2020-09-11 16:17:49

by Laurentiu Palcu

Subject: [PATCH] drm/imx/dcss: fix 64-bit divisions

Use div_s64() for the 64-bit divisions. This would allow the driver to compile
on 32-bit architectures, if needed.

Signed-off-by: Laurentiu Palcu <[email protected]>
---
drivers/gpu/drm/imx/dcss/dcss-scaler.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/imx/dcss/dcss-scaler.c b/drivers/gpu/drm/imx/dcss/dcss-scaler.c
index cd21905de580..7c1e0e461244 100644
--- a/drivers/gpu/drm/imx/dcss/dcss-scaler.c
+++ b/drivers/gpu/drm/imx/dcss/dcss-scaler.c
@@ -134,7 +134,7 @@ static int div_q(int A, int B)
else
temp -= B / 2;

- result = (int)(temp / B);
+ result = (int)(div_s64(temp, B));
return result;
}

@@ -237,7 +237,7 @@ static void dcss_scaler_gaussian_filter(int fc_q, bool use_5_taps,
ll_temp = coef[phase][i];
ll_temp <<= PSC_COEFF_PRECISION;
ll_temp += sum >> 1;
- ll_temp /= sum;
+ ll_temp = div_s64(ll_temp, sum);
coef[phase][i] = (int)ll_temp;
}
}
--
2.17.1

2020-09-11 16:36:04

by Ville Syrjälä

Subject: Re: [PATCH] drm/imx/dcss: fix 64-bit divisions

On Fri, Sep 11, 2020 at 04:48:27PM +0300, Laurentiu Palcu wrote:
> Use div_s64() for the 64-bit divisions. This would allow the driver to compile
> on 32-bit architectures, if needed.
>
> Signed-off-by: Laurentiu Palcu <[email protected]>
> ---
> drivers/gpu/drm/imx/dcss/dcss-scaler.c | 4 ++--
> 1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/gpu/drm/imx/dcss/dcss-scaler.c b/drivers/gpu/drm/imx/dcss/dcss-scaler.c
> index cd21905de580..7c1e0e461244 100644
> --- a/drivers/gpu/drm/imx/dcss/dcss-scaler.c
> +++ b/drivers/gpu/drm/imx/dcss/dcss-scaler.c
> @@ -134,7 +134,7 @@ static int div_q(int A, int B)
> else
> temp -= B / 2;
>
> - result = (int)(temp / B);
> + result = (int)(div_s64(temp, B));
> return result;
> }
>
> @@ -237,7 +237,7 @@ static void dcss_scaler_gaussian_filter(int fc_q, bool use_5_taps,
> ll_temp = coef[phase][i];
> ll_temp <<= PSC_COEFF_PRECISION;
> ll_temp += sum >> 1;
> - ll_temp /= sum;

That looks like hand rolled DIV_ROUND_CLOSEST_ULL()

> + ll_temp = div_s64(ll_temp, sum);
> coef[phase][i] = (int)ll_temp;
> }
> }
> --
> 2.17.1
>
> _______________________________________________
> dri-devel mailing list
> [email protected]
> https://lists.freedesktop.org/mailman/listinfo/dri-devel

--
Ville Syrj?l?
Intel