doc: cleanup flow mark Rx offload deprecation notice

Message ID 20211102125058.3872552-1-andrew.rybchenko@oktetlabs.ru (mailing list archive)
State Accepted, archived
Delegated to: Ferruh Yigit
Headers
Series doc: cleanup flow mark Rx offload deprecation notice |

Checks

Context Check Description
ci/checkpatch warning coding style issues
ci/iol-spell-check-testing success Testing PASS
ci/github-robot: build success github build: passed
ci/iol-mellanox-Performance success Performance Testing PASS
ci/iol-broadcom-Functional success Functional Testing PASS
ci/iol-broadcom-Performance success Performance Testing PASS
ci/iol-aarch64-unit-testing success Testing PASS
ci/Intel-compilation success Compilation OK
ci/intel-Testing success Testing PASS
ci/iol-intel-Performance success Performance Testing PASS
ci/iol-intel-Functional success Functional Testing PASS
ci/iol-aarch64-compile-testing success Testing PASS
ci/iol-x86_64-compile-testing success Testing PASS
ci/iol-x86_64-unit-testing success Testing PASS

Commit Message

Andrew Rybchenko Nov. 2, 2021, 12:50 p.m. UTC
  The problem is solved using Rx metadata delivery negotiation API [1].

[1] commit f6d8a6d3fad7 ("ethdev: negotiate delivery of packet metadata from HW to PMD")

Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
---
 doc/guides/rel_notes/deprecation.rst | 8 --------
 1 file changed, 8 deletions(-)
  

Comments

Ferruh Yigit Nov. 2, 2021, 1:47 p.m. UTC | #1
On 11/2/2021 12:50 PM, Andrew Rybchenko wrote:
> The problem is solved using Rx metadata delivery negotiation API [1].
> 

Can this API replace RX_OFFLOAD_RSS_HASH too?

> [1] commit f6d8a6d3fad7 ("ethdev: negotiate delivery of packet metadata from HW to PMD")
> 
> Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
> ---
>   doc/guides/rel_notes/deprecation.rst | 8 --------
>   1 file changed, 8 deletions(-)
> 
> diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst
> index 4366015b01..ec5073908f 100644
> --- a/doc/guides/rel_notes/deprecation.rst
> +++ b/doc/guides/rel_notes/deprecation.rst
> @@ -69,14 +69,6 @@ Deprecation Notices
>     and the related structures (``rte_fdir_*`` and ``rte_eth_fdir_*``),
>     will be removed in DPDK 20.11.
>   
> -* ethdev: New offload flags ``RTE_ETH_RX_OFFLOAD_FLOW_MARK`` will be added in 19.11.
> -  This will allow application to enable or disable PMDs from updating
> -  ``rte_mbuf::hash::fdir``.
> -  This scheme will allow PMDs to avoid writes to ``rte_mbuf`` fields on Rx and
> -  thereby improve Rx performance if application wishes do so.
> -  In 19.11 PMDs will still update the field even when the offload is not
> -  enabled.
> -
>   * ethdev: Announce moving from dedicated modify function for each field,
>     to using the general ``rte_flow_modify_field`` action.
>   
>
  
Andrew Rybchenko Nov. 2, 2021, 2:09 p.m. UTC | #2
On 11/2/21 4:47 PM, Ferruh Yigit wrote:
> On 11/2/2021 12:50 PM, Andrew Rybchenko wrote:
>> The problem is solved using Rx metadata delivery negotiation API [1].
>>
> 
> Can this API replace RX_OFFLOAD_RSS_HASH too?

It can, but it should not. RSS hash is an offload since it is a
calculation of the hash in HW and delivery of the result to SW.
Like checksums: check in HW and delivery of the result to SW.

