Re: Kernels > 1M
H. Peter Anvin (hpa@transmeta.com)
Sun, 22 Aug 1999 17:51:21 -0700
Mirian Crzig Lennox wrote:
>
> Zack Weinberg <zack@bitmover.com> writes:
> > >
> > > The 16-bit setup code *has* to run in real mode... that's the only
> > > reason it's 16 bit. As far as the APM BIOS, I'm sure you're aware of
> > > the number of problems caused by APM BIOSes that don't like to be called
> > > the way Linux calls them?
> >
> > I'll defer to you since you know more about x86 than me, but I don't
> > see why it's necessary to do any of the setup.S sequence in real
> > mode. Especially since the BSDs seem to manage it in protected mode
> > fine.
>
> BSD uses a two-stage bootloader. The second stage is usually a file
> called "boot" in the root directory of the root partition which is
> "tied" to the real-mode bootstrap by a utility program in the same way
> that vmlinuz'es are tied to the lilo bootstrap via the lilo program.
> This allows more flexibility because the second stage boot loader only
> needs to be installed once (since it knows how to walk the filesystem
> to load whatever kernel you want) HOWEVER it can't do all of the
> advanced things that linux needs, like initial ramdisk loading.
>
Furthermore, the fact that Linux contains its own real-mode code helps
avoid a major version skew problem by simplifying the boot protocol.
-hpa
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/