[v9] net/ice: optimize protocol extraction by dynamic mbuf API
diff mbox series

Message ID 20191107104402.99398-1-haiyue.wang@intel.com
State Superseded, archived
Delegated to: xiaolong ye
Headers show
Series
  • [v9] net/ice: optimize protocol extraction by dynamic mbuf API
Related show

Checks

Context Check Description
ci/Intel-compilation success Compilation OK
ci/travis-robot success Travis build: passed
ci/iol-mellanox-Performance success Performance Testing PASS
ci/iol-compilation success Compile Testing PASS
ci/iol-intel-Performance success Performance Testing PASS
ci/checkpatch success coding style OK

Commit Message

Wang, Haiyue Nov. 7, 2019, 10:44 a.m. UTC
The original design is to use rte_mbuf::udata64 to save the metadata of
protocol extraction which has network protocol data fields and type, a
private API is used to decode this metadata.

Use the dynamic mbuf field and flags to register the needed fields in
mbuf, to avoid overwriting 'rte_mbuf::udata64' if the application uses
it. It only needs 4B size to save the protocol extraction data, and its
type and validity is indicated by related bit in 'rte_mbuf::ol_flags'.

Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
---
v9: - Rename the dynamic filed and flags variables to _offs, _mask, so it
      matches the design.

v8: - Cleanup the doxgen, rename the dynamic filed and flags with ice prefix.

v7: - Change the log level from ERR to DEBUG for a successful registration,
      ERR is used to develop firstly.

v6: - Don't need to register all flags when if one dev_args is set, register
      it as required.

v5: - Remove the '_OL/_ol' in dynamic mbuf flag to make the variable clean.

v4: - Include the 'rte_pmd_ice.h' header in ICE source as needed, reduce
      its compile scope.

v3: - Use the conventions name "rte_net_<pmd>_dynfield_<name>" since it
      is defined in a in PMD.
    - Add helpers API for easiy access.

v2: - disable the protocol extraction if failed to register some ol_flags
    - rewrite the commit message

 doc/api/doxy-api-index.md               |   1 +
 doc/api/doxy-api.conf.in                |   1 +
 doc/guides/nics/ice.rst                 |  14 +-
 drivers/net/ice/ice_ethdev.c            |  84 ++++++-
 drivers/net/ice/ice_ethdev.h            |   9 +
 drivers/net/ice/ice_rxtx.c              |  88 ++++---
 drivers/net/ice/ice_rxtx.h              |   1 -
 drivers/net/ice/rte_pmd_ice.h           | 313 +++++++++++++++---------
 drivers/net/ice/rte_pmd_ice_version.map |  12 +
 9 files changed, 370 insertions(+), 153 deletions(-)

Comments

Xiaolong Ye Nov. 8, 2019, 2:35 a.m. UTC | #1
On 11/07, Haiyue Wang wrote:
>The original design is to use rte_mbuf::udata64 to save the metadata of
>protocol extraction which has network protocol data fields and type, a
>private API is used to decode this metadata.
>
>Use the dynamic mbuf field and flags to register the needed fields in
>mbuf, to avoid overwriting 'rte_mbuf::udata64' if the application uses
>it. It only needs 4B size to save the protocol extraction data, and its
>type and validity is indicated by related bit in 'rte_mbuf::ol_flags'.
>
>Signed-off-by: Haiyue Wang <haiyue.wang@intel.com>
>---
>v9: - Rename the dynamic filed and flags variables to _offs, _mask, so it
>      matches the design.
>
>v8: - Cleanup the doxgen, rename the dynamic filed and flags with ice prefix.
>
>v7: - Change the log level from ERR to DEBUG for a successful registration,
>      ERR is used to develop firstly.
>
>v6: - Don't need to register all flags when if one dev_args is set, register
>      it as required.
>
>v5: - Remove the '_OL/_ol' in dynamic mbuf flag to make the variable clean.
>
>v4: - Include the 'rte_pmd_ice.h' header in ICE source as needed, reduce
>      its compile scope.
>
>v3: - Use the conventions name "rte_net_<pmd>_dynfield_<name>" since it
>      is defined in a in PMD.
>    - Add helpers API for easiy access.
>
>v2: - disable the protocol extraction if failed to register some ol_flags
>    - rewrite the commit message
>
> doc/api/doxy-api-index.md               |   1 +
> doc/api/doxy-api.conf.in                |   1 +
> doc/guides/nics/ice.rst                 |  14 +-
> drivers/net/ice/ice_ethdev.c            |  84 ++++++-
> drivers/net/ice/ice_ethdev.h            |   9 +
> drivers/net/ice/ice_rxtx.c              |  88 ++++---
> drivers/net/ice/ice_rxtx.h              |   1 -
> drivers/net/ice/rte_pmd_ice.h           | 313 +++++++++++++++---------
> drivers/net/ice/rte_pmd_ice_version.map |  12 +
> 9 files changed, 370 insertions(+), 153 deletions(-)
>

Reviewed-by: Xiaolong Ye <xiaolong.ye@intel.com>

Applied to dpdk-next-net-intel. Thanks.
Thomas Monjalon Nov. 8, 2019, 12:34 p.m. UTC | #2
Hi,

I see this patch is already merged in next-net-intel,
but please I would prefer to have below improvements first.

07/11/2019 11:44, Haiyue Wang:
> The original design is to use rte_mbuf::udata64 to save the metadata of
> protocol extraction which has network protocol data fields and type, a
> private API is used to decode this metadata.
> 
> Use the dynamic mbuf field and flags to register the needed fields in
> mbuf, to avoid overwriting 'rte_mbuf::udata64' if the application uses
> it. It only needs 4B size to save the protocol extraction data, and its

Yes using a dynamic field is definitely more correct.

> type and validity is indicated by related bit in 'rte_mbuf::ol_flags'.

Better to say explicitly it uses a dynamic flag.

> --- a/drivers/net/ice/rte_pmd_ice.h
> +++ b/drivers/net/ice/rte_pmd_ice.h
> +/**
> + * @file rte_pmd_ice.h
> + *
> + * ice PMD specific functions.
> + *
> + * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
> + *
> + */

Adding the file in doxygen is good.
I think it could be a separate patch for doxygen + cleanups.

> +/**
> + * The supported network protocol extraction metadata format.
> + */
> +union proto_xtr_metadata {
> -struct proto_xtr_flds {

Please add a prefix rte_ice_ or rte_net_ice_ as you wish.

[...]
> +/**
> + * The mbuf dynamic flag for VLAN protocol extraction metadata, it is valid
> + * when dev_args 'proto_xtr' has 'vlan' specified.
> + */
> +#define PKT_RX_DYNF_PROTO_XTR_VLAN \
> +	(rte_net_ice_dynflag_proto_xtr_vlan_mask)
> +
> +/**
> + * The mbuf dynamic flag for IPv4 protocol extraction metadata, it is valid
> + * when dev_args 'proto_xtr' has 'ipv4' specified.
> + */
> +#define PKT_RX_DYNF_PROTO_XTR_IPV4 \
> +	(rte_net_ice_dynflag_proto_xtr_ipv4_mask)
> +
> +/**
> + * The mbuf dynamic flag for IPv6 protocol extraction metadata, it is valid
> + * when dev_args 'proto_xtr' has 'ipv6' specified.
> + */
> +#define PKT_RX_DYNF_PROTO_XTR_IPV6 \
> +	(rte_net_ice_dynflag_proto_xtr_ipv6_mask)
> +
> +/**
> + * The mbuf dynamic flag for IPv6 with flow protocol extraction metadata, it is
> + * valid when dev_args 'proto_xtr' has 'ipv6_flow' specified.
> + */
> +#define PKT_RX_DYNF_PROTO_XTR_IPV6_FLOW \
> +	(rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask)
> +
> +/**
> + * The mbuf dynamic flag for TCP protocol extraction metadata, it is valid
> + * when dev_args 'proto_xtr' has 'tcp' specified.
> + */
> +#define PKT_RX_DYNF_PROTO_XTR_TCP \
> +	(rte_net_ice_dynflag_proto_xtr_tcp_mask)

