Diego A. Fons | 2 May 14:39 2007

Problems with zlib

Hi,

I'm trying to build tinyX using buildroot snapshot 18532 with a external 
toolchain, although i was able to build zlib, it seems that installation 
step is wrong because when building tinyX i get this message:

/home/dfons/toolchain-nihilisme/arm-unknown-linux-gnu/bin/arm-unknown-linux-gnu-gcc 
-o Xfbdev -Os      -L../../exports/lib -L/usr/X11R6/lib          
dix/libdix.a os/libos.a -lXau -lXdmcp  hw/kdrive/fbdev/libfbdev.a 
miext/layer/liblayer.a hw/kdrive/libkdrive.a hw/kdrive/linux/liblinux.a  
miext/shadow/libshadow.a render/librender.a  fb/libfb.a dix/libxpstubs.a 
mi/libmi.a Xext/libext.a                            render/librender.a  
-L/home/dfons/buildroot_18532/build_arm_nofpu/staging_dir/arm-linux-uclibc/lib    
../../lib/font/libXfont.a dix/libxpstubs.a -lz
-lm
/home/dfons/toolchain-nihilisme/arm-unknown-linux-gnu/bin/../lib/gcc/arm-unknown-linux-gnu/4.1.0/../../../../arm-unknown-linux-gnu/bin/ld: 
cannot find -lz
collect2: ld returned 1 exit status
make[5]: *** [Xfbdev] Error 1
make[5]: Target `all' not remade because of errors.

It is not the headers bug which was described in previous mail, here ld 
can`t locate the zlib library. Is this another bug? how can i fix it?

I attach the .config file.

Regards,
Diego A. Fons.
(Continue reading)

Ken Cecka | 2 May 18:21 2007

wrong address when linking simple ARM test application

Hi,

May need to go to a GCC list for this, but I thought I'd start here.  I've
constructed a very minimal standalone ARM test application that I am trying
to load into SRAM on an ARM1136JFS target.  The SRAM is at address
0x1fffc000, and I am trying to link my elf file at that address, but when I
go to download it, the debugger tries to write to address 0x1fff8000.

When I look at the elf file using objdump, I see that all the sections are
at 0x1fffc000, but there's a program header at 0x1fff8000.  Why does the
header have this address, and how can I fix it?

Source files and objdump output are copied below.

Thanks,
Ken

---- Begin Makefile ----

TOOLCHAIN=/home/kcecka/projects/zasfiles/buildroot/build_arm/staging_dir
PREFIX=arm-linux-
AS=$(TOOLCHAIN)/bin/$(PREFIX)as
CC=$(TOOLCHAIN)/bin/$(PREFIX)gcc
CXX=$(TOOLCHAIN)/bin/$(PREFIX)g++
LD=$(TOOLCHAIN)/bin/$(PREFIX)ld
OBJDUMP=$(TOOLCHAIN)/bin/$(PREFIX)objdump
STRIP=$(TOOLCHAIN)/bin/$(PREFIX)strip

ASFLAGS=-gdwarf-2
CFLAGS=-gdwarf-2
(Continue reading)

jacmet | 3 May 10:25 2007

svn commit: trunk/buildroot: package/fakeroot target/cpio target/cram etc...

Author: jacmet
Date: 2007-05-03 01:25:17 -0700 (Thu, 03 May 2007)
New Revision: 18549

Log:
s/FAKEROUTE/FAKEROOT/

Modified:
   trunk/buildroot/package/fakeroot/Config.in
   trunk/buildroot/package/fakeroot/fakeroot.mk
   trunk/buildroot/target/cpio/Config.in
   trunk/buildroot/target/cramfs/Config.in
   trunk/buildroot/target/ext2/Config.in
   trunk/buildroot/target/iso9660/Config.in
   trunk/buildroot/target/jffs2/Config.in
   trunk/buildroot/target/squashfs/Config.in
   trunk/buildroot/target/tar/Config.in

Changeset:
Modified: trunk/buildroot/package/fakeroot/Config.in
===================================================================
--- trunk/buildroot/package/fakeroot/Config.in	2007-05-02 23:39:04 UTC (rev 18548)
+++ trunk/buildroot/package/fakeroot/Config.in	2007-05-03 08:25:17 UTC (rev 18549)
 <at>  <at>  -7,6 +7,6  <at>  <at> 
 	  http://joostje.op.het.net/fakeroot/

 # dummy for fakeroot-source
-config BR2_HOST_FAKEROUTE
+config BR2_HOST_FAKEROOT
 	bool
(Continue reading)

Konstantin Kletschke | 3 May 10:35 2007
Picon

Re: gcc-4.1.2 eabi softfloat toolchain, undefined reference to "raise"

Am 2007-04-30 14:57 +0200 schrieb Bernhard Fischer:

> This is a known bug with gcc. There is a bugfix (that i don't really
> understand and consequently don't know if it is really accurate)
> floating around for this one.

