Why no kernel message on my DOC system?

周 锐 san_blabla at hotmail.com
Sat Jan 26 23:00:58 EST 2002

Because in /dev/nftla1 there is only kernel image and initrd on it, so I 
have not tested to use 'root=/dev/nftla1' on my harddisk. But I do test to 
use lilo-mtd on my harddisk and boot from it, the kernel messages came out 
and the /proc/cmdline is: 
auto BOOT_IMAGE=blabla-doc rw root=100 console=tty0 console=ttyS0,9600n8
On the contrary, if boot from DiskOnChip, the /proc/cmdline is 
unrecognizable strings.

>From: David Woodhouse <dwmw2 at infradead.org>
>To: "周 锐" <san_blabla at hotmail.com>
>CC: linux-mtd at lists.infradead.org
>Subject: Re: Why no kernel message on my DOC system?
>Date: Sat, 26 Jan 2002 10:45:04 +0000
>san_blabla at hotmail.com said:
> > I've build up an X86 Linux system using MTD DiskOnChip driver, using
> > serial  port as my console. The Linux system can boot form DOC and I
> > can also login  into the system, the problem is that I can't see any
> > kernel printk  messages. But if I put the same kernel on my harddisk
> > and boot up using  harddisk, I can see all kernel messages, still
> > using serial port as my  console. my lilo.conf for DOC is:
>I suspect it's a LILO problem - when you log in after it's booted from the
>DiskOnChip, what are the contents of /proc/cmdline?
> > I don't know why it output "invalidate: busy buffer", but as I said
> > before,  the DOC can boot up and I can even login.
>LILO is asking the kernel to invalidate a buffer of the ramdisk. But on a
>ramdisk, the buffer is the _only_ copy of the data, so it's marked busy 
>cannot be invalidated. I think this is harmless.
> > My lilo.conf for my harddisk is :
> > The command to put lilo on harddisk is simple: #lilo and the output is
> > normal.
>If you boot using LILO on the hard disk, but with 'root=/dev/nftla', do
>you see the console messages then?
>Linux MTD discussion mailing list

享用世界上最大的 Web 电子邮件系统 ―― MSN Hotmail。

More information about the linux-mtd mailing list