[llvm-bugs] [Bug 32041] New: X86 xop.vpcmov trips assertion

via llvm-bugs llvm-bugs at lists.llvm.org
Wed Feb 22 14:50:50 PST 2017


            Bug ID: 32041
           Summary: X86 xop.vpcmov trips assertion
           Product: libraries
           Version: trunk
          Hardware: PC
                OS: Windows NT
            Status: NEW
          Severity: enhancement
          Priority: P
         Component: Backend: X86
          Assignee: unassignedbugs at nondot.org
          Reporter: paul_robinson at playstation.sony.com
                CC: llvm-bugs at lists.llvm.org

Created attachment 18018
  --> http://bugs.llvm.org/attachment.cgi?id=18018&action=edit
XOP intrinsics test

r295571 removed the xop.vpcmov intrinsics and added an auto-upgrade,
but we're seeing an assertion as shown below.

It looks like the vector type suffixes in general might have been missed?
I see this for .v4sf, .v2df, .v4di256, ....

(We have a long test file for XOP intrinsics, which I haven't upstreamed
because I haven't had a chance to see how much duplication it has.  If
you're interested, I've attached it.)

$ cat test.ll
target datalayout =
target triple = "x86_64-unknown-linux-gnu"

define <4 x float> @vpcmovv4sf(<4 x float> %a, <4 x float> %b, <4 x float> %c)
nounwind readnone {
  %1 = tail call <4 x float> @llvm.x86.xop.vpcmov.v4sf(<4 x float> %a, <4 x
float> %b, <4 x float>
  ret <4 x float> %1

declare <4 x float> @llvm.x86.xop.vpcmov.v4sf(<4 x float>, <4 x float>, <4 x
float>) nounwind

$ llc test.ll
llc: /home/russell/git/o/llvm/lib/IR/Instructions.cpp:2080: void
llvm::BinaryOperator::init(llvm::Instruction::BinaryOps): Assertion
`(getType()->isIntegerTy() ||
(getType()->isVectorTy() &&
cast<VectorType>(getType())->getElementType()->isIntegerTy())) &&
"Tried to create a logical operation on a non-integral type!"' failed.

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/20170222/05fab59c/attachment.html>

More information about the llvm-bugs mailing list