Well I searched like mad but I did not find one.
I digged around and added an empty raise() function and my bootloader
compiled after that and the binary runs fine so far. Hmm...

> If in doubt, please file a PR in http://gcc.gnu.org/bugzilla/ about this bug
> and perhaps add a pointer to the patch if this bug is not yet entered in
> bugzilla.

I will enter it. 

Regards, Konsti

--

-- 
synertronixx GmbH
Dipl.-Ing. (FH) Konstantin Kletschke
Vahrenwalder Str. 7
30165 Hannover
Tel.    : 0511 / 3574 - 661
Fax     : 0511 / 93 57 - 679
E-mail  : kletschke@...
Internet: http://www.synertronixx.de
Vertretungsberechtigte Geschäftsführer:
Marc-Oliver Borck, Dr. Matthias Blaschke
Registergericht: Amtsgericht Hannover
(Continue reading)

jacmet | 3 May 11:33 2007

svn commit: trunk/buildroot/package/sqlite

Author: jacmet
Date: 2007-05-03 02:33:24 -0700 (Thu, 03 May 2007)
New Revision: 18550

Log:
version bump

Modified:
   trunk/buildroot/package/sqlite/sqlite.mk

Changeset:
Modified: trunk/buildroot/package/sqlite/sqlite.mk
===================================================================
--- trunk/buildroot/package/sqlite/sqlite.mk	2007-05-03 08:25:17 UTC (rev 18549)
+++ trunk/buildroot/package/sqlite/sqlite.mk	2007-05-03 09:33:24 UTC (rev 18550)
 <at>  <at>  -4,7 +4,7  <at>  <at> 
 #
 #############################################################

-SQLITE_VERSION=3.3.16
+SQLITE_VERSION=3.3.17
 SQLITE_SOURCE=sqlite-$(SQLITE_VERSION).tar.gz
 SQLITE_SITE=http://www.sqlite.org
 SQLITE_DIR=$(BUILD_DIR)/sqlite-$(SQLITE_VERSION)
Bernhard Fischer | 3 May 11:57 2007
Picon

Re: gcc-4.1.2 eabi softfloat toolchain, undefined reference to "raise"

On Thu, May 03, 2007 at 10:35:52AM +0200, Konstantin Kletschke wrote:
>Am 2007-04-30 14:57 +0200 schrieb Bernhard Fischer:
>
>> This is a known bug with gcc. There is a bugfix (that i don't really
>> understand and consequently don't know if it is really accurate)
>> floating around for this one.
>
>Well I searched like mad but I did not find one.

A quick google for these keywords:
__div0 raise arm softfloat

would have turned up said patch, fwiw.

>I digged around and added an empty raise() function and my bootloader
>compiled after that and the binary runs fine so far. Hmm...
>
>> If in doubt, please file a PR in http://gcc.gnu.org/bugzilla/ about this bug
>> and perhaps add a pointer to the patch if this bug is not yet entered in
>> bugzilla.
>
>I will enter it. 

TIA,
Diego A. Fons | 3 May 16:48 2007

No rule to make target

Hi, it's me again...

Once i built tinyX using buildroot and an external toolchain i tried to 
build libgtk2 but i get this error:

make: *** No rule to make target `tinyx', needed by `cairo'.  Stop.

It seems that tinyX can't be built but i've already built tinyX in a 
previous make. Does anyone know how to fix it?

Thanks and regards,
Diego A. Fons.
Ryan Ordway | 4 May 00:43 2007
Picon

Problem with buildroot on ARM


I'm having some issues with a rootfs image built through buildroot for an
ARM board. Here is some basic info:

Board: Atmel AT91SAM9260-EK
CPU: ARM926EJ-S rev 5 (v5l)
Kernel: 2.6.21.1 (latest stable)

I'm using the latest buildroot from subversion, r18550.

