> -----Original Message----- > From: Stephen Hemminger [mailto:stephen at networkplumber.org] > Sent: Friday, June 13, 2014 10:59 PM > To: Doherty, Declan > Cc: dev at dpdk.org > Subject: Re: [dpdk-dev] [PATCH v3 2/5] Link Bonding PMD Library > (librte_eal/librte_ether link bonding support changes) > > On Fri, 13 Jun 2014 15:41:59 +0100 > Declan Doherty <declan.doherty at intel.com> wrote: > > > Updating functionality in EAL to support adding link bonding > > devices via ???vdev option. Link bonding devices will be > > initialized after all physical devices have been probed and > > initialized. > > > > Signed-off-by: Declan Doherty <declan.doherty at intel.com> > > The DPDK is used by both static (demo style) applications and dynamic > applications that need to be able to modify parameters while running. > The tendency of the code has been to have values that can only be set > at initialization, and this kind of inflexibility is hard to fix. > > The whole dev args style of configuration is a poor design for those > reasons. How do you support creating additional bond devices at runtime? > > Remember what works for a demo doesn't make a good API for real world. > >
Hi Stephen, I totally agree but all of the dev argument parsing / static initialization here is based on the rte_eth_bond.h public API. Actually testpmd has full support for dynamic initialization and management of bonding devices in run-time through the interactive command line options which pretty much have options corresponding to all of the public link bonding APIs. It is my intent that this would main interface that this API would be used by user applications and that the static configuration options using --vdev interface would only be used if a user wanted to use a bonded device in a existing application without having to make any changes to their code or add any explicit reference to the link bonding library.
