Hi!

From: Kuninori Morimoto <kuninori.morimoto.gx@...>

commit b7169ddea2f2a90538f606688adf4948f2da82ce upstream.

Current rsnd is using RSND_REG_xxx for register naming,
and using RSND_REG_##f style macro for read/write.
The biggest reason why it uses this style is that
we can avoid non-existing register access.
But, its demerit is sequential register access code will
be very ugly.
Current rsnd driver is well tested, so, let's remove RSND_REG_
from rsnd_reg, and cleanup sequential register access code.

+#define SRCIN_TIMSEL(i) (SRCIN_TIMSEL0 + (i))
+#define SRCOUT_TIMSEL(i) (SRCOUT_TIMSEL0 + (i))
+#define CTU_SVxxR(i, j) (CTU_SV00R + (i * 8) + (j))
+#define DVC_VOLxR(i) (DVC_VOL0R + (i))
+#define AUDIO_CLK_SEL(i) (AUDIO_CLK_SEL0 + (i))
+#define SSI_BUSIF_MODE(i) (SSI_BUSIF0_MODE + (i))
+#define SSI_BUSIF_ADINR(i) (SSI_BUSIF0_ADINR + (i))
+#define SSI_BUSIF_DALIGN(i) (SSI_BUSIF0_DALIGN + (i))
+#define SSI_SYS_STATUS(i) (SSI_SYS_STATUS0 + (i))

Would it still make sense to test that i is in expected range?

#define CHECK_RANGE(i) ({ WARN_ON(i<0 || i>4); i; })
#define SRCIN_TIMSEL(i) (SRCIN_TIMSEL0 + (i))
...

Thanks and best regards,
Pavel
--
DENX Software Engineering GmbH, Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany