[Xen-users] PCIe NIC pass-through problem with Debian Squeeze

["David MÃller (ELSOFT AG)" <d.mueller@xxxxxxxxx>]

Hello

I have installed Debian Squeeze (6.0.1) Linux-2.6.32-5-xen-686 /
Xen-4.0.1-2 on the following hardware:
- Core2 Duo
- 945GME + ICH7
- 2x 82573L
- 1GB RAM

The second Gbit NIC is passed-through to the DomU, but shortly after i
bring up this network interface in the DomU, the following problem occurs:

[   31.133830] ADDRCONF(NETDEV_UP): eth1: link is not ready
[   33.737036] e1000e: eth1 NIC Link is Up 1000 Mbps Full Duplex, Flow
Control: RX/TX
[   33.737219] ADDRCONF(NETDEV_CHANGE): eth1: link becomes ready
[   35.816679] 0000:00:00.0: eth1: Detected Tx Unit Hang:
[   35.816686]   TDH                  <4>
[   35.816690]   TDT                  <4>
[   35.816694]   next_to_use          <4>
[   35.816697]   next_to_clean        <0>
[   35.816701] buffer_info[next_to_clean]:
[   35.816705]   time_stamp           <fffefbfc>
[   35.816708]   next_to_watch        <0>
[   35.816712]   jiffies              <fffefe02>
[   35.816716]   next_to_watch.status <0>
[   37.816264] 0000:00:00.0: eth1: Detected Tx Unit Hang:
[   37.816269]   TDH                  <4>
[   37.816273]   TDT                  <4>
[   37.816277]   next_to_use          <4>
[   37.816280]   next_to_clean        <0>
[   37.816284] buffer_info[next_to_clean]:
[   37.816287]   time_stamp           <fffefbfc>
[   37.816291]   next_to_watch        <0>
[   37.816295]   jiffies              <fffefff6>
[   37.816299]   next_to_watch.status <0>
[   39.816204] 0000:00:00.0: eth1: Detected Tx Unit Hang:
[   39.816211]   TDH                  <4>
[   39.816215]   TDT                  <4>
[   39.816218]   next_to_use          <4>
[   39.816222]   next_to_clean        <0>
[   39.816226] buffer_info[next_to_clean]:
[   39.816229]   time_stamp           <fffefbfc>
[   39.816233]   next_to_watch        <0>
[   39.816237]   jiffies              <ffff01ea>
[   39.816240]   next_to_watch.status <0>
[   41.816231] 0000:00:00.0: eth1: Detected Tx Unit Hang:
[   41.816236]   TDH                  <4>
[   41.816240]   TDT                  <4>
[   41.816244]   next_to_use          <4>
[   41.816247]   next_to_clean        <0>
[   41.816251] buffer_info[next_to_clean]:
[   41.816255]   time_stamp           <fffefbfc>
[   41.816258]   next_to_watch        <0>
[   41.816262]   jiffies              <ffff03de>
[   41.816266]   next_to_watch.status <0>
[   43.816143] ------------[ cut here ]------------
[   43.816178] WARNING: at
/tmp/buildd/linux-2.6-2.6.32/debian/build/source_i386_xen/net/sched/sch_generic.c:261
dev_watchdog+0xbd/0x15d()
[   43.816208] NETDEV WATCHDOG: eth1 (e1000e): transmit queue 0 timed out
[   43.816227] Modules linked in: evdev snd_pcm snd_timer snd soundcore
snd_page_alloc pcspkr ext3 jbd mbcache e1000e xen_blkfront xen_netfront
[   43.816323] Pid: 0, comm: swapper Not tainted 2.6.32-5-xen-686 #1
[   43.816341] Call Trace:
[   43.816361]  [<c1208525>] ? dev_watchdog+0xbd/0x15d
[   43.816383]  [<c1208525>] ? dev_watchdog+0xbd/0x15d
[   43.816406]  [<c1037739>] ? warn_slowpath_common+0x5e/0x8a
[   43.816428]  [<c1208468>] ? dev_watchdog+0x0/0x15d
[   43.816449]  [<c1037797>] ? warn_slowpath_fmt+0x26/0x2a
[   43.816470]  [<c1208525>] ? dev_watchdog+0xbd/0x15d
[   43.816493]  [<c1006764>] ? check_events+0x8/0xc
[   43.816515]  [<c100675b>] ? xen_restore_fl_direct_end+0x0/0x1
[   43.816538]  [<c128f319>] ? _spin_unlock_irqrestore+0xd/0xf
[   43.816561]  [<c1034512>] ? try_to_wake_up+0x2a7/0x2b3
[   43.816584]  [<c10424a8>] ? run_timer_softirq+0x16a/0x1eb
[   43.816607]  [<c103c80a>] ? __do_softirq+0xaa/0x156
[   43.816628]  [<c103c8e7>] ? do_softirq+0x31/0x3c
[   43.816648]  [<c103c9c1>] ? irq_exit+0x26/0x58
[   43.816670]  [<c1199a8a>] ? xen_evtchn_do_upcall+0x22/0x2c
[   43.816692]  [<c1009b3f>] ? xen_do_upcall+0x7/0xc
[   43.816713]  [<c10023a7>] ? hypercall_page+0x3a7/0x1001
[   43.816734]  [<c10060b5>] ? xen_safe_halt+0xf/0x1b
[   43.816755]  [<c10041cb>] ? xen_idle+0x23/0x30
[   43.816774]  [<c1008128>] ? cpu_idle+0x89/0xa5
[   43.816795]  [<c13fb817>] ? start_kernel+0x318/0x31d
[   43.816816]  [<c13fd410>] ? xen_start_kernel+0x564/0x56b
[   43.816838]  [<c1409045>] ? check_nmi_watchdog+0xe5/0x1f2
[   43.816856] ---[ end trace 852b94cb7b0e868a ]---
[   43.960037] eth1: no IPv6 routers present
[   46.577034] e1000e: eth1 NIC Link is Up 1000 Mbps Full Duplex, Flow
Control: RX/TX



On the same system, a Debian Lenny (5.0.8) installation based on
Linux-2.6.26-2-xen-686 / Xen-3.2-1 works without problems.


I'm not sure if this is related in any way to this Debian bug:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=518182


Any ideas?


Dave

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users