[LinuxBIOS] Patch file for Fedora 8 compile problems

[Marc Karasek]

I think this got lost in noise when I sent it last..

As promised here is the patch for first the --build-on problem and also 
the final linker problem.  Although I consider the second one a patch 
rather than a fix.. :-)

I have compiled this and tested it with FILO under QEMU.   I think this 
will take care of all of the other builds as well.  At least the i386 
ones.. :-)

Files Modified:
LinuxBIOSv2/src/arch/i386/Config.lb
LinuxBIOSv2/src/arch/i386/lib/id.lds
LinuxBIOSv2/src/config/Config.lb
/*************************Patch***********************************************
diff -C 3 -r -x .svn LinuxBIOSv2.orig/src/arch/i386/Config.lb 
LinuxBIOSv2/src/arch/i386/Config.lb
*** LinuxBIOSv2.orig/src/arch/i386/Config.lb    2007-12-06 
10:06:31.000000000 -0500
--- LinuxBIOSv2/src/arch/i386/Config.lb    2007-12-04 16:25:13.000000000 
-0500
***************
*** 100,106 ****

makerule linuxbios  
          depends    "crt0.o init.o $(LINUXBIOS_APC) 
$(LINUXBIOS_RAM_ROM) ldscript.ld"
!         action    "$(CC) -nostdlib -nostartfiles -static -o $@ -T 
ldscript.ld crt0.o init.o"
          action    "$(CROSS_COMPILE)nm -n linuxbios | sort > linuxbios.map"
          end

--- 100,106 ----

makerule linuxbios  
          depends    "crt0.o init.o $(LINUXBIOS_APC) 
$(LINUXBIOS_RAM_ROM) ldscript.ld"
!         action    "$(CC) -Wl,--build-id=none -nostdlib -nostartfiles 
-static -o $@ -T ldscript.ld crt0.o init.o"
          action    "$(CROSS_COMPILE)nm -n linuxbios | sort > linuxbios.map"
          end

diff -C 3 -r -x .svn LinuxBIOSv2.orig/src/arch/i386/lib/id.lds 
LinuxBIOSv2/src/arch/i386/lib/id.lds
*** LinuxBIOSv2.orig/src/arch/i386/lib/id.lds    2007-12-06 
10:06:31.000000000 -0500
--- LinuxBIOSv2/src/arch/i386/lib/id.lds    2007-12-06 
10:36:31.000000000 -0500
***************
*** 1,5 ****
  SECTIONS {
!     . = (_ROMBASE + ROM_IMAGE_SIZE - 0x10) - (__id_end - __id_start);
      .id (.): {
          *(.id)
      }
--- 1,11 ----
  SECTIONS {
!
! /*
!  *           . = (_ROMBASE + (ROM_IMAGE_SIZE - 0x10)) - (__id_end - 
__id_start);
!  *          This is a temporary fix.  Under Fedora 8 ld does not like 
it if the .id section is above address 0xFFFF_EF00         
! */
!         _ROMTEMP = 0xffffef00;
!         . = _ROMTEMP;
      .id (.): {
          *(.id)
      }
Only in LinuxBIOSv2/src/arch/i386/lib: id.lds~
diff -C 3 -r -x .svn LinuxBIOSv2.orig/src/config/Config.lb 
LinuxBIOSv2/src/config/Config.lb
*** LinuxBIOSv2.orig/src/config/Config.lb    2007-12-06 
10:06:30.000000000 -0500
--- LinuxBIOSv2/src/config/Config.lb    2007-12-04 16:27:48.000000000 -0500
***************
*** 40,51 ****

makerule linuxbios_ram.o
      depends    "$(DRIVER) linuxbios.a $(LIBGCC_FILE_NAME)"
!     action    "$(CC) -nostdlib -r -o $@ c_start.o $(DRIVER) 
linuxbios.a $(LIBGCC_FILE_NAME)"
  end

makerule linuxbios_ram
      depends    "linuxbios_ram.o $(TOP)/src/config/linuxbios_ram.ld 
ldoptions"
!     action    "$(CC) -nostdlib -nostartfiles -static -o $@ -T 
$(TOP)/src/config/linuxbios_ram.ld linuxbios_ram.o"
      action     "$(CROSS_COMPILE)nm -n linuxbios_ram | sort > 
linuxbios_ram.map"
  end

