linux-rdma.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Leon Romanovsky <leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
To: Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>
Cc: linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	Leon Romanovsky <leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>,
	Kamal Heib <kamalh-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
Subject: [PATCH rdma-next 0/3] IPoIB TX NAPI
Date: Mon, 16 Oct 2017 08:48:58 +0300	[thread overview]
Message-ID: <20171016054901.32479-1-leon@kernel.org> (raw)

It is very common in modern network devices to use NAPI in RX and TX,
in the UD/TX flows IPoIB uses a polling mechanism (not as it does in
the CM mode).

Except of the motivation to use the same mechanism for IPoIB as for
netdevices, there are number of issues in current polling mechanism:
 * SKBs that are kept longer than they should be and there are applications
   that warn about that (some firewalls for example).
 * Statistics that are not updated to the real value. It blocks support
   for time synchronization protocols over the IPoIB protocol, like PTP
   and so on.
 * The TX in CM mode already uses NAPI, there is no reason to keep two
   different ways for TX one for UD and one for the CM.

The patches are available in the git repository at:
  git.kernel.org/pub/scm/linux/kernel/git/leon/linux-rdma.git tags/rdma-next-2017-10-16-2

	Thanks

Cc: Kamal Heib <kamalh-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
---------------------------------------

Erez Shitrit (3):
  IB/ipoib: Get rid of the tx_outstanding variable in all modes
  IB/ipoib: Use NAPI in UD/TX flows
  IB/ipoib: Change number of TX wqe to 64

 drivers/infiniband/ulp/ipoib/ipoib.h       |  14 ++--
 drivers/infiniband/ulp/ipoib/ipoib_cm.c    |  46 ++++++-----
 drivers/infiniband/ulp/ipoib/ipoib_ib.c    | 127 ++++++++++++++++++-----------
 drivers/infiniband/ulp/ipoib/ipoib_main.c  |  25 ++++--
 drivers/infiniband/ulp/ipoib/ipoib_verbs.c |  17 ++--
 5 files changed, 143 insertions(+), 86 deletions(-)

--
2.14.2

--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

             reply	other threads:[~2017-10-16  5:48 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-16  5:48 Leon Romanovsky [this message]
     [not found] ` <20171016054901.32479-1-leon-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
2017-10-16  5:48   ` [PATCH rdma-next 1/3] IB/ipoib: Get rid of the tx_outstanding variable in all modes Leon Romanovsky
2017-10-16  5:49   ` [PATCH rdma-next 2/3] IB/ipoib: Use NAPI in UD/TX flows Leon Romanovsky
2017-10-16  5:49   ` [PATCH rdma-next 3/3] IB/ipoib: Change number of TX wqe to 64 Leon Romanovsky
2017-10-25 17:24   ` [PATCH rdma-next 0/3] IPoIB TX NAPI Doug Ledford

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=20171016054901.32479-1-leon@kernel.org \
    --to=leon-dgejt+ai2ygdnm+yrofe0a@public.gmane.org \
    --cc=dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
    --cc=kamalh-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    --cc=linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.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).