[Intel-gfx] [PATCH 1/3] drm/i915: Only unwind the local pgtable layer if empty

[Chris Wilson]

Only if we allocated the layer and the lower level failed should we
remove this layer when unwinding. Otherwise we ignore the overlapping
entries by overwritting the old layer with scratch.

Fixes: c5d092a4293f ("drm/i915: Remove bitmap tracking for used-pml4")
Fixes: e2b763caa6eb ("drm/i915: Remove bitmap tracking for used-pdpes")
Reported-by: Matthew Auld <matthew.william.auld at gmail.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=99948
Testcase: igt/drv_selftest/live_gtt
Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
Cc: Matthew Auld <matthew.william.auld at gmail.com>
---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 37 ++++++++++++++++++++++++++-----------
 1 file changed, 26 insertions(+), 11 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 6fdbd5ae4fcb..c3a121ab8914 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -716,10 +716,13 @@ static bool gen8_ppgtt_clear_pd(struct i915_address_space *vm,
 	u32 pde;
 
 	gen8_for_each_pde(pt, pd, start, length, pde) {
+		GEM_BUG_ON(pt == vm->scratch_pt);
+
 		if (!gen8_ppgtt_clear_pt(vm, pt, start, length))
 			continue;
 
 		gen8_ppgtt_set_pde(vm, pd, vm->scratch_pt, pde);
+		GEM_BUG_ON(!pd->used_pdes);
 		pd->used_pdes--;
 
 		free_pt(vm, pt);
@@ -755,10 +758,13 @@ static bool gen8_ppgtt_clear_pdp(struct i915_address_space *vm,
 	unsigned int pdpe;
 
 	gen8_for_each_pdpe(pd, pdp, start, length, pdpe) {
+		GEM_BUG_ON(pd == vm->scratch_pd);
+
 		if (!gen8_ppgtt_clear_pd(vm, pd, start, length))
 			continue;
 
 		gen8_ppgtt_set_pdpe(vm, pdp, vm->scratch_pd, pdpe);
+		GEM_BUG_ON(!pdp->used_pdpes);
 		pdp->used_pdpes--;
 
 		free_pd(vm, pd);
@@ -801,6 +807,8 @@ static void gen8_ppgtt_clear_4lvl(struct i915_address_space *vm,
 	GEM_BUG_ON(!USES_FULL_48BIT_PPGTT(vm->i915));
 
 	gen8_for_each_pml4e(pdp, pml4, start, length, pml4e) {
+		GEM_BUG_ON(pdp == vm->scratch_pdp);
+
 		if (!gen8_ppgtt_clear_pdp(vm, pdp, start, length))
 			continue;
 
@@ -1089,6 +1097,7 @@ static int gen8_ppgtt_alloc_pd(struct i915_address_space *vm,
 
 			gen8_ppgtt_set_pde(vm, pd, pt, pde);
 			pd->used_pdes++;
+			GEM_BUG_ON(pd->used_pdes > I915_PDES);
 		}
 
 		pt->used_ptes += gen8_pte_count(start, length);
@@ -1118,21 +1127,25 @@ static int gen8_ppgtt_alloc_pdp(struct i915_address_space *vm,
 			gen8_initialize_pd(vm, pd);
 			gen8_ppgtt_set_pdpe(vm, pdp, pd, pdpe);
 			pdp->used_pdpes++;
+			GEM_BUG_ON(pdp->used_pdpes > I915_PDPES_PER_PDP(vm));
 
 			mark_tlbs_dirty(i915_vm_to_ppgtt(vm));
 		}
 
 		ret = gen8_ppgtt_alloc_pd(vm, pd, start, length);
-		if (unlikely(ret)) {
-			gen8_ppgtt_set_pdpe(vm, pdp, vm->scratch_pd, pdpe);
-			pdp->used_pdpes--;
-			free_pd(vm, pd);
-			goto unwind;
-		}
+		if (unlikely(ret))
+			goto unwind_pd;
 	}
 
 	return 0;
 
+unwind_pd:
+	if (!pd->used_pdes) {
+		gen8_ppgtt_set_pdpe(vm, pdp, vm->scratch_pd, pdpe);
+		GEM_BUG_ON(!pdp->used_pdpes);
+		pdp->used_pdpes--;
+		free_pd(vm, pd);
+	}
 unwind:
 	gen8_ppgtt_clear_pdp(vm, pdp, from, start - from);
 	return -ENOMEM;
@@ -1166,15 +1179,17 @@ static int gen8_ppgtt_alloc_4lvl(struct i915_address_space *vm,
 		}
 
 		ret = gen8_ppgtt_alloc_pdp(vm, pdp, start, length);
-		if (unlikely(ret)) {
-			gen8_ppgtt_set_pml4e(pml4, vm->scratch_pdp, pml4e);
-			free_pdp(vm, pdp);
-			goto unwind;
-		}
+		if (unlikely(ret))
+			goto unwind_pdp;
 	}
 
 	return 0;
 
+unwind_pdp:
+	if (!pdp->used_pdpes) {
+		gen8_ppgtt_set_pml4e(pml4, vm->scratch_pdp, pml4e);
+		free_pdp(vm, pdp);
+	}
 unwind:
 	gen8_ppgtt_clear_4lvl(vm, from, start - from);
 	return -ENOMEM;
-- 
2.11.0