On Fri, 2016-05-06 at 16:09 +1000, Stephen Rothwell wrote:
> Hi Andrew,
>
> On Thu, 5 May 2016 22:44:29 -0700 Andrew Morton <a...@linux-foundation.org>
> wrote:
> >
> > On Fri, 6 May 2016 14:58:10 +1000 Stephen Rothwell <s...@canb.auug.org.au>
> > wrote:
> >
> > > After merging the akpm-current tree, today's linux-next build (and a
> > > few earlier ones) (powerpc allnoconfig (and many others)) failed like
> > > this:
> > >
> > > lib/vsprintf.c:160:2: error: initializer element is not constant
> > > lib/vsprintf.c:160:2: error: (near initialization for 'decpair[0]')
> > > lib/vsprintf.c:160:2: error: initializer element is not constant
> > > lib/vsprintf.c:160:2: error: (near initialization for 'decpair[1]')
> > > lib/vsprintf.c:160:2: error: initializer element is not constant
> > > lib/vsprintf.c:160:2: error: (near initialization for 'decpair[2]')
> > > lib/vsprintf.c:160:2: error: initializer element is not constant
> > > lib/vsprintf.c:160:2: error: (near initialization for 'decpair[3]')
> > > lib/vsprintf.c:160:2: error: initializer element is not constant
> > >
> > > and more ... basically any big endian build of code using __swabxx
> > > or cpu_to_lexx in initailisers (in the above case cpu_to_le16()).
> > >
> > > Caused (presumably) by commit
> > >
> > > eeee46ed3cda ("byteswap: try to avoid __builtin_constant_p gcc bug")
> > >
> > > This is a buyild using gcc 4.6.3.
> > >
> > > I have revreted that commit for today to see if it fixes the overnight
> > > builds.
> >
> > hm, it works for me. powerpc allnoconfig, after setting CONFIG_PPC64=y.
>
> Interesting ... mine also works with gcc 5.2.0. I do 32 and 64 bit
> allnoconfig powerpc builds before I release linux-next and they work,
> but the 32 bit one fails overnight when gcc 4.6.3 is used :-( (we don't
> do a 64 bit allnoconfig overnight as it requires playing with the
> generated .config). Basically all the powerpc builds were failing with
> gcc 4.6.3 since next-20160502.
There's a powerpc-5.3 compiler in kisskb now, I've been testing it for a while
and it seems OK, you can start using that as well of or instead if you like for
linux-next builds.
cheers