Those fields and flags are missing a RTE_ prefix.
(Yes I know we are missing such prefix in rte_mbuf.f)
Thomas Monjalon Nov. 8, 2019, 12:55 p.m. UTC | #3
07/11/2019 11:44, Haiyue Wang:
> --- a/drivers/net/ice/rte_pmd_ice_version.map
> +++ b/drivers/net/ice/rte_pmd_ice_version.map
> +EXPERIMENTAL {
> +	global:
> +
> +	# added in 19.11
> +	rte_net_ice_dynfield_proto_xtr_metadata_offs;
> +	rte_net_ice_dynflag_proto_xtr_vlan_mask;
> +	rte_net_ice_dynflag_proto_xtr_ipv4_mask;
> +	rte_net_ice_dynflag_proto_xtr_ipv6_mask;
> +	rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask;
> +	rte_net_ice_dynflag_proto_xtr_tcp_mask;
> +};

Given that you provide some functions to access to the metadata,
why do you need to export these flags and field in the .map?

However, the functions are missing in the .map.
Did you try to compile as a shared library?
Wang, Haiyue Nov. 8, 2019, 2:01 p.m. UTC | #4
Hi Thomas,

> -----Original Message-----
> From: Thomas Monjalon <thomas@monjalon.net>
> Sent: Friday, November 8, 2019 20:55
> To: Wang, Haiyue <haiyue.wang@intel.com>
> Cc: dev@dpdk.org; olivier.matz@6wind.com; Ye, Xiaolong <xiaolong.ye@intel.com>; Yigit, Ferruh
> <ferruh.yigit@intel.com>
> Subject: Re: [dpdk-dev] [PATCH v9] net/ice: optimize protocol extraction by dynamic mbuf API
> 
> 07/11/2019 11:44, Haiyue Wang:
> > --- a/drivers/net/ice/rte_pmd_ice_version.map
> > +++ b/drivers/net/ice/rte_pmd_ice_version.map
> > +EXPERIMENTAL {
> > +	global:
> > +
> > +	# added in 19.11
> > +	rte_net_ice_dynfield_proto_xtr_metadata_offs;
> > +	rte_net_ice_dynflag_proto_xtr_vlan_mask;
> > +	rte_net_ice_dynflag_proto_xtr_ipv4_mask;
> > +	rte_net_ice_dynflag_proto_xtr_ipv6_mask;
> > +	rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask;
> > +	rte_net_ice_dynflag_proto_xtr_tcp_mask;
> > +};
> 
> Given that you provide some functions to access to the metadata,
> why do you need to export these flags and field in the .map?
>
> However, the functions are missing in the .map.
> Did you try to compile as a shared library?
> 

These functions are 'static inline', no need to be exported in the
.map. And the macros like 'PKT_RX_DYNF_PROTO_XTR_XXX', in fact, their
real definitions are global values defined in rte_pmd_ice like:
rte_net_ice_dynflag_proto_xtr_xxx_mask.

Since rte_pmd_ice are required to compiled as a shared library, so
it is needed to export these flags and field in the .map.

This design is referred to the below upstream practice about dynamic
mbuf.

commit 7743e81854944ed17df05bfdcba26556cb41ca0c
Author: Viacheslav Ovsiienko <viacheslavo@mellanox.com>
Date:   Tue Nov 5 14:19:30 2019 +0000

    ethdev: extend flow metadata

--- a/lib/librte_ethdev/rte_ethdev_version.map
+++ b/lib/librte_ethdev/rte_ethdev_version.map
@@ -289,4 +289,7 @@ EXPERIMENTAL {
        rte_eth_rx_hairpin_queue_setup;
        rte_eth_tx_burst_mode_get;
        rte_eth_tx_hairpin_queue_setup;
+       rte_flow_dynf_metadata_offs;
+       rte_flow_dynf_metadata_mask;
+       rte_flow_dynf_metadata_register;
 };
Wang, Haiyue Nov. 8, 2019, 2:08 p.m. UTC | #5
Hi Thomas,

> -----Original Message-----
> From: Thomas Monjalon <thomas@monjalon.net>
> Sent: Friday, November 8, 2019 20:34
> To: Wang, Haiyue <haiyue.wang@intel.com>
> Cc: dev@dpdk.org; olivier.matz@6wind.com; Ye, Xiaolong <xiaolong.ye@intel.com>; Yigit, Ferruh
> <ferruh.yigit@intel.com>
> Subject: Re: [dpdk-dev] [PATCH v9] net/ice: optimize protocol extraction by dynamic mbuf API
> 
> Hi,
> 
> I see this patch is already merged in next-net-intel,
> but please I would prefer to have below improvements first.
> 
> 07/11/2019 11:44, Haiyue Wang:
> > The original design is to use rte_mbuf::udata64 to save the metadata of
> > protocol extraction which has network protocol data fields and type, a
> > private API is used to decode this metadata.
> >
> > Use the dynamic mbuf field and flags to register the needed fields in
> > mbuf, to avoid overwriting 'rte_mbuf::udata64' if the application uses
> > it. It only needs 4B size to save the protocol extraction data, and its
> 
> Yes using a dynamic field is definitely more correct.
> 
> > type and validity is indicated by related bit in 'rte_mbuf::ol_flags'.
> 
> Better to say explicitly it uses a dynamic flag.
> 

Will update doc to make the description be better.

> > --- a/drivers/net/ice/rte_pmd_ice.h
> > +++ b/drivers/net/ice/rte_pmd_ice.h
> > +/**
> > + * @file rte_pmd_ice.h
> > + *
> > + * ice PMD specific functions.
> > + *
> > + * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
> > + *
> > + */
> 
> Adding the file in doxygen is good.
> I think it could be a separate patch for doxygen + cleanups.
> 
> > +/**
> > + * The supported network protocol extraction metadata format.
> > + */
> > +union proto_xtr_metadata {
> > -struct proto_xtr_flds {
> 
> Please add a prefix rte_ice_ or rte_net_ice_ as you wish.
> 

Missed, will be updated.

> [...]
> > +/**
> > + * The mbuf dynamic flag for VLAN protocol extraction metadata, it is valid
> > + * when dev_args 'proto_xtr' has 'vlan' specified.
> > + */
> > +#define PKT_RX_DYNF_PROTO_XTR_VLAN \
> > +	(rte_net_ice_dynflag_proto_xtr_vlan_mask)
> > +
> > +/**
> > + * The mbuf dynamic flag for IPv4 protocol extraction metadata, it is valid
> > + * when dev_args 'proto_xtr' has 'ipv4' specified.
> > + */
> > +#define PKT_RX_DYNF_PROTO_XTR_IPV4 \
> > +	(rte_net_ice_dynflag_proto_xtr_ipv4_mask)
> > +
> > +/**
> > + * The mbuf dynamic flag for IPv6 protocol extraction metadata, it is valid
> > + * when dev_args 'proto_xtr' has 'ipv6' specified.
> > + */
> > +#define PKT_RX_DYNF_PROTO_XTR_IPV6 \
> > +	(rte_net_ice_dynflag_proto_xtr_ipv6_mask)
> > +
> > +/**
> > + * The mbuf dynamic flag for IPv6 with flow protocol extraction metadata, it is
> > + * valid when dev_args 'proto_xtr' has 'ipv6_flow' specified.
> > + */
> > +#define PKT_RX_DYNF_PROTO_XTR_IPV6_FLOW \
> > +	(rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask)
> > +
> > +/**
> > + * The mbuf dynamic flag for TCP protocol extraction metadata, it is valid
> > + * when dev_args 'proto_xtr' has 'tcp' specified.
> > + */
> > +#define PKT_RX_DYNF_PROTO_XTR_TCP \
> > +	(rte_net_ice_dynflag_proto_xtr_tcp_mask)
> 
> Those fields and flags are missing a RTE_ prefix.
> (Yes I know we are missing such prefix in rte_mbuf.f)
> 

The "PKT_RX_" in ' lib/librte_mbuf/rte_mbuf_core.h' will be changed to
"RTE_PKT_RX_" ?

