[dpdk-stable] [PATCH 00/24] backport for 19.11.4
Wei Hu (Xavier)
huwei013 at chinasoftinc.com
Tue Aug 18 05:25:04 CEST 2020
Hi, Luca Boccassi
On 2020/8/17 21:42, Luca Boccassi wrote:
> On Mon, 2020-08-17 at 19:54 +0800, Wei Hu (Xavier) wrote:
>> Hi, Luca Boccassi
>>
>> On 2020/8/17 17:51, Luca Boccassi wrote:
>>> On Mon, 2020-08-17 at 17:25 +0800, Wei Hu (Xavier) wrote:
>>>> This series are backport for 19.11.4 about hns3 PMD driver, bonding
>>>> and testpmd.
>>>>
>>>> Chengchang Tang (6):
>>>> net/hns3: remove restriction on setting VF MTU
>>>> net/hns3: support promiscuous and allmulticast mode for VF
>>>> net/hns3: fix adding multicast MAC address
>>>> net/hns3: decrease non-nearby memory access in Rx
>>>> net/hns3: support setting VF PVID by PF driver
>>>> net/hns3: cleanup duplicated code on processing TSO in Tx
>>>>
>>>> Chengwen Feng (1):
>>>> net/hns3: get Rx/Tx queue fbd in xstats
>>>>
>>>> Hao Chen (1):
>>>> net/hns3: optimize default RSS algorithm
>>>>
>>>> Hongbo Zheng (3):
>>>> net/hns3: get link status change through mailbox
>>>> net/hns3: support TSO
>>>> net/hns3: check TSO segment size during Tx
>>>>
>>>> Lijun Ou (4):
>>>> net/hns3: remove unnecessary branch
>>>> net/hns3: support symmetric RSS
>>>> net/hns3: report Tx descriptor segment limitations
>>>> app/testpmd: remove hardcoded descriptors limit
>>>>
>>>> Wei Hu (Xavier) (8):
>>>> net/hns3: support setting VF MAC address by PF driver
>>>> net/hns3: get Tx abnormal errors in xstats
>>>> net/hns3: get PCI revision ID
>>>> net/hns3: support LRO
>>>> net/hns3: get device capability in primary process
>>>> net/hns3: support copper media type
>>>> net/hns3: fix reassembling multiple segment packets in Tx
>>>> net/hns3: fix inserted VLAN tag position in Tx
>>>>
>>>> Weifeng Li (1):
>>>> net/bonding: change state machine to defaulted
>>>>
>>>> app/test-pmd/cmdline.c | 11 +-
>>>> app/test-pmd/testpmd.c | 171 ++++++
>>>> app/test-pmd/testpmd.h | 2 +
>>>> doc/guides/nics/features/hns3.ini | 2 +
>>>> doc/guides/nics/features/hns3_vf.ini | 4 +
>>>> doc/guides/nics/hns3.rst | 2 +
>>>> drivers/net/bonding/eth_bond_8023ad_private.h | 3 +
>>>> drivers/net/bonding/rte_eth_bond_8023ad.c | 21 +-
>>>> drivers/net/hns3/hns3_ethdev.c | 269 ++++++--
>>>> drivers/net/hns3/hns3_ethdev.h | 34 +-
>>>> drivers/net/hns3/hns3_ethdev_vf.c | 578 +++++++++++++++---
>>>> drivers/net/hns3/hns3_flow.c | 12 +-
>>>> drivers/net/hns3/hns3_mbx.c | 93 +++
>>>> drivers/net/hns3/hns3_mbx.h | 13 +
>>>> drivers/net/hns3/hns3_rss.c | 35 +-
>>>> drivers/net/hns3/hns3_rss.h | 6 +-
>>>> drivers/net/hns3/hns3_rxtx.c | 525 ++++++++++++++--
>>>> drivers/net/hns3/hns3_rxtx.h | 65 +-
>>>> drivers/net/hns3/hns3_stats.c | 243 +++++---
>>>> drivers/net/hns3/hns3_stats.h | 3 +-
>>>> 20 files changed, 1779 insertions(+), 313 deletions(-)
>>> Hi,
>>>
>>> If I understand correctly, this series adds support for 6 new features:
>>>
>>> --- a/doc/guides/nics/features/hns3.ini
>>> +++ b/doc/guides/nics/features/hns3.ini
>>> @@ -8,6 +8,8 @@ Link status = Y
>>> Rx interrupt = Y
>>> MTU update = Y
>>> Jumbo frame = Y
>>> +TSO = Y
>>> +LRO = Y
>>> Promiscuous mode = Y
>>> Allmulticast mode = Y
>>> Unicast MAC filter = Y
>>>
>>> --- a/doc/guides/nics/features/hns3_vf.ini
>>> +++ b/doc/guides/nics/features/hns3_vf.ini
>>> @@ -8,6 +8,10 @@ Link status = Y
>>> Rx interrupt = Y
>>> MTU update = Y
>>> Jumbo frame = Y
>>> +TSO = Y
>>> +LRO = Y
>>> +Promiscuous mode = Y
>>> +Allmulticast mode = Y
>>> Unicast MAC filter = Y
>>> Multicast MAC filter = Y
>>> RSS hash = Y
>>>
>>> None of those are available in 20.02, and LRO is not available in
>>> 20.05.
>>>
>>> What happens if a user upgrades from 19.11.4 with any of these features
>>> enabled to 20.02 or 20.05?
>> If a user upgrades from 19.11.4 to 20.02 or 20.05 based on hns3 network
>> engine,
>>
>> some features of hns3 will not be supported. But in many cases
>> customers prefer
>>
>> to use the LTS version. We will recommend them to upgrade from 19.11.0
>> to 19.11.4,
>>
>> and will recommend that customers upgrade from 19.11.4 to 20.11.0 after the
>>
>> release of 20.11.0.
>>
>>
>> Regards
>>
>> Xavier
> Sorry, but I don't feel too comfortable adding multiple new features to
> the LTS branch, especially as they are not available in the following
> release. Stable is valuable because it is stable. Since we got ABI
> stability now, users that want new features can upgrade to
> 20.02/20.05/20.08 without any application breakages.
>
> Would it be possible to break this series into a smaller bugfix-only
> patchset?
OK, I will send a smaller V2.
Thanks, Xavier
More information about the stable
mailing list