Hi, I ran into a problem caused by this part of the OCL specs (6.1.5 Alignment of Types): "For 3-component vector data types, the size of the data type is 4 * sizeof(component)."
and the corresponding part of Khronos cl_platform.h (with all types, not just float): /* cl_float3 is identical in size, alignment and behavior to cl_float4. See section 6.1.5. */ typedef cl_float4 cl_float3; So when I try to run kernel that takes 3 element vectors as arguments I get 'invalid arg size' error. Not sure whether this is best solved in clang, libclc or clover. I tried changing float3 to have 4 elements in libclc, it caused clang to complain in thousand places. I don't think this can be handled cleanly in clang, unless we add something like __attribute__((padding)). I have attached a workaround that I use now. any advice welcome, Jan -- Jan Vesely <[email protected]>
diff --git a/src/gallium/state_trackers/clover/llvm/invocation.cpp b/src/gallium/state_trackers/clover/llvm/invocation.cpp
index a81bdf8..71ee01f 100644
--- a/src/gallium/state_trackers/clover/llvm/invocation.cpp
+++ b/src/gallium/state_trackers/clover/llvm/invocation.cpp
@@ -62,6 +62,8 @@
#include <llvm/IR/DataLayout.h>
#endif
+#include <llvm/Support/Debug.h>
+
#include "pipe/p_state.h"
#include "util/u_memory.h"
@@ -309,6 +311,13 @@ namespace {
llvm::Type *arg_type = arg.getType();
unsigned arg_size = TD.getTypeStoreSize(arg_type);
+ if (arg_type->isVectorTy() &&
+ ::llvm::cast<::llvm::VectorType>(arg_type)->getNumElements() == 3) {
+ ::llvm::dbgs() << "Fixing argument type: " << *arg_type
+ << " size: " << arg_size << "->"
+ << (arg_size / 3) * 4 << "\n";
+ arg_size = (arg_size / 3) * 4;
+ }
llvm::Type *target_type = arg_type->isIntegerTy() ?
TD.getSmallestLegalIntType(mod->getContext(), arg_size * 8) :
signature.asc
Description: This is a digitally signed message part
_______________________________________________ mesa-dev mailing list [email protected] http://lists.freedesktop.org/mailman/listinfo/mesa-dev
