Re: [dpdk-dev] [RFC] mbuf: support dynamic fields and flags

[Olivier Matz <olivier.matz@6wind.com>]

On Thu, Jul 11, 2019 at 08:04:00AM +0000, Wang, Haiyue wrote:
> > -----Original Message-----
> > From: Olivier Matz [mailto:olivier.matz@6wind.com]
> > Sent: Thursday, July 11, 2019 15:26
> > To: Wang, Haiyue <haiyue.wang@intel.com>
> > Cc: dev@dpdk.org
> > Subject: Re: [dpdk-dev] [RFC] mbuf: support dynamic fields and flags
> > 
> > Hi,
> > 
> > On Wed, Jul 10, 2019 at 05:14:33PM +0000, Wang, Haiyue wrote:
> > > Hi,
> > >
> > > Sounds cool, just have some questions inline.
> > >
> > > > -----Original Message-----
> > > > From: dev [mailto:dev-bounces@dpdk.org] On Behalf Of Olivier Matz
> > > > Sent: Wednesday, July 10, 2019 17:29
> > > > To: dev@dpdk.org
> > > > Subject: [dpdk-dev] [RFC] mbuf: support dynamic fields and flags
> > > >
> > > > Many features require to store data inside the mbuf. As the room in mbuf
> > > > structure is limited, it is not possible to have a field for each
> > > > feature. Also, changing fields in the mbuf structure can break the API
> > > > or ABI.
> > > >
> > > > This commit addresses these issues, by enabling the dynamic registration
> > > > of fields or flags:
> > > >
> > > > - a dynamic field is a named area in the rte_mbuf structure, with a
> > > >   given size (>= 1 byte) and alignment constraint.
> > > > - a dynamic flag is a named bit in the rte_mbuf structure.
> > > >
> > > > The typical use case is a PMD that registers space for an offload
> > > > feature, when the application requests to enable this feature.  As
> > > > the space in mbuf is limited, the space should only be reserved if it
> > > > is going to be used (i.e when the application explicitly asks for it).
> > > >
> > > > The registration can be done at any moment, but it is not possible
> > > > to unregister fields or flags for now.
> > > >
> > > > Signed-off-by: Olivier Matz <olivier.matz@6wind.com>
> > 
> > (...)
> > 
> > > > +/**
> > > > + * @file
> > > > + * RTE Mbuf dynamic fields and flags
> > > > + *
> > > > + * Many features require to store data inside the mbuf. As the room in
> > > > + * mbuf structure is limited, it is not possible to have a field for
> > > > + * each feature. Also, changing fields in the mbuf structure can break
> > > > + * the API or ABI.
> > > > + *
> > > > + * This module addresses this issue, by enabling the dynamic
> > > > + * registration of fields or flags:
> > > > + *
> > > > + * - a dynamic field is a named area in the rte_mbuf structure, with a
> > > > + *   given size (>= 1 byte) and alignment constraint.
> > > > + * - a dynamic flag is a named bit in the rte_mbuf structure.
> > > > + *
> > > > + * The typical use case is a PMD that registers space for an offload
> > > > + * feature, when the application requests to enable this feature.  As
> > > > + * the space in mbuf is limited, the space should only be reserved if it
> > > > + * is going to be used (i.e when the application explicitly asks for it).
> > > > + *
> > > > + * The registration can be done at any moment, but it is not possible
> > > > + * to unregister fields or flags for now.
> > > > + *
> > > > + * Example of use:
> > > > + *
> > > > + * - RTE_MBUF_DYN_<feature>_(ID|SIZE|ALIGN) are defined in this file
> > >
> > > Does it means that all PMDs define their own 'RTE_MBUF_DYN_<feature>_(ID|SIZE|ALIGN)'
> > > here ? In other words, each PMD can expose its private DYN_<feature> here for public
> > > using ?
> > 
> > For generic fields, I think they should be declared in this file. For
> > instance, if we decide to replace the current m->timestamp field by a
> > dynamic field, we should add like this:
> > 
> > #define RTE_MBUF_DYN_TIMESTAMP_ID "rte_timestamp"
> > #define RTE_MBUF_DYN_TIMESTAMP_SIZE sizeof(uint64_t)
> > #define RTE_MBUF_DYN_TIMESTAMP_ALIGN __alignof__(uint64_t)
> > 
> > If the feature is PMD-specific, the defines could be exposed in a
> > PMD header.
> > 
> 
> Now, understand the comments a little : ... must not define identifers prefixed with "rte_",
> which are reserved for standard features. Seems have big plan ?

The dynamic field can also be used by an external application or by an
external library. For instance, a field to tag a packet, like skb->mark
in linux. In this case, id, size and alignment would be defined outside
dpdk subtree.

To avoid name conflicts, I think we should define a convention for
identifiers, so they are in different namespaces:

- "rte_*" for identifiers declared inside dpdk subtree
- any other name for identifiers declared in an external application or
  library