nvc0/ir: return 0 in imageLoad on incomplete textures

We already guarded all OP_SULDP against out of bound accesses, but we
ended up just reusing whatever value was stored in the dest registers.

Fixes CTS test shader_image_load_store.incomplete_textures

v2: fix for loads not ending up with predicates (bindless_texture)
v3: fix replacing the def

Cc: <mesa-stable@lists.freedesktop.org>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
Signed-off-by: Karol Herbst <kherbst@redhat.com>

diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp b/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp
index c47d10896ce4e473823a956215ef5f6e8883c219..87ded68e5a73ce0a444244381741e443773ca0dd 100644 (file)
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.cpp
@@ -2151,13 +2151,36 @@ NVC0LoweringPass::convertSurfaceFormat(TexInstruction *su)
    }
 }
 
+void
+NVC0LoweringPass::insertOOBSurfaceOpResult(TexInstruction *su)
+{
+   if (!su->getPredicate())
+      return;
+
+   bld.setPosition(su, true);
+
+   for (unsigned i = 0; su->defExists(i); ++i) {
+      ValueDef &def = su->def(i);
+
+      Instruction *mov = bld.mkMov(bld.getSSA(), bld.loadImm(NULL, 0));
+      assert(su->cc == CC_NOT_P);
+      mov->setPredicate(CC_P, su->getPredicate());
+      Instruction *uni = bld.mkOp2(OP_UNION, TYPE_U32, bld.getSSA(), NULL, mov->getDef(0));
+
+      def.replace(uni->getDef(0), false);
+      uni->setSrc(0, def.get());
+   }
+}
+
 void
 NVC0LoweringPass::handleSurfaceOpNVE4(TexInstruction *su)
 {
    processSurfaceCoordsNVE4(su);
 
-   if (su->op == OP_SULDP)
+   if (su->op == OP_SULDP) {
       convertSurfaceFormat(su);
+      insertOOBSurfaceOpResult(su);
+   }
 
    if (su->op == OP_SUREDB || su->op == OP_SUREDP) {
       assert(su->getPredicate());
@@ -2267,8 +2290,10 @@ NVC0LoweringPass::handleSurfaceOpNVC0(TexInstruction *su)
 
    processSurfaceCoordsNVC0(su);
 
-   if (su->op == OP_SULDP)
+   if (su->op == OP_SULDP) {
       convertSurfaceFormat(su);
+      insertOOBSurfaceOpResult(su);
+   }
 
    if (su->op == OP_SUREDB || su->op == OP_SUREDP) {
       const int dim = su->tex.target.getDim();
@@ -2370,8 +2395,10 @@ NVC0LoweringPass::handleSurfaceOpGM107(TexInstruction *su)
 {
    processSurfaceCoordsGM107(su);
 
-   if (su->op == OP_SULDP)
+   if (su->op == OP_SULDP) {
       convertSurfaceFormat(su);
+      insertOOBSurfaceOpResult(su);
+   }
 
    if (su->op == OP_SUREDP) {
       Value *def = su->getDef(0);

diff --git a/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.h b/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.h
index 8724c09afd93fb7ff6fa3cb292f4a360642e50e0..5dbb3e4f009bdd7a171898441de7dd3cccbf28b0 100644 (file)
--- a/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.h
+++ b/src/gallium/drivers/nouveau/codegen/nv50_ir_lowering_nvc0.h
@@ -172,6 +172,7 @@ private:
    void processSurfaceCoordsNVE4(TexInstruction *);
    void processSurfaceCoordsNVC0(TexInstruction *);
    void convertSurfaceFormat(TexInstruction *);
+   void insertOOBSurfaceOpResult(TexInstruction *);
    Value *calculateSampleOffset(Value *sampleID);
 
 protected: