getting Linux working on IBM Walnut demo board

getting Linux working on IBM Walnut demo board

[Brian S. Park]

Hello all,

I'm new to embedded Linux and need help from all the gurus out there.

I'm trying to boot Linux on IBM Walnut demo board with PPC405GP.
I download ELDK, installed it on a RedHat Linx 7.3 machine.
I was able to build u-boot 3.0 which came with ELDK and it booted
successfully on my demo board.
However, when I tried to rebuild the kernel (2.4.4 which came with ELDK) it
gives me the following error.
I'm using the default configuration for Walnut board (i.e. make
rmproper;make walnut_config;make oldconfig;make dep; make pImage).

Also, the documentation on http://www.denx.de/doc/ seems a bit old (it
still talks about ppcboot instead u-boot). Are there new ocuments or FAQ's
for a newbie like me?

Any help will be appreciated.

Thank you.

ppc_4xx-gcc -D__KERNEL__
-I/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include -Wall
-Wstrict-prototypes -Os -fomit-frame-pointer -fno-strict-aliasing
-D__powerpc__ -fsigned-char -msoft-float -pipe -ffixed-r2
-Wno-uninitialized -mmultiple -mstring -mcpu=403 -g  -c -o init/main.o
init/main.c
In file included from
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/linux/raid/md.h:38,
                  from init/main.c:24:
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/io.h: In
function `virt_to_bus':
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/io.h:212:
`PCI_DRAM_OFFSET' undeclared (first use in this function)
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/io.h:212:
(Each undeclared identifier is reported only once
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/io.h:212:
for each function it appears in.)
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/io.h:216:
warning: control reaches end of non-void function
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/io.h: In
function `bus_to_virt':
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/io.h:223:
`PCI_DRAM_OFFSET' undeclared (first use in this function)
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/io.h:227:
warning: control reaches end of non-void function
In file included from
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/linux/bootmem.h:8,
                  from init/main.c:29:
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:
In function `enable_dma':
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:208:
`_IO_BASE' undeclared (first use in this function)
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:
In function `disable_dma':
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:224:
`_IO_BASE' undeclared (first use in this function)
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:
In function `clear_dma_ff':
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:239:
`_IO_BASE' undeclared (first use in this function)
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:
In function `set_dma_mode':
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:248:
`_IO_BASE' undeclared (first use in this function)
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:
In function `set_dma_page':
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:262:
`_IO_BASE' undeclared (first use in this function)
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:
In function `set_dma_addr':
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:308:
`_IO_BASE' undeclared (first use in this function)
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:
In function `set_dma_count':
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:336:
`_IO_BASE' undeclared (first use in this function)
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:
In function `get_dma_residue':
/mnt1/home/ELDK/ppc_4xx/usr/src/linux-2.4.4-2002-10-06/include/asm/dma.h:366:
`_IO_BASE' undeclared (first use in this function)
make: *** [init/main.o] Error 1


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: getting Linux working on IBM Walnut demo board

[Wolfgang Denk]

Dear Brian,

in message <5.1.1.6.2.20030527114004.01282240@corelis.com> you wrote:
>
> I'm trying to boot Linux on IBM Walnut demo board with PPC405GP.
> I download ELDK, installed it on a RedHat Linx 7.3 machine.
> I was able to build u-boot 3.0 which came with ELDK and it booted
> successfully on my demo board.

Fine.

> However, when I tried to rebuild the kernel (2.4.4 which came with ELDK) it
> gives me the following error.

This is a deficiency in the ELDK. For 4xx systems you need  a  recent
kernel.  I  recommend  the linuxppc_2_4_devel tree from the bitkeeper
archives, or from our CVS server.

> Also, the documentation on http://www.denx.de/doc/ seems a bit old (it
> still talks about ppcboot instead u-boot). Are there new ocuments or FAQ's
> for a newbie like me?

There aren't any significant differences between U-Boot  and  PPCboot
if  you're  working  on  PowerPC  systems.  We  are in the process of
setting up new documentation (as a wiki), but this  is  not  complete
yet.

If you have questions or problems, just ask, and we will help you.

Best regards,

Wolfgang Denk

--
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
Neckties strangle clear thinking.                       -- Lin Yutang

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: getting Linux working on IBM Walnut demo board

[Brian S. Park]

Eran,
Thank you very much. Now the kernel comes up and I can ping the target. :)

Now to the next problem. I'm booting with NFS root file system. Below is
the bootargs that I'm using.

 >bootargs=root=/dev/nfs rw nfsroot=192.168.1.78:/home/ELDK/ppc_4xx
console=ttyS0,115200
 >              console=tty0
ip=192.168.1.146:192.168.1.78:192.168.1.2:255.255.255.0:netusb::off

exports file on my development station has
 >/home/ELDK/ppc_4xx      netusb(rw,no_root_squash)

/home/ELDK/ppc_4xx is where ELDK installed the target file system on my
development machine.
Even though I can now ping the target, I can't telnet to it. I get the
following error.

 >telnet netusb
 >Trying 192.168.1.146...
 >Connected to netusb.
 >Escape character is '^]'.
 >telnetd: All network ports in use.
 >Connection closed by foreign host.

Is this normal? Do I now need to configure the target file system to enable
telnet?
Or is there still something wrong with my bootargs?

Also, after the kernel comes up, I can no longer type any command through
the serial terminal.
The following the message at the end of kernel boot. It says it was unable
to open the console.

 >IP-Config: Complete:
 >      device=eth0, addr=192.168.1.146, mask=255.255.255.0, gw=192.168.1.2,
 >     host=192.168.1.146, domain=, nis-domain=(none),
 >     bootserver=192.168.1.78, rootserver=192.168.1.78, rootpath=
 >NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
 >Looking up port of RPC 100003/2 on 192.168.1.78
 >Looking up port of RPC 100005/1 on 192.168.1.78
 >VFS: Mounted root (nfs filesystem).
 >Freeing unused kernel memory: 76k init
 >Warning: unable to open an initial console.

Any help will be appreciated.

Thank you

Brian

At 12:53 PM 5/28/2003 +0300, you wrote:
>Hello Brian,
>You can try what I suggested in:
>http://lists.linuxppc.org/linuxppc-embedded/200305/msg00222.html
>i.e. in the file arch/ppc/platforms/walnut.h to replace
>typedef struct board_info {
>...
>} bd_t;
>
>with:
>#include <asm/ppcboot.h>
>Hope that helps,
>         Eran.


===============================================================
Brian S. Park  brian@corelis.com  (562) 926-6727 x143
---------------------------------------------------------------
Everything we do helps our customers get to market
FASTER with HIGHER quality and LOWER cost
===============================================================


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

RE: getting Linux working on IBM Walnut demo board

[Kerl, John]

Brian:

You can get the "all network ports in use" error when
there do not exist /dev/ttyp* -- these are the pseudo-ttys
needed by telnet.  You can mknod them; but I would double-check
your /dev directory.  If you're using ELDK, /dev/ttyp* should
already be there.

The "Unable to open an initial console" message is in init/main.c,
and occurs when /dev/console can't be opened.  Again, this points
suspicion at your filesystem.

To peek at your RAM disk, you can do the following (as root):

# gunzip < ramdisk.image.gz > ramdisk.image
# mkdir foo
# mount -o loop ramdisk.image foo

and then cd into foo and ls, etc.  Don't forget
to then do:

# umount foo
# rmdir foo
# rm ramdisk.image



-----Original Message-----
From: Brian S. Park [mailto:brian@corelis.com]
Sent: Wednesday, May 28, 2003 11:53 AM
To: Eran Mann
Cc: linuxppc-embedded@lists.linuxppc.org
Subject: Re: getting Linux working on IBM Walnut demo board



Eran,
Thank you very much. Now the kernel comes up and I can ping the target. :)

Now to the next problem. I'm booting with NFS root file system. Below is
the bootargs that I'm using.

