linux-rdma.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Yanjun Zhu <yanjun.zhu@linux.dev>
To: "yangx.jy@fujitsu.com" <yangx.jy@fujitsu.com>
Cc: Haakon Bugge <haakon.bugge@oracle.com>,
	Jason Gunthorpe <jgg@nvidia.com>,
	"david.marchand@6wind.com" <david.marchand@6wind.com>,
	OFED mailing list <linux-rdma@vger.kernel.org>,
	"rpearsonhpe@gmail.com" <rpearsonhpe@gmail.com>
Subject: Re: [PATCH] RDMA/rxe: Check the last packet by RXE_END_MASK
Date: Wed, 19 Jan 2022 22:08:53 +0800	[thread overview]
Message-ID: <4e1e7cf6-d41b-6926-68cf-e58ca79a4559@linux.dev> (raw)
In-Reply-To: <61E623DF.5000007@fujitsu.com>


在 2022/1/18 10:20, yangx.jy@fujitsu.com 写道:
> On 2022/1/15 11:38, Yanjun Zhu wrote:
>> 在 2022/1/12 11:58, yangx.jy@fujitsu.com 写道:
>>> On 2022/1/12 9:19, Yanjun Zhu wrote:
>>>> 在 2022/1/11 23:16, Haakon Bugge 写道:
>>>>>> On 11 Jan 2022, at 15:42, Yanjun Zhu <yanjun.zhu@linux.dev> wrote:
>>>>>>
>>>>>>
>>>>>> 在 2022/1/10 13:17, yangx.jy@fujitsu.com 写道:
>>>>>>> On 2022/1/7 19:49, Yanjun Zhu wrote:
>>>>>>>> It seems that it does not mean to check the last packet. It means
>>>>>>>> that
>>>>>>>> it receives a MSN response.
>>>>>>> Hi Yanjun,
>>>>>>>
>>>>>>> Checking the last packet is a way to indicate that responder has
>>>>>>> completed an entire request(including multiple packets) and then
>>>>>>> increases a msn.
>>>>>> Hi, Xiao
>>>>>>
>>>>>> What does the msn mean?
>>>>> Message Sequence Number.
>>>> Thanks, Haakon
>>>>
>>>> I am reading the following from the spec.
>>>>
>>>> "
>>>>
>>>> C9-148: An HCA responder using Reliable Connection service shall
>>>> initialize
>>>>
>>>> its MSN value to zero. The responder shall increment its MSN
>>>> whenever it has successfully completed processing a new, valid request
>>>> message. The MSN shall not be incremented for duplicate requests. The
>>>> incremented MSN shall be returned in the last or only packet of an RDMA
>>>> READ or Atomic response. For RDMA READ requests, the responder
>>>> may increment its MSN after it has completed validating the request and
>>>> before it has begun transmitting any of the requested data, and may
>>>> return
>>>> the incremented MSN in the AETH of the first response packet. The MSN
>>>> shall be incremented only once for any given request message.
>>>>
>>>> "
>>>>
>>>> It seems that the above describe how to handle MSN increment in
>>>> details.
>>> Hi Yanjun,
>>>
>>> Sorry for the late reply.
>>>
>>> Right, 9.7.7.1 GENERATING MSN VALUE section explains Message Sequence

"

...

Since the responder may choose to coalesce acknowledges, a single 
response packet may in fact acknowledge
several request messages. Thus, when it receives a new MSN, the 
requester begins evaluating WQEs on its send queue beginning with the
oldest outstanding WQE and progressing forward.

...

"

In the above, several request messages come. From the SPEC, msn should 
increase based on the number of request messages.

Can your commit handle the above case?


Zhu Yanjun

>> Does your commit take into account of duplicate requests?
> Hi Yanjun,
>
> Responder will check duplicate requests by check_psn() and process them
> by duplicate_request().
> According to the logic of duplicate_request(), responder doesn't
> increase msn for duplicate requests.
>
> Best Regards,
> Xiao Yang
>> Zhu Yanjun
>>
>>> Number(MSN).
>>>
>>> Best Regards,
>>> Xiao Yang
>>>> Zhu Yanjun
>>>>
>>>>> Thxs, Håkon
>>>>>
>>>>>> Thanks a lot.
>>>>>>
>>>>>> Zhu Yanjun
>>>>>>
>>>>>>> Best Regards,
>>>>>>> Xiao Yang

  reply	other threads:[~2022-01-19 14:10 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-12-29  3:44 [PATCH] RDMA/rxe: Check the last packet by RXE_END_MASK Xiao Yang
2022-01-06  0:40 ` Jason Gunthorpe
2022-01-06 19:55   ` Robert Pearson
2022-01-07 11:49   ` Yanjun Zhu
2022-01-10  5:17     ` yangx.jy
2022-01-11 14:42       ` Yanjun Zhu
2022-01-11 15:16         ` Haakon Bugge
2022-01-12  1:19           ` Yanjun Zhu
2022-01-12  3:58             ` yangx.jy
2022-01-15  3:38               ` Yanjun Zhu
2022-01-18  2:20                 ` yangx.jy
2022-01-19 14:08                   ` Yanjun Zhu [this message]
2022-01-20  9:33                     ` yangx.jy
2022-01-21 13:18                       ` Yanjun Zhu
2022-02-08 16:40 ` Jason Gunthorpe

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4e1e7cf6-d41b-6926-68cf-e58ca79a4559@linux.dev \
    --to=yanjun.zhu@linux.dev \
    --cc=david.marchand@6wind.com \
    --cc=haakon.bugge@oracle.com \
    --cc=jgg@nvidia.com \
    --cc=linux-rdma@vger.kernel.org \
    --cc=rpearsonhpe@gmail.com \
    --cc=yangx.jy@fujitsu.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).