NetBSD Problem Report #34109
From jmmv@netbsd.org Sat Jul 29 08:10:59 2006
Return-Path: <jmmv@netbsd.org>
Received: from mail.netbsd.org (mail.netbsd.org [204.152.190.11])
by narn.NetBSD.org (Postfix) with ESMTP id D72B263B850
for <gnats-bugs@gnats.NetBSD.org>; Sat, 29 Jul 2006 08:10:58 +0000 (UTC)
Message-Id: <20060729081058.C84DF63B10C@mail.netbsd.org>
Date: Sat, 29 Jul 2006 08:10:58 +0000 (UTC)
From: jmmv@netbsd.org
Reply-To: jmmv@netbsd.org
To: gnats-bugs@NetBSD.org
Subject: vr(4) is painfully slow with the new rlphy(4)
X-Send-Pr-Version: 3.95
>Number: 34109
>Category: kern
>Synopsis: vr(4) is painfully slow with the new rlphy(4)
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: tsutsui
>State: closed
>Class: sw-bug
>Submitter-Id: net
>Arrival-Date: Sat Jul 29 08:15:00 +0000 2006
>Closed-Date: Mon Mar 19 12:27:10 +0000 2007
>Last-Modified: Sun Mar 09 11:10:02 +0000 2008
>Originator: Julio M. Merino Vidal
>Release: NetBSD 3.99.23
>Organization:
>Environment:
System: NetBSD dawn.home.network 3.99.23 NetBSD 3.99.23 (GENERIC) #22: Fri Jul 28 14:56:33 CEST 2006 root@max.home.network:/var/obj/usr/src-current/sys/arch/i386/compile/GENERIC i386
Architecture: i386
Machine: i386
>Description:
I have an Asus A7V8X-x motherboard, which includes the VIA KT400
chipset. The integrated network card is a VIA VT6102 (Rhine II)
10/100 Ethernet that generally works fine with the vr(4) driver.
With a current kernel (3.99.23, as shown above), the new rlphy(4)
driver is used with this card:
vr0 at pci0 dev 18 function 0: VIA VT6102 (Rhine II) 10/100 Ethernet
vr0: interrupting at irq 3
vr0: Ethernet address: 00:0c:6e:27:57:bd
rlphy0 at vr0 phy 1: RTL8201L 10/100 media interface, rev. 1
rlphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
This produces extremely slow throughput. For example, a scp of a
file between a 3.0_STABLE host in my home network and this current
machine delivers a maximum transfer rate of 150 KB/s; this "speed"
is not sustained as it fluctuates between 70 KB/s and 120 KB/s.
As you can imagine, the local network is unusable at those rates.
If I disable rlphy(4) from userconf, the kernel falls back to
ukphy(4) as it always did before rlphy(4) was added:
vr0 at pci0 dev 18 function 0: VIA VT6102 (Rhine II) 10/100 Ethernet
vr0: interrupting at irq 3
vr0: Ethernet address: 00:0c:6e:27:57:bd
ukphy0 at vr0 phy 1: Generic IEEE 802.3u media interface
ukphy0: RTL8201L 10/100 media interface (OUI 0x000004, model 0x0020), rev. 1
ukphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
With this configuration I get decent transfer speeds of the order
of 6-8 MB/s.
I'm afraid other network cards may be affected as well.
>How-To-Repeat:
Use a VIA Rhine II card with vr(4) and rlphy(4) and try to transfer
files to a machine in a local network. See how slow it is.
>Fix:
Don't know. As a workaround, use ukphy(4).
>Release-Note:
>Audit-Trail:
From: David Laight <david@l8s.co.uk>
To: gnats-bugs@NetBSD.org
Cc:
Subject: Re: kern/34109: vr(4) is painfully slow with the new rlphy(4)
Date: Mon, 31 Jul 2006 07:17:49 +0100
On Sat, Jul 29, 2006 at 08:15:01AM +0000, jmmv@netbsd.org wrote:
> >Number: 34109
> >Category: kern
> >Synopsis: vr(4) is painfully slow with the new rlphy(4)
>
> This produces extremely slow throughput. For example, a scp of a
> file between a 3.0_STABLE host in my home network and this current
> machine delivers a maximum transfer rate of 150 KB/s; this "speed"
> is not sustained as it fluctuates between 70 KB/s and 120 KB/s.
