http://qa.mandrakesoft.com/show_bug.cgi?id=3614

[EMAIL PROTECTED] changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|mkinitrd includes loop.o but|mkinitrd includes loop.o,
                   |not aes.o                   |but loop.o needs aes.o, so
                   |                            |loopback root fails to boot



------- Additional Comments From [EMAIL PROTECTED]  2003-03-31 12:45 -------
(As an aside, I've modified the summary in an attempt to improve it.)

I just attached two patches, which present two possible approaches to fixing
this bug. A third approach would be to add full-blown dependency resolution to
mkinitrd, using modules.dep, but I don't have time to do that tonight, and if a
new mkinitrd is released as a fix for Mandrake 9.1, that method would be too
invasive IMO.

FWIW, if you want me to take the second patch (which uses modules.dep) and add
some error checking to it, I'm willing to do that. Also FWIW, this bug could be
related to bug 1614, or it might even be another manifestation of the exact same
bug...



------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.



------- Reminder: -------
assigned_to: [EMAIL PROTECTED]
status: UNCONFIRMED
creation_date: 
description: 
When a loopback file is used as the root "partition," mkinitrd includes loop.o
but not aes.o in the initrd. Any attempt to use this initrd to boot the system
results in the initrd being unable to mount the root filesystem and, hence, the
system being unbootable.

I first encountered this problem during a fresh 9.1 install (I used a Knoppix
rescue CD to manually modify the initrd so I could boot after the install). It's
100% reproducible with mkinitrd from the command line after installation, too.
In any case, this bug makes Mandrake 9.1 loopback installs *totally unusable*
unless you manually hack the initrd.

I'm taking a look at the mkinitrd source code now, and I'll see if I can put
together a patch (or maybe a few patches which use different approaches) to fix
this.

Reply via email to