arm: Add VFP-accelerated version of fft16

               Before           After
               Mean    StdDev   Mean    StdDev  Change
This function   1389.3  4.2       967.8  35.1   +43.6%
Overall        15577.5 83.2     15400.0 336.4    +1.2%
Signed-off-by: Martin Storsjö <martin@martin.st>

libavcodec/arm/Makefile

@@ -54,6 +54,7 @@ ARMV6-OBJS-$(CONFIG_VP8_DECODER)       += arm/vp8_armv6.o               \
 
 VFP-OBJS-$(CONFIG_DCA_DECODER)         += arm/dcadsp_vfp.o              \
                                           arm/synth_filter_vfp.o
+VFP-OBJS-$(CONFIG_FFT)                 += arm/fft_vfp.o
 VFP-OBJS-$(CONFIG_MDCT)                += arm/mdct_vfp.o
 VFP-OBJS-$(HAVE_ARMV6)                 += arm/fmtconvert_vfp.o
 

libavcodec/arm/mdct_vfp.S

@@ -174,9 +174,8 @@ function ff_imdct_half_vfp, export=1
  .endr
 
         fmxr    FPSCR, OLDFPSCR
-        mov     ORIGOUT, OUT
-        ldr     ip, [CONTEXT, #9*4]
-        blx     ip                          @ s->fft_calc(s, output)
+        mov     a1, OUT
+        bl      ff_fft16_vfp
         ldr     lr, =0x03030000             @ RunFast mode, short vectors of length 4, stride 1
         fmxr    FPSCR, lr