nvmet: don't split large I/Os unconditionally
Subsystems (2)
Sagi Grimberg
maintainerPATCH v2 1/1 nvmet: don't split large I/Os unconditionally
If we know that the I/O size exceeds our inline bio vec, no point using it and split the rest to begin with. We could in theory reuse the inline bio and only allocate the bio_vec, but its really not worth optimizing for.
Signed-off-by Sagi Grimberg <sagi at grimberg.me>
---
Changes from v1:
- update change log
- rebased without bio_alloc failure handling patch
drivers/nvme/target/io-cmd-bdev.c | 9 +++++++-- drivers/nvme/target/nvmet.h | 1 + 2 files changed, 8 insertions(+), 2 deletions(-)
drivers/nvme/target/io-cmd-bdev.c +7 -2
@@ -58,7 +58,7 @@ static void nvmet_bio_done(struct bio *bio)static void nvmet_bdev_execute_rw (struct nvmet_req *req) {int sg_cnt = req->sg_cnt; -struct bio *bio = &req->b.inline_bio; +struct bio *bio;struct scatterlist *sg;sector_t sector;blk_qc_t cookie;
@@ -81,7 +81,12 @@ static void nvmet_bdev_execute_rw(struct nvmet_req *req) sector =le64_to_cpu (req->cmd->rw.slba); sector <<= (req->ns->blksize_shift -9 ); -bio_init (bio, req->inline_bvec,ARRAY_SIZE (req->inline_bvec)); +if (req->data_len <=NVMET_MAX_INLINE_DATA_LEN ) { + bio = &req->b.inline_bio; +bio_init (bio, req->inline_bvec,ARRAY_SIZE (req->inline_bvec)); + }else { + bio =bio_alloc (GFP_KERNEL ,min (sg_cnt,BIO_MAX_PAGES )); + }bio_set_dev (bio, req->ns->bdev); bio->bi_iter.bi_sector = sector; bio->bi_private = req;
drivers/nvme/target/nvmet.h +1 -0
@@ -273,6 +273,7 @@ struct nvmet_fabrics_ops { };#define NVMET_MAX_INLINE_BIOVEC 8 +#define NVMET_MAX_INLINE_DATA_LEN NVMET_MAX_INLINE_BIOVEC * PAGE_SIZEstruct nvmet_req {struct nvme_command *cmd;
Signature · 1 line
2.17.1
PATCH v2 1/1 nvmet: don't split large I/Os unconditionally
Thanks,
applied to nvme-4.20.