[dpdk-dev] [PATCH v11 1/6] ethdev: add Tx preparation

Thomas Monjalon thomas.monjalon at 6wind.com
Fri Oct 28 12:22:23 CEST 2016


2016-10-28 10:15, Ananyev, Konstantin:
> > From: Ananyev, Konstantin
> > > From: Thomas Monjalon [mailto:thomas.monjalon at 6wind.com]
> > > > 2016-10-27 15:52, Ananyev, Konstantin:
> > > > > > 2016-10-26 14:56, Tomasz Kulasek:
> > > > > > > --- a/config/common_base
> > > > > > > +++ b/config/common_base
> > > > > > > +CONFIG_RTE_ETHDEV_TX_PREP=y
> > > > > >
> > > > > > We cannot enable it until it is implemented in every drivers.
> > > > >
> > > > > Not sure why?
> > > > > If tx_pkt_prep == NULL, then rte_eth_tx_prep() would just act as noop.
> > > > > Right now it is not mandatory for the PMD to implement it.
> > > >
> > > > If it is not implemented, the application must do the preparation by
> > > itself.
> > > > From patch 6:
> > > > "
> > > > Removed pseudo header calculation for udp/tcp/tso packets from
> > > > application and used Tx preparation API for packet preparation and
> > > > verification.
> > > > "
> > > > So how does it behave with other drivers?
> > >
> > > Hmm so it seems that we broke testpmd csumonly mode for non-intel
> > > drivers..
> > > My bad, missed that part completely.
> > > Yes, then I suppose for now we'll need to support both (with and without)
> > > code paths for testpmd.
> > > Probably a new fwd mode or just extra parameter for the existing one?
> > > Any other suggestions?
> > >
> > 
> > I had sent txprep engine in v2 (http://dpdk.org/dev/patchwork/patch/15775/), but I'm opened on the suggestions. If you like it I can resent
> > it in place of csumonly modification.
> 
> I still not sure it is worth to have another version of csum...
> Can we introduce a new global variable in testpmd and a new command:
> testpmd> csum tx_prep
> or so? 
> Looking at current testpmd patch, I suppose the changes will be minimal.
> What do you think?

No please no!
The problem is not in testpmd.
The problem is in every applications.
Should we prepare the checksums or let tx_prep do it?
The result will depend of the driver used.


More information about the dev mailing list