Hi Biju,
On Fri, Oct 11, 2024 at 6:20 PM Biju Das <biju.das.jz@xxxxxxxxxxxxxx> wrote:
> While computing foutpostdiv_rate, the value of params->pl5_fracin
> is discarded, which results in the wrong refresh rate. Fix the formula
> for computing foutpostdiv_rate.
>
> Fixes: 1561380ee72f ("clk: renesas: rzg2l: Add FOUTPOSTDIV clk support")
> Signed-off-by: Hien Huynh <hien.huynh.px@xxxxxxxxxxx>
> Signed-off-by: Biju Das <biju.das.jz@xxxxxxxxxxxxxx>
> ---
> v1->v2:
> * Improved the precision by division of params->pl5_refdiv
> done after all multiplication.
> ---
> drivers/clk/renesas/rzg2l-cpg.c | 12 +++++++-----
> 1 file changed, 7 insertions(+), 5 deletions(-)
>
> diff --git a/drivers/clk/renesas/rzg2l-cpg.c b/drivers/clk/renesas/rzg2l-cpg.c
> index 88bf39e8c79c..a1e22d353689 100644
> --- a/drivers/clk/renesas/rzg2l-cpg.c
> +++ b/drivers/clk/renesas/rzg2l-cpg.c
> @@ -548,7 +548,7 @@ static unsigned long
> rzg2l_cpg_get_foutpostdiv_rate(struct rzg2l_pll5_param *params,
> unsigned long rate)
> {
> - unsigned long foutpostdiv_rate;
> + unsigned long foutpostdiv_rate, foutvco_rate;
While the resulting 64-bit value fits in foutvco_rate because unsigned
long is 64-bit on the target platform, I'd rather play it safe
(reuse!) and use u64 explicitly.
>
> params->pl5_intin = rate / MEGA;
> params->pl5_fracin = div_u64(((u64)rate % MEGA) << 24, MEGA);
> @@ -557,10 +557,12 @@ rzg2l_cpg_get_foutpostdiv_rate(struct rzg2l_pll5_param *params,
> params->pl5_postdiv2 = 1;
> params->pl5_spread = 0x16;
>
> - foutpostdiv_rate =
> - EXTAL_FREQ_IN_MEGA_HZ * MEGA / params->pl5_refdiv *
> - ((((params->pl5_intin << 24) + params->pl5_fracin)) >> 24) /
> - (params->pl5_postdiv1 * params->pl5_postdiv2);
> + foutvco_rate =
> + (EXTAL_FREQ_IN_MEGA_HZ * MEGA *
> + ((params->pl5_intin << 24) + params->pl5_fracin) /
> + params->pl5_refdiv) >> 24;
Shouldn't this use mul_u32_u32(EXTAL_FREQ_IN_MEGA_HZ * MEGA,
((params->pl5_intin << 24) + params->pl5_fracin)) instead of a plain
multiplication?
See also the comment for mul_u32_u32() in <linux/math64.h>.
> + foutpostdiv_rate = DIV_ROUND_CLOSEST_ULL(foutvco_rate,
> + params->pl5_postdiv1 * params->pl5_postdiv2);
Unfortunately we don't have a helper macro yet to round the result of
div_u64(), so you will have to open-code that (for now).
>
> return foutpostdiv_rate;
> }
Gr{oetje,eeting}s,
Geert
--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx
In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
-- Linus Torvalds
