Re: [RFC PATCH] clk: fractional-divider: Correct max_{m,n} handed over to rational_best_approximation()

[Andy Shevchenko]

On Wed, Jul 14, 2021 at 02:41:29PM +0800, Liu Ying wrote:
> If a fractional divider clock has the flag
> CLK_FRAC_DIVIDER_ZERO_BASED set, the maximum
> numerator and denominator handed over to
> rational_best_approximation(), in this case
> max_m and max_n, should be increased by one
> comparing to those have the flag unset.  Without
> this patch, a zero based fractional divider
> with 1-bit mwidth and 3-bit nwidth would wrongly
> generate 96MHz clock rate if the parent clock
> rate is 288MHz, while the expected clock rate
> is 115.2MHz with m = 2 and n = 5.

Make sure that your editor is configured to allow you to have lines ~70-72
characters long.

...

> The patch is RFC, because the rationale behind the below snippet in
> clk_fd_general_approximation() is unclear to Jacky and me and we are
> not sure if there is any room to improve this patch due to the snippet.
> Maybe, Andy may help shed some light here.  Thanks.
> 
> -----------------------------------8<---------------------------------
> /*
>  * Get rate closer to *parent_rate to guarantee there is no overflow
>  * for m and n. In the result it will be the nearest rate left shifted
>  * by (scale - fd->nwidth) bits.
>  */

I don't know how to rephrase above comment better.

> scale = fls_long(*parent_rate / rate - 1);
> if (scale > fd->nwidth)
> 	rate <<= scale - fd->nwidth;

This takes an advantage of the numbers be in a form of

	n = k * 2^m, (1)

where m will be scale in the snippet above. Thus, if n can be represented by
(1), we opportunistically reduce amount of bits needed for it by shifting right
by m bits.

Does it make sense?

The code looks good to me, btw, although I dunno if you need to call the newly
introduced function before or after the above mentioned snippet.

-- 
With Best Regards,
Andy Shevchenko