On 18 May 2002, Eric W. Biederman wrote:

> You may be correct, but I'm not ready to conceed that yet.  I see
> etherboot in a different light.  A set of drivers totally independent
> from the kernel, with a different set of maintainers.  With an
> independent set of drivers which hardware you support is obviously
> different.

And the fact is you know more than me on this score, since you have done
more. It just looks like more work to me.

> In summary.  Source divergence I see that as a requirement to have a
> maintainable bootloader.  Unless that bootloader is a user space
> program running on a kernel.  Lack of hardware support (or support
> divergence) I see as a problem to be fixed by having a sufficiently
> large user and developer base so you have enough people ensure the
> support is present.

Interesting thing: 9load, the plan9 loader, uses stock plan9 drivers, and
is interrupt-driven. It's an interesting approach. It definitely opened my
eyes when I saw what those guys had managed to do.

I am not completely convinced that source divergence is required.


Thanks Eric for bringing this up.

ron

Reply via email to