Invalidating the indirect state pointers might affect a previously
scheduled & still running 3DPRIMITIVE (causing page fault). So stall
on pixel scoreboard before that.
v2: Fix compile issue :(
v3: Stall on pixel scoreboard
Signed-off-by: Lionel Landwerlin <lionel.g.landwer...@intel.com>
Fixes: ca19ee33d7d39 ("i965/gen10: Ignore push constant packets during context
restore.")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=106243
---
src/mesa/drivers/dri/i965/brw_pipe_control.c | 8 ++++++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff --git a/src/mesa/drivers/dri/i965/brw_pipe_control.c
b/src/mesa/drivers/dri/i965/brw_pipe_control.c
index 02278be6d62..1a32e9a8e96 100644
--- a/src/mesa/drivers/dri/i965/brw_pipe_control.c
+++ b/src/mesa/drivers/dri/i965/brw_pipe_control.c
@@ -349,13 +349,17 @@ gen7_emit_vs_workaround_flush(struct brw_context *brw)
* context restore, so the mentioned hang doesn't happen. However,
* software must program push constant commands for all stages prior to
* rendering anything, so we flag them as dirty.
+ *
+ * Finally, we also make sure to stall at pixel scoreboard to make sure the
+ * constants have been loaded into the EUs prior to disable the push constants
+ * so that it doesn't hang a previous 3DPRIMITIVE.
*/
void
gen10_emit_isp_disable(struct brw_context *brw)
{
+ brw_emit_end_of_pipe_sync(brw, PIPE_CONTROL_STALL_AT_SCOREBOARD);
brw_emit_pipe_control(brw,
- PIPE_CONTROL_ISP_DIS |
- PIPE_CONTROL_CS_STALL,
+ PIPE_CONTROL_ISP_DIS,
NULL, 0, 0);
brw->vs.base.push_constants_dirty = true;
--
2.17.0
_______________________________________________
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev
