On Sat, 19 Feb 2000, Alan Cox wrote:
> You netboot a kernel with attached initrd. That already works on the sparc.
> We don't yet have a general solution for this for all platforms so for
> 2.4 it has to stay
I use IP autoconfig under 2.3.x on 4mb nfsroot'd PS/2 386s. Its very
slow, yes. But its very nice for testing (like watching the world in slow
motion...).
Add initrd, you not only risk blowing all your RAM on just getting an IP
address, but also at least a minute booting just initrd to get a dhcpd,
not to mention the kernel+image download time (they use boot ROMs). A 2.3
kernel is already hefty. Add an initrd it gets needlessly bigger. All I
need is simple bootp support. Having basic RARP/BOOTP support in the
kernel may not be the Right thing to do, but it makes it damned easier to
get into real userspace on these machines, using the least amount of RAM.
Kernel+initrd is not a 100% solution, even on x86. What's it harming
anyway? It breaks every once in a while. There's enough people using it
to keep it fixed. Besides the fact that it shouldn't be there, is there
anything really wrong with it? Is it breaking something else by its mere
presence?
af
---
Adam Fritzler
{ mid@auk.cx, afritz@iname.com}
http://www.auk.cx/~mid/
-
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/
This archive was generated by hypermail 2b29 : Wed Feb 23 2000 - 21:00:24 EST