From owner-freebsd-bugs  Thu Mar 11 13:29:11 1999
Delivered-To: freebsd-bugs@freebsd.org
Received: from alcanet.com.au (border.alcanet.com.au [203.62.196.10])
	by hub.freebsd.org (Postfix) with ESMTP id 750EB14F26
	for <freebsd-bugs@FreeBSD.ORG>; Thu, 11 Mar 1999 13:29:07 -0800 (PST)
	(envelope-from peter.jeremy@AUSS2.ALCATEL.COM.AU)
Received: by border.alcanet.com.au id <40398>; Fri, 12 Mar 1999 07:16:37 +1000
Date: Fri, 12 Mar 1999 07:28:35 +1000
From: Peter Jeremy <peter.jeremy@AUSS2.ALCATEL.COM.AU>
Subject: Re: kern/10535: Very poor ethernet performance with tx driver
To: dg@root.com
Cc: freebsd-bugs@FreeBSD.ORG, semen@iclub.nsu.ru
Message-Id: <99Mar12.071637est.40398@border.alcanet.com.au>
Sender: owner-freebsd-bugs@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.org

David Greenman <dg@root.com> wrote:
>   The real question in my mind is why does the driver frob with the PHY
>in a critical path? That's guaranteed to cause the performance to suck, no
>matter what you do with the for-loops.

At the beginning of epic_ifstart(), there is a check that the link is up,
otherwise the mbufs are dropped in the bit bucket.  I'm not sure why the
link status has to be read twice.  At least it is only once per packet.

If a PII-450 is intermittent with a loop count of 0x1000 and 5 insns per
loop with no memory references, that suggests that the chip takes ~~50usec
to respond.

I can't answer this, and I can't find any technical information on the SMC
website.  Maybe Semen (as the driver author) can comment.

Peter


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-bugs" in the body of the message