I'm ok with that changes. I'm not sure why kernel did not turn this optimization off and goes with compiling crt.s with sources. There could be a reasons for that. More likely to not put anything from libgcc to kernel. We should do the same for purgatory code. So maybe it's reasonable to apply my patch and go the same way as kernel goes (but it will break old compilers, that should not be huge problem.) Maxim. 2010/10/5 Matthew McClintock <msm at freescale.com>: > Purgatory code with -Os is not working on certain compilers, > instead of finding out the hard way, just compile with -O0 > > Signed-off-by: Matthew McClintock <msm at freescale.com> > --- > Tried to get a dicussion going on this issue, but for now I > suspect it is better to have a working kexec on ppc instead of > waiting > > ?purgatory/arch/ppc/Makefile | ? ?2 ++ > ?1 files changed, 2 insertions(+), 0 deletions(-) > > diff --git a/purgatory/arch/ppc/Makefile b/purgatory/arch/ppc/Makefile > index 72289a0..319870e 100644 > --- a/purgatory/arch/ppc/Makefile > +++ b/purgatory/arch/ppc/Makefile > @@ -2,6 +2,8 @@ > ?# Purgatory ppc > ?# > > +ppc_PURGATORY_EXTRA_CFLAGS = -O0 > + > ?ppc_PURGATORY_SRCS += purgatory/arch/ppc/v2wrap_32.S > ?ppc_PURGATORY_SRCS += purgatory/arch/ppc/misc.S > ?ppc_PURGATORY_SRCS += purgatory/arch/ppc/purgatory-ppc.c > -- > 1.6.6.1 > > > > _______________________________________________ > kexec mailing list > kexec at lists.infradead.org > http://lists.infradead.org/mailman/listinfo/kexec > -- Best regards, Maxim Uvarov
