linux-rdma.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Chuck Lever <chuck.lever@oracle.com>
To: linux-rdma@vger.kernel.org
Subject: [PATCH v8 0/3] Proposed trace points for RDMA/core
Date: Mon, 02 Dec 2019 11:20:54 -0500	[thread overview]
Message-ID: <20191202161518.3950.61082.stgit@manet.1015granger.net> (raw)

Hi-

Refresh of the RDMA/core trace point patches.

I've been working on improving how RPC/RDMA handles device removal
and disconnect, so I've added some trace points in the core CMA
code to report rdma_cm_id lifetime-related events.

I've also had need to record creation and release of MRs. I had an
old patch to do that, so that is also included here. The patch
description can be improved, but it's ready for review comments, at
least.


Changes since v7:
- Capture the return value from the ULP's CM event handler
- Record the lifetime of each rdma_cm_id
- Include an example patch for capturing MR lifetime

Changes since v6:
- Move include/trace/events/rmda_cma.h to drivers/infiniband/core/cma_trace.h
- Add sample trace log output to the patch descriptions
- Back to the inlined version of ib_poll_cq()

Changes since v5:
- Add low-overhead trace points in the Connection Manager
- Address #include heartburn found by lkp

Changes since v4:
- Removed __ib_poll_cq, uninlined ib_poll_cq

Changes since v3:
- Reverted unnecessary behavior change in __ib_process_cq
- Clarified what "id" is in trace point output
- Added comment before new fields in struct ib_cq
- New trace point that fires when there is a CQ allocation failure

Changes since v2:
- Removed extraneous changes to include/trace/events/rdma.h

Changes since RFC:
- Display CQ's global resource ID instead of it's pointer address

---

Chuck Lever (3):
      RDMA/cma: Add trace points in RDMA Connection Manager
      RDMA/core: Trace points for diagnosing completion queue issues
      RDMA/core: Add trace points to follow MR allocation


 drivers/infiniband/core/Makefile    |    4 
 drivers/infiniband/core/cma.c       |   68 +++++-
 drivers/infiniband/core/cma_trace.c |   16 ++
 drivers/infiniband/core/cma_trace.h |  302 ++++++++++++++++++++++++++++
 drivers/infiniband/core/cq.c        |   27 ++-
 drivers/infiniband/core/trace.c     |   14 +
 drivers/infiniband/core/verbs.c     |   39 +++-
 include/rdma/ib_verbs.h             |    5 
 include/trace/events/rdma_core.h    |  373 +++++++++++++++++++++++++++++++++++
 9 files changed, 816 insertions(+), 32 deletions(-)
 create mode 100644 drivers/infiniband/core/cma_trace.c
 create mode 100644 drivers/infiniband/core/cma_trace.h
 create mode 100644 drivers/infiniband/core/trace.c
 create mode 100644 include/trace/events/rdma_core.h

--
Chuck Lever

             reply	other threads:[~2019-12-02 16:20 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-12-02 16:20 Chuck Lever [this message]
2019-12-02 16:21 ` [PATCH v8 1/3] RDMA/cma: Add trace points in RDMA Connection Manager Chuck Lever
2019-12-02 16:21 ` [PATCH v8 2/3] RDMA/core: Trace points for diagnosing completion queue issues Chuck Lever
2019-12-02 16:21 ` [PATCH v8 3/3] RDMA/core: Add trace points to follow MR allocation Chuck Lever

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=20191202161518.3950.61082.stgit@manet.1015granger.net \
    --to=chuck.lever@oracle.com \
    --cc=linux-rdma@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 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).