Or keep the above as it is now, to keep the same ol_flags style, until all
the "PKT_RX_" are changed ?
Thomas Monjalon Nov. 8, 2019, 2:38 p.m. UTC | #6
08/11/2019 15:01, Wang, Haiyue:
> From: Thomas Monjalon <thomas@monjalon.net>
> > 07/11/2019 11:44, Haiyue Wang:
> > > --- a/drivers/net/ice/rte_pmd_ice_version.map
> > > +++ b/drivers/net/ice/rte_pmd_ice_version.map
> > > +EXPERIMENTAL {
> > > +	global:
> > > +
> > > +	# added in 19.11
> > > +	rte_net_ice_dynfield_proto_xtr_metadata_offs;
> > > +	rte_net_ice_dynflag_proto_xtr_vlan_mask;
> > > +	rte_net_ice_dynflag_proto_xtr_ipv4_mask;
> > > +	rte_net_ice_dynflag_proto_xtr_ipv6_mask;
> > > +	rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask;
> > > +	rte_net_ice_dynflag_proto_xtr_tcp_mask;
> > > +};
> > 
> > Given that you provide some functions to access to the metadata,
> > why do you need to export these flags and field in the .map?
> >
> > However, the functions are missing in the .map.
> > Did you try to compile as a shared library?
> > 
> 
> These functions are 'static inline', no need to be exported in the

Yes I missed they are inline.

> .map. And the macros like 'PKT_RX_DYNF_PROTO_XTR_XXX', in fact, their
> real definitions are global values defined in rte_pmd_ice like:
> rte_net_ice_dynflag_proto_xtr_xxx_mask.
> 
> Since rte_pmd_ice are required to compiled as a shared library, so
> it is needed to export these flags and field in the .map.
> 
> This design is referred to the below upstream practice about dynamic
> mbuf.
> 
> commit 7743e81854944ed17df05bfdcba26556cb41ca0c
> Author: Viacheslav Ovsiienko <viacheslavo@mellanox.com>
> Date:   Tue Nov 5 14:19:30 2019 +0000
> 
>     ethdev: extend flow metadata
> 
> --- a/lib/librte_ethdev/rte_ethdev_version.map
> +++ b/lib/librte_ethdev/rte_ethdev_version.map
> @@ -289,4 +289,7 @@ EXPERIMENTAL {
>         rte_eth_rx_hairpin_queue_setup;
>         rte_eth_tx_burst_mode_get;
>         rte_eth_tx_hairpin_queue_setup;
> +       rte_flow_dynf_metadata_offs;
> +       rte_flow_dynf_metadata_mask;
> +       rte_flow_dynf_metadata_register;
>  };

Yes all these internals are exported because we use macros
or inline functions.
Matter of tradeoff between API cleaning and performance...
Thomas Monjalon Nov. 8, 2019, 2:39 p.m. UTC | #7
08/11/2019 15:08, Wang, Haiyue:
> Hi Thomas,
> 
> > -----Original Message-----
> > From: Thomas Monjalon <thomas@monjalon.net>
> > Sent: Friday, November 8, 2019 20:34
> > To: Wang, Haiyue <haiyue.wang@intel.com>
> > Cc: dev@dpdk.org; olivier.matz@6wind.com; Ye, Xiaolong <xiaolong.ye@intel.com>; Yigit, Ferruh
> > <ferruh.yigit@intel.com>
> > Subject: Re: [dpdk-dev] [PATCH v9] net/ice: optimize protocol extraction by dynamic mbuf API
> > 
> > Hi,
> > 
> > I see this patch is already merged in next-net-intel,
> > but please I would prefer to have below improvements first.
> > 
> > 07/11/2019 11:44, Haiyue Wang:
> > > The original design is to use rte_mbuf::udata64 to save the metadata of
> > > protocol extraction which has network protocol data fields and type, a
> > > private API is used to decode this metadata.
> > >
> > > Use the dynamic mbuf field and flags to register the needed fields in
> > > mbuf, to avoid overwriting 'rte_mbuf::udata64' if the application uses
> > > it. It only needs 4B size to save the protocol extraction data, and its
> > 
> > Yes using a dynamic field is definitely more correct.
> > 
> > > type and validity is indicated by related bit in 'rte_mbuf::ol_flags'.
> > 
> > Better to say explicitly it uses a dynamic flag.
> > 
> 
> Will update doc to make the description be better.
> 
> > > --- a/drivers/net/ice/rte_pmd_ice.h
> > > +++ b/drivers/net/ice/rte_pmd_ice.h
> > > +/**
> > > + * @file rte_pmd_ice.h
> > > + *
> > > + * ice PMD specific functions.
> > > + *
> > > + * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
> > > + *
> > > + */
> > 
> > Adding the file in doxygen is good.
> > I think it could be a separate patch for doxygen + cleanups.
> > 
> > > +/**
> > > + * The supported network protocol extraction metadata format.
> > > + */
> > > +union proto_xtr_metadata {
> > > -struct proto_xtr_flds {
> > 
> > Please add a prefix rte_ice_ or rte_net_ice_ as you wish.
> > 
> 
> Missed, will be updated.
> 
> > [...]
> > > +/**
> > > + * The mbuf dynamic flag for VLAN protocol extraction metadata, it is valid
> > > + * when dev_args 'proto_xtr' has 'vlan' specified.
> > > + */
> > > +#define PKT_RX_DYNF_PROTO_XTR_VLAN \
> > > +	(rte_net_ice_dynflag_proto_xtr_vlan_mask)
> > > +
> > > +/**
> > > + * The mbuf dynamic flag for IPv4 protocol extraction metadata, it is valid
> > > + * when dev_args 'proto_xtr' has 'ipv4' specified.
> > > + */
> > > +#define PKT_RX_DYNF_PROTO_XTR_IPV4 \
> > > +	(rte_net_ice_dynflag_proto_xtr_ipv4_mask)
> > > +
> > > +/**
> > > + * The mbuf dynamic flag for IPv6 protocol extraction metadata, it is valid
> > > + * when dev_args 'proto_xtr' has 'ipv6' specified.
> > > + */
> > > +#define PKT_RX_DYNF_PROTO_XTR_IPV6 \
> > > +	(rte_net_ice_dynflag_proto_xtr_ipv6_mask)
> > > +
> > > +/**
> > > + * The mbuf dynamic flag for IPv6 with flow protocol extraction metadata, it is
> > > + * valid when dev_args 'proto_xtr' has 'ipv6_flow' specified.
> > > + */
> > > +#define PKT_RX_DYNF_PROTO_XTR_IPV6_FLOW \
> > > +	(rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask)
> > > +
> > > +/**
> > > + * The mbuf dynamic flag for TCP protocol extraction metadata, it is valid
> > > + * when dev_args 'proto_xtr' has 'tcp' specified.
> > > + */
> > > +#define PKT_RX_DYNF_PROTO_XTR_TCP \
> > > +	(rte_net_ice_dynflag_proto_xtr_tcp_mask)
> > 
> > Those fields and flags are missing a RTE_ prefix.
> > (Yes I know we are missing such prefix in rte_mbuf.f)
> > 
> 
> The "PKT_RX_" in ' lib/librte_mbuf/rte_mbuf_core.h' will be changed to
> "RTE_PKT_RX_" ?

