]> www.infradead.org Git - nvme.git/commitdiff
ublk: don't suggest CONFIG_BLK_DEV_UBLK=Y
authorCaleb Sander Mateos <csander@purestorage.com>
Wed, 16 Apr 2025 00:41:10 +0000 (18:41 -0600)
committerJens Axboe <axboe@kernel.dk>
Wed, 16 Apr 2025 00:59:33 +0000 (18:59 -0600)
The CONFIG_BLK_DEV_UBLK help text suggests setting the config option to
Y so task_work_add() can be used to dispatch I/O, improving performance.
However, this mechanism was removed in commit 29dc5d06613f2 ("ublk: kill
queuing request by task_work_add"). So remove this paragraph from the
config help text.

Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Reviewed-by: Uday Shankar <ushankar@purestorage.com>
Link: https://lore.kernel.org/r/20250416004111.3242817-1-csander@purestorage.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
drivers/block/Kconfig

index a97f2c40c640dd60e8916e3468aaea60197a6583..9e129fec4bb795bb71208c0b1864f0ef1ab0b809 100644 (file)
@@ -388,12 +388,6 @@ config BLK_DEV_UBLK
          definition isn't finalized yet, and might change according to future
          requirement, so mark is as experimental now.
 
-         Say Y if you want to get better performance because task_work_add()
-         can be used in IO path for replacing io_uring cmd, which will become
-         shared between IO tasks and ubq daemon, meantime task_work_add() can
-         can handle batch more effectively, but task_work_add() isn't exported
-         for module, so ublk has to be built to kernel.
-
 config BLKDEV_UBLK_LEGACY_OPCODES
        bool "Support legacy command opcode"
        depends on BLK_DEV_UBLK