[Bug 668390] New: Two 'real' ethernet cards PCI-Passthrough'd to a DomU guest report/use the *same* HWaddr
https://bugzilla.novell.com/show_bug.cgi?id=668390 https://bugzilla.novell.com/show_bug.cgi?id=668390#c0 Summary: Two 'real' ethernet cards PCI-Passthrough'd to a DomU guest report/use the *same* HWaddr Classification: openSUSE Product: openSUSE 11.3 Version: Final Platform: x86-64 OS/Version: openSUSE 11.3 Status: NEW Severity: Normal Priority: P5 - None Component: Xen AssignedTo: jdouglas@novell.com ReportedBy: pgngw+dev001+novell.com@f-m.fm QAContact: qa@suse.de Found By: --- Blocker: --- User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.2.13) Gecko/20101203 SUSE/3.6.13-3.1 Firefox/3.6.13 i'm running an openSUSE Xen Host, uname -a Linux test 2.6.34.8-7-xen #1 SMP 2011-01-20 21:58:39 +0100 x86_64 x86_64 x86_64 GNU/Linux lsb_release -d Description: openSUSE 11.3 (x86_64) rpm -qa | egrep -i "^kernel-xen|^xen-4" xen-4.0.1_21326_02-0.3.1.x86_64 kernel-xen-2.6.34.8-7.1.x86_64 cat /boot/grub/menu.lst .. title Xen root (hd0,0) kernel /xen.gz loglvl=all loglvl_guest=all apic_verbosity=debug iommu=verbose dom0_mem=1024M vga=gfx-1280x1024x32 console_timestamps console=com1 com1=57600 dom0_max_vcpus=4 dom0_vcpus_pin=true sched=credit cpuidle=1 cpufreq=xen clocksource=acpi numa=on module /vmlinuz-xen root=/dev/VG0/LV_DOM0_ROOT rootfstype=ext4 rootflags=journal_checksum resume=/dev/VG0/LV_DOM0_SWAP iommu=verbose showopts selinux=0 SELINUX_INIT=NO apparmor=0 elevator=cfq vga=0x31a xencons=ttyS0 console=ttyS0 console=xvc0,57600 clocksource=xen mce=off pci=resource_alignment=02:00.0;04:00.0 ignore_loglevel nomodeset module /initrd-xen .. i'm passing through two ethernet cards to a DomU guest, cat TEST.cfg ... vif = [ 'mac=00:16:3E:02:02:00, bridge=brINT, vifname=vifTEST.int'] vfb = [ 'type=vnc, vncdisplay=202, vnclisten=127.0.0.1' ] extra = 'textmode=1 xencons=xvc0 elevator=noop swiotlb=128,force' pci = [ '02:00.0', '04:00.0'] ... which show up in my guest as expected, lspci 00:00.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller (rev 01) 00:01.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller (rev 02) with @ guest configs of, cat /etc/sysconfig/network/ifcfg-eth1 STARTMODE='auto' BOOTPROTO='static' ETHTOOL_OPTIONS='autoneg on' IPADDR='169.254.11.202/16' PREFIXLEN='' BROADCAST='169.254.255.255' DHCPCLASS= NETWORK='169.254.0.0' MTU='1500' USERCONTROL='no' NM_CONTROLLED='no' IPV6INIT='no' IPV6_AUTOCONF='no' STARTMODE='auto' cat /etc/sysconfig/network/ifcfg-eth2 STARTMODE='auto' BOOTPROTO='static' ETHTOOL_OPTIONS='autoneg on' IPADDR='10.1.1.202/24' PREFIXLEN='' BROADCAST='10.1.1.255' DHCPCLASS= NETWORK='10.1.1.0' MTU='1500' USERCONTROL='no' NM_CONTROLLED='no' IPV6INIT='no' IPV6_AUTOCONF='no' then, checking: ifconfig | egrep "HWaddr|Loopback|inet addr" eth0 Link encap:Ethernet HWaddr 00:16:3E:02:02:00 inet addr:192.168.1.202 Bcast:192.168.1.255 Mask:255.255.255.0 eth1 Link encap:Ethernet HWaddr 00:24:8C:BC:DD:E0 inet addr:169.254.11.202 Bcast:169.254.255.255 Mask:255.255.0.0 eth2 Link encap:Ethernet HWaddr 00:24:8C:BC:DD:E0 inet addr:10.1.1.202 Bcast:10.1.1.255 Mask:255.255.255.0 lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 eth0 is the bridge from Host, and lo is loopback. both fine. note, however BOTH eth1 & eth2, the passed-through ethernet cards, are reporting the SAME HWaddr !? afaict, in the PCI-passthrough case, Xen Guest gets its HWaddr from the actual hardware. is the two-cards-report/use-same-HWaddr a bug? or a misconfig somewhere? Reproducible: Always Steps to Reproduce: 1. 2. 3. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=668390
