Re: [PATCH RFC 13/25] drm: Add VRR target frame rate properties
From: Leo Li <sunpeng.li@amd.com>
Date: 2026-09-25 18:42:22
Also in:
dri-devel, linux-fbdev, linux-rockchip, lkml
On 2026-09-22 11:26, Nicolas Frattaroli wrote:
quoted
quoted
+ * VRR Limiter/Target Properties + * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ + * + * The ``VRR_{MIN,MAX}_{NUMERATOR,DENOMINATOR}`` properties expose a mechanism + * through which userspace can control the desired range of refresh rates in + * which VRR is allowed to operate. Each rate is expressed as a + * numerator/denominator fraction of refresh rates in Hz, allowing for rational + * target rates like 24/1.001 Hz with no loss of precision or ambiguity. + * + * If the minimum and maximum rate are set to the same value (and not 0), they + * are understood as a fixed target rate. This is especially useful for media + * playback, where the content's frame rate is both constant and known in + * advance. In such cases, a refresh rate that is not an integer multiple of the + * content's frame rate will introduce judder, since not every frame is + * displayed for the same amount of time. A modeset of the display with a + * compatible rate may in those cases be either undesirable or impossible, but + * the rate can still effectively be reached through VRR. + * + * .. _VRR-MIN-NUMERATOR: + * + * "VRR_MIN_NUMERATOR": + * Default &drm_crtc integer property forming the numerator of a + * numerator/denominator pair of a frame rate to set as the minimum VRR + * target rate. Set to 0 to disable. + * + * "VRR_MIN_DENOMINATOR": + * Default &drm_crtc integer property forming the denominator of a + * numerator/denominator pair of a frame rate to set as the minimum VRR + * target rate. If :ref:`VRR_MIN_NUMERATOR <VRR-MIN-NUMERATOR>` is not + * zero, it must be non-zero. + * Otherwise, must also be zero. + * + * .. _VRR-MAX-NUMERATOR: + * + * "VRR_MAX_NUMERATOR": + * Default &drm_crtc integer property forming the numerator of a + * numerator/denominator pair of a frame rate to set as the maximum VRR + * target rate. Set to 0 to disable. + * + * "VRR_MAX_DENOMINATOR": + * Default &drm_crtc integer property forming the denominator of a + * numerator/denominator pair of a frame rate to set as the maximum VRR + * target rate. If :ref:`VRR_MAX_NUMERATOR <VRR-MAX-NUMERATOR>` is not + * zero, it must be non-zero. Otherwise, must also be zero. */If VRR_MIN_NUMERATOR == 0 && VRR_MAX_NUMERATOR > 0, do we interpret that as vrr limiting is disabled?You can picture VRR limiting as always being active, but with a limit rational of 0 it uses the display's limit as per the EDID, which is what unlimited game mode is. So with how it's implemented right now in hdmi_validate_vrr(), your example would set a maximum target, but leave the minimum at whatever the display defaults to. Now that I'm thinking through this, a possible problem is that drm_crtc_helper_vrr_is_fixed_rate() operates on the user supplied limits, but if the display supplied lower limit is equal to the user supplied upper limit, then we have a fixed rate scenario without recognising it as such. I think I need to have a ponder on what the least surprising behaviour for userspace is in that instance. The display limit stuff gets a bit complex due to CinemaVRR and QMS TFRmin/TFRmax. I'll improve the documentation on the next revision to make the meanings more explicit.
Perhaps a simple way is to require simultaneous setting MIN and MAX pairs?
IOW, require userspace to set MIN and MAX simultaneously to >0, or =0. For example:
if ((vrr_min_n == 0 || vrr_min_d == 0 ||
vrr_max_n == 0 || vrr_max_d == 0) &&
(vrr_min_n > 0 || vrr_max_n > 0))
return -EINVAL;
That way, it's never ambiguous what userspace has requested for the range. They can copy the EDID supported range if they don't care about limiting one side, rather than leaving it at 0. It's then also clear if they requested a static Hz.
Thanks,
Leo