diff mbox series

[v7,7/7] drm/i915/dsc: Check if vblank is sufficient for dsc prefill

Message ID 20250117074124.3965392-8-mitulkumar.ajitkumar.golani@intel.com (mailing list archive)
State New
Headers show
Series Check Scaler and DSC Prefill Latency Against Vblank | expand

Commit Message

Mitul Golani Jan. 17, 2025, 7:41 a.m. UTC
High refresh rate panels which may have small line times
and vblank sizes, Check if vblank size is sufficient for
dsc prefill latency.

--v2:
- Consider chroma downscaling factor in latency calculation. [Ankit]
- Replace with appropriate function name.

--v3:
- Remove FIXME tag.[Ankit]
- Replace Ycbcr444 to Ycbcr420.[Anit]
- Correct precision. [Ankit]
- Use some local valiables like linetime_factor and latency to
adjust precision.
- Declare latency to 0 initially to avoid returning any garbage values.
- Account for second scaler downscaling factor as well. [Ankit]

--v4:
- Improvise hscale and vscale calculation. [Ankit]
- Use appropriate name for number of scaler users. [Ankit]
- Update commit message and rebase.
- Add linetime and cdclk prefill adjustment calculation. [Ankit]

--v5:
- Update bspec link in trailer. [Ankit]
- Correct hscale, vscale datatype. [Ankit]
- Use intel_crtc_compute_min_cdclk. [Ankit]

--v6:
- Use cdclk_state->logical.cdclk instead of
intel_crtc_compute_min_cdclk. [Ankit]

Bspec: 70151
Signed-off-by: Mitul Golani <mitulkumar.ajitkumar.golani@intel.com>
---
 drivers/gpu/drm/i915/display/skl_watermark.c | 34 +++++++++++++++++++-
 1 file changed, 33 insertions(+), 1 deletion(-)

Comments

Nautiyal, Ankit K Jan. 17, 2025, 12:46 p.m. UTC | #1
On 1/17/2025 1:11 PM, Mitul Golani wrote:
> High refresh rate panels which may have small line times
> and vblank sizes, Check if vblank size is sufficient for
> dsc prefill latency.
>
> --v2:
> - Consider chroma downscaling factor in latency calculation. [Ankit]
> - Replace with appropriate function name.
>
> --v3:
> - Remove FIXME tag.[Ankit]
> - Replace Ycbcr444 to Ycbcr420.[Anit]
Typo: Ankit
> - Correct precision. [Ankit]
> - Use some local valiables like linetime_factor and latency to
> adjust precision.
> - Declare latency to 0 initially to avoid returning any garbage values.
> - Account for second scaler downscaling factor as well. [Ankit]
>
> --v4:
> - Improvise hscale and vscale calculation. [Ankit]
> - Use appropriate name for number of scaler users. [Ankit]
> - Update commit message and rebase.
> - Add linetime and cdclk prefill adjustment calculation. [Ankit]
>
> --v5:
> - Update bspec link in trailer. [Ankit]
> - Correct hscale, vscale datatype. [Ankit]
> - Use intel_crtc_compute_min_cdclk. [Ankit]
>
> --v6:
> - Use cdclk_state->logical.cdclk instead of
> intel_crtc_compute_min_cdclk. [Ankit]
>
> Bspec: 70151
> Signed-off-by: Mitul Golani <mitulkumar.ajitkumar.golani@intel.com>
> ---
>   drivers/gpu/drm/i915/display/skl_watermark.c | 34 +++++++++++++++++++-
>   1 file changed, 33 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/i915/display/skl_watermark.c b/drivers/gpu/drm/i915/display/skl_watermark.c
> index c8e540dd66cc..aacda7f7174c 100644
> --- a/drivers/gpu/drm/i915/display/skl_watermark.c
> +++ b/drivers/gpu/drm/i915/display/skl_watermark.c
> @@ -2307,6 +2307,38 @@ cdclk_prefill_adjustment(const struct intel_crtc_state *crtc_state)
>   				   2 * cdclk_state->logical.cdclk));
>   }
>   
> +static int
> +dsc_prefill_latency(const struct intel_crtc_state *crtc_state)
> +{
> +	const struct intel_crtc_scaler_state *scaler_state =
> +						&crtc_state->scaler_state;
> +	int latency = 0;
> +	int num_scaler_users = hweight32(scaler_state->scaler_users);
> +	int chroma_downscaling_factor =
> +		crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR420 ? 2 : 1;
> +	u64 hscale_k[2] = {0, 0};
> +	u64 vscale_k[2] = {0, 0};

We can do away with these just compute these in the loop below and use.


