[llvm] r227645 - Add PPC test for r227481, but XFAIL because this is actually more work than it appeared to be.

[David Blaikie]

On Fri, Jan 30, 2015 at 4:04 PM, Hal Finkel <hfinkel at anl.gov> wrote:

> ----- Original Message -----
> > From: "David Blaikie" <dblaikie at gmail.com>
> > To: llvm-commits at cs.uiuc.edu
> > Sent: Friday, January 30, 2015 5:52:19 PM
> > Subject: [llvm] r227645 - Add PPC test for r227481,   but XFAIL because
> this is actually more work than it appeared to
> > be.
> >
> > Author: dblaikie
> > Date: Fri Jan 30 17:52:19 2015
> > New Revision: 227645
> >
> > URL: http://llvm.org/viewvc/llvm-project?rev=227645&view=rev
> > Log:
> > Add PPC test for r227481, but XFAIL because this is actually more
> > work than it appeared to be.
> >
> > Same sort of bug as on ARM where the cmp+branch are lowered to br_cc
> > (choosing the branch's debugloc for the br_cc's debugloc) then
> > expanded
> > out to a cmp and a br, but both using the debug loc of the br_cc,
> > thus
> > losing fidelity.
>
> Do we need to have two debug locations on br_cc? (and maybe select_cc?)
>

That would be one way to solve it, I'd imagine (echristo & grosbach were
speculating about that on IRC earlier when I brought this up - but it's not
enough of a priority for me to bother investigating that right now). I
don't know enough about SelectionDAG to know if that's the right path
forward, or if we should avoid coalescing those options if we can, etc.


>
>  -Hal
>
> >
> > Added:
> >     llvm/trunk/test/DebugInfo/PowerPC/line.test
> >
> > Added: llvm/trunk/test/DebugInfo/PowerPC/line.test
> > URL:
> >
> http://llvm.org/viewvc/llvm-project/llvm/trunk/test/DebugInfo/PowerPC/line.test?rev=227645&view=auto
> >
> ==============================================================================
> > --- llvm/trunk/test/DebugInfo/PowerPC/line.test (added)
> > +++ llvm/trunk/test/DebugInfo/PowerPC/line.test Fri Jan 30 17:52:19
> > 2015
> > @@ -0,0 +1,7 @@
> > +; RUN: llc -mtriple=powerpc-unknown-linux -O0 -filetype=asm <
> > %S/../Inputs/line.ll | FileCheck %S/../Inputs/line.ll
> > +
> > +; This is more complex than it looked. It's mixed up somewhere in
> > SelectionDAG
> > +; (legalized as br_cc, losing the separation between the comparison
> > and the
> > +; branch, then further lowered to cmplwi + brcc but without the
> > fidelity that
> > +; those two instructions are on separate lines)
> > +; XFAIL: *
> >
> >
> > _______________________________________________
> > llvm-commits mailing list
> > llvm-commits at cs.uiuc.edu
> > http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
> >
>
> --
> Hal Finkel
> Assistant Computational Scientist
> Leadership Computing Facility
> Argonne National Laboratory
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/llvm-commits/attachments/20150130/5afff087/attachment.html>