I started to do such change a long time ago and never finished.
Yes it will be changed for 20.11.

> Or keep the above as it is now, to keep the same ol_flags style, until all
> the "PKT_RX_" are changed ?

Please I prefer you change yours, without waiting global change.
Xiaolong Ye Nov. 8, 2019, 3:03 p.m. UTC | #8
On 11/08, Thomas Monjalon wrote:
>Hi,
>
>I see this patch is already merged in next-net-intel,
>but please I would prefer to have below improvements first.
>

Got it, I'll drop v9 in next-net-intel and wait for the improved new version.

Thanks,
Xiaolong
Wang, Haiyue Nov. 8, 2019, 3:04 p.m. UTC | #9
> -----Original Message-----
> From: Thomas Monjalon <thomas@monjalon.net>
> Sent: Friday, November 8, 2019 22:40
> To: Wang, Haiyue <haiyue.wang@intel.com>
> Cc: dev@dpdk.org; olivier.matz@6wind.com; Ye, Xiaolong <xiaolong.ye@intel.com>; Yigit, Ferruh
> <ferruh.yigit@intel.com>
> Subject: Re: [dpdk-dev] [PATCH v9] net/ice: optimize protocol extraction by dynamic mbuf API
> 
> 08/11/2019 15:08, Wang, Haiyue:
> > Hi Thomas,
> >
> > > -----Original Message-----
> > > From: Thomas Monjalon <thomas@monjalon.net>
> > > Sent: Friday, November 8, 2019 20:34
> > > To: Wang, Haiyue <haiyue.wang@intel.com>
> > > Cc: dev@dpdk.org; olivier.matz@6wind.com; Ye, Xiaolong <xiaolong.ye@intel.com>; Yigit, Ferruh
> > > <ferruh.yigit@intel.com>
> > > Subject: Re: [dpdk-dev] [PATCH v9] net/ice: optimize protocol extraction by dynamic mbuf API
> > >
> > > Hi,
> > >
> > > I see this patch is already merged in next-net-intel,
> > > but please I would prefer to have below improvements first.
> > >
> > > 07/11/2019 11:44, Haiyue Wang:
> > > > The original design is to use rte_mbuf::udata64 to save the metadata of
> > > > protocol extraction which has network protocol data fields and type, a
> > > > private API is used to decode this metadata.
> > > >
> > > > Use the dynamic mbuf field and flags to register the needed fields in
> > > > mbuf, to avoid overwriting 'rte_mbuf::udata64' if the application uses
> > > > it. It only needs 4B size to save the protocol extraction data, and its
> > >
> > > Yes using a dynamic field is definitely more correct.
> > >
> > > > type and validity is indicated by related bit in 'rte_mbuf::ol_flags'.
> > >
> > > Better to say explicitly it uses a dynamic flag.
> > >
> >
> > Will update doc to make the description be better.
> >
> > > > --- a/drivers/net/ice/rte_pmd_ice.h
> > > > +++ b/drivers/net/ice/rte_pmd_ice.h
> > > > +/**
> > > > + * @file rte_pmd_ice.h
> > > > + *
> > > > + * ice PMD specific functions.
> > > > + *
> > > > + * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
> > > > + *
> > > > + */
> > >
> > > Adding the file in doxygen is good.
> > > I think it could be a separate patch for doxygen + cleanups.
> > >
> > > > +/**
> > > > + * The supported network protocol extraction metadata format.
> > > > + */
> > > > +union proto_xtr_metadata {
> > > > -struct proto_xtr_flds {
> > >
> > > Please add a prefix rte_ice_ or rte_net_ice_ as you wish.
> > >
> >
> > Missed, will be updated.
> >
> > > [...]
> > > > +/**
> > > > + * The mbuf dynamic flag for VLAN protocol extraction metadata, it is valid
> > > > + * when dev_args 'proto_xtr' has 'vlan' specified.
> > > > + */
> > > > +#define PKT_RX_DYNF_PROTO_XTR_VLAN \
> > > > +	(rte_net_ice_dynflag_proto_xtr_vlan_mask)
> > > > +
> > > > +/**
> > > > + * The mbuf dynamic flag for IPv4 protocol extraction metadata, it is valid
> > > > + * when dev_args 'proto_xtr' has 'ipv4' specified.
> > > > + */
> > > > +#define PKT_RX_DYNF_PROTO_XTR_IPV4 \
> > > > +	(rte_net_ice_dynflag_proto_xtr_ipv4_mask)
> > > > +
> > > > +/**
> > > > + * The mbuf dynamic flag for IPv6 protocol extraction metadata, it is valid
> > > > + * when dev_args 'proto_xtr' has 'ipv6' specified.
> > > > + */
> > > > +#define PKT_RX_DYNF_PROTO_XTR_IPV6 \
> > > > +	(rte_net_ice_dynflag_proto_xtr_ipv6_mask)
> > > > +
> > > > +/**
> > > > + * The mbuf dynamic flag for IPv6 with flow protocol extraction metadata, it is
> > > > + * valid when dev_args 'proto_xtr' has 'ipv6_flow' specified.
> > > > + */
> > > > +#define PKT_RX_DYNF_PROTO_XTR_IPV6_FLOW \
> > > > +	(rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask)
> > > > +
> > > > +/**
> > > > + * The mbuf dynamic flag for TCP protocol extraction metadata, it is valid
> > > > + * when dev_args 'proto_xtr' has 'tcp' specified.
> > > > + */
> > > > +#define PKT_RX_DYNF_PROTO_XTR_TCP \
> > > > +	(rte_net_ice_dynflag_proto_xtr_tcp_mask)
> > >
> > > Those fields and flags are missing a RTE_ prefix.
> > > (Yes I know we are missing such prefix in rte_mbuf.f)
> > >
> >
> > The "PKT_RX_" in ' lib/librte_mbuf/rte_mbuf_core.h' will be changed to
> > "RTE_PKT_RX_" ?
> 
> I started to do such change a long time ago and never finished.
> Yes it will be changed for 20.11.
> 
> > Or keep the above as it is now, to keep the same ol_flags style, until all
> > the "PKT_RX_" are changed ?
> 
> Please I prefer you change yours, without waiting global change.
> 

OK, will update it.

Patch
diff mbox series

diff --git a/doc/api/doxy-api-index.md b/doc/api/doxy-api-index.md
index 28a5dd37e..dff496be0 100644
--- a/doc/api/doxy-api-index.md
+++ b/doc/api/doxy-api-index.md
@@ -42,6 +42,7 @@  The public API headers are grouped by topics:
   [KNI]                (@ref rte_kni.h),
   [ixgbe]              (@ref rte_pmd_ixgbe.h),
   [i40e]               (@ref rte_pmd_i40e.h),
+  [ice]                (@ref rte_pmd_ice.h),
   [bnxt]               (@ref rte_pmd_bnxt.h),
   [dpaa]               (@ref rte_pmd_dpaa.h),
   [dpaa2]              (@ref rte_pmd_dpaa2.h),
diff --git a/doc/api/doxy-api.conf.in b/doc/api/doxy-api.conf.in
index cb67eb728..350b442a1 100644
--- a/doc/api/doxy-api.conf.in
+++ b/doc/api/doxy-api.conf.in
@@ -12,6 +12,7 @@  INPUT                   = @TOPDIR@/doc/api/doxy-api-index.md \
                           @TOPDIR@/drivers/net/dpaa \
                           @TOPDIR@/drivers/net/dpaa2 \
                           @TOPDIR@/drivers/net/i40e \
+                          @TOPDIR@/drivers/net/ice \
                           @TOPDIR@/drivers/net/ixgbe \
                           @TOPDIR@/drivers/net/softnic \
                           @TOPDIR@/drivers/raw/dpaa2_cmdif \
diff --git a/doc/guides/nics/ice.rst b/doc/guides/nics/ice.rst
index 933f63480..6c60494c7 100644
--- a/doc/guides/nics/ice.rst
+++ b/doc/guides/nics/ice.rst
@@ -82,8 +82,8 @@  Runtime Config Options
 
 - ``Protocol extraction for per queue``
 
-  Configure the RX queues to do protocol extraction into ``rte_mbuf::udata64``
-  for protocol handling acceleration, like checking the TCP SYN packets quickly.
+  Configure the RX queues to do protocol extraction into mbuf for protocol
+  handling acceleration, like checking the TCP SYN packets quickly.
 
   The argument format is::
 
@@ -111,7 +111,8 @@  Runtime Config Options
   This setting means queues 1, 2-3, 8-9 are TCP extraction, queues 10-23 are
   IPv6 extraction, other queues use the default VLAN extraction.
 
-  The extraction will be copied into the lower 32 bit of ``rte_mbuf::udata64``.
+  The extraction metadata is copied into the registered dynamic mbuf field, and
+  the related dynamic mbuf flags is set.
 
   .. table:: Protocol extraction : ``vlan``
 
@@ -175,10 +176,11 @@  Runtime Config Options
 
   TCPHDR2 - Reserved
 
-  Use ``get_proto_xtr_flds(struct rte_mbuf *mb)`` to access the protocol
-  extraction, do not use ``rte_mbuf::udata64`` directly.
+  Use ``rte_net_ice_dynf_proto_xtr_metadata_get`` to access the protocol
+  extraction metadata, and use ``PKT_RX_DYNF_PROTO_XTR_*`` to get the
+  metadata type of ``struct rte_mbuf::ol_flags``.
 
-  The ``dump_proto_xtr_flds(struct rte_mbuf *mb)`` routine shows how to
+  The ``rte_net_ice_dump_proto_xtr_metadata`` routine shows how to
   access the protocol extraction result in ``struct rte_mbuf``.
 
 Driver compilation and testing
diff --git a/drivers/net/ice/ice_ethdev.c b/drivers/net/ice/ice_ethdev.c
index 9d2210a45..5c39ea724 100644
--- a/drivers/net/ice/ice_ethdev.c
+++ b/drivers/net/ice/ice_ethdev.c
@@ -14,6 +14,8 @@ 
 #include "base/ice_flow.h"
 #include "base/ice_dcb.h"
 #include "base/ice_common.h"
+
+#include "rte_pmd_ice.h"
 #include "ice_ethdev.h"
 #include "ice_rxtx.h"
 #include "ice_generic_flow.h"
@@ -30,6 +32,37 @@  static const char * const ice_valid_args[] = {
 	NULL
 };
 
+static const struct rte_mbuf_dynfield ice_proto_xtr_metadata_param = {
+	.name = "ice_dynfield_proto_xtr_metadata",
+	.size = sizeof(uint32_t),
+	.align = __alignof__(uint32_t),
+	.flags = 0,
+};
+
+struct proto_xtr_ol_flag {
+	const struct rte_mbuf_dynflag param;
+	uint64_t *ol_flag;
+	bool required;
+};
+
+static struct proto_xtr_ol_flag ice_proto_xtr_ol_flag_params[] = {
+	[PROTO_XTR_VLAN] = {
+		.param = { .name = "ice_dynflag_proto_xtr_vlan" },
+		.ol_flag = &rte_net_ice_dynflag_proto_xtr_vlan_mask },
+	[PROTO_XTR_IPV4] = {
+		.param = { .name = "ice_dynflag_proto_xtr_ipv4" },
+		.ol_flag = &rte_net_ice_dynflag_proto_xtr_ipv4_mask },
+	[PROTO_XTR_IPV6] = {
+		.param = { .name = "ice_dynflag_proto_xtr_ipv6" },
+		.ol_flag = &rte_net_ice_dynflag_proto_xtr_ipv6_mask },
+	[PROTO_XTR_IPV6_FLOW] = {
+		.param = { .name = "ice_dynflag_proto_xtr_ipv6_flow" },
+		.ol_flag = &rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask },
+	[PROTO_XTR_TCP] = {
+		.param = { .name = "ice_dynflag_proto_xtr_tcp" },
+		.ol_flag = &rte_net_ice_dynflag_proto_xtr_tcp_mask },
+};
+
 #define ICE_DFLT_OUTER_TAG_TYPE ICE_AQ_VSI_OUTER_TAG_VLAN_9100
 
 /* DDP package search path */
@@ -1385,6 +1418,9 @@  ice_init_proto_xtr(struct rte_eth_dev *dev)
 			ICE_DEV_PRIVATE_TO_ADAPTER(dev->data->dev_private);
 	struct ice_pf *pf = ICE_DEV_PRIVATE_TO_PF(dev->data->dev_private);
 	struct ice_hw *hw = ICE_PF_TO_HW(pf);
+	const struct proto_xtr_ol_flag *ol_flag;
+	bool proto_xtr_enable = false;
+	int offset;
 	uint16_t i;
 
 	if (!ice_proto_xtr_support(hw)) {
@@ -1398,10 +1434,56 @@  ice_init_proto_xtr(struct rte_eth_dev *dev)
 		return;
 	}
 
-	for (i = 0; i < pf->lan_nb_qps; i++)
+	for (i = 0; i < pf->lan_nb_qps; i++) {
 		pf->proto_xtr[i] = ad->devargs.proto_xtr[i] != PROTO_XTR_NONE ?
 				   ad->devargs.proto_xtr[i] :
 				   ad->devargs.proto_xtr_dflt;
+
+		if (pf->proto_xtr[i] != PROTO_XTR_NONE) {
+			uint8_t type = pf->proto_xtr[i];
+
+			ice_proto_xtr_ol_flag_params[type].required = true;
+			proto_xtr_enable = true;
+		}
+	}
+
+	if (likely(!proto_xtr_enable))
+		return;
+
+	offset = rte_mbuf_dynfield_register(&ice_proto_xtr_metadata_param);
+	if (unlikely(offset == -1)) {
+		PMD_DRV_LOG(ERR,
+			    "Protocol extraction metadata is disabled in mbuf with error %d",
+			    -rte_errno);
+		return;
+	}
+
+	PMD_DRV_LOG(DEBUG,
+		    "Protocol extraction metadata offset in mbuf is : %d",
+		    offset);
+	rte_net_ice_dynfield_proto_xtr_metadata_offs = offset;
+
+	for (i = 0; i < RTE_DIM(ice_proto_xtr_ol_flag_params); i++) {
+		ol_flag = &ice_proto_xtr_ol_flag_params[i];
+
+		if (!ol_flag->required)
+			continue;
+
+		offset = rte_mbuf_dynflag_register(&ol_flag->param);
+		if (unlikely(offset == -1)) {
+			PMD_DRV_LOG(ERR,
+				    "Protocol extraction offload '%s' failed to register with error %d",
+				    ol_flag->param.name, -rte_errno);
+
+			rte_net_ice_dynfield_proto_xtr_metadata_offs = -1;
+			break;
+		}
+
+		PMD_DRV_LOG(DEBUG,
+			    "Protocol extraction offload '%s' offset in mbuf is : %d",
+			    ol_flag->param.name, offset);
+		*ol_flag->ol_flag = 1ULL << offset;
+	}
 }
 
