replace DIV_ROUND with DIV_ROUND_CLOSEST

The Linux-compatible macro DIV_ROUND_CLOSEST is a bit more flexible
and safer than DIV_ROUND.

For example,
  foo = DIV_ROUND_CLOSEST(x, y++)
works expectedly, but
  foo = DIV_ROUND(x, y++)
does not. (y is incremented twice.)

Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com>
diff --git a/arch/arm/cpu/armv7/omap-common/abb.c b/arch/arm/cpu/armv7/omap-common/abb.c
index 423aeb9..a0add66 100644
--- a/arch/arm/cpu/armv7/omap-common/abb.c
+++ b/arch/arm/cpu/armv7/omap-common/abb.c
@@ -48,9 +48,9 @@
 	 */
 
 	/* calculate SR2_WTCNT_VALUE */
-	sys_rate = DIV_ROUND(V_OSCK, 1000000);
-	clk_cycles = DIV_ROUND(OMAP_ABB_CLOCK_CYCLES * 10, sys_rate);
-	sr2_cnt = DIV_ROUND(OMAP_ABB_SETTLING_TIME * 10, clk_cycles);
+	sys_rate = DIV_ROUND_CLOSEST(V_OSCK, 1000000);
+	clk_cycles = DIV_ROUND_CLOSEST(OMAP_ABB_CLOCK_CYCLES * 10, sys_rate);
+	sr2_cnt = DIV_ROUND_CLOSEST(OMAP_ABB_SETTLING_TIME * 10, clk_cycles);
 
 	setbits_le32(setup,
 		     sr2_cnt << (ffs(OMAP_ABB_SETUP_SR2_WTCNT_VALUE_MASK) - 1));