 >bootargs=root=/dev/nfs rw nfsroot=192.168.1.78:/home/ELDK/ppc_4xx
console=ttyS0,115200
 >              console=tty0
ip=192.168.1.146:192.168.1.78:192.168.1.2:255.255.255.0:netusb::off

exports file on my development station has
 >/home/ELDK/ppc_4xx      netusb(rw,no_root_squash)

/home/ELDK/ppc_4xx is where ELDK installed the target file system on my
development machine.
Even though I can now ping the target, I can't telnet to it. I get the
following error.

 >telnet netusb
 >Trying 192.168.1.146...
 >Connected to netusb.
 >Escape character is '^]'.
 >telnetd: All network ports in use.
 >Connection closed by foreign host.

Is this normal? Do I now need to configure the target file system to enable
telnet?
Or is there still something wrong with my bootargs?

Also, after the kernel comes up, I can no longer type any command through
the serial terminal.
The following the message at the end of kernel boot. It says it was unable
to open the console.

 >IP-Config: Complete:
 >      device=eth0, addr=192.168.1.146, mask=255.255.255.0, gw=192.168.1.2,
 >     host=192.168.1.146, domain=, nis-domain=(none),
 >     bootserver=192.168.1.78, rootserver=192.168.1.78, rootpath=
 >NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
 >Looking up port of RPC 100003/2 on 192.168.1.78
 >Looking up port of RPC 100005/1 on 192.168.1.78
 >VFS: Mounted root (nfs filesystem).
 >Freeing unused kernel memory: 76k init
 >Warning: unable to open an initial console.

Any help will be appreciated.

Thank you

Brian

At 12:53 PM 5/28/2003 +0300, you wrote:
>Hello Brian,
>You can try what I suggested in:
>http://lists.linuxppc.org/linuxppc-embedded/200305/msg00222.html
>i.e. in the file arch/ppc/platforms/walnut.h to replace
>typedef struct board_info {
>...
>} bd_t;
>
>with:
>#include <asm/ppcboot.h>
>Hope that helps,
>         Eran.


===============================================================
Brian S. Park  brian@corelis.com  (562) 926-6727 x143
---------------------------------------------------------------
Everything we do helps our customers get to market
FASTER with HIGHER quality and LOWER cost
===============================================================


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

RE: getting Linux working on IBM Walnut demo board

[Kerl, John]

Ooooops, not reading your message well enough.
Of course you're using NFS, not RAM disk.

You can do a couple things:

*	cd into your export directory and make sure
	stuff is really there -- this is the Linux
	host looking at itself.

*	A nice diagnostic is to use another Linux PC,
	mounting the first Linux host's export directory
	over NFS.  I.e. the second PC is the NFS client,
	rather than the board.

*	You can resort to printk's ... you might do the
	following to assist in this:
	-	cd to kernel directory
	-	make tags
	-	Set up your editor for tags, e.g. in vim,
		.vimrc has a 'set tags=./tags' line in it.
	-	vim init/main.c, find the function you want,
		then control-] to dive in, control-t to
		come back out.

-----Original Message-----
From: Kerl, John [mailto:John.Kerl@avnet.com]
Sent: Wednesday, May 28, 2003 11:57 AM
To: 'Brian S. Park'
Cc: 'linuxppc-embedded@lists.linuxppc.org'
Subject: RE: getting Linux working on IBM Walnut demo board



Brian:

You can get the "all network ports in use" error when
there do not exist /dev/ttyp* -- these are the pseudo-ttys
needed by telnet.  You can mknod them; but I would double-check
your /dev directory.  If you're using ELDK, /dev/ttyp* should
already be there.

The "Unable to open an initial console" message is in init/main.c,
and occurs when /dev/console can't be opened.  Again, this points
suspicion at your filesystem.

