MILO status (was Re: Test boot floppies (from CVS) and others)
I'm currenly building new 2.0.35 MILO images with Digital BIOS emulator
and a small ext2fs patch for the new filesystems (ported from aboot).
Could people please test these images so they could make their way into
frozen (this is a bugfix-only release).
Read on for the full story...
On Mon, Jan 31, 2000 at 12:26:47PM -0500, David Huggins-Daines wrote:
>
> Can the 2.0 milos cope with 2.2-formatted filesystems? (i.e. sparse
> superblock, other features) In my experience they can't, so that was one
> reason to use the 2.2 MILO.
Neither can 2.2 - MILO doesn't use the kernel's filesystem code.
I have patched 2.0.35 MILOs to cope with the new directory entry structure.
Didn't do anything with sparse superblocks and other stuff yet.
On Tue, Feb 01, 2000 at 10:27:10AM +0100, Paul Slootman wrote:
>
> FWIW: I've had zero success with the 2.2 MILOs on my XLT. The old 2.0
> MILO I've had for years works fine.
>
I just checked - the 2.2 images are all built with -mcpu=ev56, so they for
sure won't work on any ev4/ev5 machines. I tried -mcpu=ev4 and the
resulting MILO worked just fine for me.
On Tue, Feb 08, 2000 at 05:40:31PM +0000, Nikita Schmidt wrote:
>
> I shall see if I can build the new (c7b) MILO with the emulator - the
> 2.0.35 MILO/kernel code has been having troubles with the new toolchain
> lately.
>
Turned out to be a linker bug. I changed MILO so it should no longer
trigger it.
If the new images (milo-000209) happen to be satisfactory, I'll send the
ext2 patch to the 2.2 MILO maintainer along with the linker and -mcpu fixes.
Nikita
Reply to: