<div dir="ltr"><br><div class="gmail_quote"><div dir="ltr"><span style="font-size:13px">Hi,</span><div style="font-size:13px">I'm interested in finding out if anyone has successfully tested PCI passthrough functionality for 40G interfaces in Openstack(KVM hypervisor).</div><div style="font-size:13px"><br></div><div style="font-size:13px">I am trying this out on a Fedora 21 host with Fedora 21 VM image.(3.18.7-200.fc21.x86_64)</div><div style="font-size:13px"><br></div><div style="font-size:13px">Was able to successfully test PCI passthrough of 10 G interfaces:</div><div style="font-size:13px"> Ethernet controller: Intel Corporation 82599ES 10-Gigabit SFI/SFP+ Network Connection (rev 01)</div><div style="font-size:13px"><br></div><div style="font-size:13px">With 40G interface testing, the PCI device is passed through to the VM but data transfer is failing.</div><div style="font-size:13px"> 0a:00.1 Ethernet controller: Intel Corporation Ethernet Controller XL710 for 40GbE QSFP+ (rev 01)</div><div style="font-size:13px"><br></div><div style="font-size:13px">Tried this with both the i40e driver and latest dpdk driver but no luck so far.</div><div style="font-size:13px"><br></div><div style="font-size:13px">With the i40e driver, the data transfer fails.</div><div style="font-size:13px"> Relevant dmesg output:</div><div style="font-size:13px"> [ 11.544088] i40e 0000:00:05.0 eth1: NIC Link is Up 40 Gbps Full Duplex, Flow Control: None</div><div style="font-size:13px">[ 11.689178] i40e 0000:00:06.0 eth2: NIC Link is Up 40 Gbps Full Duplex, Flow Control: None</div><div style="font-size:13px">[ 16.704071] ------------[ cut here ]------------</div><div style="font-size:13px">[ 16.705053] WARNING: CPU: 1 PID: 0 at net/sched/sch_generic.c:303 dev_watchdog+0x23e/0x250()</div><div style="font-size:13px">[ 16.705053] NETDEV WATCHDOG: eth1 (i40e): transmit queue 1 timed out</div><div style="font-size:13px">[ 16.705053] Modules linked in: cirrus ttm drm_kms_helper i40e drm ppdev serio_raw i2c_piix4 virtio_net parport_pc ptp virtio_balloon crct10dif_pclmul pps_core parport pvpanic crc32_pclmul ghash_clmulni_intel virtio_blk crc32c_intel virtio_pci virtio_ring virtio ata_generic pata_acpi</div><div style="font-size:13px">[ 16.705053] CPU: 1 PID: 0 Comm: swapper/1 Not tainted 3.18.7-200.fc21.x86_64 #1</div><div style="font-size:13px">[ 16.705053] Hardware name: Fedora Project OpenStack Nova, BIOS 1.7.5-20140709_153950- 04/01/2014</div><div style="font-size:13px">[ 16.705053] 0000000000000000 2e5932b294d0c473 ffff88043fc83d48 ffffffff8175e686</div><div style="font-size:13px">[ 16.705053] 0000000000000000 ffff88043fc83da0 ffff88043fc83d88 ffffffff810991d1</div><div style="font-size:13px">[ 16.705053] ffff88042958f5c0 0000000000000001 ffff88042865f000 0000000000000001</div><div style="font-size:13px">[ 16.705053] Call Trace:</div><div style="font-size:13px">[ 16.705053] <IRQ> [<ffffffff8175e686>] dump_stack+0x46/0x58</div><div style="font-size:13px">[ 16.705053] [<ffffffff810991d1>] warn_slowpath_common+0x81/0xa0</div><div style="font-size:13px">[ 16.705053] [<ffffffff81099245>] warn_slowpath_fmt+0x55/0x70</div><div style="font-size:13px">[ 16.705053] [<ffffffff8166e62e>] dev_watchdog+0x23e/0x250</div><div style="font-size:13px">[ 16.705053] [<ffffffff8166e3f0>] ? dev_graft_qdisc+0x80/0x80</div><div style="font-size:13px">[ 16.705053] [<ffffffff810fd52a>] call_timer_fn+0x3a/0x120</div><div style="font-size:13px">[ 16.705053] [<ffffffff8166e3f0>] ? dev_graft_qdisc+0x80/0x80</div><div style="font-size:13px">[ 16.705053] [<ffffffff810ff692>] run_timer_softirq+0x212/0x2f0</div><div style="font-size:13px">[ 16.705053] [<ffffffff8109d7a4>] __do_softirq+0x124/0x2d0</div><div style="font-size:13px">[ 16.705053] [<ffffffff8109db75>] irq_exit+0x125/0x130</div><div style="font-size:13px">[ 16.705053] [<ffffffff817681d8>] smp_apic_timer_interrupt+0x48/0x60</div><div style="font-size:13px">[ 16.705053] [<ffffffff817662bd>] apic_timer_interrupt+0x6d/0x80</div><div style="font-size:13px">[ 16.705053] <EOI> [<ffffffff811005c8>] ? hrtimer_start+0x18/0x20</div><div style="font-size:13px">[ 16.705053] [<ffffffff8105ca96>] ? native_safe_halt+0x6/0x10</div><div style="font-size:13px">[ 16.705053] [<ffffffff810f81d3>] ? rcu_eqs_enter+0xa3/0xb0</div><div style="font-size:13px">[ 16.705053] [<ffffffff8101ec7f>] default_idle+0x1f/0xc0</div><div style="font-size:13px">[ 16.705053] [<ffffffff8101f64f>] arch_cpu_idle+0xf/0x20</div><div