[Bug 1563134] Re: ICE in gimple_expand_builtin_pow with -O3 on ppc64el

[Launchpad Bug Tracker]

This bug was fixed in the package gcc-5 - 5.3.1-13ubuntu4

---------------
gcc-5 (5.3.1-13ubuntu4) xenial; urgency=medium

  * Update to SVN 20160404 (r234723, 5.3.1) from the gcc-5-branch.
    - Fix PR c++/70393, PR target/70453, PR rtl-optimization/69307,
      PR rtl-optimization/69102, PR rtl-optimization/69032, PR target/66660,
      PR target/64411, PR rtl-optimization/70460, PR target/68973,
      PR middle-end/70457 (LP: #1563134), PR rtl-optimization/68236,
      PR target/67896 (AArch64), PR fortran/66911, PR fortran/65795,
      PR target/67172 (MingW), PR target/70525 (x86), PR libgfortran/70235.
  * Update the ibm branch to 20160324.

 -- Matthias Klose <doko at ubuntu.com>  Tue, 05 Apr 2016 13:43:35 +0200

** Changed in: gcc-5 (Ubuntu)
       Status: Confirmed => Fix Released

-- 
You received this bug notification because you are a member of Ubuntu
Foundations Bugs, which is subscribed to gcc-5 in Ubuntu.
https://bugs.launchpad.net/bugs/1563134

Title:
  ICE in gimple_expand_builtin_pow with -O3 on ppc64el

Status in gcc:
  Unknown
Status in gcc-5 package in Ubuntu:
  Fix Released

Bug description:
  elk FTBFS in Ubuntu on ppc64el because ppc64el is using -O3, and elk
  internally uses a so-called "Generic_Function" that wraps calls to
  function pointers that might take either 1 or 2 args.

  The code in question will never actually try to call pow() with a
  single arg, but enforcement of this is several layers away, stymying
  the optimizer.

  Will attach a reduced test case to reproduce this issue.  Problem can
  be reproduced by calling:

    gcc-5 -O3 -c math2.i -o libelk_la-math.o

  It is also reproducible with gcc-6.

  It is reproducible cross-architecture, ppc64el is just the only
  architecture using -O3 by default.

To manage notifications about this bug go to:
https://bugs.launchpad.net/gcc/+bug/1563134/+subscriptions