The transcoder M/N values are double buffered on the transcoder's
undelayed vblank. So when doing seamless M/N fastsets we need to
evade also that.
Note that currently the pipe's delayed vblank == transcoder's
undelayed vblank, so this is still a nop change. But in the
future when we may have to delay the pipe's vblank to create
a register programming window ("window2") for the DSB.
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230404175431.23064-2-ville.syrjala@linux.intel.com
Reviewed-by: Mitul Golani <mitulkumar.ajitkumar.golani@intel.com
VBLANK_EVASION_TIME_US);
max = vblank_start - 1;
+ /*
+ * M/N is double buffered on the transcoder's undelayed vblank,
+ * so with seamless M/N we must evade both vblanks.
+ */
+ if (new_crtc_state->seamless_m_n && intel_crtc_needs_fastset(new_crtc_state))
+ min -= adjusted_mode->crtc_vblank_start - adjusted_mode->crtc_vdisplay;
+
if (min <= 0 || max <= 0)
goto irq_disable;