> As you can imagine, the local network is unusable at those rates.
>
> If I disable rlphy(4) from userconf, the kernel falls back to
> ukphy(4) as it always did before rlphy(4) was added:
At a guess full/half duplex is being incorrectly detected.
David
--
David Laight: david@l8s.co.uk
From: Juan RP <juan@xtrarom.org>
To: gnats-bugs@netbsd.org
Cc: jmmv@netbsd.org
Subject: Re: kern/34109 vr(4) is painfully slow with the new rlphy(4)
Date: Fri, 9 Feb 2007 08:59:41 +0100
Hi, I have applied two revisions: one from FreeBSD (rev 1.29) and
another one from OpenBSD (rev 1.26) which should fix the problems you
experienced, could you please confirm?
http://www.xtrarom.org/patches/rlphy.diff
--
http://plog.xtrarom.org/
Juan RP's blog - NetBSD/pkgsrc news in Spanish
From: Izumi Tsutsui <tsutsui@ceres.dti.ne.jp>
To: jmmv@NetBSD.org, gnats-bugs@NetBSD.org
Cc: kern-bug-people@NetBSD.org, gnats-admin@NetBSD.org,
netbsd-bugs@NetBSD.org, tsutsui@ceres.dti.ne.jp
Subject: Re: kern/34109 vr(4) is painfully slow with the new rlphy(4)
Date: Sun, 18 Mar 2007 14:31:25 +0900
- Could you please show "ifconfig vr0" output during slowness?
- Does it work if you set "ifconfig vr0 media 10baseT" manually?
- If it works, what happens once after "ifconfig vr0 media auto" again?
With some quick tests on my 8139 cards, rlphy on RTK8139
doesn't set BMCR_AUTOEN after reset but some of them set BMCR_FDX.
If that is the problem, could you try the attached patch?
---
Izumi Tsutsui
Index: rlphy.c
===================================================================
RCS file: /cvsroot/src/sys/dev/mii/rlphy.c,v
retrieving revision 1.12
diff -u -r1.12 rlphy.c
--- rlphy.c 16 Nov 2006 01:33:06 -0000 1.12
+++ rlphy.c 18 Mar 2007 05:15:57 -0000
@@ -65,8 +65,10 @@
int rlphy_service(struct mii_softc *, struct mii_data *, int);
void rlphy_status(struct mii_softc *);
+static void rlphy_reset(struct mii_softc *);
+
const struct mii_phy_funcs rlphy_funcs = {
- rlphy_service, rlphy_status, mii_phy_reset,
+ rlphy_service, rlphy_status, rlphy_reset,
};
static const struct mii_phydesc rlphys[] = {
@@ -331,3 +333,16 @@
} else
mii->mii_media_active = ife->ifm_media;
}
+
+static void
+rlphy_reset(struct mii_softc *sc)
+{
+
+ mii_phy_reset(sc);
+
+ /*
+ * XXX RealTek PHY doesn't set the BMCR properly after
+ * XXX reset, which breaks autonegotiation.
+ */
+ PHY_WRITE(sc, MII_BMCR, BMCR_AUTOEN);
+}
Responsible-Changed-From-To: kern-bug-people->tsutsui
Responsible-Changed-By: tsutsui@netbsd.org
Responsible-Changed-When: Sun, 18 Mar 2007 15:48:55 +0900
Responsible-Changed-Why:
I'll handle this.
State-Changed-From-To: open->feedback
State-Changed-By: tsutsui@netbsd.org
State-Changed-When: Sun, 18 Mar 2007 15:48:55 +0900
State-Changed-Why:
Could you please try my patch?
From: "Julio M. Merino Vidal" <jmmv84@gmail.com>
To: Izumi Tsutsui <tsutsui@ceres.dti.ne.jp>
Cc: gnats-bugs@NetBSD.org, netbsd-bugs@NetBSD.org
Subject: Re: kern/34109 vr(4) is painfully slow with the new rlphy(4)
Date: Sun, 18 Mar 2007 09:21:09 +0100
On 18/03/2007, at 6:31, Izumi Tsutsui wrote:
> - Could you please show "ifconfig vr0" output during slowness?
