| Message ID | f2863702-e54c-cd74-efcf-8cb238be1a7c@kernel.dk (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [for-next] block: enable per-cpu bio caching for the fs bio set | expand |
On Fri, Sep 02, 2022 at 10:42:45AM -0600, Jens Axboe wrote: >This is useful for polled IO on a file, or for polled IO with the >io_uring passthrough mechanism. If bio allocations are done with >REQ_POLLED for those cases, then initializing the bio set with >BIOSET_PERCPU_CACHE enables the local per-cpu cache which eliminates >allocations (and frees) of bio structs when possible. > >Signed-off-by: Jens Axboe <axboe@kernel.dk> Reviewed-by: Kanchan Joshi <joshi.k@samsung.com>
diff --git a/block/bio.c b/block/bio.c index 3d3a2678fea2..d3154d8beed7 100644 --- a/block/bio.c +++ b/block/bio.c @@ -1754,7 +1754,8 @@ static int __init init_bio(void) cpuhp_setup_state_multi(CPUHP_BIO_DEAD, "block/bio:dead", NULL, bio_cpu_dead); - if (bioset_init(&fs_bio_set, BIO_POOL_SIZE, 0, BIOSET_NEED_BVECS)) + if (bioset_init(&fs_bio_set, BIO_POOL_SIZE, 0, + BIOSET_NEED_BVECS | BIOSET_PERCPU_CACHE)) panic("bio: can't allocate bios\n"); if (bioset_integrity_create(&fs_bio_set, BIO_POOL_SIZE))
This is useful for polled IO on a file, or for polled IO with the io_uring passthrough mechanism. If bio allocations are done with REQ_POLLED for those cases, then initializing the bio set with BIOSET_PERCPU_CACHE enables the local per-cpu cache which eliminates allocations (and frees) of bio structs when possible. Signed-off-by: Jens Axboe <axboe@kernel.dk> ---