LinuxQuestions.org
Download your favorite Linux distribution at LQ ISO.
Go Back   LinuxQuestions.org > Forums > Linux Forums > Linux - Software > Linux - Kernel
User Name
Password
Linux - Kernel This forum is for all discussion relating to the Linux kernel.

Notices


Reply
  Search this Thread
Old 05-29-2008, 06:09 AM   #1
tilman1
Member
 
Registered: Mar 2007
Location: Stuttgart, Germany
Distribution: gentoo
Posts: 72

Rep: Reputation: 15
2.4.36.3: busybox's Init hangs at "Freeing unused kernel memory: "


Hello

I have a 2.4.36.3 kernel compiled for an embedded target (i386).
Init (busybox) however does not start up. Using a linux-2.6.24 kernel,
it starts up.

My hypothesis is currently that busybox's libc is not working properly as it is compiled on a host running linux-2.6.18-gentoo.

The library is statically linked to busybox. It should hence not matter on which host system the binary was compiled (as long as the processor architecture does not change). Libc however uses kernel system calls -- that makes me think it might matter.

Besides a solution, I would appreciate suggestion how to trace this down further.

Setup
=====
a) Kernel: 2.4.36.3

b) Rootfs: via nfs

c) Init: replaced by a hello world program (to exclude busybox as the culprit)
Code:
#include <stdio.h>

int main(int argc, char *argv)
{
  printf("Hello world!\n");
  sleep(999999999);
}
The executable is statically linked (using libc of the host system running linux 2.6.18-gentoo)

> file /var/nfs/vmic7750/usr/src/init
/var/nfs/vmic7750/usr/src/init: ELF 32-bit LSB executable, Intel 80386, version 1 (SYSV), for GNU/Linux 2.6.9, statically linked, for GNU/Linux 2.6.9, not stripped

d) Added debug output to init/main.c:run_init_process
Code:
static void run_init_process(char *init_filename)
{
        printk("Starting init (%s)\n",init_filename); <=== printk statement is my addition
        argv_init[0] = init_filename;
        execve(init_filename, argv_init, envp_init);
}
Console Output
==============
Code:
    Secondary interface chip i82555.
  General self-test: passed.
  Serial sub-system self-test: passed.
  Internal registers self-test: passed.
  ROM checksum self-test: passed (0x04f4518b).
  Receiver lock-up workaround activated.
NET4: Linux TCP/IP 1.0 for NET4.0
IP Protocols: ICMP, UDP, TCP, IGMP
IP: routing cache hash table of 2048 buckets, 16Kbytes
TCP: Hash tables configured (established 16384 bind 32768)
Sending DHCP requests . OK
IP-Config: Got DHCP answer from 192.168.1.1, my address is 192.168.1.49
IP-Config: Complete:
      device=eth0, addr=192.168.1.49, mask=255.255.255.0, gw=192.168.1.2,
     host=192.168.1.49, domain=home.gloetzner.net, nis-domain=(none),
     bootserver=192.168.1.1, rootserver=192.168.1.1, rootpath=/var/nfs/vmic7750
NET4: Unix domain sockets 1.0/SMP for Linux NET4.0.
Looking up port of RPC 100003/2 on 192.168.1.1
Looking up port of RPC 100005/1 on 192.168.1.1
VFS: Mounted root (nfs filesystem).
Mounted devfs on /dev
Freeing unused kernel memory: 84k freed
Starting init (/usr/src/init)  <===== I added this printk statement in the kernel code 
Thanks

Tilman
 
Old 05-29-2008, 07:49 AM   #2
pinniped
Senior Member
 
Registered: May 2008
Location: planet earth
Distribution: Debian
Posts: 1,732

Rep: Reputation: 50
You did build libc against the headers for the kernel you're booting, didn't you?

I also suggest you call 'fflush(stdout)' or you can't really be guaranteed that your message will be sent to the video buffer before the sleep() - otherwise you may have to wait about 32 years before you see your message displayed. Also, it's usually safer to send a message over something like the serial port; don't trust the video.

Otherwise, it looks like the rest of the boot procedure (initrd etc) went fine - can you confirm that?
 
Old 05-29-2008, 08:42 AM   #3
tilman1
Member
 
Registered: Mar 2007
Location: Stuttgart, Germany
Distribution: gentoo
Posts: 72

Original Poster
Rep: Reputation: 15
> You did build libc against the headers for the kernel you're booting, didn't you?
No, I did not -- and I suspect this is why the boot process does not complete. I am currently compiling libc against the header files of 2.4.36.3. Will see how that goes.

I used the default options while compiling , i.e. the libc of the gentoo distribution is used for linking. Interestingly enough this is V 2.6.9 -- even though I am using a 2.6.18-gentoo kernel. I read a little bit in the libc docs by now, and I understand that libc is compatible down to a kernel version 2.6.9.

