Sure
I've cross compiled everything for arm-linux from a Debian 3.0 PC.
I'm using
linux-2.4.18-rmk7 + patches for our boards
binutils 2.13 + a patch to resolve a linking problem on arm
uClibc-cvs-Aug.26-morning + patch for uClibc_config_fix.pl to add locale
gcc 3.2
busybox 0.60.3 + patch for virtual console and serial console to both
work
+ patch for sys_errlist to strerror
uClibc was configured with
${UCLIBC_PATH}/extra/Configs/ --arch=${TARGET_CPU} ${CROSSSTUFF}
--devel_prefix=${TARGET_PATH} --kernel_dir=${KERNEL_PATH}
--large_file=true --rpc_support=true --c99_math=true --float=true
--threads=true --shadow=true --ldso_path=/lib --has_locale=true
--mmu=true --shared_support=true --file=${UCLIBC_PATH}/Config~ >
${UCLIBC_PATH}/Config;
to cross compile locales I cross compiled the gen_mmap program, ran it
on an arm machine, and then transfered the output back to the PC, and
continued compiling.
$ arm-linux-ldd bin/busybox
libc.so.0 => /home/david/techsol/arm-cross-v0.9/lib/libc.so.0
/lib/ld-uClibc.so.0 => /lib/ld-uClibc.so.0
BusyBox v0.60.3 (2002.08.27-00:09+0000) Built-in shell (ash)
Enter 'help' for a list of built-in commands.
# df -h
Filesystem Size Used Available Use% Mounted on
/dev/doc1 30.9M 7.2M 23.7M 23% /
/dev/doc1 30.9M 7.2M 23.7M 23% /
# df
Filesystem 1k-blocks Used Available Use% Mounted on
pc : [<40031d48>] lr : [<40031e04>] Not tainted
sp : bffffd54 ip : 40031d28 fp : 40000000
r10: 00086484 r9 : 00000000 r8 : 00000001
r7 : bffffd6c r6 : 00000000 r5 : 00000000 r4 : 00003dc5
r3 : 0000ffff r2 : 00000490 r1 : bffffd6c r0 : 00003dc5
Flags: NzCv IRQs on FIQs on Mode USER_32 Segment user
Control: 217D Table: 4033C015 DAC: 00000015
Segmentation fault
#
Okay, that's interesting. What does the -h option do? Oh, and the
reason the root file system is mounted twice is because /etc/mtab is
getting screwed up when I boot the root as read-only, and I don't get
around to updating it.
I hope that is enough
Dave
On Tue, 2002-08-27 at 15:25, Erik Andersen wrote:
>
On Tue Aug 27, 2002 at 01:21:10PM -0700, David Meggy wrote:
>
>
>
> I haven't had a chance to look into this, but because of Erik's request
>
> to test CVS I have tried rebuilding all my tools, and I noticed that
>
> busybox's df command just gives a seg fault.
>
>
>
> As I said I haven't looked into it, but it is something that should be
>
> looked into before uClibc is released.
>
>
Thanks for the testing. I just tried it and it works for me...
>
Can you provide some details? Arch? uClibc config? busybox
>
version?
>
>
$ ldd ./busybox
>
libc.so.0 => /usr/i386-linux-uclibc/lib/libc.so.0 (0x0x40006000)
>
ld-uClibc.so.0 => /usr/i386-linux-uclibc/lib/ld-uClibc.so.0
>
(0x0x40000000)
>
$ ./busybox df -h
>
Filesystem Size Used Available Use% Mounted on
>
/dev/hda2 75.0G 32.0G 39.9G 45% /
>
>
-Erik
>
>
--
>
Erik B. Andersen http://codepoet-consulting.com/
>
--This message was written using 73% post-consumer electrons--
--
~~~~~~~~~~~~~~~~~~~~~~~~
David Meggy
Engineering
Technical Solutions Inc.
Unit #1 7157 Honeyman St
Delta BC Canada, V4G 1E2
www.techsol.ca
eMail: dmeggy@xxxxxxxxxx
Tel: 604 946 TECH (8324)
Fax: 604 946 6445
~~~~~~~~~~~~~~~~~~~~~~~~
Thread at a glance:
Previous Message by Date:
click to view message preview
Re: my system hangs with new shared libraries
On Tue, Aug 27, 2002 at 10:14:07PM +0200, Tobias Stark wrote:
> init (1): undefined instruction: pc=40014b20
Helps to give the full error message.
I guess you're missing a floating point emulator in the kernel.
Configure nwfpe to be built-in.
Next Message by Date:
click to view message preview
Re: uclibc-cvs breaks busybox df command
On Tue Aug 27, 2002 at 03:38:14PM -0700, David Meggy wrote:
> uClibc-cvs-Aug.26-morning + patch for uClibc_config_fix.pl to add locale
Hmm. Maybe that would explain it. There was a bug yesterday
with struct stat being defined wrong when largefile support was
enabled. I just tested today's version of uClibc with busybox on
arm, and everything was working fine... So I think we're still
in good shape.
-Erik
--
Erik B. Andersen http://codepoet-consulting.com/
--This message was written using 73% post-consumer electrons--
Previous Message by Thread:
click to view message preview
Re: uclibc-cvs breaks busybox df command
On Tue Aug 27, 2002 at 01:21:10PM -0700, David Meggy wrote:
>
> I haven't had a chance to look into this, but because of Erik's request
> to test CVS I have tried rebuilding all my tools, and I noticed that
> busybox's df command just gives a seg fault.
>
> As I said I haven't looked into it, but it is something that should be
> looked into before uClibc is released.
Thanks for the testing. I just tried it and it works for me...
Can you provide some details? Arch? uClibc config? busybox
version?
$ ldd ./busybox
libc.so.0 => /usr/i386-linux-uclibc/lib/libc.so.0 (0x0x40006000)
ld-uClibc.so.0 => /usr/i386-linux-uclibc/lib/ld-uClibc.so.0
(0x0x40000000)
$ ./busybox df -h
Filesystem Size Used Available Use% Mounted on
/dev/hda2 75.0G 32.0G 39.9G 45% /
-Erik
--
Erik B. Andersen http://codepoet-consulting.com/
--This message was written using 73% post-consumer electrons--
Next Message by Thread:
click to view message preview
Re: uclibc-cvs breaks busybox df command
On Tue Aug 27, 2002 at 03:38:14PM -0700, David Meggy wrote:
> uClibc-cvs-Aug.26-morning + patch for uClibc_config_fix.pl to add locale
Hmm. Maybe that would explain it. There was a bug yesterday
with struct stat being defined wrong when largefile support was
enabled. I just tested today's version of uClibc with busybox on
arm, and everything was working fine... So I think we're still
in good shape.
-Erik
--
Erik B. Andersen http://codepoet-consulting.com/
--This message was written using 73% post-consumer electrons--