[PATCH] UML - support checkstack

[PATCH] UML - support checkstack

[Jeff Dike]

Make checkstack work for UML.  We need to pass the underlying architecture
name, rather than "um" to checkstack.pl.

Signed-off-by: Jeff Dike <jdike@addtoit.com>

Index: linux-2.6.18-mm/Makefile
===================================================================
--- linux-2.6.18-mm.orig/Makefile	2006-08-07 13:49:52.000000000 -0400
+++ linux-2.6.18-mm/Makefile	2006-08-07 13:53:34.000000000 -0400
@@ -1315,9 +1315,13 @@ endif #ifeq ($(config-targets),1)
 endif #ifeq ($(mixed-targets),1)
 
 PHONY += checkstack kernelrelease kernelversion
+
+# Use $(SUBARCH) here instead of $(ARCH) so that this works for UML.
+# In the UML case, $(SUBARCH) is the name of the underlying
+# architecture, while for all other arches, it is the same as $(ARCH).
 checkstack:
 	$(OBJDUMP) -d vmlinux $$(find . -name '*.ko') | \
-	$(PERL) $(src)/scripts/checkstack.pl $(ARCH)
+	$(PERL) $(src)/scripts/checkstack.pl $(SUBARCH)
 
 kernelrelease:
 	$(if $(wildcard include/config/kernel.release), $(Q)echo $(KERNELRELEASE), \

Re: [PATCH] UML - support checkstack

[Matt Mackall]

On Wed, Aug 09, 2006 at 02:15:24PM -0400, Jeff Dike wrote:
> Make checkstack work for UML.  We need to pass the underlying architecture
> name, rather than "um" to checkstack.pl.

Does this do the right thing with something like Voyager?

Or should we just get together a small fund to send the remaining
Voyager users proper computers?

-- 
Mathematics is the supreme nostalgia of our time.

Re: [PATCH] UML - support checkstack

[Jeff Dike]

On Wed, Aug 09, 2006 at 09:09:22PM -0500, Matt Mackall wrote:
> On Wed, Aug 09, 2006 at 02:15:24PM -0400, Jeff Dike wrote:
> > Make checkstack work for UML.  We need to pass the underlying architecture
> > name, rather than "um" to checkstack.pl.
> 
> Does this do the right thing with something like Voyager?

SUBARCH has a different meaning here.  For UML, it's the underlying,
host, architecture, not a variant architecture like Voyager.

> 
> Or should we just get together a small fund to send the remaining
> Voyager users proper computers?

Yeah, that's a plan :-)

				Jeff

Re: [PATCH] UML - support checkstack

[Matt Mackall]

On Thu, Aug 10, 2006 at 12:22:16AM -0400, Jeff Dike wrote:
> On Wed, Aug 09, 2006 at 09:09:22PM -0500, Matt Mackall wrote:
> > On Wed, Aug 09, 2006 at 02:15:24PM -0400, Jeff Dike wrote:
> > > Make checkstack work for UML.  We need to pass the underlying architecture
> > > name, rather than "um" to checkstack.pl.
> > 
> > Does this do the right thing with something like Voyager?
> 
> SUBARCH has a different meaning here.  For UML, it's the underlying,
> host, architecture, not a variant architecture like Voyager.

Right, so it sounds like this breaks Voyager. Which I think means we
ought to pass ARCH and SUBARCH and do the right thing inside
checkstack.

-- 
Mathematics is the supreme nostalgia of our time.

Re: [PATCH] UML - support checkstack

[Jeff Dike]

On Thu, Aug 10, 2006 at 11:45:48AM -0500, Matt Mackall wrote:
> > SUBARCH has a different meaning here.  For UML, it's the underlying,
> > host, architecture, not a variant architecture like Voyager.
> 
> Right, so it sounds like this breaks Voyager. Which I think means we
> ought to pass ARCH and SUBARCH and do the right thing inside
> checkstack.

There is no use of the symbol SUBARCH in arch/i386.  While this may be
jarring to people who know and love Voyager, it doesn't break
anything.

We could do what you suggest, but that sounds unnecessary.

I'd rather either
	leave things as they are
	rename SUBARCH

				Jeff

Re: [uml-devel] [PATCH] UML - support checkstack

[Blaisorblade]

On Tuesday 15 August 2006 05:17, Jeff Dike wrote:
> On Thu, Aug 10, 2006 at 11:45:48AM -0500, Matt Mackall wrote:
> > > SUBARCH has a different meaning here.  For UML, it's the underlying,
> > > host, architecture, not a variant architecture like Voyager.
> >
> > Right, so it sounds like this breaks Voyager. Which I think means we
> > ought to pass ARCH and SUBARCH and do the right thing inside
> > checkstack.
>
> There is no use of the symbol SUBARCH in arch/i386.  While this may be
> jarring to people who know and love Voyager, it doesn't break
> anything.
>
> We could do what you suggest, but that sounds unnecessary.
>
> I'd rather either
> 	leave things as they are
Yes, and make the script check if it ARCH=um or not (which is obvious for 
now - nobody really wants a clear abstraction here).
> 	rename SUBARCH

-- 
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade
http://www.user-mode-linux.org/~blaisorblade
Chiacchiera con i tuoi amici in tempo reale! 
 http://it.yahoo.com/mail_it/foot/*http://it.messenger.yahoo.com 

Re: [PATCH] UML - support checkstack

[Jörn Engel]

On Wed, 9 August 2006 14:15:24 -0400, Jeff Dike wrote:
> 
> Make checkstack work for UML.  We need to pass the underlying architecture
> name, rather than "um" to checkstack.pl.
> 
> Signed-off-by: Jeff Dike <jdike@addtoit.com>

Acked-by: Joern Engel <joern@wohnheim.fh-wedel.de>

Jörn

-- 
Geld macht nicht glücklich.
Glück macht nicht satt.