From: Paolo Bonzini <pbonzini@redhat.com>
To: "Michael S. Tsirkin" <mst@redhat.com>
Cc: aliguori@us.ibm.com, qemu-devel@nongnu.org,
"Michael R. Hines" <mrhines@linux.vnet.ibm.com>,
owasserm@redhat.com, abali@us.ibm.com, mrhines@us.ibm.com,
gokul@us.ibm.com
Subject: Re: [Qemu-devel] [RFC PATCH RDMA support v5: 03/12] comprehensive protocol documentation
Date: Thu, 11 Apr 2013 16:50:21 +0200 [thread overview]
Message-ID: <5166CDAD.8060807@redhat.com> (raw)
In-Reply-To: <20130411143718.GC24942@redhat.com>
Il 11/04/2013 16:37, Michael S. Tsirkin ha scritto:
>
> pg1 -> pin -> req -> res -> rdma -> done
> pg2 -> pin -> req -> res -> rdma -> done
> pg3 -> pin -> req -> res -> rdma -> done
> pg4 -> pin -> req -> res -> rdma -> done
> pg4 -> pin -> req -> res -> rdma -> done
>
> It's like a assembly line see? So while software does the registration
> roundtrip dance, hardware is processing rdma requests for previous
> chunks.
Does this only affects the implementation, or also the wire protocol?
Does the destination have to be aware that the source is doing pipelining?
Paolo
>
> ....
>
> When do you have to stall? when you run out of rx buffer credits so you
> can not start a new req. Your protocol has 2 outstanding buffers,
> so you can only have one req in the air. Do more and
> you will not need to stall - possibly at all.
>
> One other minor point is that your protocol requires extra explicit
> ready commands. You can pass the number of rx buffers as extra payload
> in the traffic you are sending anyway, and reduce that overhead.
next prev parent reply other threads:[~2013-04-11 14:51 UTC|newest]
Thread overview: 97+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-04-09 3:04 [Qemu-devel] [RFC PATCH RDMA support v5: 00/12] new formal protocol design mrhines
2013-04-09 3:04 ` [Qemu-devel] [RFC PATCH RDMA support v5: 01/12] ./configure with and without --enable-rdma mrhines
2013-04-09 17:05 ` Paolo Bonzini
2013-04-09 18:07 ` Michael R. Hines
2013-04-09 3:04 ` [Qemu-devel] [RFC PATCH RDMA support v5: 02/12] check for CONFIG_RDMA mrhines
2013-04-09 16:46 ` Paolo Bonzini
2013-04-09 3:04 ` [Qemu-devel] [RFC PATCH RDMA support v5: 03/12] comprehensive protocol documentation mrhines
2013-04-10 5:27 ` Michael S. Tsirkin
2013-04-10 13:04 ` Michael R. Hines
2013-04-10 13:34 ` Michael S. Tsirkin
2013-04-10 15:29 ` Michael R. Hines
2013-04-10 17:41 ` Michael S. Tsirkin
2013-04-10 20:05 ` Michael R. Hines
2013-04-11 7:19 ` Michael S. Tsirkin
2013-04-11 13:12 ` Michael R. Hines
2013-04-11 13:48 ` Michael S. Tsirkin
2013-04-11 13:58 ` Michael R. Hines
2013-04-11 14:37 ` Michael S. Tsirkin
2013-04-11 14:50 ` Paolo Bonzini [this message]
2013-04-11 14:56 ` Michael S. Tsirkin
2013-04-11 17:49 ` Michael R. Hines
2013-04-11 19:15 ` Michael S. Tsirkin
2013-04-11 20:33 ` Michael R. Hines
2013-04-12 10:48 ` Michael S. Tsirkin
2013-04-12 10:53 ` Paolo Bonzini
2013-04-12 11:25 ` Michael S. Tsirkin
2013-04-12 14:43 ` Paolo Bonzini
2013-04-14 11:59 ` Michael S. Tsirkin
2013-04-14 14:09 ` Paolo Bonzini
2013-04-14 14:40 ` Michael R. Hines
2013-04-14 14:27 ` Michael R. Hines
2013-04-14 16:03 ` Michael S. Tsirkin
2013-04-14 16:07 ` Michael R. Hines
2013-04-14 16:40 ` Michael R. Hines
2013-04-14 18:30 ` Michael S. Tsirkin
2013-04-14 19:06 ` Michael R. Hines
2013-04-14 21:10 ` Michael S. Tsirkin
2013-04-15 1:06 ` Michael R. Hines
2013-04-15 6:00 ` Michael S. Tsirkin
2013-04-15 13:07 ` Michael R. Hines
2013-04-15 22:20 ` Michael S. Tsirkin
2013-04-15 8:28 ` Paolo Bonzini
2013-04-15 13:08 ` Michael R. Hines
2013-04-15 8:26 ` Paolo Bonzini
2013-04-12 13:47 ` Michael R. Hines
2013-04-14 8:28 ` Michael S. Tsirkin
2013-04-14 14:31 ` Michael R. Hines
2013-04-14 18:51 ` Michael S. Tsirkin
2013-04-14 19:43 ` Michael R. Hines
2013-04-14 21:16 ` Michael S. Tsirkin
2013-04-15 1:10 ` Michael R. Hines
2013-04-15 6:10 ` Michael S. Tsirkin
2013-04-15 8:34 ` Paolo Bonzini
2013-04-15 13:24 ` Michael R. Hines
2013-04-15 13:30 ` Paolo Bonzini
2013-04-15 19:55 ` Michael R. Hines
2013-04-11 15:01 ` Michael R. Hines
2013-04-11 15:18 ` Michael R. Hines
2013-04-11 15:33 ` Paolo Bonzini
2013-04-11 15:46 ` Michael S. Tsirkin
2013-04-11 15:47 ` Paolo Bonzini
2013-04-11 15:58 ` Michael S. Tsirkin
2013-04-11 16:06 ` Michael R. Hines
2013-04-12 5:10 ` Michael R. Hines
2013-04-12 5:26 ` Paolo Bonzini
2013-04-12 5:54 ` Michael R. Hines
2013-04-11 15:44 ` Michael S. Tsirkin
2013-04-11 16:09 ` Michael R. Hines
2013-04-11 17:04 ` Michael S. Tsirkin
2013-04-11 17:27 ` Michael R. Hines
2013-04-11 16:13 ` Michael R. Hines
2013-04-09 3:04 ` [Qemu-devel] [RFC PATCH RDMA support v5: 04/12] introduce qemu_ram_foreach_block() mrhines
2013-04-09 3:04 ` [Qemu-devel] [RFC PATCH RDMA support v5: 05/12] core RDMA migration logic w/ new protocol mrhines
2013-04-09 16:57 ` Paolo Bonzini
2013-04-09 3:04 ` [Qemu-devel] [RFC PATCH RDMA support v5: 06/12] connection-establishment for RDMA mrhines
2013-04-09 3:04 ` [Qemu-devel] [RFC PATCH RDMA support v5: 07/12] additional savevm.c accessors " mrhines
2013-04-09 17:03 ` Paolo Bonzini
2013-04-09 17:31 ` Peter Maydell
2013-04-09 18:04 ` Michael R. Hines
2013-04-09 3:04 ` [Qemu-devel] [RFC PATCH RDMA support v5: 08/12] new capabilities added and check for QMP string 'rdma' mrhines
2013-04-09 17:01 ` Paolo Bonzini
2013-04-10 1:11 ` Michael R. Hines
2013-04-10 8:07 ` Paolo Bonzini
2013-04-10 10:35 ` Michael S. Tsirkin
2013-04-10 12:24 ` Michael R. Hines
2013-04-09 17:02 ` Paolo Bonzini
2013-04-09 3:04 ` [Qemu-devel] [RFC PATCH RDMA support v5: 09/12] transmit pc.ram using RDMA mrhines
2013-04-09 16:50 ` Paolo Bonzini
2013-04-09 3:04 ` [Qemu-devel] [RFC PATCH RDMA support v5: 10/12] new header file prototypes for savevm.c mrhines
2013-04-09 16:43 ` Paolo Bonzini
2013-04-09 3:04 ` [Qemu-devel] [RFC PATCH RDMA support v5: 11/12] update schema to define new capabilities mrhines
2013-04-09 16:43 ` Paolo Bonzini
2013-04-09 3:04 ` [Qemu-devel] [RFC PATCH RDMA support v5: 12/12] don't set nonblock on invalid file descriptor mrhines
2013-04-09 16:45 ` Paolo Bonzini
2013-04-09 4:24 ` [Qemu-devel] [RFC PATCH RDMA support v5: 00/12] new formal protocol design Michael R. Hines
2013-04-09 12:44 ` Michael S. Tsirkin
2013-04-09 14:23 ` Michael R. Hines
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=5166CDAD.8060807@redhat.com \
--to=pbonzini@redhat.com \
--cc=abali@us.ibm.com \
--cc=aliguori@us.ibm.com \
--cc=gokul@us.ibm.com \
--cc=mrhines@linux.vnet.ibm.com \
--cc=mrhines@us.ibm.com \
--cc=mst@redhat.com \
--cc=owasserm@redhat.com \
--cc=qemu-devel@nongnu.org \
/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).