[dpdk-users] new_device is never created at the vhost-switch sample app

[dpdk-users] new_device is never created at the vhost-switch sample app

[edgar helmut]

Hi,
I am following http://dpdk.org/doc/guides/sample_app_ug/vhost.html using
16.11 in order to have a simple VM forwarding packets from one NIC to
another efficiently, but it doesn't work... so I need some help to
understand what am i missing.
I am using x520 (intel 82599).
host is ubuntu 16.04 and dpdk 16.11.

my steps at host:
1. enabling VHOST defines and building x86_64-native-linuxapp-gcc.

2. both interfaces are bind like:
0000:04:00.0 'Ethernet 10G 2P X520 Adapter' drv=igb_uio unused=ixgbe
0000:04:00.1 'Ethernet 10G 2P X520 Adapter' drv=igb_uio unused=ixgbe

3. then making the project examples/vhost and executing:
./build/vhost-switch -c 0x03 -n 4 --socket-mem 1024 --file-prefix p1  -- -p
1 --vm2vm 0 --socket-file /tmp/sock1 --client -P
./build/vhost-switch -c 0x30 -n 4 --socket-mem 1024 --file-prefix p2  -- -p
2 --vm2vm 0 --socket-file /tmp/sock2 --client -P

4. creating the guest
qemu- system-x86_64 -chardev socket,id=char1,path=/tmp/sock1,server -netdev
type=vhost-user,id=hostnet1,chardev=char1,queues=2 -device
virtio-net-pci,mq=on,vectors=2,netdev=hostnet1,id=net1,mac=00:00:00:00:00:01
-chardev socket,id=char2,path=/tmp/sock2,server -netdev
type=vhost-user,id=hostnet2,chardev=char2,queues=2 -device
virtio-net-pci,mq=on,vectors=2,netdev=hostnet2,id=net2,mac=00:00:00:00:00:02
-object
memory-backend-file,id=mem2,size=2048M,mem-path=/run/hugepages,share=on  ...

my steps on the guest:
1. enabling VHOST defines and building x86_64-native-linuxapp-gcc.

2. successfully setting up environment:
modprobe uio_pci_generic
./tools/dpdk-devbind.py -b uio_pci_generic 0000:00:07.0
./tools/dpdk-devbind.py -b uio_pci_generic 0000:00:08.0
./examples/l3fwd/build/l3fwd -c 0x3 -n 3 -- --parse-ptype -p 0x3 -P -L
--config="(0,0,0),(0,1,1),(1,0,0),(1,1,1)" --no-numa
it looks like the l3fwd app initialized successfully but it receives no
traffic.

I see at the host that connection is made however no traffic arrives to the
host.
digging into the code i see that the message VHOST_USER_SET_VRING_KICK
never arrives to the host hence the callback new_device is never called and
I can't figure out why.

Any ideas?

Thanks,
Edgar

Re: [dpdk-users] new_device is never created at the vhost-switch sample app

[edgar helmut]

Hi,
the silence confuses me, at least I would like to understand if it is well
supported or not for the release so I can decide whether to find an
alternative or not.
what is the trigger for the guest to send the VHOST_USER_SET_VRING_KICK?

Thanks
Edgar

On Fri, Nov 25, 2016 at 9:58 AM edgar helmut <helmut.edgar100@gmail.com>
wrote:

