AArch64: force i1 to be zero-extended at an ABI boundary.

This commit is debatable. There are two possible approaches, neither
of which is really satisfactory:

1. Use "@foo(i1 zeroext)" to mean an extension to 32-bits on Darwin,
   and 8 bits otherwise.
2. Redefine "@foo(i1)" to mean that the i1 is extended by the caller
   to 8 bits. This goes against the spirit of "zeroext" I think, but
   it's a bit of a vague construct anyway (by definition you're going
   to extend to the amount required by the ABI, that's why it's the
   ABI!).

This implements option 2. The DAG machinery really isn't setup for the
first (there's a fairly strong assumption that "zeroext" goes to at
least the smallest register size), and even if it was the resulting
DAG looks like it would be inferior in many cases.

Theoretically we could add AssertZext nodes in the consumers of
ABI-passed values too now, but this actually seems to make the code
worse in practice by making truncation proceed in two steps. The code
produced is equally valid if we continue to assume only the low bit is
defined.

Should fix PR19850

llvm-svn: 209637

diff --git a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
index 5c504d1..80d6669 100644 (file)
--- a/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
+++ b/llvm/lib/Target/AArch64/AArch64ISelLowering.cpp
@@ -2263,6 +2263,11 @@ AArch64TargetLowering::LowerCall(CallLoweringInfo &CLI,
       Arg = DAG.getNode(ISD::ZERO_EXTEND, DL, VA.getLocVT(), Arg);
       break;
     case CCValAssign::AExt:
+      if (Outs[realArgIdx].ArgVT == MVT::i1) {
+        // AAPCS requires i1 to be zero-extended to 8-bits by the caller.
+        Arg = DAG.getNode(ISD::TRUNCATE, DL, MVT::i1, Arg);
+        Arg = DAG.getNode(ISD::ZERO_EXTEND, DL, MVT::i8, Arg);
+      }
       Arg = DAG.getNode(ISD::ANY_EXTEND, DL, VA.getLocVT(), Arg);
       break;
     case CCValAssign::BCvt:
@@ -2503,6 +2508,13 @@ AArch64TargetLowering::LowerReturn(SDValue Chain, CallingConv::ID CallConv,
     default:
       llvm_unreachable("Unknown loc info!");
     case CCValAssign::Full:
+      if (Outs[i].ArgVT == MVT::i1) {
+        // AAPCS requires i1 to be zero-extended to i8 by the producer of the
+        // value. This is strictly redundant on Darwin (which uses "zeroext
+        // i1"), but will be optimised out before ISel.
+        Arg = DAG.getNode(ISD::TRUNCATE, DL, MVT::i1, Arg);
+        Arg = DAG.getNode(ISD::ZERO_EXTEND, DL, VA.getLocVT(), Arg);
+      }
       break;
     case CCValAssign::BCvt:
       Arg = DAG.getNode(ISD::BITCAST, DL, VA.getLocVT(), Arg);

diff --git a/llvm/test/CodeGen/AArch64/i1-contents.ll b/llvm/test/CodeGen/AArch64/i1-contents.ll
new file mode 100644 (file)
index 0000000..7f133fc
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/i1-contents.ll
@@ -0,0 +1,55 @@
+; RUN: llc -mtriple=aarch64-linux-gnu -o - %s | FileCheck %s
+%big = type i32
+
+@var = global %big 0
+
+; AAPCS: low 8 bits of %in (== w0) will be either 0 or 1. Need to extend to
+; 32-bits.
+define void @consume_i1_arg(i1 %in) {
+; CHECK-LABEL: consume_i1_arg:
+; CHECK: and [[BOOL32:w[0-9]+]], w0, #{{0x1|0xff}}
+; CHECK: str [[BOOL32]], [{{x[0-9]+}}, :lo12:var]
+  %val = zext i1 %in to %big
+  store %big %val, %big* @var
+  ret void
+}
+
+; AAPCS: low 8 bits of %val1 (== w0) will be either 0 or 1. Need to extend to
+; 32-bits (doesn't really matter if it's from 1 or 8 bits).
+define void @consume_i1_ret() {
+; CHECK-LABEL: consume_i1_ret:
+; CHECK: bl produce_i1_ret
+; CHECK: and [[BOOL32:w[0-9]+]], w0, #{{0x1|0xff}}
+; CHECK: str [[BOOL32]], [{{x[0-9]+}}, :lo12:var]
+  %val1 = call i1 @produce_i1_ret()
+  %val = zext i1 %val1 to %big
+  store %big %val, %big* @var
+  ret void
+}
+
+; AAPCS: low 8 bits of w0 must be either 0 or 1. Need to mask them off.
+define i1 @produce_i1_ret() {
+; CHECK-LABEL: produce_i1_ret:
+; CHECK: ldr [[VAR32:w[0-9]+]], [{{x[0-9]+}}, :lo12:var]
+; CHECK: and w0, [[VAR32]], #{{0x1|0xff}}
+  %val = load %big* @var
+  %val1 = trunc %big %val to i1
+  ret i1 %val1
+}
+
+define void @produce_i1_arg() {
+; CHECK-LABEL: produce_i1_arg:
+; CHECK: ldr [[VAR32:w[0-9]+]], [{{x[0-9]+}}, :lo12:var]
+; CHECK: and w0, [[VAR32]], #{{0x1|0xff}}
+; CHECK: bl consume_i1_arg
+  %val = load %big* @var
+  %val1 = trunc %big %val to i1
+  call void @consume_i1_arg(i1 %val1)
+  ret void
+}
+
+
+;define zeroext i1 @foo(i8 %in) {
+;  %val = trunc i8 %in to i1
+;  ret i1 %val
+;}