[Intel-gfx] [PATCH] drm/i915/vdsc: Fix first_line_bpg_offset calculation

Nautiyal, Ankit K ankit.k.nautiyal at intel.com
Fri Aug 4 07:55:02 UTC 2023


On 8/3/2023 9:32 PM, Suraj Kandpal wrote:
> On checking DSC1.1 Errata and DSC 1.2 spec the current formula
> we were using was incorrect to calculate first_line_bpg_offset.
> The new fixed formula is derived from C model.
>
> Signed-off-by: Suraj Kandpal <suraj.kandpal at intel.com>
> ---
>   drivers/gpu/drm/i915/display/intel_vdsc.c | 15 ++++++++++++---
>   1 file changed, 12 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/display/intel_vdsc.c b/drivers/gpu/drm/i915/display/intel_vdsc.c
> index 9d76c2756784..a344c6ab8767 100644
> --- a/drivers/gpu/drm/i915/display/intel_vdsc.c
> +++ b/drivers/gpu/drm/i915/display/intel_vdsc.c
> @@ -65,6 +65,9 @@ intel_vdsc_set_min_max_qp(struct drm_dsc_config *vdsc_cfg, int buf,
>   		intel_lookup_range_max_qp(bpc, buf, bpp, vdsc_cfg->native_420);
>   }
>   
> +#define CLAMP_TO(val, _min, _max) \
> +	max(_min, min(val, _max))

lets use clamp(), already defined in linux/minmax.h

With that fixed, this is:

Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal at intel.com>

> +
>   /*
>    * We are using the method provided in DSC 1.2a C-Model in codec_main.c
>    * Above method use a common formula to derive values for any combination of DSC
> @@ -80,13 +83,19 @@ calculate_rc_params(struct drm_dsc_config *vdsc_cfg)
>   	int bpc = vdsc_cfg->bits_per_component;
>   	int bpp = vdsc_cfg->bits_per_pixel >> 4;
>   	int qp_bpc_modifier = (bpc - 8) * 2;
> +	int uncompressed_bpg_rate;
> +	int first_line_bpg_offset;
>   	u32 res, buf_i, bpp_i;
>   
>   	if (vdsc_cfg->slice_height >= 8)
> -		vdsc_cfg->first_line_bpg_offset =
> -			12 + DIV_ROUND_UP((9 * min(34, vdsc_cfg->slice_height - 8)), 100);
> +		first_line_bpg_offset =
> +			12 + (9 * min(34, vdsc_cfg->slice_height - 8)) / 100;
>   	else
> -		vdsc_cfg->first_line_bpg_offset = 2 * (vdsc_cfg->slice_height - 1);
> +		first_line_bpg_offset = 2 * (vdsc_cfg->slice_height - 1);
> +
> +	uncompressed_bpg_rate = (3 * bpc + (vdsc_cfg->convert_rgb ? 0 : 2)) * 3;
> +	vdsc_cfg->first_line_bpg_offset = CLAMP_TO(first_line_bpg_offset, 0,
> +						   uncompressed_bpg_rate - 3 * bpp);
>   
>   	/*
>   	 * According to DSC 1.2 spec in Section 4.1 if native_420 is set:


More information about the Intel-gfx mailing list