To peek at your RAM disk, you can do the following (as root):

# gunzip < ramdisk.image.gz > ramdisk.image
# mkdir foo
# mount -o loop ramdisk.image foo

and then cd into foo and ls, etc.  Don't forget
to then do:

# umount foo
# rmdir foo
# rm ramdisk.image



-----Original Message-----
From: Brian S. Park [mailto:brian@corelis.com]
Sent: Wednesday, May 28, 2003 11:53 AM
To: Eran Mann
Cc: linuxppc-embedded@lists.linuxppc.org
Subject: Re: getting Linux working on IBM Walnut demo board



Eran,
Thank you very much. Now the kernel comes up and I can ping the target. :)

Now to the next problem. I'm booting with NFS root file system. Below is
the bootargs that I'm using.

 >bootargs=root=/dev/nfs rw nfsroot=192.168.1.78:/home/ELDK/ppc_4xx
console=ttyS0,115200
 >              console=tty0
ip=192.168.1.146:192.168.1.78:192.168.1.2:255.255.255.0:netusb::off

exports file on my development station has
 >/home/ELDK/ppc_4xx      netusb(rw,no_root_squash)

/home/ELDK/ppc_4xx is where ELDK installed the target file system on my
development machine.
Even though I can now ping the target, I can't telnet to it. I get the
following error.

 >telnet netusb
 >Trying 192.168.1.146...
 >Connected to netusb.
 >Escape character is '^]'.
 >telnetd: All network ports in use.
 >Connection closed by foreign host.

Is this normal? Do I now need to configure the target file system to enable
telnet?
Or is there still something wrong with my bootargs?

Also, after the kernel comes up, I can no longer type any command through
the serial terminal.
The following the message at the end of kernel boot. It says it was unable
to open the console.

 >IP-Config: Complete:
 >      device=eth0, addr=192.168.1.146, mask=255.255.255.0, gw=192.168.1.2,
 >     host=192.168.1.146, domain=, nis-domain=(none),
 >     bootserver=192.168.1.78, rootserver=192.168.1.78, rootpath=
 >NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
 >Looking up port of RPC 100003/2 on 192.168.1.78
 >Looking up port of RPC 100005/1 on 192.168.1.78
 >VFS: Mounted root (nfs filesystem).
 >Freeing unused kernel memory: 76k init
 >Warning: unable to open an initial console.

Any help will be appreciated.

Thank you

Brian

At 12:53 PM 5/28/2003 +0300, you wrote:
>Hello Brian,
>You can try what I suggested in:
>http://lists.linuxppc.org/linuxppc-embedded/200305/msg00222.html
>i.e. in the file arch/ppc/platforms/walnut.h to replace
>typedef struct board_info {
>...
>} bd_t;
>
>with:
>#include <asm/ppcboot.h>
>Hope that helps,
>         Eran.


===============================================================
Brian S. Park  brian@corelis.com  (562) 926-6727 x143
---------------------------------------------------------------
Everything we do helps our customers get to market
FASTER with HIGHER quality and LOWER cost
===============================================================


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

RE: getting Linux working on IBM Walnut demo board

[Kerl, John]

Also, 'set tagstack' in your ~/.vimrc.

-----Original Message-----
From: Kerl, John
Sent: Wednesday, May 28, 2003 12:02 PM
To: Kerl, John; 'Brian S. Park'
Cc: 'linuxppc-embedded@lists.linuxppc.org'
Subject: RE: getting Linux working on IBM Walnut demo board


Ooooops, not reading your message well enough.
Of course you're using NFS, not RAM disk.

You can do a couple things:

*	cd into your export directory and make sure
	stuff is really there -- this is the Linux
	host looking at itself.

*	A nice diagnostic is to use another Linux PC,
	mounting the first Linux host's export directory
	over NFS.  I.e. the second PC is the NFS client,
	rather than the board.

*	You can resort to printk's ... you might do the
	following to assist in this:
	-	cd to kernel directory
	-	make tags
	-	Set up your editor for tags, e.g. in vim,
		.vimrc has a 'set tags=./tags' line in it.
	-	vim init/main.c, find the function you want,
		then control-] to dive in, control-t to
		come back out.