All is well until after the ramdisk is loaded, then I get a kernel panic.
Any ideas what might be causing this?

Uncompressing 
Linux.............................................................
................... done, booting the kernel.
Linux version 2.6.21.1-pml1 (rordway@...) (gcc version
4.1.2) #2
 Thu May 3 13:47:43 PDT 2007
CPU: ARM926EJ-S [41069265] revision 5 (ARMv5TEJ), cr=00053177
Machine: Atmel AT91SAM9260-EK
Memory policy: ECC disabled, Data cache writeback
Clocks: CPU 198 MHz, master 99 MHz, main 18.432 MHz
CPU0: D VIVT write-back cache
CPU0: I cache: 8192 bytes, associativity 4, 32 byte lines, 64 sets
CPU0: D cache: 8192 bytes, associativity 4, 32 byte lines, 64 sets
Built 1 zonelists.  Total pages: 16256
Kernel command line: ram=64M console=ttyS0,115200 initrd=0x21000000
root=/dev/ram0 init=/linuxrc rw
AT91: 96 gpio irqs in 3 banks
(Continue reading)

Anoop Krishnan | 4 May 04:01 2007

U-boot 1.1.4 build error

Hi,
 
 
 
I have used crosstool-0.42 to build tool chain and it builds the toolchain without any problems
 
and I can even compile the kernel Image with the same cross-compiler perfectly
 
The probelm I am facing is that when I use the same toolchain to compile U-boot 1.1.4, I get
the following error
 
/opt/crosstool/gcc-4.1.0-glibc-2.4/arm-926ejs-linux-gnueabi/lib/gcc/arm-926ejs-linux-gnueabi/4.1.0/libgcc.a(_dvmd_lnx.o): In function `__div0':
/home/anoop/Tool_chain/crosstool-0.42/build/arm-926ejs-linux-gnueabi/gcc-4.1.0-glibc-2.4/gcc-4.1.0/gcc/config/arm/lib1funcs.asm:1000: undefined reference to `raise'
make: *** [u-boot] Error 1
 
I found that raise is defined in raise.c which is under the following path
/home/anoop/Tool_chain/crosstool-0.42/build/arm-926ejs-linux-gnueabi/gcc-4.1.0-glibc-2.4/glibc-2.4/signal/raise.c
 
and its contents are as follows
 
#include <signal.h>
#include <errno.h>
 
/* Raise the signal SIG.  */
int
raise (sig)
     int sig;
{
  __set_errno (ENOSYS);
  return -1;
}
weak_alias (raise, gsignal)
 
stub_warning (raise)
stub_warning (gsignal)
#include <stub-tag.h>
 

How can I solve this issue? Is there any patch for glibc-2.4 , applying which the issue can be solved?
Thanks in advance
Anoop
_______________________________________________
buildroot mailing list
buildroot@...
http://busybox.net/mailman/listinfo/buildroot
Konstantin Kletschke | 4 May 09:34 2007
Picon

Re: gcc-4.1.2 eabi softfloat toolchain, undefined reference to "raise"

Am 2007-05-03 11:57 +0200 schrieb Bernhard Fischer:

> A quick google for these keywords:
> __div0 raise arm softfloat

Well, one trace leads to an uClibc-gcc-41-raise-error.patch which yields
into a 404. But as I understand this "fixes" uClibc to compile but does
not change the toolchain itself.

Another trace yields into a discussion suggesting replacing u-boot's
__div(0) by a raise() stub. May be I did something like that with my
very ugly hack.

So did I understand right, that the libgcc calls raise() which is
provided by glibc not uClibc?

How comes only u-boot triggers that? I compiled many other stuff with
this toolchain, kernel and userspace leading into perfect working
binaries.

Kind regards, Konsti

--

-- 
synertronixx GmbH
Dipl.-Ing. (FH) Konstantin Kletschke
Vahrenwalder Str. 7
30165 Hannover
Tel.    : 0511 / 3574 - 661
Fax     : 0511 / 93 57 - 679
E-mail  : kletschke@...
Internet: http://www.synertronixx.de
Vertretungsberechtigte Geschäftsführer:
Marc-Oliver Borck, Dr. Matthias Blaschke
Registergericht: Amtsgericht Hannover
Registernummer: HRB 59670
USt-ID: DE 216 559 901

Gmane