 /*  Initialize SW parameters of PF */
diff --git a/drivers/net/ice/ice_ethdev.h b/drivers/net/ice/ice_ethdev.h
index de67e5934..c55713cc1 100644
--- a/drivers/net/ice/ice_ethdev.h
+++ b/drivers/net/ice/ice_ethdev.h
@@ -241,6 +241,15 @@  struct ice_vsi {
 	bool offset_loaded;
 };
 
+enum proto_xtr_type {
+	PROTO_XTR_NONE,
+	PROTO_XTR_VLAN,
+	PROTO_XTR_IPV4,
+	PROTO_XTR_IPV6,
+	PROTO_XTR_IPV6_FLOW,
+	PROTO_XTR_TCP,
+};
+
 enum ice_fdir_tunnel_type {
 	ICE_FDIR_TUNNEL_TYPE_NONE = 0,
 	ICE_FDIR_TUNNEL_TYPE_VXLAN,
diff --git a/drivers/net/ice/ice_rxtx.c b/drivers/net/ice/ice_rxtx.c
index 53e4eea7b..18c02979e 100644
--- a/drivers/net/ice/ice_rxtx.c
+++ b/drivers/net/ice/ice_rxtx.c
@@ -5,6 +5,7 @@ 
 #include <rte_ethdev_driver.h>
 #include <rte_net.h>
 
+#include "rte_pmd_ice.h"
 #include "ice_rxtx.h"
 
 #define ICE_TX_CKSUM_OFFLOAD_MASK (		 \
@@ -13,18 +14,36 @@ 
 		PKT_TX_TCP_SEG |		 \
 		PKT_TX_OUTER_IP_CKSUM)
 
-static inline uint8_t
-ice_rxdid_to_proto_xtr_type(uint8_t rxdid)
-{
-	static uint8_t xtr_map[] = {
-		[ICE_RXDID_COMMS_AUX_VLAN]      = PROTO_XTR_VLAN,
-		[ICE_RXDID_COMMS_AUX_IPV4]      = PROTO_XTR_IPV4,
-		[ICE_RXDID_COMMS_AUX_IPV6]      = PROTO_XTR_IPV6,
-		[ICE_RXDID_COMMS_AUX_IPV6_FLOW] = PROTO_XTR_IPV6_FLOW,
-		[ICE_RXDID_COMMS_AUX_TCP]       = PROTO_XTR_TCP,
+/* Offset of mbuf dynamic field for protocol extraction data */
+int rte_net_ice_dynfield_proto_xtr_metadata_offs = -1;
+
+/* Mask of mbuf dynamic flags for protocol extraction type */
+uint64_t rte_net_ice_dynflag_proto_xtr_vlan_mask;
+uint64_t rte_net_ice_dynflag_proto_xtr_ipv4_mask;
+uint64_t rte_net_ice_dynflag_proto_xtr_ipv6_mask;
+uint64_t rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask;
+uint64_t rte_net_ice_dynflag_proto_xtr_tcp_mask;
+
+static inline uint64_t
+ice_rxdid_to_proto_xtr_ol_flag(uint8_t rxdid)
+{
+	static uint64_t *ol_flag_map[] = {
+		[ICE_RXDID_COMMS_AUX_VLAN] =
+				&rte_net_ice_dynflag_proto_xtr_vlan_mask,
+		[ICE_RXDID_COMMS_AUX_IPV4] =
+				&rte_net_ice_dynflag_proto_xtr_ipv4_mask,
+		[ICE_RXDID_COMMS_AUX_IPV6] =
+				&rte_net_ice_dynflag_proto_xtr_ipv6_mask,
+		[ICE_RXDID_COMMS_AUX_IPV6_FLOW] =
+				&rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask,
+		[ICE_RXDID_COMMS_AUX_TCP] =
+				&rte_net_ice_dynflag_proto_xtr_tcp_mask,
 	};
+	uint64_t *ol_flag;
 
-	return rxdid < RTE_DIM(xtr_map) ? xtr_map[rxdid] : PROTO_XTR_NONE;
+	ol_flag = rxdid < RTE_DIM(ol_flag_map) ? ol_flag_map[rxdid] : NULL;
+
+	return ol_flag != NULL ? *ol_flag : 0ULL;
 }
 
 static inline uint8_t
@@ -1325,10 +1344,38 @@  ice_rxd_to_vlan_tci(struct rte_mbuf *mb, volatile union ice_rx_flex_desc *rxdp)
 		   mb->vlan_tci, mb->vlan_tci_outer);
 }
 
+#ifndef RTE_LIBRTE_ICE_16BYTE_RX_DESC
 #define ICE_RX_PROTO_XTR_VALID \
 	((1 << ICE_RX_FLEX_DESC_STATUS1_XTRMD4_VALID_S) | \
 	 (1 << ICE_RX_FLEX_DESC_STATUS1_XTRMD5_VALID_S))
 
+static void
+ice_rxd_to_proto_xtr(struct rte_mbuf *mb,
+		     volatile struct ice_32b_rx_flex_desc_comms *desc)
+{
+	uint16_t stat_err = rte_le_to_cpu_16(desc->status_error1);
+	uint32_t metadata;
+	uint64_t ol_flag;
+
+	if (unlikely(!(stat_err & ICE_RX_PROTO_XTR_VALID)))
+		return;
+
+	ol_flag = ice_rxdid_to_proto_xtr_ol_flag(desc->rxdid);
+	if (unlikely(!ol_flag))
+		return;
+
+	mb->ol_flags |= ol_flag;
+
+	metadata = stat_err & (1 << ICE_RX_FLEX_DESC_STATUS1_XTRMD4_VALID_S) ?
+				rte_le_to_cpu_16(desc->flex_ts.flex.aux0) : 0;
+
+	if (likely(stat_err & (1 << ICE_RX_FLEX_DESC_STATUS1_XTRMD5_VALID_S)))
+		metadata |= rte_le_to_cpu_16(desc->flex_ts.flex.aux1) << 16;
+
+	*RTE_NET_ICE_DYNF_PROTO_XTR_METADATA(mb) = metadata;
+}
+#endif
+
 static inline void
 ice_rxd_to_pkt_fields(struct rte_mbuf *mb,
 		      volatile union ice_rx_flex_desc *rxdp)
@@ -1344,28 +1391,13 @@  ice_rxd_to_pkt_fields(struct rte_mbuf *mb,
 	}
 
 #ifndef RTE_LIBRTE_ICE_16BYTE_RX_DESC
-	init_proto_xtr_flds(mb);
-
-	stat_err = rte_le_to_cpu_16(desc->status_error1);
-	if (stat_err & ICE_RX_PROTO_XTR_VALID) {
-		struct proto_xtr_flds *xtr = get_proto_xtr_flds(mb);
-
-		if (stat_err & (1 << ICE_RX_FLEX_DESC_STATUS1_XTRMD4_VALID_S))
-			xtr->u.raw.data0 =
-				rte_le_to_cpu_16(desc->flex_ts.flex.aux0);
-
-		if (stat_err & (1 << ICE_RX_FLEX_DESC_STATUS1_XTRMD5_VALID_S))
-			xtr->u.raw.data1 =
-				rte_le_to_cpu_16(desc->flex_ts.flex.aux1);
-
-		xtr->type = ice_rxdid_to_proto_xtr_type(desc->rxdid);
-		xtr->magic = PROTO_XTR_MAGIC_ID;
-	}
-
 	if (desc->flow_id != 0xFFFFFFFF) {
 		mb->ol_flags |= PKT_RX_FDIR | PKT_RX_FDIR_ID;
 		mb->hash.fdir.hi = rte_le_to_cpu_32(desc->flow_id);
 	}
+
+	if (unlikely(rte_net_ice_dynf_proto_xtr_metadata_avail()))
+		ice_rxd_to_proto_xtr(mb, desc);
 #endif
 }
 
diff --git a/drivers/net/ice/ice_rxtx.h b/drivers/net/ice/ice_rxtx.h
index 5de618976..9e3d2cd07 100644
--- a/drivers/net/ice/ice_rxtx.h
+++ b/drivers/net/ice/ice_rxtx.h
@@ -5,7 +5,6 @@ 
 #ifndef _ICE_RXTX_H_
 #define _ICE_RXTX_H_
 
