public inbox for linux-rdma@vger.kernel.org
 help / color / mirror / Atom feed
From: Bob Pearson <rpearsonhpe@gmail.com>
To: jgg@nvidia.com, zyjzyj2000@gmail.com, leon@kernel.org,
	jhack@hpe.com, ian.ziemba@hpe.com, matsuda-daisuke@fujitsu.com,
	lizhijian@fujitsu.com, haris.phnx@gmail.com,
	linux-rdma@vger.kernel.org
Cc: Bob Pearson <rpearsonhpe@gmail.com>
Subject: [PATCH v2 00/18] Implement work queues for rdma_rxe
Date: Fri, 21 Oct 2022 15:01:01 -0500	[thread overview]
Message-ID: <20221021200118.2163-1-rpearsonhpe@gmail.com> (raw)

This patch series implements work queues as an alternative for
the main tasklets in the rdma_rxe driver. The first few patches
perform some cleanups of the current tasklet code followed by a
patch that makes the internal API for task execution pluggable and
implements an inline and a tasklet based set of functions.
The remaining patches cleanup the qp reset and error code in the
three tasklets and modify the locking logic to prevent making
multiple calls to the tasklet scheduling routine. Finally after
this preparation the work queue equivalent set of functions is
added and module parameters are implemented to allow tuning the
task types.

The advantages of the work queue version of deferred task execution
is mainly that the work queue variant has much better scalability
and overall performance than the tasklet variant. The tasklet
performance saturates with one connected queue pair and stays constant.
The work queue performance is slightly better for one queue pair but
scales up with the number of connected queue pairs. The perftest
microbenchmarks in local loopback mode (not a very realistic test
case) can reach approximately 100Gb/sec with work queues compared to
about 16Gb/sec for tasklets.

It is expected that the work queue version will take the place of
the tasklet version once this patch series is accepted and moved
upstream. However, for now it is convenient to keep the option of
easily switching between the versions to help benchmarking and
debugging.

This patch series is derived from an earlier patch set developed by
Ian Ziemba at HPE which is used in some Lustre storage clients attached
to Lustre servers with hard RoCE v2 NICs.

This v2 version of the patch set has some minor changes that address
comments from Leon Romanovsky regarding locking of the valid parameter
and the setup parameters for alloc_workqueue. It also has one
additional cleanup patch.

Bob Pearson (18):
  RDMA/rxe: Remove redundant header files
  RDMA/rxe: Remove init of task locks from rxe_qp.c
  RDMA/rxe: Removed unused name from rxe_task struct
  RDMA/rxe: Split rxe_run_task() into two subroutines
  RDMA/rxe: Make rxe_do_task static
  RDMA/rxe: Rename task->state_lock to task->lock
  RDMA/rxe: Make task interface pluggable
  RDMA/rxe: Split rxe_drain_resp_pkts()
  RDMA/rxe: Simplify reset state handling in rxe_resp.c
  RDMA/rxe: Handle qp error in rxe_resp.c
  RDMA/rxe: Cleanup comp tasks in rxe_qp.c
  RDMA/rxe: Remove __rxe_do_task()
  RDMA/rxe: Make tasks schedule each other
  RDMA/rxe: Implement disable/enable_task()
  RDMA/rxe: Replace TASK_STATE_START by TASK_STATE_IDLE
  RDMA/rxe: Replace task->destroyed by task state INVALID.
  RDMA/rxe: Add workqueue support for tasks
  RDMA/rxe: Add parameters to control task type

 drivers/infiniband/sw/rxe/rxe.c       |   9 +-
 drivers/infiniband/sw/rxe/rxe_comp.c  |  35 ++-
 drivers/infiniband/sw/rxe/rxe_net.c   |   4 +-
 drivers/infiniband/sw/rxe/rxe_qp.c    |  87 +++----
 drivers/infiniband/sw/rxe/rxe_req.c   |  10 +-
 drivers/infiniband/sw/rxe/rxe_resp.c  |  75 ++++--
 drivers/infiniband/sw/rxe/rxe_task.c  | 339 +++++++++++++++++++-------
 drivers/infiniband/sw/rxe/rxe_task.h  |  77 +++---
 drivers/infiniband/sw/rxe/rxe_verbs.c |   8 +-
 9 files changed, 427 insertions(+), 217 deletions(-)


base-commit: 9abf2313adc1ca1b6180c508c25f22f9395cc780
-- 
2.34.1


             reply	other threads:[~2022-10-21 20:03 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-10-21 20:01 Bob Pearson [this message]
2022-10-21 20:01 ` [PATCH for-next v2 01/18] RDMA/rxe: Remove redundant header files Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 02/18] RDMA/rxe: Remove init of task locks from rxe_qp.c Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 03/18] RDMA/rxe: Removed unused name from rxe_task struct Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 04/18] RDMA/rxe: Split rxe_run_task() into two subroutines Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 05/18] RDMA/rxe: Make rxe_do_task static Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 06/18] RDMA/rxe: Rename task->state_lock to task->lock Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 07/18] RDMA/rxe: Make task interface pluggable Bob Pearson
2022-10-25  8:02   ` Daisuke Matsuda (Fujitsu)
2022-10-25 14:16     ` Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 08/18] RDMA/rxe: Split rxe_drain_resp_pkts() Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 09/18] RDMA/rxe: Simplify reset state handling in rxe_resp.c Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 10/18] RDMA/rxe: Handle qp error " Bob Pearson
2022-10-21 21:22   ` kernel test robot
2022-10-21 20:01 ` [PATCH for-next v2 11/18] RDMA/rxe: Cleanup comp tasks in rxe_qp.c Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 12/18] RDMA/rxe: Remove __rxe_do_task() Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 13/18] RDMA/rxe: Make tasks schedule each other Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 14/18] RDMA/rxe: Implement disable/enable_task() Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 15/18] RDMA/rxe: Replace TASK_STATE_START by TASK_STATE_IDLE Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 16/18] RDMA/rxe: Replace task->destroyed by task state INVALID Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 17/18] RDMA/rxe: Add workqueue support for tasks Bob Pearson
2022-10-21 20:01 ` [PATCH for-next v2 18/18] RDMA/rxe: Add parameters to control task type Bob Pearson
2022-10-25  9:35   ` Daisuke Matsuda (Fujitsu)
2022-10-25 11:18     ` Jason Gunthorpe
2022-10-25 14:31       ` Bob Pearson
2022-10-25 16:27         ` Jason Gunthorpe
2022-10-25 14:49     ` Bob Pearson
2022-10-26  7:07       ` Daisuke Matsuda (Fujitsu)
2022-10-28 17:04 ` [PATCH v2 00/18] Implement work queues for rdma_rxe Jason Gunthorpe
2022-10-28 18:16   ` Bob Pearson
2022-10-28 18:17     ` Jason Gunthorpe
2022-10-28 18:58       ` Bob Pearson
2022-10-28 19:16         ` Jason Gunthorpe
2022-10-28 19:29           ` Pearson, Robert B

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=20221021200118.2163-1-rpearsonhpe@gmail.com \
    --to=rpearsonhpe@gmail.com \
    --cc=haris.phnx@gmail.com \
    --cc=ian.ziemba@hpe.com \
    --cc=jgg@nvidia.com \
    --cc=jhack@hpe.com \
    --cc=leon@kernel.org \
    --cc=linux-rdma@vger.kernel.org \
    --cc=lizhijian@fujitsu.com \
    --cc=matsuda-daisuke@fujitsu.com \
    --cc=zyjzyj2000@gmail.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