> - Does it work if you set "ifconfig vr0 media 10baseT" manually?
> - If it works, what happens once after "ifconfig vr0 media auto"
> again?
>
> With some quick tests on my 8139 cards, rlphy on RTK8139
> doesn't set BMCR_AUTOEN after reset but some of them set BMCR_FDX.
> If that is the problem, could you try the attached patch?
No, sorry, I cannot test it. That motherboard died a couple of
months ago and I don't have any other hardware supporting the vr
driver :-(
--
Julio M. Merino Vidal <jmmv84@gmail.com>
From: Izumi Tsutsui <tsutsui@ceres.dti.ne.jp>
To: jmmv84@gmail.com
Cc: gnats-bugs@NetBSD.org, netbsd-bugs@NetBSD.org,
tsutsui@ceres.dti.ne.jp
Subject: Re: kern/34109 vr(4) is painfully slow with the new rlphy(4)
Date: Sun, 18 Mar 2007 17:47:12 +0900
jmmv84@gmail.com wrote:
> No, sorry, I cannot test it. That motherboard died a couple of
> months ago and I don't have any other hardware supporting the vr
> driver :-(
Okay, anyway my patch fixes a problem that autoselect doesn't
choose 100baseTX but 10baseT after reboot until media is set
manually on my landisk (8139C+), so I'll commit it and
close this PR.
---
Izumi Tsutsui
From: Izumi Tsutsui <tsutsui@netbsd.org>
To: gnats-bugs@NetBSD.org
Cc:
Subject: PR/34109 CVS commit: src/sys/dev/mii
Date: Mon, 19 Mar 2007 12:23:30 +0000 (UTC)
Module Name: src
Committed By: tsutsui
Date: Mon Mar 19 12:23:30 UTC 2007
Modified Files:
src/sys/dev/mii: rlphy.c
Log Message:
Prepare rlphy_reset() and set BMCR_AUTOEN in it.
This makes autonegotiation on rlphy(4) at re(4)
on landisk (8139C+) select 100baseTX full-duplex
at boot time (before first ifconfig) properly,
and there is no bad side effect on rlphy(4) at nfe(4).
This may also fix PR kern/34109.
To generate a diff of this commit:
cvs rdiff -r1.12 -r1.13 src/sys/dev/mii/rlphy.c
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
State-Changed-From-To: feedback->closed
State-Changed-By: tsutsui@netbsd.org
State-Changed-When: Mon, 19 Mar 2007 21:27:10 +0900
State-Changed-Why:
Possible fix committed, and submitter no longer has a hardware.
From: Manuel Bouyer <bouyer@netbsd.org>
To: gnats-bugs@gnats.NetBSD.org
Cc:
Subject: PR/34109 CVS commit: [netbsd-4] src/sys/dev/mii
Date: Sun, 9 Mar 2008 11:06:54 +0000 (UTC)
Module Name: src
Committed By: bouyer
Date: Sun Mar 9 11:06:54 UTC 2008
Modified Files:
src/sys/dev/mii [netbsd-4]: rlphy.c
Log Message:
Pull up following revision(s) (requested by tsutsui in ticket #1098):
sys/dev/mii/rlphy.c: revision 1.13
Prepare rlphy_reset() and set BMCR_AUTOEN in it.
This makes autonegotiation on rlphy(4) at re(4)
on landisk (8139C+) select 100baseTX full-duplex
at boot time (before first ifconfig) properly,
and there is no bad side effect on rlphy(4) at nfe(4).
This may also fix PR kern/34109.
To generate a diff of this commit:
cvs rdiff -r1.12 -r1.12.2.1 src/sys/dev/mii/rlphy.c
Please note that diffs are not public domain; they are subject to the
copyright notices on the relevant files.
>Unformatted:
(Contact us)
$NetBSD: query-full-pr,v 1.51 2026/08/10 02:28:17 riastradh Exp $
$NetBSD: gnats_config.sh,v 1.10 2026/05/13 22:00:09 riastradh Exp $
Copyright © 1994-2026
The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.