Re: [RFC 8/9] clk: qcom: gcc-sdm845: Switch to macros to collapse rcg2 clocks definitions

[Dmitry Baryshkov]

On Tue, 26 Jul 2022 at 17:24, Abel Vesa <abel.vesa@xxxxxxxxxx> wrote:
>
> Switch from the expanded rcg2 clocks definitions to the more compact
> macros.
>
> Signed-off-by: Abel Vesa <abel.vesa@xxxxxxxxxx>
> ---
>  drivers/clk/qcom/gcc-sdm845.c | 712 +++-------------------------------
>  1 file changed, 51 insertions(+), 661 deletions(-)
>
> diff --git a/drivers/clk/qcom/gcc-sdm845.c b/drivers/clk/qcom/gcc-sdm845.c
> index 2e66256599d3..d9751d7e617c 100644
> --- a/drivers/clk/qcom/gcc-sdm845.c
> +++ b/drivers/clk/qcom/gcc-sdm845.c

[skipped]

> +DEFINE_QCOM_CC_CLK(RCG2, gcc_cpuss_ahb_clk_src, 0x48014, 0, 5, gcc_parent_map_0, ftbl_gcc_cpuss_ahb_clk_src, gcc_parent_data_7_ao);
> +DEFINE_QCOM_CC_CLK(RCG2, gcc_cpuss_rbcpr_clk_src, 0x4815c, 0, 5, gcc_parent_map_3, ftbl_gcc_cpuss_rbcpr_clk_src, gcc_parent_data_8_ao);

Well.. Stephen & Bjorn might have other opinions here. But for me
there are just too many pieces of information on a single line. And
different values start to meld into each other.
So, e.g. I find it harder to comprehend what is going on here. What do
0 and 5 mean, etc. So while it takes less space, it looks harder to
read. Compare this with the interconnect drivers, where a few releases
ago we made an opposite switch: from using QNODE macros to the
expanded structures.

[skipped]

>
>  DEFINE_QCOM_CC_CLK(BRANCH, HALT, gcc_aggre_noc_pcie_tbu_clk, 0, 0x90014, 0, 0, 0x90014, BIT(0), 0);

And this is the perfect example. a series of 0, 1 and BIT(0). What do
they mean? And where is the parent, which is defined for the next
clocks?

>  DEFINE_QCOM_CC_CLK(BRANCH, HALT, gcc_aggre_ufs_card_axi_clk, 1, 0x82028, 0x82028, 1, 0x82028, BIT(0), CLK_SET_RATE_PARENT, &gcc_ufs_card_axi_clk_src.clkr.hw);


-- 
With best wishes
Dmitry