-----Original Message-----
From: Kerl, John [mailto:John.Kerl@avnet.com]
Sent: Wednesday, May 28, 2003 11:57 AM
To: 'Brian S. Park'
Cc: 'linuxppc-embedded@lists.linuxppc.org'
Subject: RE: getting Linux working on IBM Walnut demo board



Brian:

You can get the "all network ports in use" error when
there do not exist /dev/ttyp* -- these are the pseudo-ttys
needed by telnet.  You can mknod them; but I would double-check
your /dev directory.  If you're using ELDK, /dev/ttyp* should
already be there.

The "Unable to open an initial console" message is in init/main.c,
and occurs when /dev/console can't be opened.  Again, this points
suspicion at your filesystem.

To peek at your RAM disk, you can do the following (as root):

# gunzip < ramdisk.image.gz > ramdisk.image
# mkdir foo
# mount -o loop ramdisk.image foo

and then cd into foo and ls, etc.  Don't forget
to then do:

# umount foo
# rmdir foo
# rm ramdisk.image



-----Original Message-----
From: Brian S. Park [mailto:brian@corelis.com]
Sent: Wednesday, May 28, 2003 11:53 AM
To: Eran Mann
Cc: linuxppc-embedded@lists.linuxppc.org
Subject: Re: getting Linux working on IBM Walnut demo board



Eran,
Thank you very much. Now the kernel comes up and I can ping the target. :)

Now to the next problem. I'm booting with NFS root file system. Below is
the bootargs that I'm using.

 >bootargs=root=/dev/nfs rw nfsroot=192.168.1.78:/home/ELDK/ppc_4xx
console=ttyS0,115200
 >              console=tty0
ip=192.168.1.146:192.168.1.78:192.168.1.2:255.255.255.0:netusb::off

exports file on my development station has
 >/home/ELDK/ppc_4xx      netusb(rw,no_root_squash)

/home/ELDK/ppc_4xx is where ELDK installed the target file system on my
development machine.
Even though I can now ping the target, I can't telnet to it. I get the
following error.

 >telnet netusb
 >Trying 192.168.1.146...
 >Connected to netusb.
 >Escape character is '^]'.
 >telnetd: All network ports in use.
 >Connection closed by foreign host.

Is this normal? Do I now need to configure the target file system to enable
telnet?
Or is there still something wrong with my bootargs?

Also, after the kernel comes up, I can no longer type any command through
the serial terminal.
The following the message at the end of kernel boot. It says it was unable
to open the console.

 >IP-Config: Complete:
 >      device=eth0, addr=192.168.1.146, mask=255.255.255.0, gw=192.168.1.2,
 >     host=192.168.1.146, domain=, nis-domain=(none),
 >     bootserver=192.168.1.78, rootserver=192.168.1.78, rootpath=
 >NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
 >Looking up port of RPC 100003/2 on 192.168.1.78
 >Looking up port of RPC 100005/1 on 192.168.1.78
 >VFS: Mounted root (nfs filesystem).
 >Freeing unused kernel memory: 76k init
 >Warning: unable to open an initial console.

Any help will be appreciated.

Thank you

Brian

At 12:53 PM 5/28/2003 +0300, you wrote:
>Hello Brian,
>You can try what I suggested in:
>http://lists.linuxppc.org/linuxppc-embedded/200305/msg00222.html
>i.e. in the file arch/ppc/platforms/walnut.h to replace
>typedef struct board_info {
>...
>} bd_t;
>
>with:
>#include <asm/ppcboot.h>
>Hope that helps,
>         Eran.


===============================================================
Brian S. Park  brian@corelis.com  (562) 926-6727 x143
---------------------------------------------------------------
Everything we do helps our customers get to market
FASTER with HIGHER quality and LOWER cost
===============================================================


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: getting Linux working on IBM Walnut demo board

[Eran Mann]

Brian, see some suggestions below.

