This is the mail archive of the
systemtap@sourceware.org
mailing list for the systemtap project.
Re: [RFC][ PATCH -tip v2 3/7] Kbuild: disable gcc crossjumping
- From: Masami Hiramatsu <mhiramat at redhat dot com>
- To: Steven Rostedt <rostedt at goodmis dot org>, Ingo Molnar <mingo at elte dot hu>, Ananth N Mavinakayanahalli <ananth at in dot ibm dot com>, lkml <linux-kernel at vger dot kernel dot org>, systemtap <systemtap at sources dot redhat dot com>, DLE <dle-develop at lists dot sourceforge dot net>, Jim Keniston <jkenisto at us dot ibm dot com>, Srikar Dronamraju <srikar at linux dot vnet dot ibm dot com>, Christoph Hellwig <hch at infradead dot org>, Frederic Weisbecker <fweisbec at gmail dot com>, "H. Peter Anvin" <hpa at zytor dot com>, Anders Kaseorg <andersk at ksplice dot com>, Tim Abbott <tabbott at ksplice dot com>
- Date: Fri, 17 Jul 2009 17:13:22 -0400
- Subject: Re: [RFC][ PATCH -tip v2 3/7] Kbuild: disable gcc crossjumping
- References: <20090622212255.5384.53732.stgit@localhost.localdomain> <20090622212313.5384.84166.stgit@localhost.localdomain> <alpine.DEB.2.00.0906271611280.12592@gandalf.stny.rr.com> <20090717202825.GA6872@merkur.ravnborg.org>
Sam Ravnborg wrote:
> On Sat, Jun 27, 2009 at 04:13:04PM -0400, Steven Rostedt wrote:
>> A change like this requires an ACK from Sam Ravnborg.
>>
>> -- Steve
>
> Overdue review...
>
> I do not know the gcc option so I cannot comment on theeffect of it.
Thanks for comment. however, as far as I can see, I couldn't find
in the difference between binary with -fno-crossjump and without it.
So I decided to drop this patch...
>
>>> diff --git a/Makefile b/Makefile
>>> index 2903e13..f73b139 100644
>>> --- a/Makefile
>>> +++ b/Makefile
>>> @@ -524,6 +524,10 @@ else
>>> KBUILD_CFLAGS += -O2
>>> endif
>>>
>>> +ifdef CONFIG_DISABLE_CROSSJUMP
>>> +KBUILD_CFLAGS += -fno-crossjumping
>>> +endif
>>> +
>>> include $(srctree)/arch/$(SRCARCH)/Makefile
>
> Why do we add this option _before_ we include the arch specific Makefile?
> If we do not want the arch specific Makefile to undo this then move it lower.
>
> Also please add a comment what the effect of the disabling is.
> There is no reason I should fire up info gcc to find out
> what -fno-crossjumping is about.
Sure, I'll write a comment on, next time.
>
>
>>>
>>> ifneq ($(CONFIG_FRAME_WARN),0)
>>> diff --git a/lib/Kconfig.debug b/lib/Kconfig.debug
>>> index 8da7467..f88e6b8 100644
>>> --- a/lib/Kconfig.debug
>>> +++ b/lib/Kconfig.debug
>>> @@ -673,6 +673,13 @@ config FRAME_POINTER
>>> larger and slower, but it gives very useful debugging information
>>> in case of kernel bugs. (precise oopses/stacktraces/warnings)
>>>
>>> +config DISABLE_CROSSJUMP
>>> + bool "Disable cross-function jump optimization"
>>> + help
>>> + This build option disables cross-function jump optimization
>>> + (crossjumping) of gcc. Disabling crossjumping might increase
>>> + kernel binary size a little.
>
> So the above comment tells me that is I disable crossjumping I will
> increase the kernel image a little.
> That is by no means a good explanation.
>
> Please explain the potential advantage of disabling crossjumping.
Agreed.
Thank you,
--
Masami Hiramatsu
Software Engineer
Hitachi Computer Products (America), Inc.
Software Solutions Division
e-mail: mhiramat@redhat.com