--- 40,51 ----

makerule linuxbios_ram.o
      depends    "$(DRIVER) linuxbios.a $(LIBGCC_FILE_NAME)"
!     action    "$(CC) -Wl,--build-id=none -nostdlib -r -o $@ c_start.o 
$(DRIVER) linuxbios.a $(LIBGCC_FILE_NAME)"
  end

makerule linuxbios_ram
      depends    "linuxbios_ram.o $(TOP)/src/config/linuxbios_ram.ld 
ldoptions"
!     action    "$(CC) -Wl,--build-id=none -nostdlib -nostartfiles 
-static -o $@ -T $(TOP)/src/config/linuxbios_ram.ld linuxbios_ram.o"
      action     "$(CROSS_COMPILE)nm -n linuxbios_ram | sort > 
linuxbios_ram.map"
  end
/*************************Patch***********************************************

Signed-off-by: Marc Karasek mailto:[e-mail]

[Carl-Daniel Hailfinger]

Hi Marc,

thanks for your patch. Could you resend it in unified diff format (diff
-u)? That makes it easier to see the changes.

Regards,
Carl-Daniel

[Marc Karasek]

Here it is in -u format...

=================================================================
diff -r -u -x .svn LinuxBIOSv2.orig/src/arch/i386/Config.lb 
LinuxBIOSv2/src/arch/i386/Config.lb
--- LinuxBIOSv2.orig/src/arch/i386/Config.lb 2007-12-06 
10:06:31.000000000 -0500
+++ LinuxBIOSv2/src/arch/i386/Config.lb 2007-12-04 16:25:13.000000000 -0500
@@ -100,7 +100,7 @@

makerule linuxbios
depends "crt0.o init.o $(LINUXBIOS_APC) $(LINUXBIOS_RAM_ROM) ldscript.ld"
- action "$(CC) -nostdlib -nostartfiles -static -o $@ -T ldscript.ld 
crt0.o init.o"
+ action "$(CC) -Wl,--build-id=none -nostdlib -nostartfiles -static -o 
$@ -T ldscript.ld crt0.o init.o"
action "$(CROSS_COMPILE)nm -n linuxbios | sort > linuxbios.map"
end

diff -r -u -x .svn LinuxBIOSv2.orig/src/arch/i386/lib/id.lds 
LinuxBIOSv2/src/arch/i386/lib/id.lds
--- LinuxBIOSv2.orig/src/arch/i386/lib/id.lds 2007-12-06 
10:06:31.000000000 -0500
+++ LinuxBIOSv2/src/arch/i386/lib/id.lds 2007-12-06 10:36:31.000000000 -0500
@@ -1,5 +1,11 @@
SECTIONS {
- . = (_ROMBASE + ROM_IMAGE_SIZE - 0x10) - (__id_end - __id_start);
+
+/*
+ * . = (_ROMBASE + (ROM_IMAGE_SIZE - 0x10)) - (__id_end - __id_start);
+ * This is a temporary fix. Under Fedora 8 ld does not like it if the 
.id section is above address 0xFFFF_EF00
+*/
+ _ROMTEMP = 0xffffef00;
+ . = _ROMTEMP;
.id (.): {
*(.id)
}
diff -r -u -x .svn LinuxBIOSv2.orig/src/config/Config.lb 
LinuxBIOSv2/src/config/Config.lb
--- LinuxBIOSv2.orig/src/config/Config.lb 2007-12-06 10:06:30.000000000 
-0500
+++ LinuxBIOSv2/src/config/Config.lb 2007-12-04 16:27:48.000000000 -0500
@@ -40,12 +40,12 @@

makerule linuxbios_ram.o
depends "$(DRIVER) linuxbios.a $(LIBGCC_FILE_NAME)"
- action "$(CC) -nostdlib -r -o $@ c_start.o $(DRIVER) linuxbios.a 
$(LIBGCC_FILE_NAME)"
+ action "$(CC) -Wl,--build-id=none -nostdlib -r -o $@ c_start.o 
$(DRIVER) linuxbios.a $(LIBGCC_FILE_NAME)"
end

makerule linuxbios_ram
depends "linuxbios_ram.o $(TOP)/src/config/linuxbios_ram.ld ldoptions"
- action "$(CC) -nostdlib -nostartfiles -static -o $@ -T 
$(TOP)/src/config/linuxbios_ram.ld linuxbios_ram.o"
+ action "$(CC) -Wl,--build-id=none -nostdlib -nostartfiles -static -o 
$@ -T $(TOP)/src/config/linuxbios_ram.ld linuxbios_ram.o"
action "$(CROSS_COMPILE)nm -n linuxbios_ram | sort > linuxbios_ram.map"
end
=================================================================

