Mesa (master): intel: Use the new DRI2 flush invalidate entrypoint to signal frame done.
Eric Anholt
anholt at kemper.freedesktop.org
Tue Jan 19 19:08:23 UTC 2010
Module: Mesa
Branch: master
Commit: 7d4e674b212c9dc6408c13913a399bd4a2b9a1e3
URL: http://cgit.freedesktop.org/mesa/mesa/commit/?id=7d4e674b212c9dc6408c13913a399bd4a2b9a1e3
Author: Eric Anholt <eric at anholt.net>
Date: Tue Jan 19 10:55:21 2010 -0800
intel: Use the new DRI2 flush invalidate entrypoint to signal frame done.
Previously for frame throttling we would wait on the first batch after
a swap before emitting another swap, because we had no hook after a
swap was emitted. This meant that if an app managed to squeeze
everything it for a frame had into one batch, it would lock-step with
the GPU. With the swapbuffers changes, we now have the entrypoint we
want.
This takes the WoW intro screen from 25% GPU idle and visibly jerky to
4-5% GPU idle and rather smooth. Other apps such as OpenArena have
run into this problem as well.
---
src/mesa/drivers/dri/intel/intel_context.c | 20 --------------------
src/mesa/drivers/dri/intel/intel_screen.c | 21 +++++++++++++++++++++
2 files changed, 21 insertions(+), 20 deletions(-)
diff --git a/src/mesa/drivers/dri/intel/intel_context.c b/src/mesa/drivers/dri/intel/intel_context.c
index 3f6634c..d52fe2e 100644
--- a/src/mesa/drivers/dri/intel/intel_context.c
+++ b/src/mesa/drivers/dri/intel/intel_context.c
@@ -506,27 +506,7 @@ intelFlush(GLcontext * ctx)
static void
intel_glFlush(GLcontext *ctx)
{
- struct intel_context *intel = intel_context(ctx);
-
intel_flush(ctx, GL_TRUE);
-
- /* We're using glFlush as an indicator that a frame is done, which is
- * what DRI2 does before calling SwapBuffers (and means we should catch
- * people doing front-buffer rendering, as well)..
- *
- * Wait for the swapbuffers before the one we just emitted, so we don't
- * get too many swaps outstanding for apps that are GPU-heavy but not
- * CPU-heavy.
- *
- * Unfortunately, we don't have a handle to the batch containing the swap,
- * and getting our hands on that doesn't seem worth it, so we just us the
- * first batch we emitted after the last swap.
- */
- if (intel->first_post_swapbuffers_batch != NULL) {
- drm_intel_bo_wait_rendering(intel->first_post_swapbuffers_batch);
- drm_intel_bo_unreference(intel->first_post_swapbuffers_batch);
- intel->first_post_swapbuffers_batch = NULL;
- }
}
void
diff --git a/src/mesa/drivers/dri/intel/intel_screen.c b/src/mesa/drivers/dri/intel/intel_screen.c
index e240957..6c2cb3b 100644
--- a/src/mesa/drivers/dri/intel/intel_screen.c
+++ b/src/mesa/drivers/dri/intel/intel_screen.c
@@ -128,8 +128,29 @@ intelDRI2Flush(__DRIdrawable *drawable)
static void
intelDRI2FlushInvalidate(__DRIdrawable *drawable)
{
+ struct intel_context *intel = drawable->driContextPriv->driverPrivate;
+
intelDRI2Flush(drawable);
drawable->validBuffers = GL_FALSE;
+
+ /* We're using FlushInvalidate as an indicator that a frame is
+ * done. It's only called immediately after SwapBuffers, so it
+ * won't affect front-buffer rendering or applications explicitly
+ * managing swap regions using MESA_copy_buffer.
+ *
+ * Wait for the swapbuffers before the one we just emitted, so we don't
+ * get too many swaps outstanding for apps that are GPU-heavy but not
+ * CPU-heavy.
+ *
+ * Unfortunately, we don't have a handle to the batch containing the swap,
+ * and getting our hands on that doesn't seem worth it, so we just use the
+ * first batch we emitted after the last swap.
+ */
+ if (intel->first_post_swapbuffers_batch != NULL) {
+ drm_intel_bo_wait_rendering(intel->first_post_swapbuffers_batch);
+ drm_intel_bo_unreference(intel->first_post_swapbuffers_batch);
+ intel->first_post_swapbuffers_batch = NULL;
+ }
}
static const struct __DRI2flushExtensionRec intelFlushExtension = {
More information about the mesa-commit
mailing list