Subject: Re: bin/29774: sysctl can't raise kern.mbuf.nmbclusters
To: None <gnats-bugs@NetBSD.org>
From: Manuel Bouyer <bouyer@antioche.eu.org>
List: netbsd-bugs
Date: 03/23/2005 19:49:35
--k+w/mQv8wyuph6w0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
On Wed, Mar 23, 2005 at 05:39:00PM +0000, tecneeq@tecneeq.de wrote:
> Architecture: i386
> Machine: i386
> >Description:
> I try to raise kern.mbuf.nmbclusters from 1024 to 2048 wich,
> according to the manpage, should be possible. It is not.
sysctl(3) says:
MBUF_NMBCLUSTERS
The limit on the number of mbuf clusters. The variable
can only be increased, and only increased on machines
with direct-mapped pool pages
So it's not supposed to work on all machines, and i386 doesn't have
direct-mapped pool pages. It's a bug in sysctl(8).
How to document it in sysctl(8) ? Change "raise only" to "machine dependant" ?
--
Manuel Bouyer <bouyer@antioche.eu.org>
NetBSD: 26 ans d'experience feront toujours la difference
--
--k+w/mQv8wyuph6w0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: attachment; filename=diff
Index: sysctl.8
===================================================================
RCS file: /cvsroot/src/sbin/sysctl/sysctl.8,v
retrieving revision 1.117
diff -u -r1.117 sysctl.8
--- sysctl.8 21 Nov 2004 22:18:10 -0000 1.117
+++ sysctl.8 23 Mar 2005 18:49:29 -0000
@@ -254,7 +254,7 @@
.Xr sysctl 3 .
The changeable column indicates whether a process with appropriate
privilege can change the value.
-.Bl -column proc.xpidx.rlimit.coredumpsize.hardxxxxxx integerxxx
+.Bl -column proc.xpidx.rlimit.coredumpsize.hardxxxx integerx
.It Sy Name Type Changeable
.It ddb.fromconsole integer yes
.It ddb.lines integer yes
@@ -317,7 +317,7 @@
.It kern.mbuf.mcllowat integer yes
.It kern.mbuf.mclsize integer no
.It kern.mbuf.msize integer no
-.It kern.mbuf.nmbclusters integer raise only
+.It kern.mbuf.nmbclusters integer machine dependant
.It kern.monotonic_clock integer no
.It kern.msgbuf struct no
.It kern.msgbufsize integer no
--k+w/mQv8wyuph6w0--