I'm hoping to actually go smaller once FreeBSD has a funtional
pivot_root() in the VFS layer (we can steal it from them) Then I can do a
512k RD/MD image (init, tar, gzip, mount_mfs(8), sh, etc.), then
transition to a 64 to 96 meg MFS root.
I'm talking to a consultant tomorrow about some kernel work. If
pivot_root() discussion goes no where, I'll bounty it here.
On another system, I can have a full userland (/bin, /sbin) with as little
as a 8 MB RD/MD (with dynamic linking + crunchgen + some pruning).
It's tougher on OpenBSD, but I want it to work with my bsd-appliance
framework.
~BAS
On Wed, 6 Jun 2007, Theo de Raadt wrote:
It can't be that big. And in our tree, it isn't that big.
You're doing things on your own, hitting constraints that exist in the
code. You get to fix those things, since you've gone outside OpenBSD
parameters.
Not everything can work, you know.
The 1st stage loader just resets the prom before the kernel load.
Can anyone else confirm this? You don't even need to elfrdsetroot(8) to
test. Just compile bsd.rd with MINIROOTSIZE=16384. I've been using 32768
on my 4.0 systems for the bsd-appliance project.
I've tested it on an AMD Athalon, an AMD Geode, and a VMWare machine.
l8*
-lava (Brian A. Seklecki - Pittsburgh, PA, USA)
http://www.spiritual-machines.org/
"Guilty? Yeah. But he knows it. I mean, you're guilty.
You just don't know it. So who's really in jail?"
~James Maynard Keenan
PS:
Also, am I imagining this:
.if !${IDENT:M-DI386_CPU}
CMACHFLAGS= -march=i486
.else
CMACHFLAGS=
.endif
Somewhere this logic must be getting short-circuited in my config. I'm
seeting "-march=i486".
l8*
-lava (Brian A. Seklecki - Pittsburgh, PA, USA)
http://www.spiritual-machines.org/
"Guilty? Yeah. But he knows it. I mean, you're guilty.
You just don't know it. So who's really in jail?"
~James Maynard Keenan