On 2014-11-16 18:13:33 +0200, Martin Storsjö wrote: > Don't include the function pointer table in the code segment > in arm mode. > > This shouldn't have any significant performance effect. It does > end up as a few more instructions than before, for ARM, but > only at the entry to this function, not within the fft functions > themselves. > --- > libavcodec/arm/fft_vfp.S | 15 +++++---------- > 1 file changed, 5 insertions(+), 10 deletions(-) > > diff --git a/libavcodec/arm/fft_vfp.S b/libavcodec/arm/fft_vfp.S > index 91b833b..9c7b54e 100644 > --- a/libavcodec/arm/fft_vfp.S > +++ b/libavcodec/arm/fft_vfp.S > @@ -30,14 +30,10 @@ > function ff_fft_calc_vfp, export=1 > ldr ip, [a1, #0] @ nbits > mov a1, a2 > -A ldr pc, [pc, ip, lsl #2] > -A .word 0 > -A .word 0 > -A .word 0 > -T movrel a2, (fft_tab_vfp - 8) > -T ldr pc, [a2, ip, lsl #2] > -T endfunc > -T const fft_tab_vfp > + movrel a2, (fft_tab_vfp - 8) > + ldr pc, [a2, ip, lsl #2] > +endfunc > +const fft_tab_vfp > .word fft4_vfp > .word fft8_vfp > .word X(ff_fft16_vfp) @ this one alone is exported > @@ -53,8 +49,7 @@ T const fft_tab_vfp > .word fft16384_vfp > .word fft32768_vfp > .word fft65536_vfp > -A endfunc > -T endconst > +endconst
ok Janne _______________________________________________ libav-devel mailing list [email protected] https://lists.libav.org/mailman/listinfo/libav-devel