> +
> +	if (!crtc_state->dsc.compression_enable || !num_scaler_users)
> +		return latency;
> +
> +	for (int i = 0; i < num_scaler_users; i++) {
> +		hscale_k[i] =
> +			max(1000, mul_u32_u32(scaler_state->scalers[i].hscale, 1000) >> 16);
> +		vscale_k[i] =
> +			max(1000, mul_u32_u32(scaler_state->scalers[i].vscale, 1000) >> 16);
> +	}
> +
> +	latency  = DIV_ROUND_UP_ULL(hscale_k[0] * vscale_k[0], 1000000);
> +
> +	if (num_scaler_users > 1)
> +		latency *= DIV_ROUND_UP_ULL(hscale_k[1] * vscale_k[1], 1000000);
> +
> +	latency *= DIV_ROUND_UP(15 * crtc_state->linetime, 10) * chroma_downscaling_factor;
Need to compute linetime, like the previous patch.
> +
> +	return latency * cdclk_prefill_adjustment(crtc_state);

This function can return error.

Regards,

Ankit

> +}
> +
>   static int
>   scaler_prefill_latency(const struct intel_crtc_state *crtc_state)
>   {
> @@ -2346,10 +2378,10 @@ skl_is_vblank_too_short(const struct intel_crtc_state *crtc_state,
>   	const struct drm_display_mode *adjusted_mode =
>   		&crtc_state->hw.adjusted_mode;
>   
> -	/* FIXME missing DSC pre-fill time */
>   	return crtc_state->framestart_delay +
>   		intel_usecs_to_scanlines(adjusted_mode, latency) +
>   		scaler_prefill_latency(crtc_state) +
> +		dsc_prefill_latency(crtc_state) +
>   		wm0_lines >
>   		adjusted_mode->crtc_vtotal - adjusted_mode->crtc_vblank_start;
>   }
diff mbox series

Patch

diff --git a/drivers/gpu/drm/i915/display/skl_watermark.c b/drivers/gpu/drm/i915/display/skl_watermark.c
index c8e540dd66cc..aacda7f7174c 100644
--- a/drivers/gpu/drm/i915/display/skl_watermark.c
+++ b/drivers/gpu/drm/i915/display/skl_watermark.c
@@ -2307,6 +2307,38 @@  cdclk_prefill_adjustment(const struct intel_crtc_state *crtc_state)
 				   2 * cdclk_state->logical.cdclk));
 }
 
+static int
+dsc_prefill_latency(const struct intel_crtc_state *crtc_state)
+{
+	const struct intel_crtc_scaler_state *scaler_state =
+						&crtc_state->scaler_state;
+	int latency = 0;
+	int num_scaler_users = hweight32(scaler_state->scaler_users);
+	int chroma_downscaling_factor =
+		crtc_state->output_format == INTEL_OUTPUT_FORMAT_YCBCR420 ? 2 : 1;
+	u64 hscale_k[2] = {0, 0};
+	u64 vscale_k[2] = {0, 0};
+
+	if (!crtc_state->dsc.compression_enable || !num_scaler_users)
+		return latency;
+
+	for (int i = 0; i < num_scaler_users; i++) {
+		hscale_k[i] =
+			max(1000, mul_u32_u32(scaler_state->scalers[i].hscale, 1000) >> 16);
+		vscale_k[i] =
+			max(1000, mul_u32_u32(scaler_state->scalers[i].vscale, 1000) >> 16);
+	}
+
+	latency  = DIV_ROUND_UP_ULL(hscale_k[0] * vscale_k[0], 1000000);
+
+	if (num_scaler_users > 1)
+		latency *= DIV_ROUND_UP_ULL(hscale_k[1] * vscale_k[1], 1000000);
+
+	latency *= DIV_ROUND_UP(15 * crtc_state->linetime, 10) * chroma_downscaling_factor;
+
+	return latency * cdclk_prefill_adjustment(crtc_state);
+}
+
 static int
 scaler_prefill_latency(const struct intel_crtc_state *crtc_state)
 {
@@ -2346,10 +2378,10 @@  skl_is_vblank_too_short(const struct intel_crtc_state *crtc_state,
 	const struct drm_display_mode *adjusted_mode =
 		&crtc_state->hw.adjusted_mode;
 
-	/* FIXME missing DSC pre-fill time */
 	return crtc_state->framestart_delay +
 		intel_usecs_to_scanlines(adjusted_mode, latency) +
 		scaler_prefill_latency(crtc_state) +
+		dsc_prefill_latency(crtc_state) +
 		wm0_lines >
 		adjusted_mode->crtc_vtotal - adjusted_mode->crtc_vblank_start;
 }