bybrooklyn/linux
1
0
Fork 0
mirror of https://github.com/torvalds/linux.git synced 2026-04-18 06:44:00 -04:00
Code Issues Packages Projects Releases Wiki Activity

blk-mq: remove hybrid polling

Browse Source 
io_uring provides the only way user space can poll completions, and that
always sets BLK_POLL_NOSLEEP. This effectively makes hybrid polling dead
code, so remove it and everything supporting it.

Hybrid polling was effectively killed off with 9650b453a3, "block:
ignore RWF_HIPRI hint for sync dio", but still potentially reachable
through io_uring until d729cf9acb, "io_uring: don't sleep when
polling for I/O", but hybrid polling probably should not have been
reachable through that async interface from the beginning.

Fixes: 9650b453a3 ("block: ignore RWF_HIPRI hint for sync dio")
Fixes: d729cf9acb ("io_uring: don't sleep when polling for I/O")
Signed-off-by: Keith Busch <kbusch@kernel.org>
Link: https://lore.kernel.org/r/20230320194926.3353144-1-kbusch@meta.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This commit is contained in:
Keith Busch
2023-03-20 12:49:26 -07:00
committed by Jens Axboe
parent 4cf2c3ab2c
commit 54bdd67d0f
9 changed files with 12 additions and 299 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
Documentation/ABI/stable/sysfs-block
View File

@@ -336,18 +336,11 @@ What: /sys/block/<disk>/queue/io_poll_delay
Date: November 2016
Contact: linux-block@vger.kernel.org
Description:
[RW] If polling is enabled, this controls what kind of polling
will be performed. It defaults to -1, which is classic polling.
[RW] This was used to control what kind of polling will be
performed. It is now fixed to -1, which is classic polling.
In this mode, the CPU will repeatedly ask for completions
without giving up any time. If set to 0, a hybrid polling mode
is used, where the kernel will attempt to make an educated guess
at when the IO will complete. Based on this guess, the kernel
will put the process issuing IO to sleep for an amount of time,
before entering a classic poll loop. This mode might be a little
slower than pure classic polling, but it will be more efficient.
If set to a value larger than 0, the kernel will put the process
issuing IO to sleep for this amount of microseconds before
entering classic polling.
without giving up any time.
<deprecated>
What: /sys/block/<disk>/queue/io_timeout