> 
>> [1] commit f6d8a6d3fad7 ("ethdev: negotiate delivery of packet 
>> metadata from HW to PMD")
>>
>> Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
>> ---
>>   doc/guides/rel_notes/deprecation.rst | 8 --------
>>   1 file changed, 8 deletions(-)
>>
>> diff --git a/doc/guides/rel_notes/deprecation.rst 
>> b/doc/guides/rel_notes/deprecation.rst
>> index 4366015b01..ec5073908f 100644
>> --- a/doc/guides/rel_notes/deprecation.rst
>> +++ b/doc/guides/rel_notes/deprecation.rst
>> @@ -69,14 +69,6 @@ Deprecation Notices
>>     and the related structures (``rte_fdir_*`` and ``rte_eth_fdir_*``),
>>     will be removed in DPDK 20.11.
>> -* ethdev: New offload flags ``RTE_ETH_RX_OFFLOAD_FLOW_MARK`` will be 
>> added in 19.11.
>> -  This will allow application to enable or disable PMDs from updating
>> -  ``rte_mbuf::hash::fdir``.
>> -  This scheme will allow PMDs to avoid writes to ``rte_mbuf`` fields 
>> on Rx and
>> -  thereby improve Rx performance if application wishes do so.
>> -  In 19.11 PMDs will still update the field even when the offload is not
>> -  enabled.
>> -
>>   * ethdev: Announce moving from dedicated modify function for each 
>> field,
>>     to using the general ``rte_flow_modify_field`` action.
>>
  
Ferruh Yigit Nov. 2, 2021, 2:34 p.m. UTC | #3
On 11/2/2021 2:09 PM, Andrew Rybchenko wrote:
> On 11/2/21 4:47 PM, Ferruh Yigit wrote:
>> On 11/2/2021 12:50 PM, Andrew Rybchenko wrote:
>>> The problem is solved using Rx metadata delivery negotiation API [1].
>>>
>>
>> Can this API replace RX_OFFLOAD_RSS_HASH too?
> 
> It can, but it should not. RSS hash is an offload since it is a
> calculation of the hash in HW and delivery of the result to SW.
> Like checksums: check in HW and delivery of the result to SW.
> 

There is a slight difference, RSS hash is always calculated if RSS is enabled,
independent from RSS offload enabled or not.

That is why some PMDs always provides RSS hash and force enables this offload,
I would be happy to get rid of this quirk if possible.

>>
>>> [1] commit f6d8a6d3fad7 ("ethdev: negotiate delivery of packet metadata from HW to PMD")
>>>
>>> Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
>>> ---
>>>   doc/guides/rel_notes/deprecation.rst | 8 --------
>>>   1 file changed, 8 deletions(-)
>>>
>>> diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst
>>> index 4366015b01..ec5073908f 100644
>>> --- a/doc/guides/rel_notes/deprecation.rst
>>> +++ b/doc/guides/rel_notes/deprecation.rst
>>> @@ -69,14 +69,6 @@ Deprecation Notices
>>>     and the related structures (``rte_fdir_*`` and ``rte_eth_fdir_*``),
>>>     will be removed in DPDK 20.11.
>>> -* ethdev: New offload flags ``RTE_ETH_RX_OFFLOAD_FLOW_MARK`` will be added in 19.11.
>>> -  This will allow application to enable or disable PMDs from updating
>>> -  ``rte_mbuf::hash::fdir``.
>>> -  This scheme will allow PMDs to avoid writes to ``rte_mbuf`` fields on Rx and
>>> -  thereby improve Rx performance if application wishes do so.
>>> -  In 19.11 PMDs will still update the field even when the offload is not
>>> -  enabled.
>>> -
>>>   * ethdev: Announce moving from dedicated modify function for each field,
>>>     to using the general ``rte_flow_modify_field`` action.
>>>
>
  
Ferruh Yigit Nov. 5, 2021, 4:03 p.m. UTC | #4
On 11/2/2021 12:50 PM, Andrew Rybchenko wrote:
> The problem is solved using Rx metadata delivery negotiation API [1].
> 
> [1] commit f6d8a6d3fad7 ("ethdev: negotiate delivery of packet metadata from HW to PMD")
> 
> Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
> ---
>   doc/guides/rel_notes/deprecation.rst | 8 --------
>   1 file changed, 8 deletions(-)
> 
> diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst
> index 4366015b01..ec5073908f 100644
> --- a/doc/guides/rel_notes/deprecation.rst
> +++ b/doc/guides/rel_notes/deprecation.rst
> @@ -69,14 +69,6 @@ Deprecation Notices
>     and the related structures (``rte_fdir_*`` and ``rte_eth_fdir_*``),
>     will be removed in DPDK 20.11.
>   
> -* ethdev: New offload flags ``RTE_ETH_RX_OFFLOAD_FLOW_MARK`` will be added in 19.11.
> -  This will allow application to enable or disable PMDs from updating
> -  ``rte_mbuf::hash::fdir``.
> -  This scheme will allow PMDs to avoid writes to ``rte_mbuf`` fields on Rx and
> -  thereby improve Rx performance if application wishes do so.
> -  In 19.11 PMDs will still update the field even when the offload is not
> -  enabled.
> -
>   * ethdev: Announce moving from dedicated modify function for each field,
>     to using the general ``rte_flow_modify_field`` action.
>   
> 