https://bugzilla.novell.com/show_bug.cgi?id=668390#c1
--- Comment #1 from dev001x _
https://bugzilla.novell.com/show_bug.cgi?id=668390
https://bugzilla.novell.com/show_bug.cgi?id=668390#c2
--- Comment #2 from dev001x _
https://bugzilla.novell.com/show_bug.cgi?id=668390
https://bugzilla.novell.com/show_bug.cgi?id=668390#c3
--- Comment #3 from dev001x _
https://bugzilla.novell.com/show_bug.cgi?id=668390
https://bugzilla.novell.com/show_bug.cgi?id=668390#c4
Jan Beulich
https://bugzilla.novell.com/show_bug.cgi?id=668390
https://bugzilla.novell.com/show_bug.cgi?id=668390#c5
dev001x _
see full kernel boot logs for both the case where the devices remain in Dom0 and the DomU case mentioned above
in an attempt to get those ...
(for which it would also be nice to know that we're talking about a PV one).
yes, it's a PV DomU
individually pass through only one of the two, each time noting the address used once the guest is up.
i did that, and the Mac remained @ 00:24:8C:BC:DD:E0 for each individually. ?? on a whim, i removed/reseated all my PCI cards -- no change, same Mac addr, as above. i next removed all my cards, and juggled their slot order. to my surprise, the Mac addr issue is gone. each card has a unique Mac addr, whether one/all cards are seated &/or passed through; @ guest, dmesg | grep RTL81 | grep eth [ 1.852320] r8169 0000:01:00.0: eth1: RTL8168b/8111b at 0xffffc90010908000, 00:0a:cd:17:56:c6, XID 18000000 IRQ 436 [ 1.854451] r8169 0000:02:00.0: eth2: RTL8168b/8111b at 0xffffc9001090c000, 00:52:31:47:13:ef, XID 1c4000c0 IRQ 435 unlike prior to the order-juggle, where reproducing the problem was 100%, now i can't reproduce the problem above at all. finally, i returned the cards to their original order, and still everything is OK, now. it acts like the reorder/juggles cleared a cache somewhere ... but, tbh, huh? so, if noting this issue/resolution is of some value, great; but at this point i simply can't further reproduce this at all. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=668390
https://bugzilla.novell.com/show_bug.cgi?id=668390#c6
--- Comment #6 from Olaf Hering
https://bugzilla.novell.com/show_bug.cgi?id=668390
https://bugzilla.novell.com/show_bug.cgi?id=668390#c7
--- Comment #7 from dev001x _
Were the MAC addresses also identical before you started with Xen, in other words: did they ever differ before?
yes, they differed. originally, i'd used all the cards at Dom0, and Mac addr's were, as usual/expected, different. around (i can't tell you exactly when) the time i started working on pci-passthrough of these cards, this behavior started happening. i'd really like to backtrack/reproduce to figure this out, but simply can't make it misbehave now. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.
https://bugzilla.novell.com/show_bug.cgi?id=668390
https://bugzilla.novell.com/show_bug.cgi?id=668390#c8
Jan Beulich
participants (1)
-
bugzilla_noreply@novell.com