Re: PPC CPU speed detection [was: Re: amigappc test kernel for CSPPC and BPPC]

[Alan Hourihane <alanh%fairlite.co.uk@localhost>]

On Mon, 2010-12-13 at 20:39 +0100, Radosław Kujawa wrote:
> On Dec 13, 2010, at 7:46 PM, Alan Hourihane wrote:
> 
> > On Mon, 2010-12-13 at 19:12 +0100, Radosław Kujawa wrote:
> >> On Dec 13, 2010, at 1:58 PM, Frank Wille wrote:
> >> 
> >>> Those new values all match with the 604e PLLs, but some, like 0010 and 
> >>> 0101,
> >>> are still wrong. So you are probably right that we could use the 604e PLL
> >>> table for the 603e as well.
> >>> 
> >> 
> >> Attached patch simplifies amigappc_identify() by merging PLL tables into 
> >> one. I think that we can safely use values from 604e PLLs, as 603e's 0010 
> >> and 0101 corresponds to high multipliers of 604e. Probability that 603e 
> >> with these PLL_CFGs exist is quite low (and I've never seen 603e faster 
> >> than 300MHz). Anyway, It works for me, 300MHz CPU is detected OK.
> > 
> > There are 603e's faster than 300MHz. People have been upgrading the
> > BlizzardPPC to 330MHz, and CSPPC to 363MHz.
> 
> All these BlizzardPPCs I've seen were 300MHz clocked at 330MHz. In fact, my 
> patch fixed detection for 330MHz boards too. I highly doubt that faster 603e 
> chips were ever manufactured (and that's not the problem here). CSPPC is a 
> completely different story, as there's 604e. PLL table for 604e was always 
> fine. It should even work for a guy who upgraded his CSPPC with 466MHz 604e.

Cool. Thanks.

Alan.