From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: BUG: IPv6 stops working after a while, needs ip ne del command to reset Date: Wed, 01 Sep 2010 15:19:28 +0200 Message-ID: <1283347168.2556.319.camel@edumazet-laptop> References: <1282024802.2487.687.camel@edumazet-laptop> <1282050920.2448.47.camel@edumazet-laptop> <1282055659.2448.58.camel@edumazet-laptop> <20100817171115.GA4134@mcarlson.broadcom.com> <20100817183112.GA4351@mcarlson.broadcom.com> <20100818012318.GA4630@mcarlson.broadcom.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Matt Carlson , "linux-kernel@vger.kernel.org" , netdev , Michael Chan To: Thomas Habets Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org Le mercredi 01 septembre 2010 =C3=A0 11:21 +0200, Thomas Habets a =C3=A9= crit : > I've continued this a bit off-list but thought I would summarize for = the=20 > archives. >=20 >=20 > Summary > ------- > It looks like a firmware issue on the network card. When ILO is enabl= ed it=20 > shares the first network card with the OS. When it does this multicas= t=20 > is broken. When multicast (on a L2 level) is broken IPv6 neighbor=20 > discovery breaks. Only eth0 breaks, eth1 is unaffected. >=20 >=20 > System > ------ > HP Proliant DL320 G5p > Xeon 3GHz > 1GB RAM > Arch: amd64 > NIC: Broadcom Corporation NetXtreme BCM5715 Gigabit Ethernet (rev a3) > Debian Lenny (5.0.5) > Kernels: 2.6.35 mainline, 2.6.33.6 > Config: http://pastebin.com/raw.php?i=3DY6S8iKW7 >=20 >=20 > Problem > ------- > Buggy box will not answer IPv6 ND or ping to ff02::1. May work at som= e=20 > point in the boot process, but once box is fully booted it does not. >=20 > If I on the neighboring Cisco router run "clear ipv6 neighbors" (or i= t=20 > times out) that router cannot re-acquire the neigborship with the bug= gy=20 > box. Instant IPv6 breakage until I do one of: > * Turn on promisc mode long enough for IPv6 ND to do its thing > * ip ne del
on the buggy host. >=20 >=20 > Workarounds > ----------- > Either one of these will hide the problem: > * Set promisc mode on interface (ip link set promisc on eth0) forever > * Disable ILO > * Use eth1 instead of eth0. >=20 >=20 > Troubleshooting > --------------- > Got patch for kernel from Eric Dumazet (eric.dumazet@gmail.com) to ou= tput=20 > what MAC addresses are being subscribed to, and some registers from t= he=20 > card. Output is earlier in this thread, along with "ethtool -i eth0" = and=20 > some other data. >=20 > Managed to get diagnostic tool[1] booting from stick (no CD drive in=20 > server), but did not set up memory (himem.sys etc..). Running b57udia= g=20 > it therefore failed due to insufficient memory at test "Group D. Driv= er=20 > Associated tests". Card is assumed to be OK anyway. >=20 > Matt Carlson (mcarlson@broadcom.com) suspected firmware bug and asked= me=20 > to try disabling ASF and/or IPMI using the diagnostic tool, but runni= ng=20 > "setasf -d" and "setipmi -d" inside "b57udiag -cmd" did not seem to s= tick=20 > across reboot. It stuck properly before reboot (confirmed with setasf= -q).=20 > Also tried "b57udiag -u 0". Tried both C-A-D reboot and powercycling = (by=20 > power cord). >=20 > At boot Linux still said ASF[1] for eth0 and ASF[0] for eth1: > tg3 0000:03:04.0: eth0: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[1] TSOc= ap[1] > tg3 0000:03:04.1: eth1: RXcsums[1] LinkChgREG[0] MIirq[0] ASF[0] TSOc= ap[1] > (this output never changed throughout the process) > ethtool -d eth1 | grep 0x047 did not change either. >=20 > Then I disabled ILO and PXE in ILO bios and BIOS respectively. That f= ixed=20 > it. eth0 now works with multicast. >=20 > I don't use ILO on this server so in this case that fixes it for me, = but=20 > the bug is still there. >=20 > At this point Matt thinks I should file a bug report with HP. I will=20 > attempt to do that. >=20 > I have more detailed logs of what I did and when, and what the effect= was. >=20 >=20 > Related > ------- > May be the same issue as this: > https://bugs.launchpad.net/ubuntu/+source/linux/+bug/263260 > Which means it's the same with Ubuntu kernels 2.6.26.3, 2.6.26-5-gene= ric=20 > and 2.6.27-2-generic, and mainline kernels 2.6.25, 2.6.26 and 2.6.27. >=20 >=20 > [1] http://www.broadcom.com/support/ethernet_nic/netxtreme_server.php >=20 Thanks a lot Thomas for this very detailed report !