Subject: Re: kernel & libkvm [was IIci success]
To: der Mouse <mouse@Collatz.McRCIM.McGill.EDU>
From: Eduardo E. Horvath email@example.com <firstname.lastname@example.org>
Date: 01/12/1996 14:53:19
On Thu, 11 Jan 1996, der Mouse wrote:
> Um, doing that is equivalent to having the kernel symbol table around.
> You'll have to either have the kernel grovel through its own symbol
> table, or else restrict the structures you can ask for to a compiled-in
> list (which is equivalent to having the kernel know a (probably tiny)
> fraction of its own symbol table).
Aren't we all going overboard a bit here? I personally find it extremely
annoying that things like ps use the symbol table of a kernel that may not
match the image of the kernel that's actually running, especially since my
kernel usually has DDB and a full symbol table. Now what if we modify kvm
to open the running kernel's symbol table first, then if it doesn't exist,
start grovelling through /netbsd? That way we temporarily wind back the
pages that are locked into the kernel symbol table. However, this does
not solve the problem of needing to upgrade ps, etc. with every time the
kernel symbols change.