[Mesa-dev] [PATCH 9/9] radeonsi: don't do any flushes at the end of IBs
Marek Olšák
maraeo at gmail.com
Fri Jun 16 12:58:05 UTC 2017
From: Marek Olšák <marek.olsak at amd.com>
---
src/gallium/drivers/radeonsi/si_hw_context.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/src/gallium/drivers/radeonsi/si_hw_context.c b/src/gallium/drivers/radeonsi/si_hw_context.c
index 5d930a6..b420c9f 100644
--- a/src/gallium/drivers/radeonsi/si_hw_context.c
+++ b/src/gallium/drivers/radeonsi/si_hw_context.c
@@ -122,20 +122,23 @@ void si_context_gfx_flush(void *context, unsigned flags,
}
ctx->gfx_flush_in_progress = true;
/* This CE dump should be done in parallel with the last draw. */
if (ctx->ce_ib)
si_ce_save_all_descriptors_at_ib_end(ctx);
r600_preflush_suspend_features(&ctx->b);
+ /* The kernel-side fence will flush everything. */
+ ctx->b.flags = 0;
+
/* DRM 3.1.0 doesn't flush TC for VI correctly. */
if (ctx->b.chip_class == VI && ctx->b.screen->info.drm_minor <= 1) {
ctx->b.flags |= SI_CONTEXT_PS_PARTIAL_FLUSH |
SI_CONTEXT_CS_PARTIAL_FLUSH |
SI_CONTEXT_INV_GLOBAL_L2 |
SI_CONTEXT_INV_VMEM_L1;
}
si_emit_cache_flush(ctx);
--
2.7.4
More information about the mesa-dev
mailing list