[PATCH RFT V2 1/8] clk: divider: add explicit big endian support

Jonas Gorski jonas.gorski at gmail.com
Thu Apr 18 19:00:21 AEST 2019


On Thu, 18 Apr 2019 at 01:32, Stephen Boyd <sboyd at kernel.org> wrote:
>
> Quoting Jonas Gorski (2019-04-15 03:10:39)
> > @@ -370,7 +388,7 @@ static long clk_divider_round_rate(struct clk_hw *hw, unsigned long rate,
> >         if (divider->flags & CLK_DIVIDER_READ_ONLY) {
> >                 u32 val;
> >
> > -               val = clk_readl(divider->reg) >> divider->shift;
> > +               val = clk_div_readl(divider->reg) >> divider->shift;
>
> Good deal that kbuild running sparse found that this was supposed to be
> divider and not divider->reg. If you can fix that and remove the else in
> all the basic type readl wrappers then this series looks good to merge
> for me.

Yeah, I'm quite glad about that. I was first confused why I didn't
catch it in my test build, but then remembered only C++ complains
about implicit void* casts (IIRC). I have it already fixed locally and
did a sparse build myself, was just waiting for at least one non-bot
comment before resending the series.


Regards
Jonas


More information about the Linuxppc-dev mailing list