[PATCH] libpayload: fix cross compilation (somewhat)

[Stefan Reinauer]

See patch

[Peter Stuge]

On Thu, Aug 07, 2008 at 01:16:03PM +0200, Stefan Reinauer wrote:
> +++ libpayload/util/kconfig/lxdialog/check-lxdialog.sh	(working copy)
> @@ -19,6 +19,11 @@
>    	echo '-lcurses'
>    	exit
>    fi
> $cc -print-file-name=libcurses.dylib | grep -q /

Really lowercase cc here?

Otherwise:

Acked-by: Peter Stuge

[Stefan Reinauer]

Peter Stuge wrote:
> On Thu, Aug 07, 2008 at 01:16:03PM +0200, Stefan Reinauer wrote:
> > +++ libpayload/util/kconfig/lxdialog/check-lxdialog.sh	(working copy)
> > @@ -19,6 +19,11 @@
> >    	echo '-lcurses'
> >    	exit
> >    fi
> > $cc -print-file-name=libcurses.dylib | grep -q /
>
> Really lowercase cc here?

that variable is internally set. I know it looks weird, but I did not
invent that. Though, that one change slipped in, it should not be
committed, instead Uwe's upcoming port of the coreinfo kconfig should be
used. Plus the patch I comitted for that today.
> Otherwise:
>
> Acked-by: Peter Stuge

Great, thanks!

I'll wait some more to see whether Jordan and/or Uwe have something to
say to that stuff. It fixes my problem, but it may not be the direction
the want to see this going...?

libpayload: fix cross compilation (somewhat)

[Jordan Crouse]

On 07/08/08 13:16 +0200, Stefan Reinauer wrote:
> See patch
> Attempt to fix cross compilation for libpayload / coreinfo by honoring
> the setting of CC in the payload (coreinfo) when calling
>
> make CC=i386-elf-gcc AS=i386-elf-as AR=i386-elf-ar STRIP=i386-elf-strip
>
> This still does not cope with the hardcoded -fno-stack-protector in
> libpayload's Makefile.
>
> Signed-off-by: Stefan Reinauer
> Index: libpayload/bin/lpgcc
>
> --- libpayload/bin/lpgcc	(revision 3473)
> +++ libpayload/bin/lpgcc	(working copy)
> @@ -28,8 +28,7 @@
>
> # GCC wrapper for libpayload
>
> -DEFAULT_PREFIX=/opt
> -DEFAULT_CC=gcc
> +DEFAULT_CC=$CC

You should put a DEFAULT_CC ?= gcc after this line to account for
CC being empty. Other then that, this looks good.

Jordan

libpayload: fix cross compilation (somewhat)

[Stefan Reinauer]

Jordan Crouse wrote:
> > Index: libpayload/bin/lpgcc
> >
> > --- libpayload/bin/lpgcc	(revision 3473)
> > +++ libpayload/bin/lpgcc	(working copy)
> > @@ -28,8 +28,7 @@
> >
> > # GCC wrapper for libpayload
> >
> > -DEFAULT_PREFIX=/opt
> > -DEFAULT_CC=gcc
> > +DEFAULT_CC=$CC
>
> You should put a DEFAULT_CC ?= gcc after this line to account for
> CC being empty. Other then that, this looks good.

ok, done.

Next problem: How do I probe for -fno-stack-protector in the libpayload
makefile?

should we use lpgcc internally?

It doesn't work as easy as specifying it with CC. Anyone got a simple idea?

Stefan

libpayload: fix cross compilation (somewhat)

[Jordan Crouse]

On 07/08/08 17:20 +0200, Stefan Reinauer wrote:
> Jordan Crouse wrote:
> > > Index: libpayload/bin/lpgcc
> > >
> > > --- libpayload/bin/lpgcc	(revision 3473)
> > > +++ libpayload/bin/lpgcc	(working copy)
> > > @@ -28,8 +28,7 @@
> > >
> > > # GCC wrapper for libpayload
> > >
> > > -DEFAULT_PREFIX=/opt
> > > -DEFAULT_CC=gcc
> > > +DEFAULT_CC=$CC
> >
> > You should put a DEFAULT_CC ?= gcc after this line to account for
> > CC being empty. Other then that, this looks good.
>
> ok, done.
>
> Next problem: How do I probe for -fno-stack-protector in the libpayload
> makefile?

We don't have the -fno-stack-protector logic in the libpayload makefile?
Was I really that stupid? /me checks - Yep, i was that stupid.

Use the logic from buildrom:

try-run= $(shell set -e; \
TMP=".$$$$.tmp"; \
if ($(1)) > /dev/null 2>&1; \
then echo "$(2)"; \
else echo "$(3)"; \
fi; rm -rf "$$TMP")

cc-option= $(call
try-run,\
$(CC) $(1) -S
-xc /dev/null
-o "$$TMP",
$(1), $(2))

STACKPROTECT += $(call cc-option, -fno-stack-protector,)
> should we use lpgcc internally?

If it works, but I bet it wouldn't.

Jordan

libpayload: fix cross compilation (somewhat)

[Stefan Reinauer]

Jordan Crouse wrote:
> On 07/08/08 17:20 +0200, Stefan Reinauer wrote:
> > Jordan Crouse wrote:
> > > > Index: libpayload/bin/lpgcc
> > > >
> > > > --- libpayload/bin/lpgcc	(revision 3473)
> > > > +++ libpayload/bin/lpgcc	(working copy)
> > > > @@ -28,8 +28,7 @@
> > > >
> > > > # GCC wrapper for libpayload
> > > >
> > > > -DEFAULT_PREFIX=/opt
> > > > -DEFAULT_CC=gcc
> > > > +DEFAULT_CC=$CC
> > >
> > > You should put a DEFAULT_CC ?= gcc after this line to account for
> > > CC being empty. Other then that, this looks good.
> >
> > ok, done.
> >
> > Next problem: How do I probe for -fno-stack-protector in the libpayload
> > makefile?
>
> We don't have the -fno-stack-protector logic in the libpayload makefile?
> Was I really that stupid? /me checks - Yep, i was that stupid.
>
> Use the logic from buildrom:
>
> try-run= $(shell set -e; \
> TMP=".$$$$.tmp"; \
> if ($(1)) > /dev/null 2>&1; \
> then echo "$(2)"; \
> else echo "$(3)"; \
> fi; rm -rf "$$TMP")
>
> cc-option= $(call
> try-run,\
> $(CC) $(1) -S
> -xc /dev/null
> -o "$$TMP",
> $(1), $(2))
>
> STACKPROTECT += $(call cc-option, -fno-stack-protector,)

Cool, works and checked in...

libpayload: fix cross compilation (somewhat)

[Uwe Hermann]

On Thu, Aug 07, 2008 at 05:31:01PM +0200, Stefan Reinauer wrote:
> > Use the logic from buildrom:
> >
> > try-run= $(shell set -e; \
> > TMP=".$$$$.tmp"; \
> > if ($(1)) > /dev/null 2>&1; \
> > then echo "$(2)"; \
> > else echo "$(3)"; \
> > fi; rm -rf "$$TMP")
> >
> > cc-option= $(call
> > try-run,\
> > $(CC) $(1) -S
> > -xc /dev/null
> > -o "$$TMP",
> > $(1), $(2))
> >
> > STACKPROTECT += $(call cc-option, -fno-stack-protector,)
>
> Cool, works and checked in...

r3478.

Uwe.