[llvm] r276101 - [ARM] Skip inline asm memory operands in DAGToDAGISel

[Diana Picus via llvm-commits]

Author: rovka
Date: Wed Jul 20 04:48:24 2016
New Revision: 276101

URL: http://llvm.org/viewvc/llvm-project?rev=276101&view=rev
Log:
[ARM] Skip inline asm memory operands in DAGToDAGISel

Retry r275776 (no changes, we suspect the issue was with another commit).

The current logic for handling inline asm operands in DAGToDAGISel interprets
the operands by looking for constants, which should represent the flags
describing the kind of operand we're dealing with (immediate, memory, register
def etc). The operands representing actual data are skipped only if they are
non-const, with the exception of immediate operands which are skipped explicitly
when a flag describing an immediate is found.

The oversight is that memory operands may be const too (e.g. for device drivers
reading a fixed address), so we should explicitly skip the operand following a
flag describing a memory operand. If we don't, we risk interpreting that
constant as a flag, which is definitely not intended.

Fixes PR26038

Differential Revision: https://reviews.llvm.org/D22103

Modified:
    llvm/trunk/lib/Target/ARM/ARMISelDAGToDAG.cpp
    llvm/trunk/test/CodeGen/ARM/inlineasm3.ll

Modified: llvm/trunk/lib/Target/ARM/ARMISelDAGToDAG.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/ARM/ARMISelDAGToDAG.cpp?rev=276101&r1=276100&r2=276101&view=diff
==============================================================================
--- llvm/trunk/lib/Target/ARM/ARMISelDAGToDAG.cpp (original)
+++ llvm/trunk/lib/Target/ARM/ARMISelDAGToDAG.cpp Wed Jul 20 04:48:24 2016
@@ -4264,6 +4264,17 @@ bool ARMDAGToDAGISel::tryInlineAsm(SDNod
     if (Changed && InlineAsm::isUseOperandTiedToDef(Flag, DefIdx))
       IsTiedToChangedOp = OpChanged[DefIdx];
 
+    // Memory operands to inline asm in the SelectionDAG are modeled with two
+    // operands: a constant of value InlineAsm::Kind_Mem followed by the input
+    // operand. If we get here and we have a Kind_Mem, skip the next operand (so
+    // it doesn't get misinterpreted), and continue. We do this here because
+    // it's important to update the OpChanged array correctly before moving on.
+    if (Kind == InlineAsm::Kind_Mem) {
+      SDValue op = N->getOperand(++i);
+      AsmNodeOperands.push_back(op);
+      continue;
+    }
+
     if (Kind != InlineAsm::Kind_RegUse && Kind != InlineAsm::Kind_RegDef
         && Kind != InlineAsm::Kind_RegDefEarlyClobber)
       continue;

Modified: llvm/trunk/test/CodeGen/ARM/inlineasm3.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/ARM/inlineasm3.ll?rev=276101&r1=276100&r2=276101&view=diff
==============================================================================
--- llvm/trunk/test/CodeGen/ARM/inlineasm3.ll (original)
+++ llvm/trunk/test/CodeGen/ARM/inlineasm3.ll Wed Jul 20 04:48:24 2016
@@ -121,3 +121,14 @@ entry:
   %0 = tail call <4 x i32> asm "vld1.s32 {${0:e}[], ${0:f}[]}, [$1]", "=w,r"(i32* %p) nounwind
   ret <4 x i32> %0
 }
+
+; Bugzilla PR26038
+
+define i32 @fn1() local_unnamed_addr nounwind {
+; CHECK-LABEL: fn1
+entry:
+; CHECK: mov [[addr:r[0-9]+]], #5
+; CHECK: ldrh {{.*}}[[addr]]
+  %0 = tail call i32 asm "ldrh  $0, $1", "=r,*Q"(i8* inttoptr (i32 5 to i8*)) nounwind
+  ret i32 %0
+}