[Bug math/19545] 'sqrt' on e6500 target evaluates to an invalid opcode

["jsm28 at gcc dot gnu.org" <sourceware-bugzilla at sourceware dot org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=19545

Joseph Myers <jsm28 at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |WAITING

--- Comment #2 from Joseph Myers <jsm28 at gcc dot gnu.org> ---
You can't possibly be using the reported version of 2.24, because glibc 2.23
hasn't released or branched yet.  Please confirm exactly what glibc version you
are using, in particular whether it includes:

commit 08cee2a464f614a6d4275b5af6c52481f1aa16e6
Author: Adhemerval Zanella <azanella@linux.vnet.ibm.com>
Date:   Tue Jan 27 13:16:39 2015 -0500

    powerpc: Fix fsqrt build in libm [BZ#16576]

    Some powerpc64 processors (e5500 core for instance) does not provide the
    fsqrt instruction, however current check to use in math_private.h is
    __WORDSIZE and _ARCH_PWR4 (ISA 2.02).  This is patch change it to use
    the compiler flag _ARCH_PPCSQ (which is the same condition GCC uses to
    decide whether to generate fsqrt instruction).

    It fixes BZ#16576.

If it does, you'll need to look in more detail at exactly what glibc source
files get built to produce the problem sqrt implementation.  Is _ARCH_PPCSQ
defined at the relevant point in those source files?  If so, that's a problem
with the GCC configuration used to build those files; it's not correctly
configured for those processors.  If it's not defined, where do those fsqrt
instructions come from that's not conditioned on _ARCH_PPCSQ?

-- 
You are receiving this mail because:
You are on the CC list for the bug.