[ovs-dev] [PATCH v3 0/6] MFEX Optimizations IPv6 + Hashing

Ilya Maximets i.maximets at ovn.org
Wed Oct 13 14:54:21 UTC 2021


On 9/28/21 19:13, Van Haaren, Harry wrote:
>> -----Original Message-----
>> From: Ilya Maximets <i.maximets at ovn.org>
>> Sent: Tuesday, September 21, 2021 1:45 PM
>> To: Amber, Kumar <kumar.amber at intel.com>; ovs-dev at openvswitch.org
>> Cc: ktraynor at redhat.com; i.maximets at ovn.org; Stokes, Ian
>> <ian.stokes at intel.com>; fbl at sysclose.org; echaudro at redhat.com; Van Haaren,
>> Harry <harry.van.haaren at intel.com>
>> Subject: Re: [PATCH v3 0/6] MFEX Optimizations IPv6 + Hashing
>>
>> On 9/21/21 12:23, Kumar Amber wrote:
>>> ---
>>> v3:
>>> - rebase to master.
>>> v2:
>>> - fix the CI build.
>>> - fix check-patch for co-author.
>>> ---
>>>
>>> The patch-set introduces AVX512 optimizations of IPv6
>>> traffic profiles and hashing improvements for all AVX512
>>> supported traffic profiles for IPv4 and IPv6.
>>>
>>> Kumar Amber (6):
>>>   dpif-netdev/mfex: Add AVX512 basic ipv6 traffic profiles
>>>   dpif-netdev/mfex: Add AVX512 vlan ipv6 traffic profiles
>>>   dpif-netdev/mfex: Add packet hash check to autovalidator
>>>   dpif-netdev/mfex: Add ipv4 profile based hashing
>>>   dpif-netdev/mfex: Add ipv6 profile based hashing
>>>   dpif-netdev/mfex: Avoid hashing when opt mfex called
>>>
>>>  NEWS                              |   7 +
>>>  lib/automake.mk                   |   1 +
>>>  lib/dpif-netdev-avx512.c          |   6 +-
>>>  lib/dpif-netdev-extract-avx512.c  | 348 +++++++++++++++++++++++++++++-
>>>  lib/dpif-netdev-private-extract.c |  63 +++++-
>>>  lib/dpif-netdev-private-extract.h |  12 ++
>>>  tests/pcap/mfex_test.pcap         | Bin 416 -> 632 bytes
>>>  7 files changed, 432 insertions(+), 5 deletions(-)
>>>
>>
>> Hi.  A few months ago I was told that it's easy for Intel to set up CI
>> to test upstream patches with AVX512 features enabled.  Is there any
>> progress on that front?
> 
> Yes there is progress on that front, as you know Ian and Aaron are working on that,

I didn't know that as this information is not public.

> and status updates available if you're particularly interested in its progress.

I was interested, so I asked in this email.

> 
>> My point is that we should refrain from adding new features in this
>> area until we have a proper CI.
> 
> There is already various CI efforts for OVS, and as you know there is ongoing efforts 
> to add AVX512 specifically, and report back on patchwork.
> 
>> Especially considering the unit test failure you reported yesterday, which is
>> supposedly related to AVX512 optimizations.
> 
> I don't know why you say this is related to AVX512 - it is not. See the detailed reply
> Amber sent with details of how the test-case assumed SW based murmur hash output.

Sure, now I know that problem was unrelated.  But the more or less detailed
reply was sent a week after my previous email, so I had no clue what was the
actual problem with a test at the moment of writing and the week after.

> 
>> // Marking this patch-set as deferred for now.
> 
> This is not acceptable, deferring patchsets just because the AVX512 CI isn't in place was never agreed on.
> I do not like that AVX512 was "blamed" for the above unit-test failure, and now other AVX512 patchsets
> are being deferred and ignored as a result of that mistaken blame.
> 
> @Amber, Kumar, please mark this v3 patchset as "new".
> 
>> Best regards, Ilya Maximets.
> 
> Regards, -Harry
> 



More information about the dev mailing list