Brian S. Park wrote:

> Now to the next problem. I'm booting with NFS root file system. Below
> is the bootargs that I'm using.
>
> >bootargs=root=/dev/nfs rw nfsroot=192.168.1.78:/home/ELDK/ppc_4xx
> console=ttyS0,115200
> >              console=tty0
> ip=192.168.1.146:192.168.1.78:192.168.1.2:255.255.255.0:netusb::off
>
> exports file on my development station has
> >/home/ELDK/ppc_4xx      netusb(rw,no_root_squash)
>
> /home/ELDK/ppc_4xx is where ELDK installed the target file system on
> my development machine.
> Even though I can now ping the target, I can't telnet to it. I get the
> following error.
>
> >telnet netusb
> >Trying 192.168.1.146...
> >Connected to netusb.
> >Escape character is '^]'.
> >telnetd: All network ports in use.
> >Connection closed by foreign host.
>
> Is this normal? Do I now need to configure the target file system to
> enable telnet?
> Or is there still something wrong with my bootargs?

It's not normal. Perhaps it's a problem with inetd or xinetd
configuration? is there any more information in the logs or in dmesg
(well, once you get a working console..)?

> Also, after the kernel comes up, I can no longer type any command
> through the serial terminal.
> The following the message at the end of kernel boot. It says it was
> unable to open the console.
>
> >IP-Config: Complete:
> >      device=eth0, addr=192.168.1.146, mask=255.255.255.0,
> gw=192.168.1.2,
> >     host=192.168.1.146, domain=, nis-domain=(none),
> >     bootserver=192.168.1.78, rootserver=192.168.1.78, rootpath=
> >NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
> >Looking up port of RPC 100003/2 on 192.168.1.78
> >Looking up port of RPC 100005/1 on 192.168.1.78
> >VFS: Mounted root (nfs filesystem).
> >Freeing unused kernel memory: 76k init
> >Warning: unable to open an initial console.

Perhaps you dont have a proper /dev/ttyS0 device in the nfsroot?

>
> Any help will be appreciated.
>
> Thank you
>
> Brian
> ===============================================================
> Brian S. Park  brian@corelis.com  (562) 926-6727 x143
> ---------------------------------------------------------------
> Everything we do helps our customers get to market
> FASTER with HIGHER quality and LOWER cost
> ===============================================================


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: getting Linux working on IBM Walnut demo board

[Brian S. Park]

Thank you all for good suggestions.

It turned out that ELDK default setup does not have any dev files. How do I
create all the dev files I need?

 >[root@karla2 dev]# ls
 >initctl  log  null

Also, I have a general question about Embedded Linux.
What is the reasonable amount of flash to have for a minimally configured
Embedded Linux system?

My target will be running small (<<1MB) program which will take commands
over TCP/IP network.
So, I want to know how much flash memory we should put on our custom target.
We do not need support for PCI/Video/Sound/HD/SCSI/Parallel port/...
All we need from the OS is network support. Probably using Linux is an
overkill but since it's free. :)

Thank you for any input.

Brian


Is this normal? Do I now need to configure the target file system to enable
telnet?
>>Or is there still something wrong with my bootargs?
>
>It's not normal. Perhaps it's a problem with inetd or xinetd
>configuration? is there any more information in the logs or in dmesg
>(well, once you get a working console..)?
>
>>Also, after the kernel comes up, I can no longer type any command through
>>the serial terminal.
>>The following the message at the end of kernel boot. It says it was
>>unable to open the console.
>>
>> >IP-Config: Complete:
>> >      device=eth0, addr=192.168.1.146, mask=255.255.255.0, gw=192.168.1.2,
>> >     host=192.168.1.146, domain=, nis-domain=(none),
>> >     bootserver=192.168.1.78, rootserver=192.168.1.78, rootpath=
>> >NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
>> >Looking up port of RPC 100003/2 on 192.168.1.78
>> >Looking up port of RPC 100005/1 on 192.168.1.78
>> >VFS: Mounted root (nfs filesystem).
>> >Freeing unused kernel memory: 76k init
>> >Warning: unable to open an initial console.
>
>Perhaps you dont have a proper /dev/ttyS0 device in the nfsroot?

