bpf: Extends zero extension elimination beyond comparison instructions

The current zero extension elimination was restricted to operands of
comparison. It actually could be extended to more cases.

For example:

  int *inc_p (int *p, unsigned a)
  {
    return p + a;
  }

'a' will be promoted to i64 during addition, and the zero extension could
be eliminated as well.

For the elimination optimization, it should be much better to start
recognizing the candidate sequence from the SRL instruction instead of J*
instructions.

This patch makes it an generic zero extension elimination pass instead of
one restricted with comparison.

Signed-off-by: Jiong Wang <jiong.wang@netronome.com>
Signed-off-by: Yonghong Song <yhs@fb.com>
llvm-svn: 327367

1 files changed, 16 insertions, 0 deletions

diff --git a/llvm/test/CodeGen/BPF/32-bit-subreg-peephole.ll b/llvm/test/CodeGen/BPF/32-bit-subreg-peephole.ll
index 78c344b1577..ed5268458f8 100644
--- a/llvm/test/CodeGen/BPF/32-bit-subreg-peephole.ll
+++ b/llvm/test/CodeGen/BPF/32-bit-subreg-peephole.ll
@@ -35,6 +35,11 @@
 ;   else
 ;     return c;
 ; }
+;
+; int *inc_p (int *p, unsigned a)
+; {
+;   return p + a;
+; }
 
 ; Function Attrs: norecurse nounwind readnone
 define dso_local i64 @select_u(i32 %a, i32 %b, i64 %c, i64 %d) local_unnamed_addr #0 {
@@ -89,3 +94,14 @@ entry:
 }
 
 declare dso_local i64 @bar(...) local_unnamed_addr #1
+
+; Function Attrs: norecurse nounwind readnone
+define dso_local i32* @inc_p(i32* readnone %p, i32 %a) local_unnamed_addr #0 {
+; CHECK-LABEL: inc_p:
+entry:
+  %idx.ext = zext i32 %a to i64
+; CHECK-NOT: r{{[0-9]+}} <<= 32
+; CHECK-NOT: r{{[0-9]+}} >>= 32
+  %add.ptr = getelementptr inbounds i32, i32* %p, i64 %idx.ext
+  ret i32* %add.ptr
+}