#327 mke2fs.conf needs ext2 tag in [fstypes] with inode_size=128

[Paul Franklin]

Hello. I apologize if this is in the wrong format, or the
wrong place (maybe it's a "feature request"?). I've never
tried to report a bug "upstream" in my life, before now.

I tried reporting this on the Fedora bug tracker but the
maintainer said it had been that way for four years. (See
https://bugzilla.redhat.com/show_bug.cgi?id=875605 )

I will admit that I doubt that it hits very many people, or
that very many people care about it, but I'm hoping you do.

Basically, for years I created an empty ext2 filesystem,
using an old operating system (which I wanted to be able to
access that ext2 partition) and then installed Fedora on it.

The Fedora installer no longer allows this (to prevent
idiots from installing a new Fedora on top of an old one)
and requires you to make and format a new root filesystem.

For reasons which don't matter, I have been trying for weeks
to install a Fedora 18 (pre-alpha, pre-beta, anything,) and
then filing bugs. Once their custom partitioning section
reached the point where it would let me, I tried creating a
new ext2 partition and then installed F18 onto it.

But my old grub wouldn't boot it, and my old operating
system wouldn't mount it, so I asked them to allow an
install onto an empty filesystem. They wouldn't, but did
mention the file /etc/mke2fs.conf in their reply.

So I looked at /etc/mke2fs.conf (in their initial ramdisk)
and learned that if I added an "ext2" tag to the [fstypes]
stanza, with an inode_size of 128, that the ext2 filesystem
it created was bootable by my ancient grub and mountable by
my ancient operating system (RH7.3, 2.4.20 -- don't ask).

So that's what I'm hoping you'll consider a bug, and fix.

Namely by adding something like this to mke2fs.conf's
[fstypes] stanza:

ext2 = {
inode_size = 128
}

(and I put it ahead of the ext3 line, if it matters).

I doubt that many people will make ext2 partitions these
days, but I am guessing that if they do it will be because
they have something old which needs one. Thus it should be
made in the old way, so it can be accessed by old systems.

I doubt that very many such newly-created ext2 filesystems
will ever be migrated to ext3 or ext4, so that in practice
there will be no performance hit (as none will be upgraded).

It seems to me that this is required for older kernels (and
so on) to be able to access newly-created ext2 partitions.

I am hoping you will agree.

In any case, thank you for taking the time to read this.