diff options
| author | Duncan Sands <baldrick@free.fr> | 2009-04-18 20:16:54 +0000 |
|---|---|---|
| committer | Duncan Sands <baldrick@free.fr> | 2009-04-18 20:16:54 +0000 |
| commit | e4ff21ba4be9fae5c77a9f4d8e2187cf5d8f9688 (patch) | |
| tree | 2c5aab269955e7b4dc40c2127d6ff5a6876c0d9d /llvm/test/CodeGen/PowerPC/vec_splat.ll | |
| parent | b0cb5b2db2c1c688073f312406d99164dd2b04bf (diff) | |
| download | bcm5719-llvm-e4ff21ba4be9fae5c77a9f4d8e2187cf5d8f9688.tar.gz bcm5719-llvm-e4ff21ba4be9fae5c77a9f4d8e2187cf5d8f9688.zip | |
Don't try to make BUILD_VECTOR operands have the same
type as the vector element type: allow them to be of
a wider integer type than the element type all the way
through the system, and not just as far as LegalizeDAG.
This should be safe because it used to be this way
(the old type legalizer would produce such nodes), so
backends should be able to handle it. In fact only
targets which have legal vector types with an illegal
promoted element type will ever see this (eg: <4 x i16>
on ppc). This fixes a regression with the new type
legalizer (vec_splat.ll). Also, treat SCALAR_TO_VECTOR
the same as BUILD_VECTOR. After all, it is just a
special case of BUILD_VECTOR.
llvm-svn: 69467
Diffstat (limited to 'llvm/test/CodeGen/PowerPC/vec_splat.ll')
| -rw-r--r-- | llvm/test/CodeGen/PowerPC/vec_splat.ll | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/llvm/test/CodeGen/PowerPC/vec_splat.ll b/llvm/test/CodeGen/PowerPC/vec_splat.ll index bc860cbf024..a6311370003 100644 --- a/llvm/test/CodeGen/PowerPC/vec_splat.ll +++ b/llvm/test/CodeGen/PowerPC/vec_splat.ll @@ -1,7 +1,7 @@ ; Test that vectors are scalarized/lowered correctly. ; RUN: llvm-as < %s | llc -march=ppc32 -mcpu=g3 | \ ; RUN: grep stfs | count 4 -; RUN: llvm-as < %s | llc -march=ppc32 -mcpu=g5 -o %t -f -disable-legalize-types +; RUN: llvm-as < %s | llc -march=ppc32 -mcpu=g5 -o %t -f ; RUN: grep vspltw %t | count 2 ; RUN: grep vsplti %t | count 3 ; RUN: grep vsplth %t | count 1 |
