Date:   Thu, 20 Oct 2022 09:24:32 +0200
From:   Jan Karcher <jaka@...ux.ibm.com>
To:     "D. Wythe" <alibuda@...ux.alibaba.com>
Cc:     kuba@...nel.org, davem@...emloft.net, netdev@...r.kernel.org,
        linux-s390@...r.kernel.org, linux-rdma@...r.kernel.org
Subject: Re: [PATCH net-next v3 00/10] optimize the parallelism of SMC-R
 connections



On 20.10.2022 09:00, D. Wythe wrote:
> 
> Hi Jan,
> 
> Sorry for the long delay, The main purpose of v3 is to put optimizes 
> also works on SMC-D, dues to the environment,
> I can only tests it in SMC-R, so please help us to verify the stability 
> and functional in SMC-D,
> Thanks a lot.
> 
> If you have any problems, please let us know.
> 
> Besides, PATCH bug fixes need to be reordered. After the code review 
> passes and the SMC-D test goes stable, I will adjust it
> in next serial.

Hi D. Wythe,

no problem and thank you. I'm going to test your changes and let you 
know as soon as I'm done.

- Jan

> 
> 
> On 10/20/22 2:43 PM, D.Wythe wrote:
>> From: "D.Wythe" <alibuda@...ux.alibaba.com>
>>
>> This patch set attempts to optimize the parallelism of SMC-R connections,
>> mainly to reduce unnecessary blocking on locks, and to fix exceptions 
>> that
>> occur after thoses optimization.
>>
>> According to Off-CPU graph, SMC worker's off-CPU as that:
>>
>> smc_close_passive_work                  (1.09%)
>>          smcr_buf_unuse                  (1.08%)
>>                  smc_llc_flow_initiate   (1.02%)
>>
>> smc_listen_work                         (48.17%)
>>          __mutex_lock.isra.11            (47.96%)
>>
>>
>> An ideal SMC-R connection process should only block on the IO events
>> of the network, but it's quite clear that the SMC-R connection now is
>> queued on the lock most of the time.
>>
>> The goal of this patchset is to achieve our ideal situation where
>> network IO events are blocked for the majority of the connection 
>> lifetime.
>>
>> There are three big locks here:
>>
>> 1. smc_client_lgr_pending & smc_server_lgr_pending
>>
>> 2. llc_conf_mutex
>>
>> 3. rmbs_lock & sndbufs_lock
>>
>> And an implementation issue:
>>
>> 1. confirm/delete rkey msg can't be sent concurrently while
>> protocol allows indeed.
>>
>> Unfortunately,The above problems together affect the parallelism of
>> SMC-R connection. If any of them are not solved. our goal cannot
>> be achieved.
>>
>> After this patch set, we can get a quite ideal off-CPU graph as
>> following:
>>
>> smc_close_passive_work                                  (41.58%)
>>          smcr_buf_unuse                                  (41.57%)
>>                  smc_llc_do_delete_rkey                  (41.57%)
>>
>> smc_listen_work                                         (39.10%)
>>          smc_clc_wait_msg                                (13.18%)
>>                  tcp_recvmsg_locked                      (13.18)
>>          smc_listen_find_device                          (25.87%)
>>                  smcr_lgr_reg_rmbs                       (25.87%)
>>                          smc_llc_do_confirm_rkey         (25.87%)
>>
>> We can see that most of the waiting times are waiting for network IO
>> events. This also has a certain performance improvement on our
>> short-lived conenction wrk/nginx benchmark test:
>>
>> +--------------+------+------+-------+--------+------+--------+
>> |conns/qps     |c4    | c8   |  c16  |  c32   | c64  |  c200  |
>> +--------------+------+------+-------+--------+------+--------+
>> |SMC-R before  |9.7k  | 10k  |  10k  |  9.9k  | 9.1k |  8.9k  |
>> +--------------+------+------+-------+--------+------+--------+
>> |SMC-R now     |13k   | 19k  |  18k  |  16k   | 15k  |  12k   |
>> +--------------+------+------+-------+--------+------+--------+
>> |TCP           |15k   | 35k  |  51k  |  80k   | 100k |  162k  |
>> +--------------+------+------+-------+--------+------+--------+
>>
>> The reason why the benefit is not obvious after the number of connections
>> has increased dues to workqueue. If we try to change workqueue to 
>> UNBOUND,
>> we can obtain at least 4-5 times performance improvement, reach up to 
>> half
>> of TCP. However, this is not an elegant solution, the optimization of it
>> will be much more complicated. But in any case, we will submit relevant
>> optimization patches as soon as possible.
>>
>> Please note that the premise here is that the lock related problem
>> must be solved first, otherwise, no matter how we optimize the workqueue,
>> there won't be much improvement.
>>
>> Because there are a lot of related changes to the code, if you have
>> any questions or suggestions, please let me know.
>>
>> Thanks
>> D. Wythe
>>
>> v1 -> v2:
>>
>> 1. Fix panic in SMC-D scenario
>> 2. Fix lnkc related hashfn calculation exception, caused by operator
>> priority
>> 3. Only wake up one connection if the lnk is not active
>> 4. Delete obsolete unlock logic in smc_listen_work()
>> 5. PATCH format, do Reverse Christmas tree
>> 6. PATCH format, change all xxx_lnk_xxx function to xxx_link_xxx
>> 7. PATCH format, add correct fix tag for the patches for fixes.
>> 8. PATCH format, fix some spelling error
>> 9. PATCH format, rename slow to do_slow
>>
>> v2 -> v3:
>>
>> 1. add SMC-D support, remove the concept of link cluster since SMC-D has
>> no link at all. Replace it by lgr decision maker, who provides 
>> suggestions
>> to SMC-D and SMC-R on whether to create new link group.
>>
>> 2. Fix the corruption problem described by PATCH 'fix application
>> data exception' on SMC-D.
>>
>> D. Wythe (10):
>>    net/smc: remove locks smc_client_lgr_pending and
>>      smc_server_lgr_pending
>>    net/smc: fix SMC_CLC_DECL_ERR_REGRMB without smc_server_lgr_pending
>>    net/smc: allow confirm/delete rkey response deliver multiplex
>>    net/smc: make SMC_LLC_FLOW_RKEY run concurrently
>>    net/smc: llc_conf_mutex refactor, replace it with rw_semaphore
>>    net/smc: use read semaphores to reduce unnecessary blocking in
>>      smc_buf_create() & smcr_buf_unuse()
>>    net/smc: reduce unnecessary blocking in smcr_lgr_reg_rmbs()
>>    net/smc: replace mutex rmbs_lock and sndbufs_lock with rw_semaphore
>>    net/smc: Fix potential panic dues to unprotected
>>      smc_llc_srv_add_link()
>>    net/smc: fix application data exception
>>
>>   net/smc/af_smc.c   |  70 ++++----
>>   net/smc/smc_core.c | 478 
>> +++++++++++++++++++++++++++++++++++++++++++++++------
>>   net/smc/smc_core.h |  36 +++-
>>   net/smc/smc_llc.c  | 277 ++++++++++++++++++++++---------
>>   net/smc/smc_llc.h  |   6 +
>>   net/smc/smc_wr.c   |  10 --
>>   net/smc/smc_wr.h   |  10 ++
>>   7 files changed, 712 insertions(+), 175 deletions(-)
>>

Powered by blists - more mailing lists