NetBSD Problem Report #46900

From campbell@mumble.net  Tue Sep  4 15:33:07 2012
Return-Path: <campbell@mumble.net>
Received: from mail.netbsd.org (mail.netbsd.org [149.20.53.66])
	by www.NetBSD.org (Postfix) with ESMTP id 6234063B86D
	for <gnats-bugs@gnats.NetBSD.org>; Tue,  4 Sep 2012 15:33:07 +0000 (UTC)
Message-Id: <20120904153300.1BC3A604CE@jupiter.mumble.net>
Date: Tue,  4 Sep 2012 15:32:59 +0000 (UTC)
From: Taylor R Campbell <campbell+netbsd@mumble.net>
Reply-To: Taylor R Campbell <campbell+netbsd@mumble.net>
To: gnats-bugs@gnats.NetBSD.org
Subject: can't receive packets on wm interface
X-Send-Pr-Version: 3.95

>Number:         46900
>Category:       kern
>Synopsis:       can't receive packets on wm interface
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    msaitoh
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Tue Sep 04 15:35:00 +0000 2012
>Closed-Date:    Sun Apr 18 20:37:23 +0000 2021
>Last-Modified:  Sun Apr 18 20:37:23 +0000 2021
>Originator:     Taylor R Campbell <campbell+netbsd@mumble.net>
>Release:        NetBSD 6.99.10
>Organization:
>Environment:
Architecture: amd64
Machine: amd64
>Description:

	I have two on-board wm interfaces, one of which, wm0, is
	somehow shared with a BMC, and the other of which, wm1, does
	not appear to be shared with a BMC although it is hard to rule
	that possibility out.  I can transmit and receive packets on
	wm0, but I can only transmit packets on wm1: on the host with
	wm1, I can see packets going out, but I see no packets coming
	in; on its peer, I can see packets coming out, and I can see
	packets going in, but the packets the peer transmits never
	seems to get there.

wm0 at pci9 dev 0 function 0: Intel i82574L (rev. 0x00)
wm0: interrupting at ioapic0 pin 16
wm0: PCI-Express bus
wm0: 256 word (8 address bits) SPI EEPROM
wm0: Ethernet address 00:25:90:..:..:..
ukphy0 at wm0 phy 1: OUI 0x000ac2, model 0x000b, rev. 1
ukphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto
...
wm1 at pci10 dev 0 function 0: Intel i82574L (rev. 0x00)
wm1: interrupting at ioapic0 pin 16
wm1: PCI-Express bus
wm1: 256 word (8 address bits) SPI EEPROM
wm1: Ethernet address 00:25:90:..:..:..
ukphy1 at wm1 phy 1: OUI 0x000ac2, model 0x000b, rev. 1
ukphy1: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto

	I don't think this is a hardware problem because FreeBSD 9.0
	can transmit and receive on both interfaces, although it
	wouldn't surprise me if the brain-damaged crossing of wires
	with the BMC to share ethernet ports is bollocksing the thing
	up.

>How-To-Repeat:

	Try to transmit packets on wm1.  Fail.

>Fix:

	Yes, please!

>Release-Note:

>Audit-Trail:
From: Manuel Bouyer <bouyer@antioche.eu.org>
To: gnats-bugs@NetBSD.org
Cc: kern-bug-people@NetBSD.org, gnats-admin@NetBSD.org, netbsd-bugs@NetBSD.org
Subject: Re: kern/46900: can't receive packets on wm interface
Date: Tue, 4 Sep 2012 20:43:57 +0200

 On Tue, Sep 04, 2012 at 03:35:00PM +0000, Taylor R Campbell wrote:
 > 	I have two on-board wm interfaces, one of which, wm0, is
 > 	somehow shared with a BMC, and the other of which, wm1, does
 > 	not appear to be shared with a BMC although it is hard to rule
 > 	that possibility out.  I can transmit and receive packets on
 > 	wm0, but I can only transmit packets on wm1: on the host with
 > 	wm1, I can see packets going out, but I see no packets coming
 > 	in; on its peer, I can see packets coming out, and I can see
 > 	packets going in, but the packets the peer transmits never
 > 	seems to get there.
 > 
 > wm0 at pci9 dev 0 function 0: Intel i82574L (rev. 0x00)
 > wm0: interrupting at ioapic0 pin 16
 > wm0: PCI-Express bus
 > wm0: 256 word (8 address bits) SPI EEPROM
 > wm0: Ethernet address 00:25:90:..:..:..
 > ukphy0 at wm0 phy 1: OUI 0x000ac2, model 0x000b, rev. 1
 > ukphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto
 > ...
 > wm1 at pci10 dev 0 function 0: Intel i82574L (rev. 0x00)
 > wm1: interrupting at ioapic0 pin 16
 > wm1: PCI-Express bus
 > wm1: 256 word (8 address bits) SPI EEPROM
 > wm1: Ethernet address 00:25:90:..:..:..
 > ukphy1 at wm1 phy 1: OUI 0x000ac2, model 0x000b, rev. 1
 > ukphy1: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto
 > 
 > 	I don't think this is a hardware problem because FreeBSD 9.0
 > 	can transmit and receive on both interfaces, although it
 > 	wouldn't surprise me if the brain-damaged crossing of wires
 > 	with the BMC to share ethernet ports is bollocksing the thing
 > 	up.

 what does netstat -i show ?

 -- 
 Manuel Bouyer <bouyer@antioche.eu.org>
      NetBSD: 26 ans d'experience feront toujours la difference
 --

From: Taylor R Campbell <campbell+netbsd@mumble.net>
To: Manuel Bouyer <bouyer@antioche.eu.org>
Cc: gnats-bugs@NetBSD.org, kern-bug-people@netbsd.org, gnats-admin@netbsd.org,
	netbsd-bugs@netbsd.org
Subject: Re: kern/46900: can't receive packets on wm interface
Date: Wed, 5 Sep 2012 01:31:54 +0000

    Date: Tue, 4 Sep 2012 20:43:57 +0200
    From: Manuel Bouyer <bouyer@antioche.eu.org>

    what does netstat -i show ?

 For a while it showed some ipkts and opkts on wm0, and opkts but no
 ipkts on wm1.  But then after futzing around for a while and switching
 ethernet cables and tcpdumping, I saw some ipkts on wm1.  I don't
 remember exactly what I did while futzing around, but it seems now
 that if I have *both* ethernet connected to my LAN, then they both
 work.  However, if only one is connected, then it's only wm0 that
 works; wm1 doesn't.

 I also messed around with the BMC to find why it is sharing wm0 when
 it has its own MAC and its own PHY and its own ethernet port.  I
 frobbed some switch in the IPMI nonsense that I think means it will
 really only try to use its own dedicated ethernet port and not the two
 ports for wm0 and wm1 on the mainboard.

From: Taylor R Campbell <campbell+netbsd@mumble.net>
To: Manuel Bouyer <bouyer@antioche.eu.org>
Cc: gnats-bugs@NetBSD.org, kern-bug-people@netbsd.org, gnats-admin@netbsd.org, netbsd-bugs@netbsd.org
Subject: Re: kern/46900: can't receive packets on wm interface
Date: Wed, 5 Sep 2012 05:54:11 +0000

 Another observation: pinging wm0 over my LAN takes about twice as much
 time as I'd expect, consistently over a millisecond when it should be
 around half a millisecond; but pinging wm1 over my LAN takes well over
 half a second (!).  The two interfaces are connected to different
 ports in my wall and I forget how the whole thing is wired, but I see
 the same behaviour if I switch the connectors, so obviously this
 abysmal performance of wm1 is independent of the wiring of my LAN.

 I haven't tried this with FreeBSD but I will soon if I find time.

 If you have any other tests you'd like me to run, let me know soon,
 because I may cease to have physical access to this machine in the
 next few days.

From: mlelstv@serpens.de (Michael van Elst)
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/46900: can't receive packets on wm interface
Date: Wed, 5 Sep 2012 08:55:28 +0000 (UTC)

 campbell+netbsd@mumble.net (Taylor R Campbell) writes:

 >I also messed around with the BMC to find why it is sharing wm0 when
 >it has its own MAC and its own PHY and its own ethernet port.

 That's a VERY SILLY feature to automatically detect the used port.

 -- 
 -- 
                                 Michael van Elst
 Internet: mlelstv@serpens.de
                                 "A potential Snark may lurk in every tree."

From: mlelstv@serpens.de (Michael van Elst)
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/46900: can't receive packets on wm interface
Date: Wed, 5 Sep 2012 08:56:50 +0000 (UTC)

 campbell+netbsd@mumble.net (Taylor R Campbell) writes:

 >Another observation: pinging wm0 over my LAN takes about twice as much
 >time as I'd expect, consistently over a millisecond when it should be
 >around half a millisecond; but pinging wm1 over my LAN takes well over
 >half a second (!).

 Sounds like wm1 doesn't produce receive interrupts or that NetBSD
 doesn't route them correctly.

 -- 
 -- 
                                 Michael van Elst
 Internet: mlelstv@serpens.de
                                 "A potential Snark may lurk in every tree."

Responsible-Changed-From-To: kern-bug-people->msaitoh
Responsible-Changed-By: msaitoh@NetBSD.org
Responsible-Changed-When: Tue, 25 Jun 2013 03:13:18 +0000
Responsible-Changed-Why:
mine.


State-Changed-From-To: open->closed
State-Changed-By: riastradh@NetBSD.org
State-Changed-When: Sun, 18 Apr 2021 20:37:23 +0000
State-Changed-Why:
seems to be fixed now


>Unformatted:

Home	PR Database Search