From owner-freebsd-smp  Tue Jun 24 02:10:27 1997
Return-Path: <owner-smp>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id CAA04920
          for smp-outgoing; Tue, 24 Jun 1997 02:10:27 -0700 (PDT)
Received: from gw.itfs.nsk.su (ns.nsk.ru [193.124.36.33])
          by hub.freebsd.org (8.8.5/8.8.5) with SMTP id CAA04912
          for <smp@freebsd.org>; Tue, 24 Jun 1997 02:10:22 -0700 (PDT)
Received: from itfs.UUCP (uucp@localhost) by gw.itfs.nsk.su (8.6.12/8.6.12) with UUCP id QAA26400 for smp@freebsd.org; Tue, 24 Jun 1997 16:10:19 +0700
Received: by itfs.nsk.su; Tue, 24 Jun 97 16:20:49 +0700 (NST)
Received: (from daemon@localhost) by news.itfs.nsk.su (8.7.5/8.6.12) id PAA23468; Tue, 24 Jun 1997 15:52:38 +0700 (NSD)
From: nnd@itfs.nsk.su
To: smp@freebsd.org
Subject: Re: SMP_PRIVPAGES
Date: 24 Jun 1997 08:52:37 GMT
Message-ID: <5oo1sl$mdv@news.itfs.nsk.su>
References: <199706232018.OAA07315@Ilsa.StevesCafe.com>
Sender: owner-smp@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

Steve Passe <smp@csn.net> wrote:
> Peter has committed the code for enabling SMP_PRIVPAGES.  Not that it
> is automatic, ie you don't need to add the option to your config file.
> I successfully used it to "make world" with -current as of earlier this
> morning.

> Thanx, Peter!

	I also make world and make kernel with this
code and after that discover that famous 'dd if=/dev/zero of=/dev/null...'
bencmark shows even smaller results than previous SMP-kernel.

	After that I've decide to try the next patch on
/sys/i386/isa/npx.c:

--- npx.c	Tue Jun 24 14:42:03 1997
+++ npx.c.orig	Tue Jun 24 09:02:56 1997
@@ -413,7 +413,7 @@
 	}
 	npxinit(__INITIAL_NPXCW__);
 
-#if defined(I586_CPU)
+#if defined(I586_CPU) && !defined(SMP)
 	/* FPU not working under SMP yet */
 	if (cpu_class == CPUCLASS_586 && npx_ex16) {
 		if (!(dvp->id_flags & NPX_DISABLE_I586_OPTIMIZED_BCOPY)) {

and ... this is a result:

(before my patch)

dd if=/dev/zero of=/dev/null bs=1m count=1000
1000+0 records in
1000+0 records out
1048576000 bytes transferred in 21.219967 secs (49414591 bytes/sec)

(and after it)

dd if=/dev/zero of=/dev/null bs=1m count=1000
1000+0 records in
1000+0 records out
1048576000 bytes transferred in 11.985571 secs (87486529 bytes/sec)

 The last number is still lose to NON-SMP case, but it seems
to me that there is another place to gain speed for bzero/bcopy -

in file /sys/i386/i386/support.s there is a "label" kernel_fpu_lock:

As I can (NOT?) understand it can be duplicated for each CPU
and may be this can give us some more speedup of bzero/bcopy ?

	N.Dudorov