+ Pavan, who added the deprecation notice at first place, plus people who ack'ed
the patch are cc'ed.
  
Ferruh Yigit Nov. 19, 2021, 5:13 p.m. UTC | #5
On 11/5/2021 4:03 PM, Ferruh Yigit wrote:
> On 11/2/2021 12:50 PM, Andrew Rybchenko wrote:
>> The problem is solved using Rx metadata delivery negotiation API [1].
>>
>> [1] commit f6d8a6d3fad7 ("ethdev: negotiate delivery of packet metadata from HW to PMD")
>>
>> Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
>> ---
>>   doc/guides/rel_notes/deprecation.rst | 8 --------
>>   1 file changed, 8 deletions(-)
>>
>> diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst
>> index 4366015b01..ec5073908f 100644
>> --- a/doc/guides/rel_notes/deprecation.rst
>> +++ b/doc/guides/rel_notes/deprecation.rst
>> @@ -69,14 +69,6 @@ Deprecation Notices
>>     and the related structures (``rte_fdir_*`` and ``rte_eth_fdir_*``),
>>     will be removed in DPDK 20.11.
>> -* ethdev: New offload flags ``RTE_ETH_RX_OFFLOAD_FLOW_MARK`` will be added in 19.11.
>> -  This will allow application to enable or disable PMDs from updating
>> -  ``rte_mbuf::hash::fdir``.
>> -  This scheme will allow PMDs to avoid writes to ``rte_mbuf`` fields on Rx and
>> -  thereby improve Rx performance if application wishes do so.
>> -  In 19.11 PMDs will still update the field even when the offload is not
>> -  enabled.
>> -
>>   * ethdev: Announce moving from dedicated modify function for each field,
>>     to using the general ``rte_flow_modify_field`` action.
>>
> 
> + Pavan, who added the deprecation notice at first place, plus people who ack'ed
> the patch are cc'ed.

Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>

Pavan, Jerin, Hemant, Konstantin, any objection to remove the deprecation notice?
  
Jerin Jacob Nov. 22, 2021, 4:10 a.m. UTC | #6
On Fri, Nov 19, 2021 at 11:23 PM Ferruh Yigit <ferruh.yigit@intel.com> wrote:
>
> On 11/5/2021 4:03 PM, Ferruh Yigit wrote:
> > On 11/2/2021 12:50 PM, Andrew Rybchenko wrote:
> >> The problem is solved using Rx metadata delivery negotiation API [1].
> >>
> >> [1] commit f6d8a6d3fad7 ("ethdev: negotiate delivery of packet metadata from HW to PMD")
> >>
> >> Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
> >> ---
> >>   doc/guides/rel_notes/deprecation.rst | 8 --------
> >>   1 file changed, 8 deletions(-)
> >>
> >> diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst
> >> index 4366015b01..ec5073908f 100644
> >> --- a/doc/guides/rel_notes/deprecation.rst
> >> +++ b/doc/guides/rel_notes/deprecation.rst
> >> @@ -69,14 +69,6 @@ Deprecation Notices
> >>     and the related structures (``rte_fdir_*`` and ``rte_eth_fdir_*``),
> >>     will be removed in DPDK 20.11.
> >> -* ethdev: New offload flags ``RTE_ETH_RX_OFFLOAD_FLOW_MARK`` will be added in 19.11.
> >> -  This will allow application to enable or disable PMDs from updating
> >> -  ``rte_mbuf::hash::fdir``.
> >> -  This scheme will allow PMDs to avoid writes to ``rte_mbuf`` fields on Rx and
> >> -  thereby improve Rx performance if application wishes do so.
> >> -  In 19.11 PMDs will still update the field even when the offload is not
> >> -  enabled.
> >> -
> >>   * ethdev: Announce moving from dedicated modify function for each field,
> >>     to using the general ``rte_flow_modify_field`` action.
> >>
> >
> > + Pavan, who added the deprecation notice at first place, plus people who ack'ed
> > the patch are cc'ed.
>
> Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
>
> Pavan, Jerin, Hemant, Konstantin, any objection to remove the deprecation notice?