-#include "rte_pmd_ice.h"
 #include "ice_ethdev.h"
 
 #define ICE_ALIGN_RING_DESC  32
diff --git a/drivers/net/ice/rte_pmd_ice.h b/drivers/net/ice/rte_pmd_ice.h
index 719487e1e..61201f785 100644
--- a/drivers/net/ice/rte_pmd_ice.h
+++ b/drivers/net/ice/rte_pmd_ice.h
@@ -5,144 +5,223 @@ 
 #ifndef _RTE_PMD_ICE_H_
 #define _RTE_PMD_ICE_H_
 
+/**
+ * @file rte_pmd_ice.h
+ *
+ * ice PMD specific functions.
+ *
+ * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
+ *
+ */
+
 #include <stdio.h>
 #include <rte_mbuf.h>
-#include <rte_ethdev.h>
+#include <rte_mbuf_dyn.h>
 
 #ifdef __cplusplus
 extern "C" {
 #endif
 
-enum proto_xtr_type {
-	PROTO_XTR_NONE,
-	PROTO_XTR_VLAN,
-	PROTO_XTR_IPV4,
-	PROTO_XTR_IPV6,
-	PROTO_XTR_IPV6_FLOW,
-	PROTO_XTR_TCP,
+/**
+ * The supported network protocol extraction metadata format.
+ */
+union proto_xtr_metadata {
+	uint32_t metadata;
+
+	struct {
+		uint16_t data0;
+		uint16_t data1;
+	} raw;
+
+	struct {
+		uint16_t stag_vid:12,
+			 stag_dei:1,
+			 stag_pcp:3;
+		uint16_t ctag_vid:12,
+			 ctag_dei:1,
+			 ctag_pcp:3;
+	} vlan;
+
+	struct {
+		uint16_t protocol:8,
+			 ttl:8;
+		uint16_t tos:8,
+			 ihl:4,
+			 version:4;
+	} ipv4;
+
+	struct {
+		uint16_t hoplimit:8,
+			 nexthdr:8;
+		uint16_t flowhi4:4,
+			 tc:8,
+			 version:4;
+	} ipv6;
+
+	struct {
+		uint16_t flowlo16;
+		uint16_t flowhi4:4,
+			 tc:8,
+			 version:4;
+	} ipv6_flow;
+
+	struct {
+		uint16_t fin:1,
+			 syn:1,
+			 rst:1,
+			 psh:1,
+			 ack:1,
+			 urg:1,
+			 ece:1,
+			 cwr:1,
+			 res1:4,
+			 doff:4;
+		uint16_t rsvd;
+	} tcp;
 };
 
-struct proto_xtr_flds {
-	union {
-		struct {
-			uint16_t data0;
-			uint16_t data1;
-		} raw;
-		struct {
-			uint16_t stag_vid:12,
-				 stag_dei:1,
-				 stag_pcp:3;
-			uint16_t ctag_vid:12,
-				 ctag_dei:1,
-				 ctag_pcp:3;
-		} vlan;
-		struct {
-			uint16_t protocol:8,
-				 ttl:8;
-			uint16_t tos:8,
-				 ihl:4,
-				 version:4;
-		} ipv4;
-		struct {
-			uint16_t hoplimit:8,
-				 nexthdr:8;
-			uint16_t flowhi4:4,
-				 tc:8,
-				 version:4;
-		} ipv6;
-		struct {
-			uint16_t flowlo16;
-			uint16_t flowhi4:4,
-				 tc:8,
-				 version:4;
-		} ipv6_flow;
-		struct {
-			uint16_t fin:1,
-				 syn:1,
-				 rst:1,
-				 psh:1,
-				 ack:1,
-				 urg:1,
-				 ece:1,
-				 cwr:1,
-				 res1:4,
-				 doff:4;
-			uint16_t rsvd;
-		} tcp;
-	} u;
-
-	uint16_t rsvd;
-
-	uint8_t type;
-
-#define PROTO_XTR_MAGIC_ID	0xCE
-	uint8_t magic;
-};
+/* Offset of mbuf dynamic field for protocol extraction data */
+extern int rte_net_ice_dynfield_proto_xtr_metadata_offs;
 
-static inline void
-init_proto_xtr_flds(struct rte_mbuf *mb)
+/* Mask of mbuf dynamic flags for protocol extraction type */
+extern uint64_t rte_net_ice_dynflag_proto_xtr_vlan_mask;
+extern uint64_t rte_net_ice_dynflag_proto_xtr_ipv4_mask;
+extern uint64_t rte_net_ice_dynflag_proto_xtr_ipv6_mask;
+extern uint64_t rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask;
+extern uint64_t rte_net_ice_dynflag_proto_xtr_tcp_mask;
+
+/**
+ * The mbuf dynamic field pointer for protocol extraction metadata.
+ */
+#define RTE_NET_ICE_DYNF_PROTO_XTR_METADATA(m) \
+	RTE_MBUF_DYNFIELD((m), \
+			  rte_net_ice_dynfield_proto_xtr_metadata_offs, \
+			  uint32_t *)
+
+/**
+ * The mbuf dynamic flag for VLAN protocol extraction metadata, it is valid
+ * when dev_args 'proto_xtr' has 'vlan' specified.
+ */
+#define PKT_RX_DYNF_PROTO_XTR_VLAN \
+	(rte_net_ice_dynflag_proto_xtr_vlan_mask)
+
+/**
+ * The mbuf dynamic flag for IPv4 protocol extraction metadata, it is valid
+ * when dev_args 'proto_xtr' has 'ipv4' specified.
+ */
+#define PKT_RX_DYNF_PROTO_XTR_IPV4 \
+	(rte_net_ice_dynflag_proto_xtr_ipv4_mask)
+
+/**
+ * The mbuf dynamic flag for IPv6 protocol extraction metadata, it is valid
+ * when dev_args 'proto_xtr' has 'ipv6' specified.
+ */
+#define PKT_RX_DYNF_PROTO_XTR_IPV6 \
+	(rte_net_ice_dynflag_proto_xtr_ipv6_mask)
+
+/**
+ * The mbuf dynamic flag for IPv6 with flow protocol extraction metadata, it is
+ * valid when dev_args 'proto_xtr' has 'ipv6_flow' specified.
+ */
+#define PKT_RX_DYNF_PROTO_XTR_IPV6_FLOW \
+	(rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask)
+
+/**
+ * The mbuf dynamic flag for TCP protocol extraction metadata, it is valid
+ * when dev_args 'proto_xtr' has 'tcp' specified.
+ */
+#define PKT_RX_DYNF_PROTO_XTR_TCP \
+	(rte_net_ice_dynflag_proto_xtr_tcp_mask)
+
+/**
+ * Check if mbuf dynamic field for protocol extraction metadata is registered.
+ *
+ * @return
+ *   True if registered, false otherwise.
+ */
+__rte_experimental
+static __rte_always_inline int
+rte_net_ice_dynf_proto_xtr_metadata_avail(void)
 {
-	mb->udata64 = 0;
+	return rte_net_ice_dynfield_proto_xtr_metadata_offs != -1;
 }
 
