This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Don't use -Wno-uninitialized in math/
- From: Steve Ellcey <sellcey at imgtec dot com>
- To: Joseph Myers <joseph at codesourcery dot com>
- Cc: Chris Metcalf <cmetcalf at ezchip dot com>, <libc-alpha at sourceware dot org>
- Date: Wed, 26 Aug 2015 15:12:51 -0700
- Subject: Re: Don't use -Wno-uninitialized in math/
- Authentication-results: sourceware.org; auth=none
- References: <alpine dot DEB dot 2 dot 10 dot 1508201800380 dot 30940 at digraph dot polyomino dot org dot uk> <1440173079 dot 23512 dot 36 dot camel at ubuntu-sellcey> <55DE1A17 dot 5060706 at ezchip dot com> <alpine dot DEB dot 2 dot 10 dot 1508262116390 dot 6767 at digraph dot polyomino dot org dot uk>
- Reply-to: <sellcey at imgtec dot com>
On Wed, 2015-08-26 at 21:18 +0000, Joseph Myers wrote:
> On Wed, 26 Aug 2015, Chris Metcalf wrote:
>
> > Doing a tilegx build just now against the tip, I see something similar,
> > using gcc 4.8.2:
> >
> > In file included from ../soft-fp/soft-fp.h:321:0,
> > from ../soft-fp/fmadf4.c:30,
> > from ../sysdeps/tile/s_fma.c:1:
> > ../soft-fp/fmadf4.c: In function ʽ__fmaʼ:
> > ../soft-fp/op-common.h:274:10: error: ʽR_eʼ may be used uninitialized in this
> > function [-Werror=maybe-uninitialized]
>
> I wonder if it's something to do with macro locations, and moving the push
> / ignore right up above the include of soft-fp.h helps (in which case
> there isn't much point in the pop)? It works for me as-is with GCC 5, but
> given the various reports of problems, maybe it's too fragile and the
> warning needs disabling for the whole file in soft-fp/fma*.c.
I tried moving the diag/ignore to just after the include of
libc-internal.h and I got errors about __GNUC_PREREQ not being defined.
If I put the diag/ignore's after the include of math.h (and
libc-internal.h) then things seemed to work. Any idea why I need to
include math.h to get __GNUC_PREREQ defined?
Steve Ellcey
sellcey@imgtec.com