renderer: fix dgpu directscanout explicit sync (#13229)

* directscanout: fix dgpu directscanout explicit sync

without setting an infence, AQ doesnt explicit sync, nor recreate the
dgpu fence for the blit work. and as such attemptdirectscanout path
artifacts and breaks. create a dummy CEGLSync even tho we dont really
have any pending glwork to get a proper fence, and set it.

* monitor: dont use new scheduling if direct scanout

using the new_render_scheduling makes no sense in the direct scanout
path, add a if guard against it.

1 files changed, 12 insertions, 1 deletions

diff --git a/src/helpers/Monitor.cpp b/src/helpers/Monitor.cpp
index be0b78fd..bb710269 100644
--- a/src/helpers/Monitor.cpp
+++ b/src/helpers/Monitor.cpp
@@ -1893,7 +1893,18 @@ bool CMonitor::attemptDirectScanout() {
     PSURFACE->presentFeedback(Time::steadyNow(), m_self.lock());
 
     m_output->state->addDamage(PSURFACE->m_current.accumulateBufferDamage());
-    m_output->state->resetExplicitFences();
+
+    // multigpu needs a fence to trigger fence syncing blits and also committing with the recreated dgpu fence
+    if (m_output->getBackend()->preferredAllocator()->drmFD() != g_pCompositor->m_drm.fd && g_pHyprOpenGL->explicitSyncSupported()) {
+        auto sync = CEGLSync::create();
+
+        if (sync->fd().isValid()) {
+            m_inFence = sync->takeFd();
+            m_output->state->setExplicitInFence(m_inFence.get());
+        } else
+            m_output->state->resetExplicitFences(); // good luck.
+    } else
+        m_output->state->resetExplicitFences();
 
     // no need to do explicit sync here as surface current can only ever be ready to read