When a display update triggers a DDB re-allocation, we should start by
assuming that only the updated pipes need to be re-allocated (we have
logic later that may add additional pipes if, e.g., a modeset triggers a
change to the global allocation).
We were erroneously using the _active_ pipes as our starting point
rather than the changed pipes.  This causes us to grab CRTC locks that
we didn't actually need, reducing parallelism.  Given the recent
non-blocking atomic changes, it also causes legacy pageflips against one
CRTC to return -EBUSY if there's an outstanding pageflip against a
different CRTC (a situation easily triggered via compositors like
Weston).
Fixes: 98d39494d3 ("drm/i915/gen9: Compute DDB allocation at atomic check time (v4)")
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/1467070964-14864-1-git-send-email-matthew.d.roper@intel.com
Reviewed-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
 
        return 0;
 }
 
+static uint32_t
+pipes_modified(struct drm_atomic_state *state)
+{
+       struct drm_crtc *crtc;
+       struct drm_crtc_state *cstate;
+       uint32_t i, ret = 0;
+
+       for_each_crtc_in_state(state, crtc, cstate, i)
+               ret |= drm_crtc_mask(crtc);
+
+       return ret;
+}
+
 static int
 skl_compute_ddb(struct drm_atomic_state *state)
 {
        struct intel_atomic_state *intel_state = to_intel_atomic_state(state);
        struct intel_crtc *intel_crtc;
        struct skl_ddb_allocation *ddb = &intel_state->wm_results.ddb;
-       unsigned realloc_pipes = dev_priv->active_crtcs;
+       uint32_t realloc_pipes = pipes_modified(state);
        int ret;
 
        /*