-static inline struct proto_xtr_flds *
-get_proto_xtr_flds(struct rte_mbuf *mb)
+/**
+ * Get the mbuf dynamic field for protocol extraction metadata.
+ *
+ * @param m
+ *    The pointer to the mbuf.
+ * @return
+ *   The saved protocol extraction metadata.
+ */
+__rte_experimental
+static __rte_always_inline uint32_t
+rte_net_ice_dynf_proto_xtr_metadata_get(struct rte_mbuf *m)
 {
-	RTE_BUILD_BUG_ON(sizeof(struct proto_xtr_flds) > sizeof(mb->udata64));
-
-	return (struct proto_xtr_flds *)&mb->udata64;
+	return *RTE_NET_ICE_DYNF_PROTO_XTR_METADATA(m);
 }
 
+/**
+ * Dump the mbuf dynamic field for protocol extraction metadata.
+ *
+ * @param m
+ *    The pointer to the mbuf.
+ */
+__rte_experimental
 static inline void
-dump_proto_xtr_flds(struct rte_mbuf *mb)
+rte_net_ice_dump_proto_xtr_metadata(struct rte_mbuf *m)
 {
-	struct proto_xtr_flds *xtr = get_proto_xtr_flds(mb);
+	union proto_xtr_metadata data;
 
-	if (xtr->magic != PROTO_XTR_MAGIC_ID || xtr->type == PROTO_XTR_NONE)
+	if (!rte_net_ice_dynf_proto_xtr_metadata_avail())
 		return;
 
-	printf(" - Protocol Extraction:[0x%04x:0x%04x],",
-	       xtr->u.raw.data0, xtr->u.raw.data1);
-
-	if (xtr->type == PROTO_XTR_VLAN)
-		printf("vlan,stag=%u:%u:%u,ctag=%u:%u:%u ",
-		       xtr->u.vlan.stag_pcp,
-		       xtr->u.vlan.stag_dei,
-		       xtr->u.vlan.stag_vid,
-		       xtr->u.vlan.ctag_pcp,
-		       xtr->u.vlan.ctag_dei,
-		       xtr->u.vlan.ctag_vid);
-	else if (xtr->type == PROTO_XTR_IPV4)
-		printf("ipv4,ver=%u,hdrlen=%u,tos=%u,ttl=%u,proto=%u ",
-		       xtr->u.ipv4.version,
-		       xtr->u.ipv4.ihl,
-		       xtr->u.ipv4.tos,
-		       xtr->u.ipv4.ttl,
-		       xtr->u.ipv4.protocol);
-	else if (xtr->type == PROTO_XTR_IPV6)
-		printf("ipv6,ver=%u,tc=%u,flow_hi4=0x%x,nexthdr=%u,hoplimit=%u ",
-		       xtr->u.ipv6.version,
-		       xtr->u.ipv6.tc,
-		       xtr->u.ipv6.flowhi4,
-		       xtr->u.ipv6.nexthdr,
-		       xtr->u.ipv6.hoplimit);
-	else if (xtr->type == PROTO_XTR_IPV6_FLOW)
-		printf("ipv6_flow,ver=%u,tc=%u,flow=0x%x%04x ",
-		       xtr->u.ipv6_flow.version,
-		       xtr->u.ipv6_flow.tc,
-		       xtr->u.ipv6_flow.flowhi4,
-		       xtr->u.ipv6_flow.flowlo16);
-	else if (xtr->type == PROTO_XTR_TCP)
-		printf("tcp,doff=%u,flags=%s%s%s%s%s%s%s%s ",
-		       xtr->u.tcp.doff,
-		       xtr->u.tcp.cwr ? "C" : "",
-		       xtr->u.tcp.ece ? "E" : "",
-		       xtr->u.tcp.urg ? "U" : "",
-		       xtr->u.tcp.ack ? "A" : "",
-		       xtr->u.tcp.psh ? "P" : "",
-		       xtr->u.tcp.rst ? "R" : "",
-		       xtr->u.tcp.syn ? "S" : "",
-		       xtr->u.tcp.fin ? "F" : "");
+	data.metadata = rte_net_ice_dynf_proto_xtr_metadata_get(m);
+
+	if (m->ol_flags & PKT_RX_DYNF_PROTO_XTR_VLAN)
+		printf(" - Protocol Extraction:[0x%04x:0x%04x],vlan,stag=%u:%u:%u,ctag=%u:%u:%u",
+		       data.raw.data0, data.raw.data1,
+		       data.vlan.stag_pcp,
+		       data.vlan.stag_dei,
+		       data.vlan.stag_vid,
+		       data.vlan.ctag_pcp,
+		       data.vlan.ctag_dei,
+		       data.vlan.ctag_vid);
+	else if (m->ol_flags & PKT_RX_DYNF_PROTO_XTR_IPV4)
+		printf(" - Protocol Extraction:[0x%04x:0x%04x],ipv4,ver=%u,hdrlen=%u,tos=%u,ttl=%u,proto=%u",
+		       data.raw.data0, data.raw.data1,
+		       data.ipv4.version,
+		       data.ipv4.ihl,
+		       data.ipv4.tos,
+		       data.ipv4.ttl,
+		       data.ipv4.protocol);
+	else if (m->ol_flags & PKT_RX_DYNF_PROTO_XTR_IPV6)
+		printf(" - Protocol Extraction:[0x%04x:0x%04x],ipv6,ver=%u,tc=%u,flow_hi4=0x%x,nexthdr=%u,hoplimit=%u",
+		       data.raw.data0, data.raw.data1,
+		       data.ipv6.version,
+		       data.ipv6.tc,
+		       data.ipv6.flowhi4,
+		       data.ipv6.nexthdr,
+		       data.ipv6.hoplimit);
+	else if (m->ol_flags & PKT_RX_DYNF_PROTO_XTR_IPV6_FLOW)
+		printf(" - Protocol Extraction:[0x%04x:0x%04x],ipv6_flow,ver=%u,tc=%u,flow=0x%x%04x",
+		       data.raw.data0, data.raw.data1,
+		       data.ipv6_flow.version,
+		       data.ipv6_flow.tc,
+		       data.ipv6_flow.flowhi4,
+		       data.ipv6_flow.flowlo16);
+	else if (m->ol_flags & PKT_RX_DYNF_PROTO_XTR_TCP)
+		printf(" - Protocol Extraction:[0x%04x:0x%04x],tcp,doff=%u,flags=%s%s%s%s%s%s%s%s",
+		       data.raw.data0, data.raw.data1,
+		       data.tcp.doff,
+		       data.tcp.cwr ? "C" : "",
+		       data.tcp.ece ? "E" : "",
+		       data.tcp.urg ? "U" : "",
+		       data.tcp.ack ? "A" : "",
+		       data.tcp.psh ? "P" : "",
+		       data.tcp.rst ? "R" : "",
+		       data.tcp.syn ? "S" : "",
+		       data.tcp.fin ? "F" : "");
 }
 
 #ifdef __cplusplus
diff --git a/drivers/net/ice/rte_pmd_ice_version.map b/drivers/net/ice/rte_pmd_ice_version.map
index 7b23b609d..0677a1c69 100644
--- a/drivers/net/ice/rte_pmd_ice_version.map
+++ b/drivers/net/ice/rte_pmd_ice_version.map
@@ -2,3 +2,15 @@  DPDK_19.02 {
 
 	local: *;
 };
+
+EXPERIMENTAL {
+	global:
+
+	# added in 19.11
+	rte_net_ice_dynfield_proto_xtr_metadata_offs;
+	rte_net_ice_dynflag_proto_xtr_vlan_mask;
+	rte_net_ice_dynflag_proto_xtr_ipv4_mask;
+	rte_net_ice_dynflag_proto_xtr_ipv6_mask;
+	rte_net_ice_dynflag_proto_xtr_ipv6_flow_mask;
+	rte_net_ice_dynflag_proto_xtr_tcp_mask;
+};