Acked-by: Jerin Jacob <jerinj@marvell.com>
  
Ferruh Yigit Nov. 24, 2021, 12:17 p.m. UTC | #7
On 11/22/2021 4:10 AM, Jerin Jacob wrote:
> On Fri, Nov 19, 2021 at 11:23 PM Ferruh Yigit <ferruh.yigit@intel.com> wrote:
>>
>> On 11/5/2021 4:03 PM, Ferruh Yigit wrote:
>>> On 11/2/2021 12:50 PM, Andrew Rybchenko wrote:
>>>> The problem is solved using Rx metadata delivery negotiation API [1].
>>>>
>>>> [1] commit f6d8a6d3fad7 ("ethdev: negotiate delivery of packet metadata from HW to PMD")
>>>>
>>>> Signed-off-by: Andrew Rybchenko <andrew.rybchenko@oktetlabs.ru>
>>>> ---
>>>>    doc/guides/rel_notes/deprecation.rst | 8 --------
>>>>    1 file changed, 8 deletions(-)
>>>>
>>>> diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst
>>>> index 4366015b01..ec5073908f 100644
>>>> --- a/doc/guides/rel_notes/deprecation.rst
>>>> +++ b/doc/guides/rel_notes/deprecation.rst
>>>> @@ -69,14 +69,6 @@ Deprecation Notices
>>>>      and the related structures (``rte_fdir_*`` and ``rte_eth_fdir_*``),
>>>>      will be removed in DPDK 20.11.
>>>> -* ethdev: New offload flags ``RTE_ETH_RX_OFFLOAD_FLOW_MARK`` will be added in 19.11.
>>>> -  This will allow application to enable or disable PMDs from updating
>>>> -  ``rte_mbuf::hash::fdir``.
>>>> -  This scheme will allow PMDs to avoid writes to ``rte_mbuf`` fields on Rx and
>>>> -  thereby improve Rx performance if application wishes do so.
>>>> -  In 19.11 PMDs will still update the field even when the offload is not
>>>> -  enabled.
>>>> -
>>>>    * ethdev: Announce moving from dedicated modify function for each field,
>>>>      to using the general ``rte_flow_modify_field`` action.
>>>>
>>>
>>> + Pavan, who added the deprecation notice at first place, plus people who ack'ed
>>> the patch are cc'ed.
>>
>> Acked-by: Ferruh Yigit <ferruh.yigit@intel.com>
>>
>> Pavan, Jerin, Hemant, Konstantin, any objection to remove the deprecation notice?
> 
> Acked-by: Jerin Jacob <jerinj@marvell.com>
> 

There is no objection, so proceeding with the patch, thanks for cleanup.

Applied to dpdk-next-net/main, thanks.
  

Patch

diff --git a/doc/guides/rel_notes/deprecation.rst b/doc/guides/rel_notes/deprecation.rst
index 4366015b01..ec5073908f 100644
--- a/doc/guides/rel_notes/deprecation.rst
+++ b/doc/guides/rel_notes/deprecation.rst
@@ -69,14 +69,6 @@  Deprecation Notices
   and the related structures (``rte_fdir_*`` and ``rte_eth_fdir_*``),
   will be removed in DPDK 20.11.
 
-* ethdev: New offload flags ``RTE_ETH_RX_OFFLOAD_FLOW_MARK`` will be added in 19.11.
-  This will allow application to enable or disable PMDs from updating
-  ``rte_mbuf::hash::fdir``.
-  This scheme will allow PMDs to avoid writes to ``rte_mbuf`` fields on Rx and
-  thereby improve Rx performance if application wishes do so.
-  In 19.11 PMDs will still update the field even when the offload is not
-  enabled.
-
 * ethdev: Announce moving from dedicated modify function for each field,
   to using the general ``rte_flow_modify_field`` action.