Failing to load a freshly compiled kernel 2.6.13.1 with swsusp2 4925

CLEAR ENOUGH.

Awups. Yerright. Better?

That message is not in the kernel source. THESE messages are:

.-drivers-mtd-inftlmount.c: printk(KERNWARNING "INFTL: Unable to read spare " .-drivers-usb-core-inode.c: err ("Unable to get usbfs mount"); .-fs-devfs-base.c: PRINTK("(): unable to mount devfs, err: %d-n", err); .-fs-hfs-super.c: hfswarn("hfsfs: unable to parse mount options.-n"); .-fs-hfsplus-super.c: printk("HFS+-fs: unable to parse mount options-n"); .-fs-nfs-nfsroot.c: printk(KERNERR "Root-NFS: Unable to get mountd port " .-fs-reiserfs-journal.c: reiserfswarning(pssb, "journal-1168 plus 1: kmalloc failed, unable to mount FS") ; .-fs-reiserfs-journal.c: reiserfswarning(pssb, "Replay Failure, unable to mount") ; .-init-domounts.c: panic("VFS: Unable to mount root fs on %s", b); .-init-domounts.c: panic("VFS: Unable to mount root fs on %s", bdevname(ROOTDEV, b)); .-init-domounts.c: printk(KERNERR "VFS: Unable to mount root fs via NFS, trying floppy.-n");

I would bet that you have the second from bottom:

"VFS: Unable to mount root fs on %s", bdevname(ROOTDEV, b)

bdevname does produce odd results when it doesn't know know what is up. Anyway, the code in domounts.c does what you expect.

for (p = fsnames; *p; p += strlen (p) + 1) { .... * * Allow the user to distinguish between failed sysopen * and bad superblock on root device. *

bdevname (ROOTDEV, b);

printk ("VFS: Cannot open root device"%s-" or % sn", rootdevicename, b); printk ("Please append a correct"root=-" boot optionn");

panic ("VFS: Unable to mount root fs on %s ", b); }

panic ("VFS: Unable to mount root fs on %s ", bdevname (ROOTDEV, b));

So you got one of the last two messages. Telling us what was the PRIOR message would allow us to distinguish which instance it was.

You did. Ipso facto ...

(please be HELPFUL! You can look up the message in the source as easily as I, and you can see as easily as I that the immediately prior context is essential).

Show us the imediately previous messages.

bdevname is defined as follows:

const char * bdevname (devt dev, char *buffer) { scnprintf (buffer, BDEVNAMESIZE, "unknown-block(%u,%u)", MAJOR (dev), MINOR (dev)); return buffer; }

so you are write about the unknown-block(foo,bar) part. I'd send in a kernel correction! It didn't use to be that way.

There is no name lookup attempted, so "unknown" is a misnomer.

Boot from a cd using THAT KERNEL, and see if you can determine what it can and can't do.

My deduction is that since you have support for IDE devices (and presumably for the fs on them), that that is not the device on which your fs is. Are you sure it is not on anther device? Sda something (SATA?). Hdg?

Thanks - they ain't free. Any more lip and you get to pay.

Peter