GBE: fixup/refine a bug for image1D array's extra binding index handling.

Due to hardware limitation on Gen7/Gen75 when sampling a
surface with clamp address mode and nearest filter mode
on a integer image1Darray type surface, we have to bind
one buffer to to bti. The previous implementation hard
coded it to 128 + original index and when check whether
it is such type bti in driver layer, assume the bti reserved
is 3 which is wrong now.

This patch fixed those hard coded functions and use the
macros defined in the program.h.

Signed-off-by: Zhigang Gong <zhigang.gong@intel.com>
Reviewed-by: "Song, Ruiling" <ruiling.song@intel.com>

3 files changed, 12 insertions, 5 deletions

diff --git a/backend/src/backend/gen_insn_selection.cpp b/backend/src/backend/gen_insn_selection.cpp
index abbff77..d631579 100644
--- a/backend/src/backend/gen_insn_selection.cpp
+++ b/backend/src/backend/gen_insn_selection.cpp
@@ -3862,7 +3862,7 @@ namespace gbe
         msgLen = srcNum;
       }
       // We switch to a fixup bti for linear filter on a image1d array sampling.
-      uint32_t bti = insn.getImageIndex() + (insn.getSamplerOffset() == 2 ? 128 : 0);
+      uint32_t bti = insn.getImageIndex() + (insn.getSamplerOffset() == 2 ? BTI_MAX_IMAGE_NUM : 0);
       if (bti > 253) {
         std::cerr << "Too large bti " << bti;
         return false;
diff --git a/backend/src/backend/program.h b/backend/src/backend/program.h
index 254df92..c63ae6a 100644
--- a/backend/src/backend/program.h
+++ b/backend/src/backend/program.h
@@ -66,6 +66,8 @@ enum gbe_get_arg_info_value {
 #define BTI_CONSTANT 0
 #define BTI_PRIVATE 1
 #define BTI_RESERVED_NUM 2
+#define BTI_MAX_IMAGE_NUM 128
+#define BTI_MAX_ID (BTI_MAX_IMAGE_NUM + BTI_RESERVED_NUM - 1)
 
 /*! Constant buffer values (ie values to setup in the constant buffer) */
 enum gbe_curbe_type {
diff --git a/backend/src/llvm/llvm_gen_backend.cpp b/backend/src/llvm/llvm_gen_backend.cpp
index 1604ede..738f7d3 100644
--- a/backend/src/llvm/llvm_gen_backend.cpp
+++ b/backend/src/llvm/llvm_gen_backend.cpp
@@ -500,6 +500,11 @@ namespace gbe
     ir::ImmediateIndex processConstantImmIndex(Constant *CPV, int32_t index = 0u);
     const ir::Immediate &processConstantImm(Constant *CPV, int32_t index = 0u);
 
+    uint32_t incBtiBase() {
+      GBE_ASSERT(btiBase <= BTI_MAX_ID);
+      return btiBase++;
+    }
+
     bool runOnFunction(Function &F) {
      // Do not codegen any 'available_externally' functions at all, they have
      // definitions outside the translation unit.
@@ -1363,7 +1368,7 @@ namespace gbe
                 globalPointer.insert(std::make_pair(I, btiBase));
                 ctx.appendSurface(btiBase, reg);
                 ctx.input(argName, ir::FunctionArgument::GLOBAL_POINTER, reg, llvmInfo, ptrSize, align, btiBase);
-                btiBase++;
+                incBtiBase();
               break;
               case ir::MEM_LOCAL:
                 ctx.input(argName, ir::FunctionArgument::LOCAL_POINTER, reg,  llvmInfo, ptrSize, align, 0xfe);
@@ -1374,7 +1379,7 @@ namespace gbe
               break;
               case ir::IMAGE:
                 ctx.input(argName, ir::FunctionArgument::IMAGE, reg, llvmInfo, ptrSize, align, 0x0);
-                ctx.getFunction().getImageSet()->append(reg, &ctx, btiBase++);
+                ctx.getFunction().getImageSet()->append(reg, &ctx, incBtiBase());
               break;
               default: GBE_ASSERT(addrSpace != ir::MEM_PRIVATE);
             }
@@ -1720,12 +1725,12 @@ namespace gbe
         if(v.getName().equals(StringRef("__gen_ocl_printf_buf"))) {
           ctx.appendSurface(btiBase, ir::ocl::printfbptr);
           ctx.getFunction().getPrintfSet()->setBufBTI(btiBase);
-          globalPointer.insert(std::make_pair(&v, btiBase++));
+          globalPointer.insert(std::make_pair(&v, incBtiBase()));
           regTranslator.newScalarProxy(ir::ocl::printfbptr, const_cast<GlobalVariable*>(&v));
         } else if(v.getName().equals(StringRef("__gen_ocl_printf_index_buf"))) {
           ctx.appendSurface(btiBase, ir::ocl::printfiptr);
           ctx.getFunction().getPrintfSet()->setIndexBufBTI(btiBase);
-          globalPointer.insert(std::make_pair(&v, btiBase++));
+          globalPointer.insert(std::make_pair(&v, incBtiBase()));
           regTranslator.newScalarProxy(ir::ocl::printfiptr, const_cast<GlobalVariable*>(&v));
 	} else if(v.getName().str().substr(0, 4) == ".str") {
           /* When there are multi printf statements in multi kernel fucntions within the same