Signed-off-by: Marc Karasek mailto:[e-mail]

/*********************
Marc Karasek
MTS
Sun Microsystems
mailto:[e-mail]
ph:770.360.6415
*********************/

Carl-Daniel Hailfinger wrote:
> Hi Marc,
>
> thanks for your patch. Could you resend it in unified diff format (diff
> -u)? That makes it easier to see the changes.
>
> Regards,
> Carl-Daniel

[Uwe Hermann]

On Fri, Dec 07, 2007 at 12:43:57PM -0500, Marc Karasek wrote:
> Here it is in -u format...

Thanks, but I think your Thunderbird broke the patch, probably because
you copy-pasted it into the mailer and it clipped long lines.

I recommend to use 'svn diff > foo.patch' in your svn tree (easier IMO,
and that already produces patches in the -u format) and then just attach
that file to the mail.
> =================================================================
> diff -r -u -x .svn LinuxBIOSv2.orig/src/arch/i386/Config.lb 
> LinuxBIOSv2/src/arch/i386/Config.lb
> --- LinuxBIOSv2.orig/src/arch/i386/Config.lb 2007-12-06 
> 10:06:31.000000000 -0500
> +++ LinuxBIOSv2/src/arch/i386/Config.lb 2007-12-04 16:25:13.000000000 -0500
> @@ -100,7 +100,7 @@
>
> makerule linuxbios
> depends "crt0.o init.o $(LINUXBIOS_APC) $(LINUXBIOS_RAM_ROM) ldscript.ld"
> action "$(CC) -nostdlib -nostartfiles -static -o $@ -T ldscript.ld
>
> crt0.o init.o"
> action "$(CC) -Wl,--build-id=none -nostdlib -nostartfiles -static -o

We should make sure this doesn't break any other distros / compiler
versions. Do all reasonably new gcc/ld versions know about '--build-id'?
> $@ -T ldscript.ld crt0.o init.o"
> action "$(CROSS_COMPILE)nm -n linuxbios | sort > linuxbios.map"
> end
>
> diff -r -u -x .svn LinuxBIOSv2.orig/src/arch/i386/lib/id.lds 
> LinuxBIOSv2/src/arch/i386/lib/id.lds
> --- LinuxBIOSv2.orig/src/arch/i386/lib/id.lds 2007-12-06 
> 10:06:31.000000000 -0500
> +++ LinuxBIOSv2/src/arch/i386/lib/id.lds 2007-12-06 10:36:31.000000000 -0500
> @@ -1,5 +1,11 @@
> SECTIONS {
> . = (_ROMBASE + ROM_IMAGE_SIZE - 0x10) - (__id_end - __id_start);
>
> +/*
> . = (_ROMBASE + (ROM_IMAGE_SIZE - 0x10)) - (__id_end - __id_start);
> This is a temporary fix. Under Fedora 8 ld does not like it if the
>
> .id section is above address 0xFFFF_EF00
> +*/
> _ROMTEMP = 0xffffef00;
> . = _ROMTEMP;

I cannot parse this changeset completely, but won't this
break the code for other distros or boards?

Please resend the 'svn diff' output so we can test this on some
non-Fedora distros.

Thanks, Uwe.

[Marc Karasek]

I have attached the patch file. You are right I did cut and paste it 
from this file.

/*********************
Marc Karasek
MTS
Sun Microsystems
mailto:[e-mail]
ph:770.360.6415
*********************/

