After checking the code and the datasheet, it seems like we are handling
the clock inversion (SND_SOC_DAIFMT_NB_IF and SND_SOC_DAIFMT_IB_IF) not
correctly.

>From the datasheet (Table 58):
R5 Format Control, BITS[5:4], [BCP:LRP]:
  (0) 00 =  normal   BCLK, normal   LRCLK
  (1) 01 =  normal   BCLK, inverted LRCLK <-- Fix this
  (2) 10 =  inverted BCLK, normal   LRCLK
  (3) 11 =  inverted BCLK, inverted LRCLK <-- Fix this
Signed-off-by: Sergej Sawazki <sergej@taudac.com>
Acked-by: Charles Keepax <ckeepax@opensource.cirrus.com>
Signed-off-by: Mark Brown <broonie@kernel.org>