I finally tracked down the cause of the weird mis-tunings I've seen
occasionally with my cx24123-based card -- you can find border cases
where one frequency would tune fine, +1MHz would be a totally different
transport, but +2MHz would be back to the correct transport, etc.
It comes back to the PLL divisor setup -- the cx24109 datasheet says this:
NOTE: if A=0, then N=N+1
and the implementation interpreted that as this:
if (adiv == 0)
ndiv++;
which seems reasonable at first glance, but it's exactly the opposite of
what is required -- the datasheet means that the actual value of N is 1
greater than the value written when A is 0, so 1 needs to be
*subtracted* from it to compensate.
Simple patch attached which eliminates the weird mis-tunings.
Signed-off-by: Yeasah Pell <yeasah at schwide dot net>
diff -r 323d46fbb593 linux/drivers/media/dvb/frontends/cx24123.c
--- a/linux/drivers/media/dvb/frontends/cx24123.c Wed Aug 23 18:21:35 2006 -0300
+++ b/linux/drivers/media/dvb/frontends/cx24123.c Sat Sep 16 23:06:15 2006 -0400
@@ -564,8 +562,8 @@ static int cx24123_pll_calculate(struct
ndiv = ( ((p->frequency * vco_div * 10) / (2 * XTAL / 1000)) / 32) & 0x1ff;
adiv = ( ((p->frequency * vco_div * 10) / (2 * XTAL / 1000)) % 32) & 0x1f;
- if (adiv == 0)
- ndiv++;
+ if (adiv == 0 && ndiv > 0)
+ ndiv--;
/* control bits 11, refdiv 11, charge pump polarity 1, charge pump current, ndiv, adiv */
state->pllarg = (3 << 19) | (3 << 17) | (1 << 16) | (pump << 14) | (ndiv << 5) | adiv;
_______________________________________________
linux-dvb mailing list
linux-dvb@xxxxxxxxxxx
http://www.linuxtv.org/cgi-bin/mailman/listinfo/linux-dvb
