On Fri, 20 Apr 2001, Charles Steinkuehler wrote:
> > 1. Kernel is no longer able to mount filesystem images on the loopback
> > device.
>
> This seems like a bad thing, but it is probably tolerable. Why not make the
> loopback device a module? Note that a loopback device or a spare ramdisk
> will be required to backup the initial ramdisk image if we migrate away from
> the initrd-archive patch and use a plain-vanilla kernel...
Yeah. This is what happens when you compile when tired. (And you all
thought I learned that lesson when I left out DOS support last time
around... =)
> > 2. There is no longer a PCI Device Database, so PCI devices are listed in
> > /proc/pci by card ID.
>
> Absolutely no problem here...
I figured that would be the way to go. That device database is something
on the order of 20k, and we don't use it enough to be worth the effort.
> > 3. The Network Block Device was removed, as I couldn't really see a need
> > for it on a secure system.
>
> Does it save a lot of space being removed over being a module?
Don't know. Something else to check out. Honestly, I don't know how much
of a difference leaving the hooks for the module in make on average...
> > 4. Modularized serial support.
>
> OK, but this prevents headless boxes controled with a serial cable...
Nope. Just toss serial.o into /etc/modules. I'm really hoping that it's
small and I can add at least that back in.
> It'd be interesting to see how much each option affected size, but overall a
> 411K 2.4 kernel is VERY COOL, and should be quite usable for floppy
> firewalls. While I'd like to see a 'one size fits all' kernel, perhaps
> there could be a floppy only, minimal kernel, and a larger kernel with all
> the 'goodies' like RAID, loopback, etc (compiled as modules, where possible)
> for folks running from CD, HDD, Flash, or what have you.
Right. There's none of the MTD stuff compiled into this kernel, and I even
went modular on the IDE support. The bitch of it all is that, for some of
the ideas I've had, IDE support is Sorta-Kinda-Necessary. I think I'll
play around a bit over the next few days and see what I come up with.
I just compiled the kernel with the loopback and network block-device
support as modules, and put the serial stuff back into the kernel itself.
Results are:
-rw-r--r-- 1 wolfstar root 489k Apr 20 13:05 kernel.standard
-rw-r--r-- 1 wolfstar root 422k Apr 20 13:06 kernel.upx
11k for the serial stuff and modules isn't too bad. I did a quick bzImage
on it after putting the serial stuff back to Module, and I got:
-rw-r--r-- 1 wolfstar root 474k Apr 20 13:08 kernel.standard
-rw-r--r-- 1 wolfstar root 410k Apr 20 13:09 kernel.upx
(This time I specified -9 when I UPX'd the kernel, hence why it's a bit
smaller.)
I guess that I'll be updating the kernel again and putting it up on
Sourceforge. I REALLY need to write a script that does all this; doing by
hand is TEDIOUS.
--
George Metz
Commercial Routing Engineer
[EMAIL PROTECTED]
"We know what deterrence was with 'mutually assured destruction' during
the Cold War. But what is deterrence in information warfare?" -- Brigadier
General Douglas Richardson, USAF, Commander - Space Warfare Center
_______________________________________________
Leaf-devel mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/leaf-devel