> Hi,
> I am following http://dpdk.org/doc/guides/sample_app_ug/vhost.html using
> 16.11 in order to have a simple VM forwarding packets from one NIC to
> another efficiently, but it doesn't work... so I need some help to
> understand what am i missing.
> I am using x520 (intel 82599).
> host is ubuntu 16.04 and dpdk 16.11.
>
> my steps at host:
> 1. enabling VHOST defines and building x86_64-native-linuxapp-gcc.
>
> 2. both interfaces are bind like:
> 0000:04:00.0 'Ethernet 10G 2P X520 Adapter' drv=igb_uio unused=ixgbe
> 0000:04:00.1 'Ethernet 10G 2P X520 Adapter' drv=igb_uio unused=ixgbe
>
> 3. then making the project examples/vhost and executing:
> ./build/vhost-switch -c 0x03 -n 4 --socket-mem 1024 --file-prefix p1  --
> -p 1 --vm2vm 0 --socket-file /tmp/sock1 --client -P
> ./build/vhost-switch -c 0x30 -n 4 --socket-mem 1024 --file-prefix p2  --
> -p 2 --vm2vm 0 --socket-file /tmp/sock2 --client -P
>
> 4. creating the guest
> qemu- system-x86_64 -chardev socket,id=char1,path=/tmp/sock1,server
> -netdev type=vhost-user,id=hostnet1,chardev=char1,queues=2 -device
> virtio-net-pci,mq=on,vectors=2,netdev=hostnet1,id=net1,mac=00:00:00:00:00:01
> -chardev socket,id=char2,path=/tmp/sock2,server -netdev
> type=vhost-user,id=hostnet2,chardev=char2,queues=2 -device
> virtio-net-pci,mq=on,vectors=2,netdev=hostnet2,id=net2,mac=00:00:00:00:00:02
> -object
> memory-backend-file,id=mem2,size=2048M,mem-path=/run/hugepages,share=on  ...
>
> my steps on the guest:
> 1. enabling VHOST defines and building x86_64-native-linuxapp-gcc.
>
> 2. successfully setting up environment:
> modprobe uio_pci_generic
> ./tools/dpdk-devbind.py -b uio_pci_generic 0000:00:07.0
> ./tools/dpdk-devbind.py -b uio_pci_generic 0000:00:08.0
> ./examples/l3fwd/build/l3fwd -c 0x3 -n 3 -- --parse-ptype -p 0x3 -P -L
> --config="(0,0,0),(0,1,1),(1,0,0),(1,1,1)" --no-numa
> it looks like the l3fwd app initialized successfully but it receives no
> traffic.
>
> I see at the host that connection is made however no traffic arrives to
> the host.
> digging into the code i see that the message VHOST_USER_SET_VRING_KICK
> never arrives to the host hence the callback new_device is never called and
> I can't figure out why.
>
> Any ideas?
>
> Thanks,
> Edgar
>

Re: [dpdk-users] new_device is never created at the vhost-switch sample app

[gowrishankar muthukrishnan]

On Friday 25 November 2016 01:28 PM, edgar helmut wrote:
> Hi,
> I am following http://dpdk.org/doc/guides/sample_app_ug/vhost.html using
> 16.11 in order to have a simple VM forwarding packets from one NIC to
> another efficiently, but it doesn't work... so I need some help to
> understand what am i missing.
> I am using x520 (intel 82599).
> host is ubuntu 16.04 and dpdk 16.11.
>
> my steps at host:
> 1. enabling VHOST defines and building x86_64-native-linuxapp-gcc.
>
> 2. both interfaces are bind like:
> 0000:04:00.0 'Ethernet 10G 2P X520 Adapter' drv=igb_uio unused=ixgbe
> 0000:04:00.1 'Ethernet 10G 2P X520 Adapter' drv=igb_uio unused=ixgbe
>
> 3. then making the project examples/vhost and executing:
> ./build/vhost-switch -c 0x03 -n 4 --socket-mem 1024 --file-prefix p1  -- -p
> 1 --vm2vm 0 --socket-file /tmp/sock1 --client -P
> ./build/vhost-switch -c 0x30 -n 4 --socket-mem 1024 --file-prefix p2  -- -p
> 2 --vm2vm 0 --socket-file /tmp/sock2 --client -P
>
> 4. creating the guest
> qemu- system-x86_64 -chardev socket,id=char1,path=/tmp/sock1,server -netdev
> type=vhost-user,id=hostnet1,chardev=char1,queues=2 -device
> virtio-net-pci,mq=on,vectors=2,netdev=hostnet1,id=net1,mac=00:00:00:00:00:01
> -chardev socket,id=char2,path=/tmp/sock2,server -netdev
> type=vhost-user,id=hostnet2,chardev=char2,queues=2 -device
> virtio-net-pci,mq=on,vectors=2,netdev=hostnet2,id=net2,mac=00:00:00:00:00:02
> -object
> memory-backend-file,id=mem2,size=2048M,mem-path=/run/hugepages,share=on  ...
>
> my steps on the guest:
> 1. enabling VHOST defines and building x86_64-native-linuxapp-gcc.
>
> 2. successfully setting up environment:
> modprobe uio_pci_generic
> ./tools/dpdk-devbind.py -b uio_pci_generic 0000:00:07.0
> ./tools/dpdk-devbind.py -b uio_pci_generic 0000:00:08.0
> ./examples/l3fwd/build/l3fwd -c 0x3 -n 3 -- --parse-ptype -p 0x3 -P -L
> --config="(0,0,0),(0,1,1),(1,0,0),(1,1,1)" --no-numa
> it looks like the l3fwd app initialized successfully but it receives no
> traffic.
How about simple io forward through testpmd here (testpmd -c 0x3 -n 3 -- 
-portmask=0x3)  ?
Are you ending up traffic not seen for this case as well ? (Just to 
eliminate cases with l3fwd).

