llvm/test/CodeGen/PowerPC/swaps-le-4.ll
Bill Schmidt 397fac95d5 [PPC64LE] Remove implicit-subreg restriction from VSX swap removal
In r241285, I removed the SUBREG_TO_REG restriction from VSX swap
removal, determining that this was overly conservative.  We have
another form of the same restriction in that we check for the presence
of implicit subregs in vector operations.  As with SUBREG_TO_REG for
partial register conversions, an implicit subreg is safe in and of
itself, provided no other operation makes a lane-sensitive assumption
about the result.  This patch removes that restriction, by removing
the HasImplicitSubreg flag and all code that relies on it.

I've added a test case that fails to optimize before this patch is
applied, and optimizes properly with the patch.  Test based on a
report from Anton Blanchard.


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@241290 91177308-0d34-0410-b5e6-96231b3b80d8
2015-07-02 19:01:22 +00:00

28 lines
825 B
LLVM

; RUN: llc -mcpu=pwr8 -mtriple=powerpc64le-unknown-linux-gnu -O3 < %s | FileCheck %s
; This test verifies that VSX swap optimization works when an implicit
; subregister is present (in this case, in the XXPERMDI associated with
; the store).
define void @bar() {
entry:
%x = alloca <2 x i64>, align 16
%0 = bitcast <2 x i64>* %x to i8*
call void @llvm.lifetime.start(i64 16, i8* %0)
%arrayidx = getelementptr inbounds <2 x i64>, <2 x i64>* %x, i64 0, i64 0
store <2 x i64> <i64 0, i64 1>, <2 x i64>* %x, align 16
call void @foo(i64* %arrayidx)
call void @llvm.lifetime.end(i64 16, i8* %0)
ret void
}
; CHECK-LABEL: @bar
; CHECK: lxvd2x
; CHECK: stxvd2x
; CHECK-NOT: xxswapd
declare void @llvm.lifetime.start(i64, i8* nocapture)
declare void @foo(i64*)
declare void @llvm.lifetime.end(i64, i8* nocapture)