Re: [dpdk-dev] [PATCH v12] net/tap: new TUN/TAP device PMD

2017-01-20 Thread Ferruh Yigit
On 12/13/2016 1:54 PM, Ferruh Yigit wrote: > On 12/12/2016 2:38 PM, Keith Wiles wrote: >> The PMD allows for DPDK and the host to communicate using a raw >> device interface on the host and in the DPDK application. The device >> created is a Tap device with a L2 packet header. >> >> v12- Fixup mino

Re: [dpdk-dev] [PATCH v12] net/tap: new TUN/TAP device PMD

2016-12-13 Thread Ferruh Yigit
On 12/12/2016 2:38 PM, Keith Wiles wrote: > The PMD allows for DPDK and the host to communicate using a raw > device interface on the host and in the DPDK application. The device > created is a Tap device with a L2 packet header. > > v12- Fixup minor changes for driver_name and version number > v1

Re: [dpdk-dev] [PATCH v12] net/tap: new TUN/TAP device PMD

2016-12-12 Thread Wiles, Keith
> On Dec 12, 2016, at 1:13 PM, Marc wrote: > > Keith, > > A bit late, but two very high level questions. Do you have performance > numbers compared to KNI? Did you consider using AF_PACKET PACKET_MMAP which > could potentially reduce the number of syscalls to 1 for RX and TX of a burst? Hi M

Re: [dpdk-dev] [PATCH v12] net/tap: new TUN/TAP device PMD

2016-12-12 Thread Marc
Keith, A bit late, but two very high level questions. Do you have performance numbers compared to KNI? Did you consider using AF_PACKET PACKET_MMAP which could potentially reduce the number of syscalls to 1 for RX and TX of a burst? Marc On 12 December 2016 at 15:38, Keith Wiles wrote: > The P