Thanks,
Gowrishankar
>
> I see at the host that connection is made however no traffic arrives to the
> host.
> digging into the code i see that the message VHOST_USER_SET_VRING_KICK
> never arrives to the host hence the callback new_device is never called and
> I can't figure out why.
>
> Any ideas?
>
> Thanks,
> Edgar
>
>

Re: [dpdk-users] new_device is never created at the vhost-switch sample app

[edgar helmut]

Thanks Gowrishankar

Using thevtestpmd resulted with the same behavior.
However it was solved after getting and building the latest qemu with
softmmu flag. Then it started to function.
Until then I was using the qemu as installed on my machine using the
default package configuration (using apt install...).

Thanks


On 05 Dec 2016 7:27 PM, "gowrishankar muthukrishnan" <
gowrishankar.m@linux.vnet.ibm.com> wrote:

On Friday 25 November 2016 01:28 PM, edgar helmut wrote:

> Hi,
> I am following http://dpdk.org/doc/guides/sample_app_ug/vhost.html using
> 16.11 in order to have a simple VM forwarding packets from one NIC to
> another efficiently, but it doesn't work... so I need some help to
> understand what am i missing.
> I am using x520 (intel 82599).
> host is ubuntu 16.04 and dpdk 16.11.
>
> my steps at host:
> 1. enabling VHOST defines and building x86_64-native-linuxapp-gcc.
>
> 2. both interfaces are bind like:
> 0000:04:00.0 'Ethernet 10G 2P X520 Adapter' drv=igb_uio unused=ixgbe
> 0000:04:00.1 'Ethernet 10G 2P X520 Adapter' drv=igb_uio unused=ixgbe
>
> 3. then making the project examples/vhost and executing:
> ./build/vhost-switch -c 0x03 -n 4 --socket-mem 1024 --file-prefix p1  -- -p
> 1 --vm2vm 0 --socket-file /tmp/sock1 --client -P
> ./build/vhost-switch -c 0x30 -n 4 --socket-mem 1024 --file-prefix p2  -- -p
> 2 --vm2vm 0 --socket-file /tmp/sock2 --client -P
>
> 4. creating the guest
> qemu- system-x86_64 -chardev socket,id=char1,path=/tmp/sock1,server
> -netdev
> type=vhost-user,id=hostnet1,chardev=char1,queues=2 -device
> virtio-net-pci,mq=on,vectors=2,netdev=hostnet1,id=net1,mac=0
> 0:00:00:00:00:01
> -chardev socket,id=char2,path=/tmp/sock2,server -netdev
> type=vhost-user,id=hostnet2,chardev=char2,queues=2 -device
> virtio-net-pci,mq=on,vectors=2,netdev=hostnet2,id=net2,mac=0
> 0:00:00:00:00:02
> -object
> memory-backend-file,id=mem2,size=2048M,mem-path=/run/hugepages,share=on
> ...
>
> my steps on the guest:
> 1. enabling VHOST defines and building x86_64-native-linuxapp-gcc.
>
> 2. successfully setting up environment:
> modprobe uio_pci_generic
> ./tools/dpdk-devbind.py -b uio_pci_generic 0000:00:07.0
> ./tools/dpdk-devbind.py -b uio_pci_generic 0000:00:08.0
> ./examples/l3fwd/build/l3fwd -c 0x3 -n 3 -- --parse-ptype -p 0x3 -P -L
> --config="(0,0,0),(0,1,1),(1,0,0),(1,1,1)" --no-numa
> it looks like the l3fwd app initialized successfully but it receives no
> traffic.
>
How about simple io forward through testpmd here (testpmd -c 0x3 -n 3 --
-portmask=0x3)  ?
Are you ending up traffic not seen for this case as well ? (Just to
eliminate cases with l3fwd).

Thanks,
Gowrishankar


> I see at the host that connection is made however no traffic arrives to the
> host.
> digging into the code i see that the message VHOST_USER_SET_VRING_KICK
> never arrives to the host hence the callback new_device is never called and
> I can't figure out why.
>
> Any ideas?
>
> Thanks,
> Edgar
>
>
>