Re: building glibc 2.15

[Allin Cottrell <cottrell at wfu dot edu>]

On Fri, 23 Mar 2012, Ryan S. Arnold wrote:

> On Fri, Mar 23, 2012 at 3:25 PM, Mike Frysinger <vapier@gentoo.org> wrote:
> > On Friday 23 March 2012 15:58:23 Allin Cottrell wrote:
> > > Thanks. Given the "check" failures I thought I'd try using
> > > slightly more conservative kernel headers -- 3.2.12 instead of
> > > 3.3. Now make check gives fewer fails: just tst-mqueue5 and
> > > check-localplt. The messages are
> > >
> > > tst-mqueue5.out:
> > > SIGRTMIN signal in child did not arrive
> > > child failed with status 256
> > >
> > > check-localplt.out:
> > > --- ../scripts/data/localplt-i386-linux-gnu.data
> > > 2012-01-01 07:16:32.000000000 -0500
> > > +++ - Â 2012-03-23 15:26:52.785977398 -0400
> > > @@ -4,4 +4,5 @@
> > > Â libc.so: malloc
> > > Â libc.so: memalign
> > > Â libc.so: realloc
> > > +libm.so: feraiseexcept
> > > Â libm.so: matherr
> > >
> > > Any judgment on whether these are things to worry about?
> >
> > should be fine. Âthe 2nd failure shouldn't cause any runtime bugs, but it is
> > something that should be looked at in latest git if no one has.
> > -mike
>
> A while ago I created a write-up on how to find and fix these:
>
> http://sourceware.org/glibc/wiki/Testing/Check-localplt

Very clear. So that means that something inside glibc 2.15 is 
invoking feraiseexcept via PLT when it shouldn't?

Allin Cottrell