From scheidell@scanner.secnap.net Tue Nov 26 15:52:28 2002 Return-Path: Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id CA55437B401 for ; Tue, 26 Nov 2002 15:52:28 -0800 (PST) Received: from hackertrap.secnap.net (secnap.secnap.net [208.237.120.135]) by mx1.FreeBSD.org (Postfix) with ESMTP id 310FE43E4A for ; Tue, 26 Nov 2002 15:52:28 -0800 (PST) (envelope-from scheidell@scanner.secnap.net) Received: from hackertrap.secnap.net (localhost [127.0.0.1]) by hackertrap.secnap.net (8.12.6/8.11.3) with ESMTP id gAQNqRZ9013152 for ; Tue, 26 Nov 2002 18:52:27 -0500 (EST) (envelope-from scheidell@scanner.secnap.net) Received: (from scheidell@localhost) by hackertrap.secnap.net (8.12.6/8.12.6/Submit) id gAQNqRgv013151; Tue, 26 Nov 2002 18:52:27 -0500 (EST) Message-Id: <200211262352.gAQNqRgv013151@hackertrap.secnap.net> Date: Tue, 26 Nov 2002 18:52:27 -0500 (EST) From: Michael Scheidell To: FreeBSD-gnats-submit@freebsd.org Subject: Softboot causes autoconf failure on Broadcom 5703 X-Send-Pr-Version: 3.113 X-GNATS-Notify: >Number: 45773 >Category: i386 >Synopsis: [bge] Softboot causes autoconf failure on Broadcom 5703 >Confidential: no >Severity: serious >Priority: high >Responsible: yongari >State: closed >Quarter: >Keywords: >Date-Required: >Class: sw-bug >Submitter-Id: current-users >Arrival-Date: Tue Nov 26 16:00:09 PST 2002 >Closed-Date: Wed Sep 22 19:20:55 UTC 2010 >Last-Modified: Wed Sep 22 19:20:55 UTC 2010 >Originator: Michael Scheidell >Release: FreeBSD 4.7-STABLE i386 >Organization: SECNAP Network Security, LLC >Environment: System: FreeBSD hackertrap.secnap.net 4.7-STABLE FreeBSD 4.7-STABLE #6: Fri Nov 22 08:40:01 EST 2002 scheidell@hackertrap.secnap.net:/usr/obj/usr/src/sys/HACKERTRAP i386 >Description: IBM x225, FBSD 4.7, built in Broadcom 5703 10/100/1000gb enet Soft boot leaves the card in unknown (1000BaseS) state, and un able to communicate. ifconfig shows 1000BaseS, link line on, but no communications. Must either hard boot or fix HUB or SWITCH into 100base mode. Also, what should be normal operation on card fails. ifconfig bge0 media 100BaseTX fails, 'device not configured' >How-To-Repeat: Softboot >Fix: Hardboot (power off/on) >Release-Note: >Audit-Trail: State-Changed-From-To: open->feedback State-Changed-By: remko State-Changed-When: Sun Sep 3 09:55:20 UTC 2006 State-Changed-Why: Hello Michael, can you tell me whether we are still facing this problem? We have had several updates on the bge driver so this might be resolved, and I want to know whether that is true or not. Responsible-Changed-From-To: freebsd-i386->remko Responsible-Changed-By: remko Responsible-Changed-When: Sun Sep 3 09:55:20 UTC 2006 Responsible-Changed-Why: Grab the PR http://www.freebsd.org/cgi/query-pr.cgi?pr=45773 State-Changed-From-To: feedback->closed State-Changed-By: linimon State-Changed-When: Fri Jun 15 10:40:48 UTC 2007 State-Changed-Why: Feedback timeout (> 6 months). http://www.freebsd.org/cgi/query-pr.cgi?pr=45773 From: "Michael Scheidell" To: Cc: Subject: Re: i386/45773: [bge] Softboot causes autoconf failure on Broadcom 5703 Date: Fri, 15 Jun 2007 07:56:24 -0400 This is a multi-part message in MIME format. ------=_NextPart_000_0002_01C7AF22.ABA366F0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Still happens. ------=_NextPart_000_0002_01C7AF22.ABA366F0 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable

Still = happens.

