diff options
author | Bill Schmidt <wschmidt@linux.vnet.ibm.com> | 2014-06-24 20:05:18 +0000 |
---|---|---|
committer | Bill Schmidt <wschmidt@linux.vnet.ibm.com> | 2014-06-24 20:05:18 +0000 |
commit | 808d878a968257a4a010ce2cd563f552dcf91147 (patch) | |
tree | f4f1d68f6f88ec3fe640d3c332f05b2b2b1f236a /test/CodeGen/PowerPC/fast-isel-conversion-p5.ll | |
parent | 031ad1b930104d89494c7d76e20bfabc6901fabf (diff) |
[PPC64] Fix PR20071 (fctiduz generated for targets lacking that instruction)
PR20071 identifies a problem in PowerPC's fast-isel implementation for
floating-point conversion to integer. The fctiduz instruction was added in
Power ISA 2.06 (i.e., Power7 and later). However, this instruction is being
generated regardless of which 64-bit PowerPC target is selected.
The intent is for fast-isel to punt to DAG selection when this instruction is
not available. This patch implements that change. For testing purposes, the
existing fast-isel-conversion.ll test adds a RUN line for -mcpu=970 and tests
for the expected code generation. Additionally, the existing test
fast-isel-conversion-p5.ll was found to be incorrectly expecting the
unavailable instruction to be generated. I've removed these test variants
since we have adequate coverage in fast-isel-conversion.ll.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@211627 91177308-0d34-0410-b5e6-96231b3b80d8
Diffstat (limited to 'test/CodeGen/PowerPC/fast-isel-conversion-p5.ll')
-rw-r--r-- | test/CodeGen/PowerPC/fast-isel-conversion-p5.ll | 23 |
1 files changed, 0 insertions, 23 deletions
diff --git a/test/CodeGen/PowerPC/fast-isel-conversion-p5.ll b/test/CodeGen/PowerPC/fast-isel-conversion-p5.ll index db0d8ed0ffa..ac41e8c2770 100644 --- a/test/CodeGen/PowerPC/fast-isel-conversion-p5.ll +++ b/test/CodeGen/PowerPC/fast-isel-conversion-p5.ll @@ -116,18 +116,6 @@ entry: ret void } -define void @fptoui_float_i64(float %a) nounwind ssp { -entry: -; ELF64: fptoui_float_i64 - %b.addr = alloca i64, align 4 - %conv = fptoui float %a to i64 -; ELF64: fctiduz -; ELF64: stfd -; ELF64: ld - store i64 %conv, i64* %b.addr, align 4 - ret void -} - define void @fptoui_double_i32(double %a) nounwind ssp { entry: ; ELF64: fptoui_double_i32 @@ -140,14 +128,3 @@ entry: ret void } -define void @fptoui_double_i64(double %a) nounwind ssp { -entry: -; ELF64: fptoui_double_i64 - %b.addr = alloca i64, align 8 - %conv = fptoui double %a to i64 -; ELF64: fctiduz -; ELF64: stfd -; ELF64: ld - store i64 %conv, i64* %b.addr, align 8 - ret void -} |