It appears now that we have the ring TLB invalidation in place, we need
only update the page directory cachelines that we have altered. A great
reduction from rewriting the whole 2MiB ppgtt on every update.
Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
---
drivers/gpu/drm/i915/i915_gem_gtt.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 4da1dfe11007..6a2183442fed 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -1693,15 +1693,18 @@ static void gen6_ppgtt_insert_entries(struct i915_address_space *vm,
vma->page_sizes.gtt = I915_GTT_PAGE_SIZE;
}
-static void gen6_flush_pd(struct gen6_ppgtt *ppgtt)
+static void gen6_flush_pd(struct gen6_ppgtt *ppgtt, u64 start, u64 end)
{
struct i915_page_directory * const pd = ppgtt->base.pd;
struct i915_page_table *pt;
unsigned int pde;
+ start = round_down(start, SZ_64K);
+ end = round_up(end, SZ_64K) - start;
+
mutex_lock(&ppgtt->flush);
- gen6_for_all_pdes(pt, pd, pde)
+ gen6_for_each_pde(pt, pd, start, end, pde)
gen6_write_pde(ppgtt, pde, pt);
ioread32(ppgtt->pd_addr + pde - 1);
@@ -1754,8 +1757,7 @@ static int gen6_alloc_va_range(struct i915_address_space *vm,
spin_unlock(&pd->lock);
if (i915_vma_is_bound(ppgtt->vma, I915_VMA_GLOBAL_BIND))
- /* Rewrite them all! Anything less misses an invalidate. */
- gen6_flush_pd(ppgtt);
+ gen6_flush_pd(ppgtt, from, start);
goto out;
@@ -1844,7 +1846,7 @@ static int pd_vma_bind(struct i915_vma *vma,
px_base(ppgtt->base.pd)->ggtt_offset = ggtt_offset * sizeof(gen6_pte_t);
ppgtt->pd_addr = (gen6_pte_t __iomem *)ggtt->gsm + ggtt_offset;
- gen6_flush_pd(ppgtt);
+ gen6_flush_pd(ppgtt, 0, ppgtt->base.vm.total);
return 0;
}
--
2.24.0
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
