On 08/22, Geert Uytterhoeven wrote:
> Hi Stephen,
>
> On Fri, Aug 19, 2016 at 9:17 PM, Stephen Boyd <sboyd@xxxxxxxxxxxxxx> wrote:
> > On 08/18, Geert Uytterhoeven wrote:
> >> diff --git a/arch/mips/txx9/generic/setup.c b/arch/mips/txx9/generic/setup.c
> >> index ada92db92f87d91a..2fdbcf91b2cc472c 100644
> >> --- a/arch/mips/txx9/generic/setup.c
> >> +++ b/arch/mips/txx9/generic/setup.c
> >> @@ -560,8 +527,39 @@ void __init plat_time_init(void)
> >> txx9_board_vec->time_init();
> >> }
> >>
> >> +static void txx9_clk_init(void)
> >> +{
> >> + struct clk *clk;
> >> + int error;
> >> +
> >> + clk = clk_register_fixed_rate(NULL, "gbus", NULL, 0, txx9_gbus_clock);
> >
> > Can we use the clk_hw_*() based variants instead please?
>
> Yes we can.
>
> BTW, is it intentional that clk_hw_register_clkdev() doesn't detect errors
> from a previous registration call, like clk_register_clkdev() does?
>
Slightly intentional, because the assumption is providers already
have a clk_hw structure that they've created themselves when they
register a clkdev, whereas a struct clk is more likely to be an
error pointer because of clk_get(), etc. But I suppose we can add
the error check to ease registration in cases where providers are
using the basic types.
--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project