------=_NextPart_000_0002_01C7AF22.ABA366F0-- State-Changed-From-To: closed->open State-Changed-By: remko State-Changed-When: Fri Jun 15 12:04:14 UTC 2007 State-Changed-Why: This still happends, I also had discussed this with Michael a lot time ago but never gotten round to put that back into the ticket: Michael sept 3 2006: I think it still does this on many occasions, especially if connected to a 100bt port. I'll check Tuesday with one connected to a HP 10/100bt HUB, and a 3com 10/100bt switch. (I think connected to HUB is biggest problem) If I do a 'ifconfig bge0 media 100basetx' it seems to help. if I just did a ifconfig to look at it, it shows no carrier, even if cable plugged in. latest drivers are on freebsd 5.5 RELENG_5_5 Sept 5: > I had been in contact with one of our broadcom > developers and he mentioned that the RELENG_5 > (including 5.5) driver for bge is "old". > > Is it possible for you to check whether the 6.X > releases (6-STABLE or 6.1 pref.) have this same > issue? > I still have boxes at 4.11! (and I finally figured out how to get them to remotely upgrade to 5.5 in a headless, hands off situation) Not ready to go to 6.1 yet, sorry. (and 5.5 is still production, right? we going to have a 5.6? with new bge drivers?) Sept 16: verified on 5.5 p4, system was dell 420 on a netlink 10/100 switch. softboot failed to reconnect to network. all lights on, 'carrier' detected, ifconfig showed 100basetx full-duplex, but had to power off dell 420 in order to actually get it to connect to network. Responsible-Changed-From-To: remko->freebsd-bugs Responsible-Changed-By: remko Responsible-Changed-When: Fri Jun 15 12:04:14 UTC 2007 Responsible-Changed-Why: Return to the pool because of the feedback, perhaps we need davidch here.. http://www.freebsd.org/cgi/query-pr.cgi?pr=45773 State-Changed-From-To: open->feedback State-Changed-By: vwe State-Changed-When: Wed Jan 14 21:21:47 UTC 2009 State-Changed-Why: Michael, can you please tell us if you're still seeing the issue with late{r|est} releases? Responsible-Changed-From-To: freebsd-bugs->freebsd-net Responsible-Changed-By: vwe Responsible-Changed-When: Wed Jan 14 21:21:47 UTC 2009 Responsible-Changed-Why: Over to maintainer(s). http://www.freebsd.org/cgi/query-pr.cgi?pr=45773 Responsible-Changed-From-To: freebsd-net->yongari Responsible-Changed-By: yongari Responsible-Changed-When: Tue Jul 6 19:45:32 UTC 2010 Responsible-Changed-Why: Take. There had been a lot of PHY related stuff changes and I guess it might be fixed. Would you try latest FreeBSD release(8.1-RC2 or 7.3-RELEASE) and let me know whether it's still issue or not? http://www.freebsd.org/cgi/query-pr.cgi?pr=45773 From: Pyun YongHyeon To: Florian Smeets Cc: yongari@FreeBSD.org, bug-followup@FreeBSD.org Subject: Re: i386/45773: [bge] Softboot causes autoconf failure on Broadcom 5703 Date: Tue, 6 Jul 2010 14:40:17 -0700 On Tue, Jul 06, 2010 at 10:24:14PM +0200, Florian Smeets wrote: > On 7/6/10 9:46 PM, yongari@FreeBSD.org wrote: > > Synopsis: [bge] Softboot causes autoconf failure on Broadcom 5703 > > > > Responsible-Changed-From-To: freebsd-net->yongari > > Responsible-Changed-By: yongari > > Responsible-Changed-When: Tue Jul 6 19:45:32 UTC 2010 > > Responsible-Changed-Why: > > Take. > > There had been a lot of PHY related stuff changes and I guess it > > might be fixed. Would you try latest FreeBSD release(8.1-RC2 or > > 7.3-RELEASE) and let me know whether it's still issue or not? > > > > Hi Pyun, > > i saw this on a > > bge0@pci0:1:2:0: class=0x020000 card=0x00e30e11 chip=0x165414e4 rev=0x03 > hdr=0x00 > vendor = 'Broadcom Corporation' > device = 'NetXtreme Gigabit Ethernet (BCM5705-)' > class = network > subclass = ethernet > > when installing 8.1-RC2. During