Hi Mohsin, Thanks a lot for the tips!
Regards, Dmitry On Mon, Nov 21, 2016 at 1:26 PM, Mohsin Kazmi (sykazmi) <syka...@cisco.com> wrote: > Hello Dmitry, > > > This is processing header length which includes dpdk header + vlib header > to process packets in graph nodes. > > > While normal buffer size is 2048 bytes. MTU can be set using following > command line provided that interface should be within the range of min/max > supported values. > > > *set interface mtu* <value> infc-name > > > Bigger packets consist of multiple buffers. There is a common piece of > code used for it. For example, function *buffer_add_to_chain()* in file > *v**pp/vnet/vnet/d**evices/netmap/node.c *sets the flag > VLIB_BUFFER_NEXT_PRESENT​ along with other necessary parameters. It tells > the graph nodes about the next buffer in the processing of same packet. > > > ​ > > -br > > Mohsin > ------------------------------ > *From:* vpp-dev-boun...@lists.fd.io <vpp-dev-boun...@lists.fd.io> on > behalf of Dmitry Bufistov <dmi...@midokura.com> > *Sent:* Monday, November 21, 2016 12:54 PM > *To:* vpp-dev@lists.fd.io > *Subject:* [vpp-dev] VLIB buffer size is 512 bytes > > Hi VPP list! > > The VPP buffer size is set to 512 bytes in buffer.h > What about bigger packets? How does VPP handle them? > > Thanks a lot, > > Dmitry >
_______________________________________________ vpp-dev mailing list vpp-dev@lists.fd.io https://lists.fd.io/mailman/listinfo/vpp-dev