I reasoned that it might work as the init executable is statically linked -- assuming that the syscall table between 2.4.36.3 and 2.6.9 was only enhance but not changed.

>Also, it's usually safer to send a message over something like the serial port; don't trust the video.
the default output is redirected to the serial port -- the bios has on option for that. The box is headless. I am using minicom to interact with it.

>Otherwise, it looks like the rest of the boot procedure (initrd etc) went fine - can you >confirm that?
Everything before "Freeing unused kernel memory" went OK. I think that also the mounting of the nfs went ok -- I am however not entirely sure. The host's /var/log/messages has an entry from rpc.mount. It does not look any different if booting with V 2.6.24 -- and that works.

/var/log/messages of the nfs host:
Code:
rpc.mountd: authenticated mount request from 192.168.1.49:
800 for /var/nfs/vmic7750 (/var/nfs/vmic7750)
Thanks

Tilman
 
Old 05-30-2008, 04:55 PM   #4
tilman1
Member
 
Registered: Mar 2007
Location: Stuttgart, Germany
Distribution: gentoo
Posts: 72

Original Poster
Rep: Reputation: 15
I compiled libc 2.3.6 against the source of kernel 2.4.36.3, and then linked my init program with new libc. This works. So the issue was indeed libc.

A questions remains however: How do I tell the make of the busybox to use a different kernel source (I understand that busybox uses a smaller library than libc)

Thanks
Tilman
 
Old 05-31-2008, 01:54 AM   #5
pinniped
Senior Member
 
Registered: May 2008
Location: planet earth
Distribution: Debian
Posts: 1,732

Rep: Reputation: 50
You can look at the LFS project (CLFS sub-project) to see what they do with busybox.
Busybox in Debian is linked against the usual glibc. In CLFS it is patched and linked to 'uClibc'. I have some strange behavior from busybox on uClibc though and haven't had time to look into it - some tests fail such as md5 calculation, but OpenSSL can do md5 hashes so I ignore the broken busybox version. I also suspect that some of the file utilities aren't working quite right on my embedded system, but once again I haven't had time to track down problems.
 
Old 06-15-2008, 04:05 PM   #6
tilman1
Member
 
Registered: Mar 2007
Location: Stuttgart, Germany
Distribution: gentoo
Posts: 72

Original Poster
Rep: Reputation: 15
I compiled busybox using suse 8.0 (running on vmware) kernel source of 2.4.36.3 and libc 2.2.5. Init(=Busybox) seems to come up (kernel version 2.4.36.3). The devices are however not recognized (see console output below). Using a 2.6.24 kernel, init starts up, and the devices are recognized. Any idea why that is?


Console Output (Kernel 2.4.36.3)
================================
....
VFS: Mounted root (nfs filesystem).
Mounted devfs on /dev
Freeing unused kernel memory: 84k freed
Starting init (/sbin/init)
Mount proc FS
Mount ROOTFS RW
Set hostname
Starting syslogd
Can't open /dev/ttyS0: No such file or directory
Can't open /dev/tty1: No such file or directory


Thanks
Tilman
 
Old 06-16-2008, 02:01 AM   #7
tilman1
Member
 
Registered: Mar 2007
Location: Stuttgart, Germany
Distribution: gentoo
Posts: 72

Original Poster
Rep: Reputation: 15
Found the issue myself: I had the devfs kernel option turned on (and at the same time, I am using a classic dev file system)
 
  


Reply


Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is Off
HTML code is Off



Similar Threads
Thread Thread Starter Forum Replies Last Post
The booting hang on "Freeing unused kernel memory : 168k freed" Brian Lu Linux - Software 1 05-27-2008 08:45 PM
2.6.19 hangs after "Freeing unused kernel memory" vas77 Linux - Kernel 2 02-26-2007 03:01 PM
kernel hangs at "Freeing unused kernel memory :126k" raklo Linux - Kernel 1 09-22-2006 08:14 PM
xen hangs when freeing unused kernel memory camlinux Linux - Server 1 08-09-2006 03:34 PM
Hang after "Freeing unused Kernel Memory" tmg Linux - Distributions 0 05-06-2003 07:27 PM

LinuxQuestions.org > Forums > Linux Forums > Linux - Software > Linux - Kernel

All times are GMT -5. The time now is 07:10 AM.

Main Menu
Advertisement
My LQ
Write for LQ
LinuxQuestions.org is looking for people interested in writing Editorials, Articles, Reviews, and more. If you'd like to contribute content, let us know.
Main Menu
Syndicate
RSS1  Latest Threads
RSS1  LQ News
Twitter: @linuxquestions
Open Source Consulting | Domain Registration