installation the server was connected to > a 100MBit switch, i had to set the interface to 100BaseTX manually, then > it came up as 100BaseTX half-duplex. After i put it in the rack and it This looks like expected behavior. When link partner is set to AUTO and bge(4) is manually forced to use full-duplex media normal link resolution is half-duplex as defined by IEEE 802.3. When you have to manually set media speed/duplex, link partner also should be configured to use the same media speed/duplex. The real question is why AUTO negotiation does not work. It could be switch problem though. Would you show me verbose dmesg output? > was connected to a different 100Mbit switch where the problem did not > occur. A day later that switch was replaced by a gigabit one where we > also could not observe a problem with auto negotiation. > OK. > So it seems to be related to the combination of switch and interface... > At least, I'm not aware of AUTO negotiation issue of BCM5705. > HTH, > Florian From: Florian Smeets To: pyunyh@gmail.com Cc: bug-followup@FreeBSD.org Subject: Re: i386/45773: [bge] Softboot causes autoconf failure on Broadcom 5703 Date: Wed, 07 Jul 2010 10:56:37 +0200 On 7/6/10 11:40 PM, Pyun YongHyeon wrote: > On Tue, Jul 06, 2010 at 10:24:14PM +0200, Florian Smeets wrote: >> On 7/6/10 9:46 PM, yongari@FreeBSD.org wrote: >>> Synopsis: [bge] Softboot causes autoconf failure on Broadcom 5703 >>> >> >> bge0@pci0:1:2:0: class=0x020000 card=0x00e30e11 chip=0x165414e4 rev=0x03 >> hdr=0x00 >> vendor = 'Broadcom Corporation' >> device = 'NetXtreme Gigabit Ethernet (BCM5705-)' >> class = network >> subclass = ethernet >> >> when installing 8.1-RC2. During installation the server was connected to >> a 100MBit switch, i had to set the interface to 100BaseTX manually, then >> it came up as 100BaseTX half-duplex. After i put it in the rack and it > > This looks like expected behavior. When link partner is set to > AUTO and bge(4) is manually forced to use full-duplex media normal > link resolution is half-duplex as defined by IEEE 802.3. When you > have to manually set media speed/duplex, link partner also should > be configured to use the same media speed/duplex. > The real question is why AUTO negotiation does not work. It could > be switch problem though. Would you show me verbose dmesg output? > Sure. bge0: mem 0xfddf0000-0xfddfffff irq 17 at device 2.0 on pci1 bge0: Reserved 0x10000 bytes for rid 0x10 type 3 at 0xfddf0000 bge0: CHIP ID 0x00003003; ASIC REV 0x03; CHIP REV 0x30; PCI miibus0: on bge0 brgphy0: PHY 1 on miibus0 brgphy0: OUI 0x000818, model 0x001a, rev. 2 brgphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto bge0: bpf attached bge0: Ethernet address: 00:17:a4:XX:XX:XX ioapic0: routing intpin 17 (PCI IRQ 17) to lapic 0 vector 54 bge0: [MPSAFE] bge0: [ITHREAD] in case you need more, the complete log is available here: http://smeets.im/~flo/ml350g4p.dmesg.boot Cheers, Florian From: Pyun YongHyeon To: Florian Smeets Cc: bug-followup@FreeBSD.org Subject: Re: i386/45773: [bge] Softboot causes autoconf failure on Broadcom 5703 Date: Thu, 8 Jul 2010 12:38:17 -0700 On Wed, Jul 07, 2010 at 10:56:37AM +0200, Florian Smeets wrote: > On 7/6/10 11:40 PM, Pyun YongHyeon wrote: > > On Tue, Jul 06, 2010 at 10:24:14PM +0200, Florian Smeets wrote: > >> On 7/6/10 9:46 PM, yongari@FreeBSD.org wrote: > >>> Synopsis: [bge] Softboot causes autoconf failure on Broadcom 5703 > >>> > >> > >> bge0@pci0:1:2:0: class=0x020000 card=0x00e30e11 chip=0x165414e4 rev=0x03 > >> hdr=0x00 > >> vendor = 'Broadcom Corporation' > >> device = 'NetXtreme Gigabit Ethernet (BCM5705-)' > >> class = network > >> subclass = ethernet > >> > >> when installing 8.1-RC2. During installation the server was connected