Re: [PATCH 17/21] drm/i915/display: Move vrr.guardband/pipeline_full out of !fastset block

["Ville Syrjälä" <ville.syrjala@linux.intel.com>]

On Thu, Mar 06, 2025 at 06:40:56PM +0530, Ankit Nautiyal wrote:
> The vrr.guardband/pipeline_full depend on the vrr.vmin. Since we have
> set vrr.vmin to adjusted_mode->crtc_vtotal, this shouldn't change on the
> fly. With this we can move vrr.guardband/pipeline_full out from !fastset
> block.
> 
> Signed-off-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
> ---
>  drivers/gpu/drm/i915/display/intel_display.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c
> index 322a05648f58..a642496e366c 100644
> --- a/drivers/gpu/drm/i915/display/intel_display.c
> +++ b/drivers/gpu/drm/i915/display/intel_display.c
> @@ -5393,8 +5393,6 @@ intel_pipe_config_compare(const struct intel_crtc_state *current_config,
>  		PIPE_CONF_CHECK_I(vrr.vmin);
>  		PIPE_CONF_CHECK_I(vrr.vmax);
>  		PIPE_CONF_CHECK_I(vrr.flipline);
> -		PIPE_CONF_CHECK_I(vrr.pipeline_full);
> -		PIPE_CONF_CHECK_I(vrr.guardband);
>  		PIPE_CONF_CHECK_I(vrr.vsync_start);
>  		PIPE_CONF_CHECK_I(vrr.vsync_end);
>  		PIPE_CONF_CHECK_LLI(cmrr.cmrr_m);
> @@ -5402,6 +5400,9 @@ intel_pipe_config_compare(const struct intel_crtc_state *current_config,
>  		PIPE_CONF_CHECK_BOOL(cmrr.enable);
>  	}
>  
> +	PIPE_CONF_CHECK_I(vrr.pipeline_full);
> +	PIPE_CONF_CHECK_I(vrr.guardband);


Assuming we can't reprogram the guardband safely live 
I think this would have to become 

if (!fastset || always_use_vrr_tg()) {
        ...
}

which avoids breaking the LRR fastset on older platforms.

Whether we can still risk the fastboot exception for the new
platforms I don't know. I guess leave it out for now and
ponder it further later.

In the future I guess one option to resurrecting the fastsets
would be to somehow use a more fixed size gurdband instead of
the full vblank length, but that would need a lot of thought
and work, so definitely not something we can just do right
now.

> +
>  #undef PIPE_CONF_CHECK_X
>  #undef PIPE_CONF_CHECK_I
>  #undef PIPE_CONF_CHECK_LLI
> -- 
> 2.45.2

-- 
Ville Syrjälä
Intel