Uwe Hermann wrote:
> On Fri, Dec 07, 2007 at 12:43:57PM -0500, Marc Karasek wrote:
> > Here it is in -u format...
>
> Thanks, but I think your Thunderbird broke the patch, probably because
> you copy-pasted it into the mailer and it clipped long lines.
>
> I recommend to use 'svn diff > foo.patch' in your svn tree (easier IMO,
> and that already produces patches in the -u format) and then just attach
> that file to the mail.
> > =================================================================
> > diff -r -u -x .svn LinuxBIOSv2.orig/src/arch/i386/Config.lb 
> > LinuxBIOSv2/src/arch/i386/Config.lb
> > --- LinuxBIOSv2.orig/src/arch/i386/Config.lb 2007-12-06 
> > 10:06:31.000000000 -0500
> > +++ LinuxBIOSv2/src/arch/i386/Config.lb 2007-12-04 16:25:13.000000000 -0500
> > @@ -100,7 +100,7 @@
> >
> > makerule linuxbios
> > depends "crt0.o init.o $(LINUXBIOS_APC) $(LINUXBIOS_RAM_ROM) ldscript.ld"
> > action "$(CC) -nostdlib -nostartfiles -static -o $@ -T ldscript.ld
> >
> > crt0.o init.o"
> > action "$(CC) -Wl,--build-id=none -nostdlib -nostartfiles -static -o
>
> We should make sure this doesn't break any other distros / compiler
> versions. Do all reasonably new gcc/ld versions know about '--build-id'?
> > $@ -T ldscript.ld crt0.o init.o"
> > action "$(CROSS_COMPILE)nm -n linuxbios | sort > linuxbios.map"
> > end
> >
> > diff -r -u -x .svn LinuxBIOSv2.orig/src/arch/i386/lib/id.lds 
> > LinuxBIOSv2/src/arch/i386/lib/id.lds
> > --- LinuxBIOSv2.orig/src/arch/i386/lib/id.lds 2007-12-06 
> > 10:06:31.000000000 -0500
> > +++ LinuxBIOSv2/src/arch/i386/lib/id.lds 2007-12-06 10:36:31.000000000 -0500
> > @@ -1,5 +1,11 @@
> > SECTIONS {
> > . = (_ROMBASE + ROM_IMAGE_SIZE - 0x10) - (__id_end - __id_start);
> >
> > +/*
> > . = (_ROMBASE + (ROM_IMAGE_SIZE - 0x10)) - (__id_end - __id_start);
> > This is a temporary fix. Under Fedora 8 ld does not like it if the
> >
> > .id section is above address 0xFFFF_EF00
> > +*/
> > _ROMTEMP = 0xffffef00;
> > . = _ROMTEMP;
>
> I cannot parse this changeset completely, but won't this
> break the code for other distros or boards?
>
> Please resend the 'svn diff' output so we can test this on some
> non-Fedora distros.
>
> Thanks, Uwe.

[Ed Swierk]

On 12/7/07, Uwe Hermann  wrote:
> We should make sure this doesn't break any other distros / compiler
> versions. Do all reasonably new gcc/ld versions know about '--build-id'?

The Fedora Core 6 version of ld (2.17.50.0.6-5.fc6) doesn't support
it. Not sure if that qualifies as reasonably new, but we're still
using it.

--Ed

[Ed Swierk]

Here is another attempt that does not rely on the --build-id option,
incorporating Marc's workaround in src/arch/i386/lib/id.lds for the
.id allocation issue.

--Ed

[Uwe Hermann]

On Fri, Dec 07, 2007 at 12:39:36PM -0800, Ed Swierk wrote:
> Here is another attempt that does not rely on the --build-id option,
> incorporating Marc's workaround in src/arch/i386/lib/id.lds for the
> .id allocation issue.

The patch is missing the original Signed-off-by from Marc, and your
Signed-off-by.
> Index: LinuxBIOSv2-2571/src/config/linuxbios_ram.ld
>
> --- LinuxBIOSv2-2571.orig/src/config/linuxbios_ram.ld
> +++ LinuxBIOSv2-2571/src/config/linuxbios_ram.ld
> @@ -62,6 +62,10 @@ SECTIONS
>    	 . = ALIGN(4);
> _erodata = .;
> }
> /* ld segfaults if we give it --build-id and then discard this section */
> .note.gnu.build-id : {
> *(.note.gnu.build-id)
> }	
> /*
> After the code we place initialized data (typically initialized
> Index: LinuxBIOSv2-2571/src/arch/i386/lib/id.lds
>
> --- LinuxBIOSv2-2571.orig/src/arch/i386/lib/id.lds
> +++ LinuxBIOSv2-2571/src/arch/i386/lib/id.lds
> @@ -1,5 +1,10 @@
>  SECTIONS {
> . = (_ROMBASE + ROM_IMAGE_SIZE - 0x10) - (__id_end - __id_start);
> /*
> * . = (_ROMBASE + ROM_IMAGE_SIZE - 0x10) - (__id_end - __id_start);
> * This is a temporary fix. Under Fedora 8 ld does not like it if the .id section is above address 0xFFFF_EF00
> */
> _ROMTEMP = 0xffffef00;
> . = _ROMTEMP;

This part breaks the majority of all boards in v2 (checked with abuild).

E.g. on asus/p2b:

./buildrom linuxbios.strip linuxbios.rom payload 0x10000 0x20000
linuxbios image is 192277 bytes; only 65536 allowed
Linuxbios input file larger than allowed size!
: Success
make[1]: *** [linuxbios.rom] Error 2
make[1]: Leaving directory `/tmp/v2/targets/asus/p2b/p2b/normal'
make: *** [normal/linuxbios.rom] Error 1

There must be some other way to fix this.
> .id (.): {
>    	*(.id)
>    }
> Index: LinuxBIOSv2-2571/src/arch/i386/init/ldscript.lb
> ===================================================================
> --- LinuxBIOSv2-2571.orig/src/arch/i386/init/ldscript.lb
> +++ LinuxBIOSv2-2571/src/arch/i386/init/ldscript.lb
> @@ -53,6 +53,10 @@ SECTIONS
>    	_erom = .;
>    }
> .note.gnu.build-id : {
> *(.note.gnu.build-id)
> }
> _lrom = LOADADDR(.rom);
>  _elrom = LOADADDR(.rom) + SIZEOF(.rom);
>  _iseg = _RAMBASE;
>
> @@ -63,5 +67,6 @@ SECTIONS
>    /DISCARD/ : {
>    	*(.comment)
>    	*(.note)
> *(.note.*)
> }
>
> }

Uwe.

[Marc Karasek]

I am going to look at how to make the Makefile find out what you are 
running on.  Maybe we can use the xcompile util from V3 or look at ld 
somehow and make a decision about what options to compile with.

I will look into this in the next day or so and will post my findings...

/*********************
Marc Karasek
MTS
Sun Microsystems
mailto:[e-mail]
ph:770.360.6415
*********************/

Ed Swierk wrote:
> On 12/7/07, Uwe Hermann  wrote:
> > We should make sure this doesn't break any other distros / compiler
> > versions. Do all reasonably new gcc/ld versions know about '--build-id'?
>
> The Fedora Core 6 version of ld (2.17.50.0.6-5.fc6) doesn't support
> it. Not sure if that qualifies as reasonably new, but we're still
> using it.
>
> --Ed

[Marc Karasek]

I have come up with a awk script that could be used in the makefile to 
query the ld version and make a decision based the output to add the 
compile option or not.  I have attached the script and will incorporate 
it into the makefile and send a patch out hopefully today.

/*********************
Marc Karasek
MTS
Sun Microsystems
mailto:[e-mail]
ph:770.360.6415
*********************/

Ed Swierk wrote:
> On 12/7/07, Uwe Hermann  wrote:
> > We should make sure this doesn't break any other distros / compiler
> > versions. Do all reasonably new gcc/ld versions know about '--build-id'?
>
> The Fedora Core 6 version of ld (2.17.50.0.6-5.fc6) doesn't support
> it. Not sure if that qualifies as reasonably new, but we're still
> using it.
>
> --Ed

[Triton]

On Tue, 11 Dec 2007 11:20:36 -0500
Marc Karasek  wrote:
> I have come up with a awk script that could be used in the makefile to 
> query the ld version and make a decision based the output to add the 
> compile option or not.  I have attached the script and will incorporate 
> it into the makefile and send a patch out hopefully today.

No attachment here. :-/

[Corey Osgood]

Triton wrote:
> On Tue, 11 Dec 2007 11:20:36 -0500
> Marc Karasek  wrote:
> > I have come up with a awk script that could be used in the makefile to 
> > query the ld version and make a decision based the output to add the 
> > compile option or not.  I have attached the script and will incorporate 
> > it into the makefile and send a patch out hopefully today.
>
> No attachment here. :-/

Read the quote again ;)

-Corey

[Marc Karasek]

Forgot the attachment

/*********************
Marc Karasek
MTS
Sun Microsystems
mailto:[e-mail]
ph:770.360.6415
*********************/

Ed Swierk wrote:
> On 12/7/07, Uwe Hermann  wrote:
> > We should make sure this doesn't break any other distros / compiler
> > versions. Do all reasonably new gcc/ld versions know about '--build-id'?
>
> The Fedora Core 6 version of ld (2.17.50.0.6-5.fc6) doesn't support
> it. Not sure if that qualifies as reasonably new, but we're still
> using it.
>
> --Ed

{for (i=1;i<=NF;i++) if ($i=="2.17"){n++} }; END {print n}

(echo | ld --version) | awk '{for (i=1;i<=NF;i++) if ($i=="2.17.50.0.18-1"){n++} }; END {print n}'

[ron minnich]

I think you are adding too much complexity to the build process. it
may be time to step back a bit and  think about all this.

Just my $.02

ron

[Marc Karasek]

I did and I came up with a better solution.  I added an option called 
BUILDID_OPTION to the process.  This is a string that is set to the 
"-Wl,--build-id=none"  for newer ld versions and can be set to blank for 
older ld versions.  I have attached the patch file to add this to V2.  I 
think this should take care of any backwards compatibility. :-)

/*********************
Marc Karasek
MTS
Sun Microsystems
mailto:[e-mail]
ph:770.360.6415
*********************/

ron minnich wrote:
> I think you are adding too much complexity to the build process. it
> may be time to step back a bit and  think about all this.
>
> Just my $.02
>
> ron

[ron minnich]

OK, I am looking at this patch again, and it's still not quite the
right idea. Setting this kind of build option in the Config file is
not really right.

However ...

if you just retained this much:
Index: src/arch/i386/Config.lb
===================================================================
--- src/arch/i386/Config.lb     (revision 3003)
+++ src/arch/i386/Config.lb     (working copy)
@@ -100,7 +100,7 @@

makerule linuxbios
                depends "crt0.o init.o $(LINUXBIOS_APC)
$(LINUXBIOS_RAM_ROM) ldscript.ld"
-               action  "$(CC) -nostdlib -nostartfiles -static -o $@
-T ldscript.ld crt0.o init.o"
+               action  "$(CC) $(BUILDID_OPTION) -nostdlib
-nostartfiles -static -o $@ -T ldscript.ld crt0.o init.o"
                action  "$(CROSS_COMPILE)nm -n linuxbios | sort > linuxbios.map"
         end

Then couldn't the build pick up an environment variable?

Still thinking about how to fix this. I don't think what I've seen so
far will work.

ron

[Carl-Daniel Hailfinger]

On 16.12.2007 02:39, ron minnich wrote:
> OK, I am looking at this patch again, and it's still not quite the
> right idea. Setting this kind of build option in the Config file is
> not really right.
>
> However ...
>
> if you just retained this much:
> Index: src/arch/i386/Config.lb
> ===================================================================
> --- src/arch/i386/Config.lb     (revision 3003)
> +++ src/arch/i386/Config.lb     (working copy)
> @@ -100,7 +100,7 @@
>      makerule linuxbios
>             depends "crt0.o init.o $(LINUXBIOS_APC)
>
> $(LINUXBIOS_RAM_ROM) ldscript.ld"
>           action  "$(CC) -nostdlib -nostartfiles -static -o $@ -T ldscript.ld crt0.o init.o"
>           action  "$(CC) $(BUILDID_OPTION) -nostdlib -nostartfiles -static -o $@ -T ldscript.ld crt0.o init.o"
>           action  "$(CROSS_COMPILE)nm -n linuxbios | sort > linuxbios.map"
>    end
>
>
> Then couldn't the build pick up an environment variable?
>
> Still thinking about how to fix this. I don't think what I've seen so
> far will work.

Why not simply compile a small test.c file, look whether the output file
contains buildid, and set a Make variable based on that?

Regards,
Carl-Daniel

[ron minnich]

On Dec 18, 2007 12:51 PM, Carl-Daniel Hailfinger
 wrote:
> Why not simply compile a small test.c file, look whether the output file
> contains buildid, and set a Make variable based on that?

I'm not picky about how we figure it out, only that the test be run
out of buildtarget, and that it set variables in a makefile that is
included.

short form: no running complex tests out of the build process each
time, and for sure no setting things Options.lb.

Thanks

ron

[Marc Karasek]

How about my awk script I sent out that uses the output from ld --help 
to see if buildid is and option.  This handles the different distro  
problem.

If I can get some time,  I will put together a patch file with the 
changes to buildtarget to create a makefile.distro using awk to test ld 
for the buildid option.

/*********************
Marc Karasek
MTS
Sun Microsystems
mailto:[e-mail]
ph:770.360.6415
*********************/

ron minnich wrote:
> On Dec 18, 2007 12:51 PM, Carl-Daniel Hailfinger
>  wrote:
> > Why not simply compile a small test.c file, look whether the output file
> > contains buildid, and set a Make variable based on that?
>
> I'm not picky about how we figure it out, only that the test be run
> out of buildtarget, and that it set variables in a makefile that is
> included.
>
> short form: no running complex tests out of the build process each
> time, and for sure no setting things Options.lb.
>
> Thanks
>
> ron

[ron minnich]

On Dec 18, 2007 12:58 PM, Marc Karasek  wrote:
> If I can get some time,  I will put together a patch file with the
> changes to buildtarget to create a makefile.distro using awk to test ld
> for the buildid option.

perfect.

Thanks

ron

[Peter Stuge]

On Tue, Dec 18, 2007 at 03:58:48PM -0500, Marc Karasek wrote:
> How about my awk script I sent out that uses the output from ld
> --help to see if buildid is and option.

The awk script behavior may be a problem, depending on how it is
called from buildtarget. See my other message with the simple grep
suggestion.

A patch would be great! :)

//Peter

[ron minnich]

On Dec 18, 2007 4:11 PM, Peter Stuge  wrote:
> The awk script behavior may be a problem, depending on how it is
> called from buildtarget. See my other message with the simple grep
> suggestion.

no problem.
> A patch would be great! :)

yes, from there we can figure it out.

ron

[Stefan Reinauer]

Uwe Hermann schrieb:
> On Fri, Dec 07, 2007 at 12:43:57PM -0500, Marc Karasek wrote:
> > Here it is in -u format...
>
> Thanks, but I think your Thunderbird broke the patch, probably because
> you copy-pasted it into the mailer and it clipped long lines.
>
> I recommend to use 'svn diff > foo.patch' in your svn tree (easier IMO,
> and that already produces patches in the -u format) and then just attach
> that file to the mail.
> > =================================================================
> > diff -r -u -x .svn LinuxBIOSv2.orig/src/arch/i386/Config.lb 
> > LinuxBIOSv2/src/arch/i386/Config.lb
> > --- LinuxBIOSv2.orig/src/arch/i386/Config.lb 2007-12-06 
> > 10:06:31.000000000 -0500
> > +++ LinuxBIOSv2/src/arch/i386/Config.lb 2007-12-04 16:25:13.000000000 -0500
> > @@ -100,7 +100,7 @@
> >
> > makerule linuxbios
> > depends "crt0.o init.o $(LINUXBIOS_APC) $(LINUXBIOS_RAM_ROM) ldscript.ld"
> > action "$(CC) -nostdlib -nostartfiles -static -o $@ -T ldscript.ld
> >
> > crt0.o init.o"
> > action "$(CC) -Wl,--build-id=none -nostdlib -nostartfiles -static -o
>
> We should make sure this doesn't break any other distros / compiler
> versions. Do all reasonably new gcc/ld versions know about '--build-id'?

that's why we started using util/xcompile.

[Marc Karasek]

Is the xcompile util a V3 only util?  I did not see it in the V2 tree.

If this is so, then it will not help us, unless we add it to V2 as V3 
compiles fine, the only broken build currently is the V2 build.

/*********************
Marc Karasek
MTS
Sun Microsystems
mailto:[e-mail]
ph:770.360.6415
*********************/

Stefan Reinauer wrote:
> Uwe Hermann schrieb:
> > On Fri, Dec 07, 2007 at 12:43:57PM -0500, Marc Karasek wrote:
> > > Here it is in -u format...
> >
> > Thanks, but I think your Thunderbird broke the patch, probably because
> > you copy-pasted it into the mailer and it clipped long lines.
> >
> > I recommend to use 'svn diff > foo.patch' in your svn tree (easier IMO,
> > and that already produces patches in the -u format) and then just attach
> > that file to the mail.
> > > =================================================================
> > > diff -r -u -x .svn LinuxBIOSv2.orig/src/arch/i386/Config.lb 
> > > LinuxBIOSv2/src/arch/i386/Config.lb
> > > --- LinuxBIOSv2.orig/src/arch/i386/Config.lb 2007-12-06 
> > > 10:06:31.000000000 -0500
> > > +++ LinuxBIOSv2/src/arch/i386/Config.lb 2007-12-04 
> > > 16:25:13.000000000 -0500
> > > @@ -100,7 +100,7 @@
> > >
> > > makerule linuxbios
> > > depends "crt0.o init.o $(LINUXBIOS_APC) $(LINUXBIOS_RAM_ROM) 
> > > ldscript.ld"
> > > action "$(CC) -nostdlib -nostartfiles -static -o $@ -T ldscript.ld
> > >
> > > crt0.o init.o"
> > > action "$(CC) -Wl,--build-id=none -nostdlib -nostartfiles -static
> > >
> > > -o
> >
> > We should make sure this doesn't break any other distros / compiler
> > versions. Do all reasonably new gcc/ld versions know about '--build-id'?
>
> that's why we started using util/xcompile.

[Marc Karasek]

I have done the svn diff as you requested and attached the patch file.

This also removes the patch that Ed had submitted to first try to fix 
the --build-id option.

I saw some other posts about just adding the node-id section to the 
linker files.  This will increase the file size unless we also add the 
section into the flash part.

/*********************
Marc Karasek
MTS
Sun Microsystems
mailto:[e-mail]
ph:770.360.6415
*********************/

Uwe Hermann wrote:
> On Fri, Dec 07, 2007 at 12:43:57PM -0500, Marc Karasek wrote:
> > Here it is in -u format...
>
> Thanks, but I think your Thunderbird broke the patch, probably because
> you copy-pasted it into the mailer and it clipped long lines.
>
> I recommend to use 'svn diff > foo.patch' in your svn tree (easier IMO,
> and that already produces patches in the -u format) and then just attach
> that file to the mail.
> > =================================================================
> > diff -r -u -x .svn LinuxBIOSv2.orig/src/arch/i386/Config.lb 
> > LinuxBIOSv2/src/arch/i386/Config.lb
> > --- LinuxBIOSv2.orig/src/arch/i386/Config.lb 2007-12-06 
> > 10:06:31.000000000 -0500
> > +++ LinuxBIOSv2/src/arch/i386/Config.lb 2007-12-04 16:25:13.000000000 -0500
> > @@ -100,7 +100,7 @@
> >
> > makerule linuxbios
> > depends "crt0.o init.o $(LINUXBIOS_APC) $(LINUXBIOS_RAM_ROM) ldscript.ld"
> > action "$(CC) -nostdlib -nostartfiles -static -o $@ -T ldscript.ld
> >
> > crt0.o init.o"
> > action "$(CC) -Wl,--build-id=none -nostdlib -nostartfiles -static -o
>
> We should make sure this doesn't break any other distros / compiler
> versions. Do all reasonably new gcc/ld versions know about '--build-id'?
> > $@ -T ldscript.ld crt0.o init.o"
> > action "$(CROSS_COMPILE)nm -n linuxbios | sort > linuxbios.map"
> > end
> >
> > diff -r -u -x .svn LinuxBIOSv2.orig/src/arch/i386/lib/id.lds 
> > LinuxBIOSv2/src/arch/i386/lib/id.lds
> > --- LinuxBIOSv2.orig/src/arch/i386/lib/id.lds 2007-12-06 
> > 10:06:31.000000000 -0500
> > +++ LinuxBIOSv2/src/arch/i386/lib/id.lds 2007-12-06 10:36:31.000000000 -0500
> > @@ -1,5 +1,11 @@
> > SECTIONS {
> > . = (_ROMBASE + ROM_IMAGE_SIZE - 0x10) - (__id_end - __id_start);
> >
> > +/*
> > . = (_ROMBASE + (ROM_IMAGE_SIZE - 0x10)) - (__id_end - __id_start);
> > This is a temporary fix. Under Fedora 8 ld does not like it if the
> >
> > .id section is above address 0xFFFF_EF00
> > +*/
> > _ROMTEMP = 0xffffef00;
> > . = _ROMTEMP;
>
> I cannot parse this changeset completely, but won't this
> break the code for other distros or boards?
>
> Please resend the 'svn diff' output so we can test this on some
> non-Fedora distros.
>
> Thanks, Uwe.