[media] lgdt3306a: don't go past the buffer

As warned by smatch:
	drivers/media/dvb-frontends/lgdt3306a.c:1354 log10_x1000() error: buffer overflow 'valx_x10' 14 <= 14
	drivers/media/dvb-frontends/lgdt3306a.c:1355 log10_x1000() error: buffer overflow 'log10x_x1000' 14 <= 14

There's a potential of returning a value out of the buffer. Fix it.

While here, remove the ugly braced block.
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>

drivers/media/dvb-frontends/lgdt3306a.c

@@ -1319,6 +1319,7 @@ static u32 log10_x1000(u32 x)
 	static u32 valx_x10[]     = {  10,  11,  13,  15,  17,  20,  25,  33,  41,  50,  59,  73,  87,  100 };
 	static u32 log10x_x1000[] = {   0,  41, 114, 176, 230, 301, 398, 518, 613, 699, 771, 863, 939, 1000 };
 	static u32 nelems = sizeof(valx_x10)/sizeof(valx_x10[0]);
+	u32 diff_val, step_val, step_log10;
 	u32 log_val = 0;
 	u32 i;
 
@@ -1348,15 +1349,16 @@ static u32 log10_x1000(u32 x)
 		if (valx_x10[i] >= x)
 			break;
 	}
+	if (i == nelems)
+		return log_val + log10x_x1000[i - 1];
+
+	diff_val   = x - valx_x10[i-1];
+	step_val   = valx_x10[i] - valx_x10[i - 1];
+	step_log10 = log10x_x1000[i] - log10x_x1000[i - 1];
 
-	{
-		u32 diff_val   = x - valx_x10[i-1];
-		u32 step_val   = valx_x10[i] - valx_x10[i-1];
-		u32 step_log10 = log10x_x1000[i] - log10x_x1000[i-1];
 	/* do a linear interpolation to get in-between values */
-		return log_val + log10x_x1000[i-1] +
+	return log_val + log10x_x1000[i - 1] +
 		((diff_val*step_log10) / step_val);
-	}
 }
 
 static u32 lgdt3306a_calculate_snr_x100(struct lgdt3306a_state *state)