[LLVMbugs] [Bug 16562] New: LegalizeTypes.cpp:618: void llvm::DAGTypeLegalizer::RemapValue(llvm::SDValue&): Assertion `N.getNode()->getNodeId() != NewNode && "Mapped to new node!"' failed.
bugzilla-daemon at llvm.org
bugzilla-daemon at llvm.org
Mon Jul 8 08:04:14 PDT 2013
http://llvm.org/bugs/show_bug.cgi?id=16562
Bug ID: 16562
Summary: LegalizeTypes.cpp:618: void
llvm::DAGTypeLegalizer::RemapValue(llvm::SDValue&):
Assertion `N.getNode()->getNodeId() != NewNode &&
"Mapped to new node!"' failed.
Product: libraries
Version: trunk
Hardware: PC
OS: Linux
Status: NEW
Severity: normal
Priority: P
Component: Common Code Generator Code
Assignee: unassignedbugs at nondot.org
Reporter: hfinkel at anl.gov
CC: llvmbugs at cs.uiuc.edu
Classification: Unclassified
Created attachment 10829
--> http://llvm.org/bugs/attachment.cgi?id=10829&action=edit
bugpoint reduced test case
If you run the attached test case with:
llc -mtriple=powerpc64-unknown-linux-gnu -mcpu=pwr7
then this crashes:
Scalarize node result 0: 0x123b0dc0: v1i64 = extract_subvector 0x12392bc0,
0x1239d790 [ORD=7] [ID=0]
Scalarize node operand 0: 0x123ad790: i64 = extract_vector_elt 0x123b0dc0,
0x123929c0 [ORD=7] [ID=0]
llc: /llvm-trunk/lib/CodeGen/SelectionDAG/LegalizeTypes.cpp:618: void
llvm::DAGTypeLegalizer::RemapValue(llvm::SDValue&): Assertion
`N.getNode()->getNodeId() != NewNode && "Mapped to new node!"' failed.
This seems similar to http://llvm.org/bugs/show_bug.cgi?id=8923 (which Nadav
fixed a couple of years ago).
Also, this might just be a symptom of a more-general problem that I've noticed
before: we currently cannot replace a node with one of its operands.
(gdb) call Res.getNode()->dump()
0x123929c0: i64 = Constant<0> [ID=-3]
(gdb) call N->dump()
0x123ad790: i64 = extract_vector_elt 0x123b0dc0, 0x123929c0 [ORD=7] [ID=0]
(gdb) call N->getOperand(0).getNode()->dump()
0x123b0dc0: v1i64 = extract_subvector 0x12392bc0, 0x1239d790 [ORD=7] [ID=-3]
(gdb) call N->getOperand(1).getNode()->dump()
0x123929c0: i64 = Constant<0> [ID=-3]
(so the result is the constant 0 and so is one of the original operands)
--
You are receiving this mail because:
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-bugs/attachments/20130708/0c56073f/attachment.html>
More information about the llvm-bugs
mailing list