Subject: Re: kernel limits (??)
To: None <firstname.lastname@example.org>
From: John Darrow <John.P.Darrow@wheaton.edu>
Date: 10/14/1998 18:02:31
>> #ifndef MAX_KMAPENT
>> #define MAX_KMAPENT (50 + 2 * MAXPROC)
>Oops, should be NPROC, not MAXPROC. Plus we'd need to move the
>default definition of NPROC in param.c as
> #ifndef NPROC
> #define NPROC (20 + 16 * MAXUSERS)
>to param.h (inside #ifdef _KERNEL) so both param.c and the
>VM code can access NPROC.
>That seems like The Right Thing till we make the static table go away
>completely. Comments? Yell now if you see a problem.
Sounds good to me.
I went ahead and sent in a PR (pr kern/6293) before these last few messages
came through. It just left MAX_KMAPENT at 500 (1000 in current) but made
it overrideable. But the version using NPROC sounds better.
(and while we're in there, anyone want to hit port-i386/6226? It's just
a one-line patch, fixes an incorrect message, but it hasn't been touched
since I sent it in...)
Computing Services, Wheaton College, Wheaton, IL