Initrd - cramfs image is not being recognized by 2.4.20 kernel

From: Keith Goettert (Keith_at_EcologyEnergy.com)
Date: 12/18/03

  • Next message: Kenward Vaughan: "Re: HOWTO paste to dos" 
    To: <debian-user@lists.debian.org>
Date: Wed, 17 Dec 2003 19:07:14 -0800

    I have a kernel built from kernel.org source. I have included support for
    ramdisk, initial ramdisk, and cramfs.
     
    I built a ram disk image using mkinitrd
     
    I can mount the image and list directories and files in it by using mount -o
    loop rdimage mntpnt
     
    So, it looks like it is functional. Additionally, when I look at the mount
    I see:
     
    /boot/raidrd.img on /initrd type cramfs (rw,loop=/dev/loop0)
     
    Again, this looks fine.
     
    In my lilo.conf I have placed:
     
    image=/vmlinuz.raid
            label=RAID.2.4.20
            read-only
            append="hde=noprobe hdf=noprobe hdg=noprobe hdh=noprobe"
            initrd=/boot/raidrd.img

     
    And when I run lilo -v 3 I get:
     
     
    Boot image: /vmlinuz.raid -> /boot/vmlinuz-2.4.20-raid
    Device 0x0301: BIOS drive 0x80, 255 heads, 9732 cylinders,
                   63 sectors. Partition offset: 63 sectors.
    Using Volume ID 17E81084 on bios 80
    Setup length is 5 sectors.
    Mapped 2001 sectors.
    Mapping RAM disk /boot/raidrd.img
    Device 0x0301: BIOS drive 0x80, 255 heads, 9732 cylinders,
                   63 sectors. Partition offset: 63 sectors.
    Using Volume ID 17E81084 on bios 80
    RAM disk: 1832 sectors.
    Added RAID.2.4.20 *
        <dev=0xe0,hd=24,cyl=92,sct=182>
        "ro root=301 hde=noprobe hdf=noprobe hdg=noprobe hdh=noprobe"
     
     
     
    All indications are that Lilo is mapping the rdimage to the proper place.
     
    BUT, when I boot, I get:
     
     
     ... (lotza stuff)...
    RAMDISK driver initialized: 16 RAM disks of 4096K size 1024 blocksize
    loop: loaded (max 8 devices)
     ...(lotza stuff) ...
    RAMDISK: Couldn't find valid RAM disk image starting at 0.
    Freeing initrd memory: 916k freed

     
    I have run out of ideas. Everything appears to be according to the initrd,
    mkinitrd, and lilo man pages. Could my problem be with versions? I am
    using the latest lilo (24), a 2.4.20 kernel and whatever mkinitrd and
    mkcramfs came with debian 3. Is there a flag I can set in my lilo.conf file
    that will cause the kernel to be more chatty about why there is no love?
     
     
      

    -- 
To UNSUBSCRIBE, email to debian-user-request@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
  • Next message: Kenward Vaughan: "Re: HOWTO paste to dos"

    Relevant Pages

    • Re: [patch][rfc] rewrite ramdisk
      ... I embedded a struct address_space instead of rolled rerolled it by ... This RAM disk is designed to have filesystems created on it and mounted ... of the ramdisk driver is also the hardblocksize;) of PAGE_SIZE because ... int count; ...
      (Linux-Kernel)
    • Re: BOOT_CS
      ... the CRC32 of the decompressed RAM disk - and stop the boot process ... another kernel version / initrd to load. ... You cannot place the decompressed initrd at a maximum address before ... address minus ramdisk size if I remember ...
      (Linux-Kernel)
    • Re: Is bin ever written to?
      ... I was thinking of putting them in a ram disk with SquashFS to speed up ... frequently accessed pages in memory anyway. ... The kernel writers usually says that it's much better to NOT use a ramdisk ...
      (Ubuntu)
    • Re: Possible to setup a Ramdisk in WindowsXP?
      ... When I said he's wanting to use a Ramdisk and possibly turn off VM (my ... suggestion that if he's going to do it then do it such that it will only ... > RAM disk will hurt his performance greatly. ... > If he's currently paging a lot with the mix of apps that he runs, ...
      (microsoft.public.windowsxp.configuration_manage)
    • Re: Possible to setup a Ramdisk in WindowsXP?
      ... When I said he's wanting to use a Ramdisk and possibly turn off VM (my ... suggestion that if he's going to do it then do it such that it will only ... > RAM disk will hurt his performance greatly. ... > If he's currently paging a lot with the mix of apps that he runs, ...
      (microsoft.public.windowsxp.general)