[PATCH -next] drm/nouveau/clk: Use plain "/" for 32-bit division

[Alexandre Courbot]

Hi Geert,

On 02/05/2015 01:02 AM, Geert Uytterhoeven wrote:
> On 32-bit platforms using asm-generic/div64.h:
>
> drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c: In function 'gk20a_pllg_calc_rate':
> drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c:147:79: warning: comparison of distinct pointer types lacks a cast
>    do_div(rate, divider);
>                                                                                 ^
> drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c:147:2: warning: right shift count >= width of type
>    do_div(rate, divider);
>    ^
> drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c:147:238: warning: passing argument 1 of '__div64_32' from incompatible pointer type
>    do_div(rate, divider);
>                                                                                                                                                                                                                                                ^
> In file included from arch/parisc/include/generated/asm/div64.h:1:0,
>                   from include/linux/kernel.h:124,
>                   from include/linux/list.h:8,
>                   from include/linux/preempt.h:10,
>                   from include/linux/spinlock.h:50,
>                   from include/linux/mmzone.h:7,
>                   from include/linux/gfp.h:5,
>                   from include/linux/slab.h:14,
>                   from drivers/gpu/drm/nouveau/include/nvif/os.h:5,
>                   from drivers/gpu/drm/nouveau/include/nvkm/core/os.h:3,
>                   from drivers/gpu/drm/nouveau/include/nvkm/core/object.h:3,
>                   from drivers/gpu/drm/nouveau/include/nvkm/core/subdev.h:3,
>                   from drivers/gpu/drm/nouveau/include/nvkm/subdev/clk.h:3,
>                   from drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c:25:
> include/asm-generic/div64.h:35:17: note: expected 'uint64_t *' but argument is of type 'u32 *'
>   extern uint32_t __div64_32(uint64_t *dividend, uint32_t divisor);
>                   ^
>
> do_div() is meant for 64-bit by 32-bit division, but both the dividend
> and divisor are 32-bit here. Hence use plain "/" instead.
>
> Signed-off-by: Geert Uytterhoeven <geert at linux-m68k.org>
> ---
> Compile-tested only.
>
> parisc/allmodconfig:
> http://kisskb.ellerman.id.au/kisskb/buildresult/12358386/
> ---
>   drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c | 3 +--
>   1 file changed, 1 insertion(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c
> index 65c532742b08d1c6..022595876ea4dc85 100644
> --- a/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c
> +++ b/drivers/gpu/drm/nouveau/nvkm/subdev/clk/gk20a.c
> @@ -144,9 +144,8 @@ gk20a_pllg_calc_rate(struct gk20a_clk_priv *priv)
>
>   	rate = priv->parent_rate * priv->n;
>   	divider = priv->m * pl_to_div[priv->pl];
> -	do_div(rate, divider);
>
> -	return rate / 2;
> +	return rate / divider / 2;

I agree there is a problem here, but considering the theoretical values 
that rate can take, could we rather fix this by making rate a u64? With 
the current maximum values of priv->parent_rate and priv->n, rate might 
approach dangerously to the u32 limit and I believe casting it to u64 
would be safer.