Merge tag 'io_uring-bio-cache.5-2021-08-30' of git://git.kernel.dk/linux-block

Pull support for struct bio recycling from Jens Axboe:
 "This adds bio recycling support for polled IO, allowing quick reuse of
  a bio for high IOPS scenarios via a percpu bio_set list.

  It's good for almost a 10% improvement in performance, bumping our
  per-core IO limit from ~3.2M IOPS to ~3.5M IOPS"

* tag 'io_uring-bio-cache.5-2021-08-30' of git://git.kernel.dk/linux-block:
  bio: improve kerneldoc documentation for bio_alloc_kiocb()
  block: provide bio_clear_hipri() helper
  block: use the percpu bio cache in __blkdev_direct_IO
  io_uring: enable use of bio alloc cache
  block: clear BIO_PERCPU_CACHE flag if polling isn't supported
  bio: add allocation cache abstraction
  fs: add kiocb alloc cache flag
  bio: optimize initialization of a bio

4 files changed, 17 insertions, 0 deletions

diff --git a/include/linux/bio.h b/include/linux/bio.h
index 7b5f65a81f2b..3d67d0fbc868 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -400,6 +400,7 @@ static inline struct bio *bio_next_split(struct bio *bio, int sectors,
 enum {
 	BIOSET_NEED_BVECS = BIT(0),
 	BIOSET_NEED_RESCUER = BIT(1),
+	BIOSET_PERCPU_CACHE = BIT(2),
 };
 extern int bioset_init(struct bio_set *, unsigned int, unsigned int, int flags);
 extern void bioset_exit(struct bio_set *);
@@ -408,6 +409,8 @@ extern int bioset_init_from_src(struct bio_set *bs, struct bio_set *src);
 
 struct bio *bio_alloc_bioset(gfp_t gfp, unsigned short nr_iovecs,
 		struct bio_set *bs);
+struct bio *bio_alloc_kiocb(struct kiocb *kiocb, unsigned short nr_vecs,
+		struct bio_set *bs);
 struct bio *bio_kmalloc(gfp_t gfp_mask, unsigned short nr_iovecs);
 extern void bio_put(struct bio *);
 
@@ -657,6 +660,11 @@ struct bio_set {
 	struct kmem_cache *bio_slab;
 	unsigned int front_pad;
 
+	/*
+	 * per-cpu bio alloc cache
+	 */
+	struct bio_alloc_cache __percpu *cache;
+
 	mempool_t bio_pool;
 	mempool_t bvec_pool;
 #if defined(CONFIG_BLK_DEV_INTEGRITY)
@@ -673,6 +681,11 @@ struct bio_set {
 	struct bio_list		rescue_list;
 	struct work_struct	rescue_work;
 	struct workqueue_struct	*rescue_workqueue;
+
+	/*
+	 * Hot un-plug notifier for the per-cpu cache, if used
+	 */
+	struct hlist_node cpuhp_dead;
 };
 
 static inline bool bioset_initialized(struct bio_set *bs)
diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
index 1335efa8a1db..9e392daa1d7f 100644
--- a/include/linux/blk_types.h
+++ b/include/linux/blk_types.h
@@ -297,6 +297,7 @@ enum {
 	BIO_TRACKED,		/* set if bio goes through the rq_qos path */
 	BIO_REMAPPED,
 	BIO_ZONE_WRITE_LOCKED,	/* Owns a zoned device zone write lock */
+	BIO_PERCPU_CACHE,	/* can participate in per-cpu alloc cache */
 	BIO_FLAG_LAST
 };
 
diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 6ac543d33b66..95f88edc8f09 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -46,6 +46,7 @@ enum cpuhp_state {
 	CPUHP_ARM_OMAP_WAKE_DEAD,
 	CPUHP_IRQ_POLL_DEAD,
 	CPUHP_BLOCK_SOFTIRQ_DEAD,
+	CPUHP_BIO_DEAD,
 	CPUHP_ACPI_CPUDRV_DEAD,
 	CPUHP_S390_PFAULT_DEAD,
 	CPUHP_BLK_MQ_DEAD,
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 7eae53fff7c5..ae29027d2ac8 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -319,6 +319,8 @@ enum rw_hint {
 /* iocb->ki_waitq is valid */
 #define IOCB_WAITQ		(1 << 19)
 #define IOCB_NOIO		(1 << 20)
+/* can use bio alloc cache */
+#define IOCB_ALLOC_CACHE	(1 << 21)
 
 struct kiocb {
 	struct file		*ki_filp;