Re: [PATCH 1/2] function-graph: disable when both x86_32 and optimize for size are configured

[Ingo Molnar <mingo@elte.hu>]

* Steven Rostedt <rostedt@goodmis.org> wrote:

> On Sat, 20 Jun 2009, Ingo Molnar wrote:
> 
> > 
> > * Steven Rostedt <rostedt@goodmis.org> wrote:
> > 
> > > diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig
> > > index 4a13e5a..1eac852 100644
> > > --- a/kernel/trace/Kconfig
> > > +++ b/kernel/trace/Kconfig
> > > @@ -121,6 +121,7 @@ config FUNCTION_GRAPH_TRACER
> > >  	bool "Kernel Function Graph Tracer"
> > >  	depends on HAVE_FUNCTION_GRAPH_TRACER
> > >  	depends on FUNCTION_TRACER
> > > +	depends on !X86_32 || !CC_OPTIMIZE_FOR_SIZE
> > >  	default y
> > 
> > Hm, nice fix, but this is a quite nasty constraint - distros like to 
> > enable CC_OPTIMIZE_FOR_SIZE as it neatly trims the kernel's size by 
> > about 30%.
> > 
> > Just in case you have not checked yet: is there no way to turn off 
> > the specific gcc optimization that causes this? Or is it -Os itself 
> > that does this optimization?
> 
> It seems to me that -Os causes it for i386. I did a make V=1 to 
> capture how the files in question were being compiled, and tried 
> various disabling of flags. -Os was the only one to make a 
> difference. :-(

oh well. I think GCC should be fixed/improved to not do this 
particular (rather narrow looking) optimization if -pg is specified 
too.

	Ingo