style="font-size:13px">[ 16.705053] [<ffffffff810dad35>] cpu_startup_entry+0x3c5/0x410</div><div style="font-size:13px">[ 16.705053] [<ffffffff8104a2af>] start_secondary+0x1af/0x1f0</div><div style="font-size:13px">[ 16.705053] ---[ end trace 7bda53aeda558267 ]---</div><div style="font-size:13px">[ 16.705053] i40e 0000:00:05.0 eth1: tx_timeout recovery level 1</div><div style="font-size:13px">[ 16.705053] i40e 0000:00:05.0: i40e_vsi_control_tx: VSI seid 519 Tx ring 0 disable timeout</div><div style="font-size:13px">[ 16.744198] i40e 0000:00:05.0: i40e_vsi_control_tx: VSI seid 520 Tx ring 64 disable timeout</div><div style="font-size:13px">[ 16.779322] i40e 0000:00:05.0: i40e_ptp_init: added PHC on eth1</div><div style="font-size:13px">[ 16.791819] i40e 0000:00:05.0: PF 40 attempted to control timestamp mode on port 1, which is owned by PF 1</div><div style="font-size:13px">[ 16.933869] i40e 0000:00:05.0 eth1: NIC Link is Up 40 Gbps Full Duplex, Flow Control: None</div><div style="font-size:13px">[ 18.853624] SELinux: initialized (dev tmpfs, type tmpfs), uses transition SIDs</div><div style="font-size:13px">[ 22.720083] i40e 0000:00:05.0 eth1: tx_timeout recovery level 2</div><div style="font-size:13px">[ 22.826993] i40e 0000:00:05.0: i40e_vsi_control_tx: VSI seid 519 Tx ring 0 disable timeout</div><div style="font-size:13px">[ 22.935288] i40e 0000:00:05.0: i40e_vsi_control_tx: VSI seid 520 Tx ring 64 disable timeout</div><div style="font-size:13px">[ 23.669555] i40e 0000:00:05.0: i40e_ptp_init: added PHC on eth1</div><div style="font-size:13px">[ 23.682067] i40e 0000:00:05.0: PF 40 attempted to control timestamp mode on port 1, which is owned by PF 1</div><div style="font-size:13px">[ 23.722423] i40e 0000:00:05.0 eth1: NIC Link is Up 40 Gbps Full Duplex, Flow Control: None</div><div style="font-size:13px">[ 23.800206] i40e 0000:00:06.0: i40e_ptp_init: added PHC on eth2</div><div style="font-size:13px">[ 23.813804] i40e 0000:00:06.0: PF 48 attempted to control timestamp mode on port 0, which is owned by PF 0</div><div style="font-size:13px">[ 23.855275] i40e 0000:00:06.0 eth2: NIC Link is Up 40 Gbps Full Duplex, Flow Control: None</div><div style="font-size:13px">[ 38.720091] i40e 0000:00:05.0 eth1: tx_timeout recovery level 3</div><div style="font-size:13px">[ 38.725844] random: nonblocking pool is initialized</div><div style="font-size:13px">[ 38.729874] i40e 0000:00:06.0: HMC error interrupt</div><div style="font-size:13px">[ 38.733425] i40e 0000:00:06.0: i40e_vsi_control_tx: VSI seid 518 Tx ring 0 disable timeout</div><div style="font-size:13px">[ 38.738886] i40e 0000:00:06.0: i40e_vsi_control_tx: VSI seid 521 Tx ring 64 disable timeout</div><div style="font-size:13px">[ 39.689569] i40e 0000:00:06.0: i40e_ptp_init: added PHC on eth2</div><div style="font-size:13px">[ 39.704197] i40e 0000:00:06.0: PF 48 attempted to control timestamp mode on port 0, which is owned by PF 0</div><div style="font-size:13px">[ 39.746879] i40e 0000:00:06.0 eth2: NIC Link is Down</div><div style="font-size:13px">[ 39.838356] i40e 0000:00:05.0: i40e_ptp_init: added PHC on eth1</div><div style="font-size:13px">[ 39.851788] i40e 0000:00:05.0: PF 40 attempted to control timestamp mode on port 1, which is owned by PF 1</div><div style="font-size:13px">[ 39.892822] i40e 0000:00:05.0 eth1: NIC Link is Down</div><div style="font-size:13px">[ 43.011610] i40e 0000:00:06.0 eth2: NIC Link is Up 40 Gbps Full Duplex, Flow Control: None</div><div style="font-size:13px">[ 43.059976] i40e 0000:00:05.0 eth1: NIC Link is Up 40 Gbps Full Duplex, Flow Control: None</div><div><br></div><div style="font-size:13px"><br></div><div style="font-size:13px">Similarly, with dpdk driver, no packet transfer is happening.</div><div style="font-size:13px"><br></div><div style="font-size:13px"><br></div><div style="font-size:13px">The host was booted with iommu=pt intel_iommu=on.</div><div style="font-size:13px"><br></div><div style="font-size:13px">It is to be noted that everything works fine on the host itself. The issue is seen only when the devices are passed through to the VM.</div><div style="font-size:13px"><br></div><div style="font-size:13px">Would really appreciate any additional information on how to address or further debug this issue.</div><div style="font-size:13px"><br></div><div style="font-size:13px">Thanks</div><div style="font-size:13px"></div></div>
</div>Jacob</div>