[llvm] r199768 - CodeGen: Stop treating vectors as aggregates

Duncan P. N. Exon Smith dexonsmith at apple.com
Tue Jan 21 14:46:46 PST 2014


Author: dexonsmith
Date: Tue Jan 21 16:46:46 2014
New Revision: 199768

URL: http://llvm.org/viewvc/llvm-project?rev=199768&view=rev
Log:
CodeGen: Stop treating vectors as aggregates

Fix a crash in SjLjEHPrepare::lowerIncomingArguments caused by treating
VectorType like an aggregate.  It's first-class!

<rdar://problem/15854596>

Added:
    llvm/trunk/test/CodeGen/Thumb/sjljehprepare-lower-vector.ll
Modified:
    llvm/trunk/lib/CodeGen/SjLjEHPrepare.cpp

Modified: llvm/trunk/lib/CodeGen/SjLjEHPrepare.cpp
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/CodeGen/SjLjEHPrepare.cpp?rev=199768&r1=199767&r2=199768&view=diff
==============================================================================
--- llvm/trunk/lib/CodeGen/SjLjEHPrepare.cpp (original)
+++ llvm/trunk/lib/CodeGen/SjLjEHPrepare.cpp Tue Jan 21 16:46:46 2014
@@ -251,7 +251,7 @@ void SjLjEHPrepare::lowerIncomingArgumen
     // Aggregate types can't be cast, but are legal argument types, so we have
     // to handle them differently. We use an extract/insert pair as a
     // lightweight method to achieve the same goal.
-    if (isa<StructType>(Ty) || isa<ArrayType>(Ty) || isa<VectorType>(Ty)) {
+    if (isa<StructType>(Ty) || isa<ArrayType>(Ty)) {
       Instruction *EI = ExtractValueInst::Create(AI, 0, "", AfterAllocaInsPt);
       Instruction *NI = InsertValueInst::Create(AI, EI, 0);
       NI->insertAfter(EI);

Added: llvm/trunk/test/CodeGen/Thumb/sjljehprepare-lower-vector.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/CodeGen/Thumb/sjljehprepare-lower-vector.ll?rev=199768&view=auto
==============================================================================
--- llvm/trunk/test/CodeGen/Thumb/sjljehprepare-lower-vector.ll (added)
+++ llvm/trunk/test/CodeGen/Thumb/sjljehprepare-lower-vector.ll Tue Jan 21 16:46:46 2014
@@ -0,0 +1,23 @@
+; RUN: llc -mtriple=thumbv7-apple-ios < %s
+; SjLjEHPrepare shouldn't crash when lowering vectors.
+
+target datalayout = "e-p:32:32:32-i1:8:32-i8:8:32-i16:16:32-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:32:64-v128:32:128-a0:0:32-n32-S32"
+target triple = "thumbv7-apple-ios"
+
+define i8* @foo(<4 x i32> %c) {
+entry:
+  invoke void @bar ()
+    to label %unreachable unwind label %handler
+
+unreachable:
+  unreachable
+
+handler:
+  %tmp = landingpad { i8*, i32 } personality i8* bitcast (i32 (...)* @baz to i8*)
+  cleanup
+  resume { i8*, i32 } undef
+}
+
+declare void @bar()
+declare i32 @baz(...)
+





More information about the llvm-commits mailing list