public inbox for [email protected]
 help / color / mirror / Atom feed
* [PATCHSET 0/2] Add unbounded io_wq for io_uring
@ 2019-11-06 19:40 Jens Axboe
  2019-11-06 19:40 ` [PATCH 1/2] io-wq: pass in io_wq to work handler Jens Axboe
  2019-11-06 19:40 ` [PATCH 2/2] io_uring: use separate io_wq's for bounded and unbounded request times Jens Axboe
  0 siblings, 2 replies; 3+ messages in thread
From: Jens Axboe @ 2019-11-06 19:40 UTC (permalink / raw)
  To: io-uring; +Cc: linux-block

Currently we have just the one workqueue for each io_uring ctx, which
supports any kind of IO. This is problematic with requests that have
very different life times. For example, disk IO is always bounded in
time, while networked IO can take a long time. Ditto with POLL commands,
that also have unbounded execution time.

First patch is just a prep patch, patch number two adds support for
determining the io_wq placement based on the request type.

Patches are against my for-5.5/io_uring-test branch.

 fs/io-wq.c    |   8 +--
 fs/io-wq.h    |   2 +-
 fs/io_uring.c | 145 ++++++++++++++++++++++++++++++++++++--------------
 3 files changed, 110 insertions(+), 45 deletions(-)

-- 
 Jens Axboe



^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2019-11-06 19:40 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2019-11-06 19:40 [PATCHSET 0/2] Add unbounded io_wq for io_uring Jens Axboe
2019-11-06 19:40 ` [PATCH 1/2] io-wq: pass in io_wq to work handler Jens Axboe
2019-11-06 19:40 ` [PATCH 2/2] io_uring: use separate io_wq's for bounded and unbounded request times Jens Axboe

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox