.o           \
> >                                            x86/idct_sse2_xvid.o          \
> >                                            x86/simple_idct.o             \
> 
> Given the code was duplicated before, I am assuming that these are indeed the
> only two things which rely on it? They've passed your "build everything alone"
> script, I assume.

I did not run the script, but the functions were static before.

> > --- a/libavcodec/x86/dsputil_mmx.h
> > +++ b/libavcodec/x86/dsputil_mmx.h
> > @@ -155,6 +155,11 @@ void ff_add_pixels_clamped_mmx(const int16_t *block, 
> > uint8_t *pixels, int line_s
> >  void ff_put_pixels_clamped_mmx(const int16_t *block, uint8_t *pixels, int 
> > line_size);
> >  void ff_put_signed_pixels_clamped_mmx(const int16_t *block, uint8_t 
> > *pixels, int line_size);
> >  
> > +
> > +void ff_put_pixels8_mmx(uint8_t *block, const uint8_t *pixels,
> > +                        ptrdiff_t line_size, int h);
> > +void ff_put_pixels16_mmx(uint8_t *block, const uint8_t *pixels,
> > +                         ptrdiff_t line_size, int h);
> 
> Any particular reason they're still in dsputil's header?

More put_pixels functions are declared in there.  This header could get
cleaned up at some point as well I guess.

> > --- /dev/null
> > +++ b/libavcodec/x86/fpel_mmx.c
> > @@ -0,0 +1,94 @@
> > +/*
> > + * MMX optimized DSP utils
> 
> Kind of vague.
> 
> > + * Copyright (c) 2000, 2001 Fabrice Bellard
> > + * Copyright (c) 2002-2004 Michael Niedermayer <[email protected]>
> > + *
> > + * MMX optimization by Nick Kurshev <[email protected]>
> 
> This is confusing. Shouldn't this just be Nick Kurshev here, since
> this file only contains MMX?

I could not be bothered to investigate closer; I'll do it.

> > +#if HAVE_INLINE_ASM
> 
> HAVE_MMX_INLINE?

Yeah, cannot hurt...

> > --- a/libavcodec/x86/hpeldsp_init.c
> > +++ b/libavcodec/x86/hpeldsp_init.c
> > @@ -74,6 +74,10 @@ void ff_avg_pixels8_xy2_mmxext(uint8_t *block, const 
> > uint8_t *pixels,
> >  void ff_avg_pixels8_xy2_3dnow(uint8_t *block, const uint8_t *pixels,
> >                                ptrdiff_t line_size, int h);
> >  
> > +#define put_pixels8_mmx         ff_put_pixels8_mmx
> > +#define put_pixels16_mmx        ff_put_pixels16_mmx
> > +#define put_no_rnd_pixels8_mmx  ff_put_pixels8_mmx
> > +#define put_no_rnd_pixels16_mmx ff_put_pixels16_mmx
> 
> Eww... I know it's unrelated. But still. Ew. Oh well.

Welcome to dsputil ... :)

Diego
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to