All of lore.kernel.org
 help / color / mirror / Atom feed
From: Stefan Roesch <shr@fb.com>
To: <io-uring@vger.kernel.org>, <kernel-team@fb.com>
Cc: <shr@fb.com>
Subject: [PATCH v2 0/6] liburing: add support for large CQE sizes
Date: Wed, 20 Apr 2022 12:15:18 -0700	[thread overview]
Message-ID: <20220420191524.2906409-1-shr@fb.com> (raw)

This adds support for large CQE sizes in the liburing layer. The large CQE
sizes double the size compared to the default CQE size.

To support larger CQE sizes the mmap call needs to be modified to map a larger
memory region for large CQE's. For default CQE's the size of the mapping stays
the same.
Also the ring size calculation needs to change.

Finally when large CQE's are indexed, they need to take into account the bigger
CQE size. The index manipulation remains unchanged, only when the CQE array is
accessed, the offset is changed for large CQE's.

The nop test has been modified to test that the new values are set correctly.

Testing:
The liburing test suite has been run with the four different configurations:
- default
- large SQE
- large CQE
- large SQE & large CQE
To do this the default setting has been changed for the test run to the above
values.:

To use these
changes, also the corresponding kernel changes are required.

Changes:
  V2: the changed kernel definition of io_uring_cqe_extra has been applied to
      the first patch in this patch series.


Stefan Roesch (6):
  liburing: Update io_uring.h with large CQE kernel changes
  liburing: increase mmap size for large CQE's
  liburing: return correct ring size for large CQE's
  liburing: index large CQE's correctly
  liburing: add large CQE tests to nop test
  liburing: Test all configurations with NOP test

 src/include/liburing.h          | 18 +++++++-
 src/include/liburing/io_uring.h | 13 ++++++
 src/queue.c                     |  6 ++-
 src/setup.c                     | 13 ++++--
 test/nop.c                      | 74 ++++++++++++++++++++-------------
 test/test.h                     | 35 ++++++++++++++++
 6 files changed, 123 insertions(+), 36 deletions(-)
 create mode 100644 test/test.h


base-commit: c0ba24d1215e9f2b08266b28b35436988c6f3543
-- 
2.30.2


             reply	other threads:[~2022-04-20 19:15 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-04-20 19:15 Stefan Roesch [this message]
2022-04-20 19:15 ` [PATCH v2 1/6] liburing: Update io_uring.h with large CQE kernel changes Stefan Roesch
2022-04-20 19:15 ` [PATCH v2 2/6] liburing: increase mmap size for large CQE's Stefan Roesch
2022-04-20 19:15 ` [PATCH v2 3/6] liburing: return correct ring " Stefan Roesch
2022-04-20 19:15 ` [PATCH v2 4/6] liburing: index large CQE's correctly Stefan Roesch
2022-04-20 19:15 ` [PATCH v2 5/6] liburing: add large CQE tests to nop test Stefan Roesch
2022-04-20 19:15 ` [PATCH v2 6/6] liburing: Test all configurations with NOP test Stefan Roesch
2022-04-20 22:45 ` [PATCH v2 0/6] liburing: add support for large CQE sizes Jens Axboe

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=20220420191524.2906409-1-shr@fb.com \
    --to=shr@fb.com \
    --cc=io-uring@vger.kernel.org \
    --cc=kernel-team@fb.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 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.