to > >> a 100MBit switch, i had to set the interface to 100BaseTX manually, then > >> it came up as 100BaseTX half-duplex. After i put it in the rack and it > > > > This looks like expected behavior. When link partner is set to > > AUTO and bge(4) is manually forced to use full-duplex media normal > > link resolution is half-duplex as defined by IEEE 802.3. When you > > have to manually set media speed/duplex, link partner also should > > be configured to use the same media speed/duplex. > > The real question is why AUTO negotiation does not work. It could > > be switch problem though. Would you show me verbose dmesg output? > > > > Sure. > > bge0: mem > 0xfddf0000-0xfddfffff irq 17 at device 2.0 on pci1 > bge0: Reserved 0x10000 bytes for rid 0x10 type 3 at 0xfddf0000 > bge0: CHIP ID 0x00003003; ASIC REV 0x03; CHIP REV 0x30; PCI > miibus0: on bge0 > brgphy0: PHY 1 on miibus0 > brgphy0: OUI 0x000818, model 0x001a, rev. 2 > brgphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, > 1000baseT-FDX, auto > bge0: bpf attached > bge0: Ethernet address: 00:17:a4:XX:XX:XX > ioapic0: routing intpin 17 (PCI IRQ 17) to lapic 0 vector 54 > bge0: [MPSAFE] > bge0: [ITHREAD] > > in case you need more, the complete log is available here: > http://smeets.im/~flo/ml350g4p.dmesg.boot > Thanks for the info. I don't see any abnormal things there. If the 100Mbps switch provides some management functions, could you check resolved link state on the switch? And could you reproduce the issue with other 100Mbps switch? > Cheers, > Florian From: Pyun YongHyeon To: Florian Smeets Cc: bug-followup@FreeBSD.org Subject: Re: i386/45773: [bge] Softboot causes autoconf failure on Broadcom 5703 Date: Wed, 22 Sep 2010 11:21:44 -0700 On Thu, Jul 08, 2010 at 12:38:17PM -0700, Pyun YongHyeon wrote: > On Wed, Jul 07, 2010 at 10:56:37AM +0200, Florian Smeets wrote: > > On 7/6/10 11:40 PM, Pyun YongHyeon wrote: > > > On Tue, Jul 06, 2010 at 10:24:14PM +0200, Florian Smeets wrote: > > >> On 7/6/10 9:46 PM, yongari@FreeBSD.org wrote: > > >>> Synopsis: [bge] Softboot causes autoconf failure on Broadcom 5703 > > >>> > > >> > > >> bge0@pci0:1:2:0: class=0x020000 card=0x00e30e11 chip=0x165414e4 rev=0x03 > > >> hdr=0x00 > > >> vendor = 'Broadcom Corporation' > > >> device = 'NetXtreme Gigabit Ethernet (BCM5705-)' > > >> class = network > > >> subclass = ethernet > > >> > > >> when installing 8.1-RC2. During installation the server was connected to > > >> a 100MBit switch, i had to set the interface to 100BaseTX manually, then > > >> it came up as 100BaseTX half-duplex. After i put it in the rack and it > > > > > > This looks like expected behavior. When link partner is set to > > > AUTO and bge(4) is manually forced to use full-duplex media normal > > > link resolution is half-duplex as defined by IEEE 802.3. When you > > > have to manually set media speed/duplex, link partner also should > > > be configured to use the same media speed/duplex. > > > The real question is why AUTO negotiation does not work. It could > > > be switch problem though. Would you show me verbose dmesg output? > > > > > > > Sure. > > > > bge0: mem > > 0xfddf0000-0xfddfffff irq 17 at device 2.0 on pci1 > > bge0: Reserved 0x10000 bytes for rid 0x10 type 3 at 0xfddf0000 > > bge0: CHIP ID 0x00003003; ASIC REV 0x03; CHIP REV 0x30; PCI > > miibus0: on bge0 > > brgphy0: PHY 1 on miibus0 > > brgphy0: OUI 0x000818, model 0x001a, rev. 2 > > brgphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, > > 1000baseT-FDX, auto > > bge0: bpf attached > > bge0: Ethernet address: 00:17:a4:XX:XX:XX > > ioapic0: routing intpin 17 (PCI IRQ 17) to lapic 0 vector 54 > > bge0: [MPSAFE] > > bge0: [ITHREAD] > > > > in case you need more, the complete log is available