On Friday 30 January 2009, Kevin Hilman wrote:
> > NOTE:  on dm355, sysclkbp values (for both PLLs) are incorrect.
> >
> 
> I changed the sysclkbp calculations recently because of what I
> understood of the spec.  According to the doc I'm reading (dm355:
> sprs463e, revised july 2008, in particular Figure 3.3), SYSCLKBP comes
> out of the pll before the multiplier and dividers, ans is only
> affected by BPDIV.  It looks the same for the other devices too.

OK, that's correct then.  Agrees with sprufb3, which is what
I'm looking at, fig 6-1 ... the PLLC1 configuration picture
gives a slightly different (more detailed/accurate) story than
the device clocking picture (fig 5-1), which just shows that
as a PLL output.

Both AUXCLK and SYSCLKBP get taps *before* the PLL.  If you
want to make the time, the pll1_aux clocks could stand to be
parented from the PLL with the same kind of special casing.


> So, on my board I see pll1_sysclkbp = 8MHz, which is correct with
> ref_clk = 24MHz and PLL1.BPDIV = 3.  For PLL2, I see 3MHz which
> looks correct for PLL2.BPDIV = 8.

Whereas if it were PLL1/3 and PLL2/8 it'd be off by a lot,
and that's what i was reacting to.  Sorry for the false alarm.

- Dave

_______________________________________________
Davinci-linux-open-source mailing list
[email protected]
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source

Reply via email to