Re: [H8-uclinux-port] /bin/init process execution and busybox console

[Yoshinori S. <ys...@us...>]

At 06 Sep 2004 09:35:31 +0200,
Aleix Badia i Bosch wrote:
> Dear all, I've some problems executing uclinux (linux-2.6.7 + uclinux +
> h8300 port) using gdb debugger.
>=20
> It looks like if there's no kernel initialization problem:
>   - MTD initialization (RAM)
>   - FS (romfs) initialization
>   - console registration
>   - sys_open /dev/console
>=20
> NOTE: I'm using rootimage from uclinux-h8.sourceforge.jp with some /dev
> modifications
>=20
> crw-------    1 root     root       5,   1 gen  1  1970 console
> crw-------    1 root     root       4,   0 gen  1  1970 tty0
> crw-------    1 root     root       4,  64 gen  1  1970 ttySC0
> crw-------    1 root     root       4,  65 gen  1  1970 ttySC1
>=20
>=20
> but it fails (or at least it looks like it fail) as there's no ouput,
> when trying to execute /bin/init, just like idle task?=BF. I'm not sure if
> /bin/init is running but gdb can't trace that process that way, or if
> there's really a problem on kernel execve (-> syscall).
>=20
>=20
> ....
>     if (execute_command)
>              run_init_process(execute_command);
>=20
>      run_init_process("/sbin/init");  -> fails (OK rootimage->/bin/init)
>      run_init_process("/etc/init");   -> fails (OK rootimage->/bin/init)
>      run_init_process("/bin/init");   -> OK (MTD OK, romfs OK)
>      run_init_process("/bin/sh");
>=20
>       panic("No init found.  Try passing init=3D option to kernel.");
> ....
>=20
>=20
> Breakpoint 5, sys_execve (name=3D0x5bb000 "/bin/init", argv=3D0x202004,=20
>     envp=3D0x20202c, dummy=3D703605) at arch/h8300/kernel/process.c:256
> 256             error =3D do_execve(filename, argv, envp, regs);
> 3: error =3D 6008832
> (gdb) n
> 257             putname(filename);
> 3: error =3D 0
> (gdb)=20
>=20
>=20
> GDB should show some init starting message (perhaps not the same as
> busybox code below and rootimage are extracted from different busybox
> versions)
>=20
> extern int init_main(int argc, char **argv)
> {
>=20
> ..=20
>   /* Hello world */
>   message(MAYBE_CONSOLE | LOG, "init started:  %s",
> bb_msg_full_version);
> ...
>=20
> }
>=20
> Perhaps bad busybox console initialization ?
>=20
>=20
> I've no clear idea about gdb execution and the way console (struct
> console) gdb_console, registred as gdb_con, is used by debugger; the
> same with /dev/ttySC* or /dev/ttyS*
>=20
>=20
> I just insert gdb vmlinux debugging output and attach .config file:
>=20
> (gdb) target sim
> SCI0 =3D /dev/pts/9
> Connected to the simulator.
>=20
> (gdb) load
> Loading section .vectors, size 0x100 vma 0x0
> Loading section .text, size 0xb4410 vma 0x100
> Loading section .data, size 0xfc40 vma 0x200000
> Loading section .romfs, size 0xc400 vma 0x218a98
> Start address 0x200
> Transfer rate: 6826624 bits in <1 sec.
>=20
> (gdb) r
> Starting program: /home/usuari/linux-2.6.7/vmlinux=20
> Linux version 2.6.7-uc0 (usuari@debian) (gcc version 3.1.1) #77 Mon Sep
> 6 08:38:01 CEST 2004
>=20
>=20
> uClinux H8/300H
> Target Hardware: generic
> Flat model support (C) 1998,1999 Kenneth Albanowski, D. Jeff Dionne
> H8/300 series support by Yoshinori Sato <ys...@us...>
> On node 0 totalpages: 1536
>   DMA zone: 0 pages, LIFO batch:1
>   Normal zone: 1536 pages, LIFO batch:1
>   HighMem zone: 0 pages, LIFO batch:1
> Built 1 zonelists
> Kernel command line: console=3DttySC0,38400n81
> PID hash table entries: 16 (order 4: 128 bytes)
> Linux version 2.6.7-uc0 (usuari@debian) (gcc version 3.1.1) #77 Mon Sep
> 6 08:38:01 CEST 2004
>=20
>=20
> uClinux H8/300H
> Target Hardware: generic
> Flat model support (C) 1998,1999 Kenneth Albanowski, D. Jeff Dionne
> H8/300 series support by Yoshinori Sato <ys...@us...>
> On node 0 totalpages: 1536
>   DMA zone: 0 pages, LIFO batch:1
>   Normal zone: 1536 pages, LIFO batch:1
>   HighMem zone: 0 pages, LIFO batch:1
> Built 1 zonelists
> Kernel command line: console=3DttySC0,38400n81
> PID hash table entries: 16 (order 4: 128 bytes)
> Memory available: 3860k/999k RAM, 0k/0k ROM (720k kernel code, 98k data)
> Calibrating delay loop... 5.00 BogoMIPS
> Dentry cache hash table entries: 1024 (order: 0, 4096 bytes)
> Inode-cache hash table entries: 1024 (order: 0, 4096 bytes)
> Mount-cache hash table entries: 512 (order: 0, 4096 bytes)
> Linux NoNET1.0 for Linux 2.6
> SuperH SCI(F) driver initialized
> ttySC0 at MMIO 0xffffb0 (irq =3D 54) is a sci
> ttySC1 at MMIO 0xffffb8 (irq =3D 58) is a sci
> ttySC2 at MMIO 0xffffc0 (irq =3D 62) is a sci
> Using anticipatory io scheduler
> uclinux[mtd]: RAM probe address=3D0x218a98 size=3D0xd000
> Creating 1 MTD partitions on "RAM":
> 0x00000000-0x0000d000 : "ROMfs"
> uclinux[mtd]: set ROMfs to be root filesystem
> VFS: Mounted root (romfs filesystem) readonly.
>=20
> Program received signal SIGSEGV, Segmentation fault.
> 0x0006c5a4 in L1a5 () at lib/radix-tree.c:756
> 756     }
> (gdb)=20
>=20
>=20
> Regards,
>=20

I cannot identify a problem in my environment.
Will you identify it with new gcc?
Can you take a root image if possible?

--=20
Yoshinori Sato
<ys...@us...>