here: > > http://smeets.im/~flo/ml350g4p.dmesg.boot > > > > Thanks for the info. I don't see any abnormal things there. > If the 100Mbps switch provides some management functions, could you > check resolved link state on the switch? > And could you reproduce the issue with other 100Mbps switch? > Any progress? From: Florian Smeets To: pyunyh@gmail.com Cc: bug-followup@FreeBSD.org Subject: Re: i386/45773: [bge] Softboot causes autoconf failure on Broadcom 5703 Date: Wed, 22 Sep 2010 21:02:17 +0200 On 22.09.10 20:21, Pyun YongHyeon wrote: > On Thu, Jul 08, 2010 at 12:38:17PM -0700, Pyun YongHyeon wrote: >> On Wed, Jul 07, 2010 at 10:56:37AM +0200, Florian Smeets wrote: >>> On 7/6/10 11:40 PM, Pyun YongHyeon wrote: >>>> On Tue, Jul 06, 2010 at 10:24:14PM +0200, Florian Smeets wrote: >>>>> On 7/6/10 9:46 PM, yongari@FreeBSD.org wrote: >>>>>> Synopsis: [bge] Softboot causes autoconf failure on Broadcom 5703 >>>>>> >>>>> >>>>> bge0@pci0:1:2:0: class=0x020000 card=0x00e30e11 chip=0x165414e4 rev=0x03 >>>>> hdr=0x00 >>>>> vendor = 'Broadcom Corporation' >>>>> device = 'NetXtreme Gigabit Ethernet (BCM5705-)' >>>>> class = network >>>>> subclass = ethernet >>>>> >>>>> when installing 8.1-RC2. During installation the server was connected to >>>>> a 100MBit switch, i had to set the interface to 100BaseTX manually, then >>>>> it came up as 100BaseTX half-duplex. After i put it in the rack and it >>>> >>>> This looks like expected behavior. When link partner is set to >>>> AUTO and bge(4) is manually forced to use full-duplex media normal >>>> link resolution is half-duplex as defined by IEEE 802.3. When you >>>> have to manually set media speed/duplex, link partner also should >>>> be configured to use the same media speed/duplex. >>>> The real question is why AUTO negotiation does not work. It could >>>> be switch problem though. Would you show me verbose dmesg output? >>>> >>> >>> Sure. >>> >>> bge0: mem >>> 0xfddf0000-0xfddfffff irq 17 at device 2.0 on pci1 >>> bge0: Reserved 0x10000 bytes for rid 0x10 type 3 at 0xfddf0000 >>> bge0: CHIP ID 0x00003003; ASIC REV 0x03; CHIP REV 0x30; PCI >>> miibus0: on bge0 >>> brgphy0: PHY 1 on miibus0 >>> brgphy0: OUI 0x000818, model 0x001a, rev. 2 >>> brgphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, >>> 1000baseT-FDX, auto >>> bge0: bpf attached >>> bge0: Ethernet address: 00:17:a4:XX:XX:XX >>> ioapic0: routing intpin 17 (PCI IRQ 17) to lapic 0 vector 54 >>> bge0: [MPSAFE] >>> bge0: [ITHREAD] >>> >>> in case you need more, the complete log is available here: >>> http://smeets.im/~flo/ml350g4p.dmesg.boot >>> >> >> Thanks for the info. I don't see any abnormal things there. >> If the 100Mbps switch provides some management functions, could you >> check resolved link state on the switch? >> And could you reproduce the issue with other 100Mbps switch? >> > > Any progress? The server was connected to the 100Mbps switch only during installation, ever since it's in production it's connected to 1000Mbps switch and working fine, so i cannot test it anymore. Sorry. State-Changed-From-To: feedback->closed State-Changed-By: yongari State-Changed-When: Wed Sep 22 19:19:24 UTC 2010 State-Changed-Why: Close, submitter is not able to test against 100Mbps switch. Because there was a lot of PHY changes since 4.7-RELEASE to address several silicon bug of Broadcom I think it might be fixed long time ago. At least there was no link establishment issue for copper PHY in past several years. It's also possible that the link parter's switch is faulty but it's hard to confirm it as submitter is not able to test. If you happen to encounter the issue again please open a new PR. Thanks for reporting! http://www.freebsd.org/cgi/query-pr.cgi?pr=45773 >Unformatted: