From: Bart Van Assche <bart.vanassche@sandisk.com>
To: open-iscsi@googlegroups.com, Mike Christie <michaelc@cs.wisc.edu>,
"Nicholas A. Bellinger" <nab@linux-iscsi.org>
Cc: James Bottomley <James.Bottomley@HansenPartnership.com>,
lsf-pc@lists.linux-foundation.org,
linux-scsi <linux-scsi@vger.kernel.org>,
target-devel <target-devel@vger.kernel.org>
Subject: Re: [Lsf-pc] [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion
Date: Mon, 12 Jan 2015 13:56:30 +0100 [thread overview]
Message-ID: <54B3C47E.6010109@sandisk.com> (raw)
In-Reply-To: <54B24501.7090801@dev.mellanox.co.il>
On 01/11/15 10:40, Sagi Grimberg wrote:
> I would say there is no need for specific coordination from iSCSI PoV.
> This is exactly what flow steering is designed for. As I see it, in
> order to get the TX/RX to match rings, the user can attach 5-tuple rules
> (using standard ethtool) to steer packets to the right rings.
Hello Sagi,
Can the 5-tuple rules be chosen such that it is guaranteed that the
sockets used to implement per-CPU queues are spread evenly over MSI-X
completion vectors ? If not, would it help to add a socket option to the
Linux network stack that allows to select the TX ring explicitly, just
like ib_create_cq() in the Linux RDMA stack allows to select a
completion vector explicitly ? My concerns are as follows:
- If the number of queues exceeds the number of MSI-X vectors then I
expect that it will be much easier to guarantee even spreading by
selecting tx queues explicitly instead of relying on a hashing scheme.
- On multi-socket systems it is important to process completion
interrupts on the CPU socket from where the I/O was initiated. I'm
not sure it is possible to guarantee this when using a hashing
algorithm to select the TX ring.
Bart.
next prev parent reply other threads:[~2015-01-12 13:16 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-01-07 16:25 [LSF/MM TOPIC] iSCSI MQ adoption via MCS discussion Sagi Grimberg
[not found] ` <54AD5DDD.2090808-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2015-01-07 16:57 ` Hannes Reinecke
[not found] ` <54AD6563.4040603-l3A5Bk7waGM@public.gmane.org>
2015-01-07 21:39 ` Mike Christie
2015-01-08 7:50 ` Bart Van Assche
2015-01-08 13:45 ` Sagi Grimberg
[not found] ` <54AE8A02.1030100-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2015-01-08 14:11 ` Bart Van Assche
[not found] ` <54AE9010.5080609-HInyCGIudOg@public.gmane.org>
2015-01-08 15:57 ` Paul Koning
2015-01-09 11:39 ` Sagi Grimberg
2015-01-09 13:31 ` Bart Van Assche
[not found] ` <5EE87F5E6631894E80EB1A63198F964D040A6A8F-cXZ6iGhjG0hm/BozF5lIdDJ2aSJ780jGSxCzGc5ayCJWk0Htik3J/w@public.gmane.org>
2015-01-11 9:52 ` Sagi Grimberg
2015-01-14 4:16 ` Vladislav Bolkhovitin
2015-01-08 22:16 ` Nicholas A. Bellinger
2015-01-08 22:29 ` James Bottomley
2015-01-08 22:57 ` Nicholas A. Bellinger
[not found] ` <1420757822.2842.39.camel-XoQW25Eq2zviZyQQd+hFbcojREIfoBdhmpATvIKMPHk@public.gmane.org>
2015-01-08 23:22 ` [Lsf-pc] " James Bottomley
2015-01-09 5:03 ` Nicholas A. Bellinger
2015-01-09 6:25 ` James Bottomley
[not found] ` <1420779808.21830.21.camel-XoQW25Eq2zviZyQQd+hFbcojREIfoBdhmpATvIKMPHk@public.gmane.org>
2015-01-09 18:00 ` Michael Christie
2015-01-09 18:28 ` Hannes Reinecke
[not found] ` <54B01DBD.5020707-l3A5Bk7waGM@public.gmane.org>
2015-01-09 18:34 ` James Bottomley
2015-01-09 20:19 ` Mike Christie
[not found] ` <54B037BF.1010903-hcNo3dDEHLuVc3sceRu5cw@public.gmane.org>
2015-01-11 9:40 ` Sagi Grimberg
2015-01-12 12:56 ` Bart Van Assche [this message]
[not found] ` <54B3C47E.6010109-XdAiOPVOjttBDgjK7y7TUQ@public.gmane.org>
2015-01-13 9:46 ` Sagi Grimberg
[not found] ` <54B24501.7090801-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2015-01-12 20:14 ` Mike Christie
[not found] ` <38CE4ECA-D155-4BF9-9D6D-E1A01ADA05E4-hcNo3dDEHLuVc3sceRu5cw@public.gmane.org>
2015-01-11 9:23 ` Sagi Grimberg
[not found] ` <54B24117.7050204-LDSdmyG8hGV8YrgS2mwiifqBs+8SCbDb@public.gmane.org>
2015-01-12 20:05 ` Mike Christie
[not found] ` <54B428F2.2010507-hcNo3dDEHLuVc3sceRu5cw@public.gmane.org>
2015-01-13 9:55 ` Sagi Grimberg
2015-01-08 23:26 ` Mike Christie
[not found] ` <54AF122C.9070703-hcNo3dDEHLuVc3sceRu5cw@public.gmane.org>
2015-01-09 11:17 ` Sagi Grimberg
2015-01-08 23:01 ` Mike Christie
2015-01-08 14:50 ` James Bottomley
2015-01-08 17:25 ` Sagi Grimberg
[not found] ` <54ADA777.6090801-hcNo3dDEHLuVc3sceRu5cw@public.gmane.org>
2015-01-08 23:40 ` Mike Christie
2015-01-07 17:22 ` Lee Duncan
2015-01-07 19:11 ` [Lsf-pc] " Jan Kara
2015-01-07 16:58 ` Nicholas A. Bellinger
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=54B3C47E.6010109@sandisk.com \
--to=bart.vanassche@sandisk.com \
--cc=James.Bottomley@HansenPartnership.com \
--cc=linux-scsi@vger.kernel.org \
--cc=lsf-pc@lists.linux-foundation.org \
--cc=michaelc@cs.wisc.edu \
--cc=nab@linux-iscsi.org \
--cc=open-iscsi@googlegroups.com \
--cc=target-devel@vger.kernel.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.