bpf: fix incorrect SELECT_CC lowering

Commit 37962a331c77 ("bpf: Improve expanding logic in LowerSELECT_CC")
intended to improve code quality for certain jmp conditions. The
commit, however, has a couple of issues:
  (1). In code, just swap is not enough, ConditionalCode CC
       should also be swapped, otherwise incorrect code will
       be generated.
  (2). The ConditionalCode swap should be subject to
       getHasJmpExt(). If getHasJmpExt() is False, certain
       conditional codes will not be supported and swap
       may generate incorrect code.

The original goal for this patch is to optimize jmp operations
which does not have JmpExt turned on. If JmpExt is on,
better code could be generated. For example, the test
select_ri.ll is introduced to demonstrate the optimization.
The same result can be achieved with -mcpu=v2 flag.

Signed-off-by: Yonghong Song <yhs@fb.com>
Acked-by: Alexei Starovoitov <ast@kernel.org>
llvm-svn: 329043

diff --git a/llvm/lib/Target/BPF/BPFISelLowering.cpp b/llvm/lib/Target/BPF/BPFISelLowering.cpp
index e02cdf5..e81c2f5 100644 (file)
--- a/llvm/lib/Target/BPF/BPFISelLowering.cpp
+++ b/llvm/lib/Target/BPF/BPFISelLowering.cpp
@@ -502,13 +502,7 @@ SDValue BPFTargetLowering::LowerSELECT_CC(SDValue Op, SelectionDAG &DAG) const {
     NegateCC(LHS, RHS, CC);
 
   SDValue TargetCC = DAG.getConstant(CC, DL, LHS.getValueType());
-
   SDVTList VTs = DAG.getVTList(Op.getValueType(), MVT::Glue);
-
-  // The constant is expected at RHS in Select_Ri pattern.
-  if (isa<ConstantSDNode>(LHS.getNode()))
-    std::swap(LHS, RHS);
-
   SDValue Ops[] = {LHS, RHS, TargetCC, TrueV, FalseV};
 
   return DAG.getNode(BPFISD::SELECT_CC, DL, VTs, Ops);

diff --git a/llvm/test/CodeGen/BPF/32-bit-subreg-peephole.ll b/llvm/test/CodeGen/BPF/32-bit-subreg-peephole.ll
index ed52684..cf68ec5 100644 (file)
--- a/llvm/test/CodeGen/BPF/32-bit-subreg-peephole.ll
+++ b/llvm/test/CodeGen/BPF/32-bit-subreg-peephole.ll
@@ -1,4 +1,4 @@
-; RUN: llc -O2 -march=bpfel -mattr=+alu32 < %s | FileCheck %s
+; RUN: llc -O2 -march=bpfel -mcpu=v2 -mattr=+alu32 < %s | FileCheck %s
 ;
 ; long long select_u(unsigned a, unsigned b, long long c, long long d)
 ; {

diff --git a/llvm/test/CodeGen/BPF/cmp.ll b/llvm/test/CodeGen/BPF/cmp.ll
index d86f190..44f00ff 100644 (file)
--- a/llvm/test/CodeGen/BPF/cmp.ll
+++ b/llvm/test/CodeGen/BPF/cmp.ll
@@ -97,7 +97,7 @@ define zeroext i8 @minu(i8 zeroext %a, i8 zeroext %b) #0 {
   %a.b = select i1 %1, i8 %a, i8 %b
   ret i8 %a.b
 ; CHECK-LABEL:minu:
-; CHECK: if r{{[0-9]+}} {{<|>}} 100
+; CHECK: if r{{[0-9]+}} {{<|>}} r{{[0-9]+}}
 }
 
 ; Function Attrs: nounwind readnone uwtable

diff --git a/llvm/test/CodeGen/BPF/select_ri.ll b/llvm/test/CodeGen/BPF/select_ri.ll
index dbd2111..61e852d 100644 (file)
--- a/llvm/test/CodeGen/BPF/select_ri.ll
+++ b/llvm/test/CodeGen/BPF/select_ri.ll
@@ -1,4 +1,4 @@
-; RUN: llc < %s -march=bpf -verify-machineinstrs | FileCheck %s
+; RUN: llc < %s -march=bpf -mcpu=v2 -verify-machineinstrs | FileCheck %s
 ;
 ; Source file:
 ; int b, c;