===============================================================
Brian S. Park  brian@corelis.com  (562) 926-6727 x143
---------------------------------------------------------------
Everything we do helps our customers get to market
FASTER with HIGHER quality and LOWER cost
===============================================================


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

RE: getting Linux working on IBM Walnut demo board

[Kerl, John]

Brian:

Re flash capacity, there was a thread on this some months ago.
There are those who pride themselves on having the most compact
system possible, and who will tell you 2MB is "bloated".  There
are those who prefer to use a standard C library, and who think
5 MB is OK.  So it depends on how much you want to crunch things
down ...

However, I will say that an 8MB flash device should be plenty
for compressed kernel, compressed RAM disk which contains a nice
set of libraries + applications, firmware, etc.  And you might
be able to get by with 4MB.

We use a pair of StrataFlash parts on most all of our boards
-- these are 8MB, 16 bits wide; we usually use a pair of them
(for a 32-bit bus) which gives us 16MB of flash.  And that is
more than plenty.  All my firmware (several different tools)
fit comfortably in the last 1 MB.  I generally put some extras
in my RAM disk, e.g. some largish PDF files (product briefs)
so that a board can run a web server, dishing up some marketing
literature, etc, and I am not one of the people who wants the
tiniest C library, so I am one of the more profligate users.
And my zvmlinux.initrd is generally in the 3-5MB range.  That
leaves space in our 16MB flash for 2-3 of those zvmlinux.initrd's.

-----Original Message-----
From: Brian S. Park [mailto:brian@corelis.com]
Sent: Wednesday, May 28, 2003 1:53 PM
Subject: Re: getting Linux working on IBM Walnut demo board

Thank you all for good suggestions.

It turned out that ELDK default setup does not have any dev files. How do I
create all the dev files I need?

 >[root@karla2 dev]# ls
 >initctl  log  null

Also, I have a general question about Embedded Linux. What is the
reasonable amount of flash to have for a minimally configured Embedded
Linux system?

My target will be running small (<<1MB) program which will take
commands over TCP/IP network. So, I want to know how much flash
memory we should put on our custom target. We do not need support for
PCI/Video/Sound/HD/SCSI/Parallel port/... All we need from the OS is
network support. Probably using Linux is an overkill but since it's
free. :)

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: getting Linux working on IBM Walnut demo board

[Wolfgang Denk]

Dear Brian,

in message <5.1.1.6.2.20030528132553.0128a2f0@corelis.com> you wrote:
>
> It turned out that ELDK default setup does not have any dev files. How do I
> create all the dev files I need?

Well, I recommend you start reading the documentation now.

There is a section "7. Mounting  Target  Components  via  NFS"  which
explains  that  you  have  to  run the ELDK_MAKEDEV and ELDK_FIXOWNER
scripts.

I bet a case of beer that you didn't do that yet.

> What is the reasonable amount of flash to have for a minimally configured
> Embedded Linux system?

Depends on what you want to squeeze in. I have seen systems with just
128 kB. Obviously this is just enough for the bootloader to boot  the
kernel  over network. There are products on the market like Bluetooth
LAN Access Points or ISDN routers which have just 2 MB flash. With  4
MB  flash  you  can  set  up  a powerful system including things like
network access, web server, and Microwindows  or  Qt  embedded  based
GUIs.

I have not seen a system with less than 4 MB flash lately.

> My target will be running small (<<1MB) program which will take commands
> over TCP/IP network.
> So, I want to know how much flash memory we should put on our custom target.

If thisi s all you need (just one binary) you can plan for some 600kB
for the compressed kernel image plus some 800kB for  the  application
image.  2  MB  flash  will do for such a setup - but please note that
software tendsa to fill available space, and  I  bet  you  will  have
ideas for things to add once the box is running.

