[Oneiric] Xen HVM: transmit timeouts on emulated 8139cp device
Bug #854829 reported by
Stefan Bader
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
xen (Ubuntu) |
Fix Released
|
Medium
|
Chuck Short | ||
Oneiric |
Fix Released
|
Medium
|
Chuck Short | ||
Precise |
Fix Released
|
Undecided
|
Unassigned |
Bug Description
Using an Oneiric dom0 and an Oneiric HVM domU, the network comes initially up, DHCP succeeds and the interface has an IP address assigned. Pings usually are successful. But any higher traffic (like doing an apt-get update) causes the network to hang. After a while the net device watchdog triggers and resets the connection. In dmesg there are also messages like this:
8139cp: 0000:00:04.0: eth0: Transmit timeout, status d 3b 5 80ff
Changed in xen (Ubuntu): | |
assignee: | nobody → Stefan Bader (stefan-bader-canonical) |
importance: | Undecided → Medium |
status: | New → Confirmed |
tags: | added: server-o-ro |
Changed in xen (Ubuntu): | |
assignee: | Stefan Bader (stefan-bader-canonical) → Chuck Short (zulcss) |
status: | Confirmed → Triaged |
Changed in xen (Ubuntu Precise): | |
status: | New → Fix Released |
To post a comment you must log in.
Testing various emulated devices shows:
- ne2k_pci gets the tx problems almost immediatly
- 8139cp on any slightly higher traffic
- e1000 seems to work
When pirq's are not used, 8139cp also works again. So there seems to be a subtle bug in the delivery of those.