Subject: Re: kernel & libkvm [was IIci success]
To: None <current-users@NetBSD.ORG>
From: der Mouse <mouse@Collatz.McRCIM.McGill.EDU>
Date: 01/12/1996 07:06:38
>> Are there any ports left without working NetBSD bootblocks?
> Yes. The mac68k port still boots via MacOS.
But the booter program is NetBSD, right? I guess I meant, are there
any ports left that use someone else's boot code? Ports for which we
couldn't tweak the booter program to make sure the kernel symbol table
gets loaded into core along with the rest of the kernel?
>> [Keeping the whole kernel incore] wouldn't be that bad, though.
>> [...] The text segment perforce has to stay around anyway. The bss
>> segment can be synthesized on-demand. All that's really needed is a
>> virgin copy of everything neither text nor bss, which cannot be more
>> than 143986 bytes in this case.
> I'm confused. How much of this space is needed all the time (ie the
> actual running kernel), and how much is needed to make libkvm happy?
All libkvm needs is the symbol table, which unfortunately is most of
that space. Of that 142986 bytes, 29792 is data segment, which one
could argue should be the live data segment anyway so's to avoid having
to read /dev/kmem for it, and the rest is symbol and string tables,
with a tiny bit of a.out overhead (headers and such) thrown in.
And none of it needs to be constantly resident. kernfs could allocate
some swap space and page it on demand....