> All we need from the OS is network support. Probably using Linux is an
> overkill but since it's free. :)

Happy hacking.


Wolfgang Denk

--
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
It all seemed, he thought, to be rather a lot of  trouble  to  go  to
just sharpen a razor blade.  - Terry Pratchett, _The Light Fantastic_

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: getting Linux working on IBM Walnut demo board

[Brian S. Park]

Thank you Wolfgang,
That was it. I forgot to run those script.
After running the script, the console works and I can telnet without any
problem.

Now I can start hacking and get it customized for our custom target.

Thank you all for helping me get it up and running.

Brian


At 11:34 PM 5/28/2003 +0200, Wolfgang Denk wrote:
>Dear Brian,
>
>in message <5.1.1.6.2.20030528132553.0128a2f0@corelis.com> you wrote:
> >
> > It turned out that ELDK default setup does not have any dev files. How do I
> > create all the dev files I need?
>
>Well, I recommend you start reading the documentation now.
>
>There is a section "7. Mounting  Target  Components  via  NFS"  which
>explains  that  you  have  to  run the ELDK_MAKEDEV and ELDK_FIXOWNER
>scripts.
>

===============================================================
Brian S. Park  brian@corelis.com  (562) 926-6727 x143
---------------------------------------------------------------
Everything we do helps our customers get to market
FASTER with HIGHER quality and LOWER cost
===============================================================


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: getting Linux working on IBM Walnut demo board

[Brian S. Park]

Hello again,
I need more help!
Some time, when u-Boot is trying to download the linux kernel image over
tftp connection,
I get the following error. Any one have any idea what may be causing it?

Also, I was wondering if it's possible to make u-boot image that will work
with 256KB of ram/flash.
I tried changing TEXT_BASE to 0xFFFC0000 but I get an error during boot.
Something about the reserved memory is less than required. Is this possible?

Thank you.

Brian

ENET Speed is 100 Mbps - FULL duplex connection
ARP broadcast 1
TFTP from server 192.168.1.78; our IP address is 192.168.1.146
Filename 'pImage'.
Load address: 0x100000
Loading: ###########################################ERROR : Receive buffers are
full!

MAL error occured.... ISR = c0100010 UIC = = 220000  MAL_DEF = 2e0000  MAL_ERR=
260000

MAL error occured.... ISR = c0100010 UIC = = 220000  MAL_DEF = 2e0000  MAL_ERR=
260000

MAL error occured.... ISR = c0100010 UIC = = 220000  MAL_DEF = 2e0000  MAL_ERR=
260000
T T T T T


===============================================================
Brian S. Park  brian@corelis.com  (562) 926-6727 x143
---------------------------------------------------------------
Everything we do helps our customers get to market
FASTER with HIGHER quality and LOWER cost
===============================================================


** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/

Re: getting Linux working on IBM Walnut demo board

[Wolfgang Denk]

In message <5.1.1.6.2.20030529131257.01276478@corelis.com> you wrote:
>
> I need more help!
> Some time, when u-Boot is trying to download the linux kernel image over

I recommend to send such messages to the u-boot-users  mailing  list.
They are off topic here.

> Also, I was wondering if it's possible to make u-boot image that will work
> with 256KB of ram/flash.

You mean your board has just 256 kB RAM? Really??

Well, the U-Boot image can be made to fit easily in 128 kB  or  less;
you  will have to configure for a smaller malloc area. The U-Boot can
probably run in 256 kB RAM - but what for?  There  will  not  be  any
memory left for U-Boot's main purpose: to load an OS.

Best regards,

Wolfgang Denk

--
Software Engineering:  Embedded and Realtime Systems,  Embedded Linux
Phone: (+49)-8142-4596-87  Fax: (+49)-8142-4596-88  Email: wd@denx.de
"Obviously, a major malfunction has occurred."
              -- Steve Nesbitt, voice of Mission Control, January 28,
                 1986, as the shuttle Challenger exploded within view
                 of the grandstands.

** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/