diff mbox

[1/5] drm/i915/gtt: Move wmb inside ggtt_invalidate

Message ID 20180507165222.17009-1-mika.kuoppala@linux.intel.com (mailing list archive)
State New, archived
Headers show

Commit Message

Mika Kuoppala May 7, 2018, 4:52 p.m. UTC 
We use a pattern of wmb() along with ggtt_invalidate.
Move the wmb out from call sites into the ggtt_invalidate
as it is part of invalidation.

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

Comments

Chris Wilson May 7, 2018, 9:59 p.m. UTC | #1 
Quoting Mika Kuoppala (2018-05-07 17:52:18)
> We use a pattern of wmb() along with ggtt_invalidate.
> Move the wmb out from call sites into the ggtt_invalidate
> as it is part of invalidation.
> 
> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Matthew Auld <matthew.auld@intel.com>
> Signed-off-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

If we take this and the last which look at ggtt together, then
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>

I've a persistent dread when it comes to write through the GTT itself,
as different wc buffers seem to mess up aliasing, but this should be
simple wc vs uc mmio, so playing by concrete rules.
-Chris
diff mbox

Patch

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index c879bfd9294f..2963d3d71729 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -110,6 +110,9 @@  i915_get_ggtt_vma_pages(struct i915_vma *vma);
 
 static void gen6_ggtt_invalidate(struct drm_i915_private *dev_priv)
 {
+	/* Flush write combining buffer */
+	wmb();
+
 	/* Note that as an uncached mmio write, this should flush the
 	 * WCB of the writes into the GGTT before it triggers the invalidate.
 	 */
@@ -2418,8 +2421,6 @@  static void gen8_ggtt_insert_entries(struct i915_address_space *vm,
 	for_each_sgt_dma(addr, sgt_iter, vma->pages)
 		gen8_set_pte(gtt_entries++, pte_encode | addr);
 
-	wmb();
-
 	/* This next bit makes the above posting read even more important. We
 	 * want to flush the TLBs only after we're certain all the PTE updates
 	 * have finished.
@@ -2460,7 +2461,6 @@  static void gen6_ggtt_insert_entries(struct i915_address_space *vm,
 	dma_addr_t addr;
 	for_each_sgt_dma(addr, iter, vma->pages)
 		iowrite32(vm->pte_encode(addr, level, flags), &entries[i++]);
-	wmb();
 
 	/* This next bit makes the above posting read even more important. We
 	 * want to flush the TLBs only after we're certain all the PTE updates