DEFINE_TRACE(block_split);
-static struct kmem_cache *bio_slab __read_mostly;
+/*
+ * Test patch to inline a certain number of bi_io_vec's inside the bio
+ * itself, to shrink a bio data allocation from two mempool calls to one
+ */
+#define BIO_INLINE_VECS 4
static mempool_t *bio_split_pool __read_mostly;
* break badly! cannot be bigger than what you can fit into an
* unsigned short
*/
-
#define BV(x) { .nr_vecs = x, .name = "biovec-"__stringify(x) }
-static struct biovec_slab bvec_slabs[BIOVEC_NR_POOLS] __read_mostly = {
+struct biovec_slab bvec_slabs[BIOVEC_NR_POOLS] __read_mostly = {
BV(1), BV(4), BV(16), BV(64), BV(128), BV(BIO_MAX_PAGES),
};
#undef BV
*/
struct bio_set *fs_bio_set;
+/*
+ * Our slab pool management
+ */
+struct bio_slab {
+ struct kmem_cache *slab;
+ unsigned int slab_ref;
+ unsigned int slab_size;
+ char name[8];
+};
+static DEFINE_MUTEX(bio_slab_lock);
+static struct bio_slab *bio_slabs;
+static unsigned int bio_slab_nr, bio_slab_max;
+
+static struct kmem_cache *bio_find_or_create_slab(unsigned int extra_size)
+{
+ unsigned int sz = sizeof(struct bio) + extra_size;
+ struct kmem_cache *slab = NULL;
+ struct bio_slab *bslab;
+ unsigned int i, entry = -1;
+
+ mutex_lock(&bio_slab_lock);
+
+ i = 0;
+ while (i < bio_slab_nr) {
+ struct bio_slab *bslab = &bio_slabs[i];
+
+ if (!bslab->slab && entry == -1)
+ entry = i;
+ else if (bslab->slab_size == sz) {
+ slab = bslab->slab;
+ bslab->slab_ref++;
+ break;
+ }
+ i++;
+ }
+
+ if (slab)
+ goto out_unlock;
+
+ if (bio_slab_nr == bio_slab_max && entry == -1) {
+ bio_slab_max <<= 1;
+ bio_slabs = krealloc(bio_slabs,
+ bio_slab_max * sizeof(struct bio_slab),
+ GFP_KERNEL);
+ if (!bio_slabs)
+ goto out_unlock;
+ }
+ if (entry == -1)
+ entry = bio_slab_nr++;
+
+ bslab = &bio_slabs[entry];
+
+ snprintf(bslab->name, sizeof(bslab->name), "bio-%d", entry);
+ slab = kmem_cache_create(bslab->name, sz, 0, SLAB_HWCACHE_ALIGN, NULL);
+ if (!slab)
+ goto out_unlock;
+
+ printk("bio: create slab <%s> at %d\n", bslab->name, entry);
+ bslab->slab = slab;
+ bslab->slab_ref = 1;
+ bslab->slab_size = sz;
+out_unlock:
+ mutex_unlock(&bio_slab_lock);
+ return slab;
+}
+
+static void bio_put_slab(struct bio_set *bs)
+{
+ struct bio_slab *bslab = NULL;
+ unsigned int i;
+
+ mutex_lock(&bio_slab_lock);
+
+ for (i = 0; i < bio_slab_nr; i++) {
+ if (bs->bio_slab == bio_slabs[i].slab) {
+ bslab = &bio_slabs[i];
+ break;
+ }
+ }
+
+ if (WARN(!bslab, KERN_ERR "bio: unable to find slab!\n"))
+ goto out;
+
+ WARN_ON(!bslab->slab_ref);
+
+ if (--bslab->slab_ref)
+ goto out;
+
+ kmem_cache_destroy(bslab->slab);
+ bslab->slab = NULL;
+
+out:
+ mutex_unlock(&bio_slab_lock);
+}
+
unsigned int bvec_nr_vecs(unsigned short idx)
{
return bvec_slabs[idx].nr_vecs;
}
+void bvec_free_bs(struct bio_set *bs, struct bio_vec *bv, unsigned int idx)
+{
+ BIO_BUG_ON(idx >= BIOVEC_NR_POOLS);
+
+ if (idx == BIOVEC_MAX_IDX)
+ mempool_free(bv, bs->bvec_pool);
+ else {
+ struct biovec_slab *bvs = bvec_slabs + idx;
+
+ kmem_cache_free(bvs->slab, bv);
+ }
+}
+
struct bio_vec *bvec_alloc_bs(gfp_t gfp_mask, int nr, unsigned long *idx,
struct bio_set *bs)
{
* kzalloc() for the exact number of vecs right away.
*/
if (!bs)
- bvl = kzalloc(nr * sizeof(struct bio_vec), gfp_mask);
+ bvl = kmalloc(nr * sizeof(struct bio_vec), gfp_mask);
/*
* see comment near bvec_array define!
}
}
- if (bvl)
- memset(bvl, 0, bvec_nr_vecs(*idx) * sizeof(struct bio_vec));
-
return bvl;
}
void bio_free(struct bio *bio, struct bio_set *bs)
{
- if (bio->bi_io_vec) {
- const int pool_idx = BIO_POOL_IDX(bio);
-
- BIO_BUG_ON(pool_idx >= BIOVEC_NR_POOLS);
-
- if (pool_idx == BIOVEC_MAX_IDX)
- mempool_free(bio->bi_io_vec, bs->bvec_pool);
- else {
- struct biovec_slab *bvs = bvec_slabs + pool_idx;
+ void *p;
- kmem_cache_free(bvs->slab, bio->bi_io_vec);
- }
- }
+ if (bio_has_allocated_vec(bio))
+ bvec_free_bs(bs, bio->bi_io_vec, BIO_POOL_IDX(bio));
if (bio_integrity(bio))
- bio_integrity_free(bio, bs);
+ bio_integrity_free(bio);
+
+ /*
+ * If we have front padding, adjust the bio pointer before freeing
+ */
+ p = bio;
+ if (bs->front_pad)
+ p -= bs->front_pad;
- mempool_free(bio, bs->bio_pool);
+ mempool_free(p, bs->bio_pool);
}
/*
static void bio_kmalloc_destructor(struct bio *bio)
{
- kfree(bio->bi_io_vec);
+ if (bio_has_allocated_vec(bio))
+ kfree(bio->bi_io_vec);
kfree(bio);
}
* for a &struct bio to become free. If a %NULL @bs is passed in, we will
* fall back to just using @kmalloc to allocate the required memory.
*
- * allocate bio and iovecs from the memory pools specified by the
- * bio_set structure, or @kmalloc if none given.
+ * Note that the caller must set ->bi_destructor on succesful return
+ * of a bio, to do the appropriate freeing of the bio once the reference
+ * count drops to zero.
**/
struct bio *bio_alloc_bioset(gfp_t gfp_mask, int nr_iovecs, struct bio_set *bs)
{
- struct bio *bio;
-
- if (bs)
- bio = mempool_alloc(bs->bio_pool, gfp_mask);
- else
+ struct bio_vec *bvl = NULL;
+ struct bio *bio = NULL;
+ unsigned long idx = 0;
+ void *p = NULL;
+
+ if (bs) {
+ p = mempool_alloc(bs->bio_pool, gfp_mask);
+ if (!p)
+ goto err;
+ bio = p + bs->front_pad;
+ } else {
bio = kmalloc(sizeof(*bio), gfp_mask);
+ if (!bio)
+ goto err;
+ }
- if (likely(bio)) {
- struct bio_vec *bvl = NULL;
-
- bio_init(bio);
- if (likely(nr_iovecs)) {
- unsigned long uninitialized_var(idx);
-
- bvl = bvec_alloc_bs(gfp_mask, nr_iovecs, &idx, bs);
- if (unlikely(!bvl)) {
- if (bs)
- mempool_free(bio, bs->bio_pool);
- else
- kfree(bio);
- bio = NULL;
- goto out;
- }
- bio->bi_flags |= idx << BIO_POOL_OFFSET;
- bio->bi_max_vecs = bvec_nr_vecs(idx);
- }
- bio->bi_io_vec = bvl;
+ bio_init(bio);
+
+ if (unlikely(!nr_iovecs))
+ goto out_set;
+
+ if (nr_iovecs <= BIO_INLINE_VECS) {
+ bvl = bio->bi_inline_vecs;
+ nr_iovecs = BIO_INLINE_VECS;
+ } else {
+ bvl = bvec_alloc_bs(gfp_mask, nr_iovecs, &idx, bs);
+ if (unlikely(!bvl))
+ goto err_free;
+
+ nr_iovecs = bvec_nr_vecs(idx);
}
-out:
+ bio->bi_flags |= idx << BIO_POOL_OFFSET;
+ bio->bi_max_vecs = nr_iovecs;
+out_set:
+ bio->bi_io_vec = bvl;
+
return bio;
+
+err_free:
+ if (bs)
+ mempool_free(p, bs->bio_pool);
+ else
+ kfree(bio);
+err:
+ return NULL;
}
+/**
+ * bio_alloc - allocate a bio for I/O
+ * @gfp_mask: the GFP_ mask given to the slab allocator
+ * @nr_iovecs: number of iovecs to pre-allocate
+ *
+ * Description:
+ * bio_alloc will allocate a bio and associated bio_vec array that can hold
+ * at least @nr_iovecs entries. Allocations will be done from the
+ * fs_bio_set. Also see @bio_alloc_bioset.
+ *
+ * If %__GFP_WAIT is set, then bio_alloc will always be able to allocate
+ * a bio. This is due to the mempool guarantees. To make this work, callers
+ * must never allocate more than 1 bio at the time from this pool. Callers
+ * that need to allocate more than 1 bio must always submit the previously
+ * allocate bio for IO before attempting to allocate a new one. Failure to
+ * do so can cause livelocks under memory pressure.
+ *
+ **/
struct bio *bio_alloc(gfp_t gfp_mask, int nr_iovecs)
{
struct bio *bio = bio_alloc_bioset(gfp_mask, nr_iovecs, fs_bio_set);
if (bio_integrity(bio)) {
int ret;
- ret = bio_integrity_clone(b, bio, fs_bio_set);
+ ret = bio_integrity_clone(b, bio, gfp_mask);
- if (ret < 0)
+ if (ret < 0) {
+ bio_put(b);
return NULL;
+ }
}
return b;
int i, ret;
int nr_pages = 0;
unsigned int len = 0;
+ unsigned int offset = map_data ? map_data->offset & ~PAGE_MASK : 0;
for (i = 0; i < iov_count; i++) {
unsigned long uaddr;
bio->bi_rw |= (!write_to_vm << BIO_RW);
ret = 0;
- i = 0;
+
+ if (map_data) {
+ nr_pages = 1 << map_data->page_order;
+ i = map_data->offset / PAGE_SIZE;
+ }
while (len) {
- unsigned int bytes;
+ unsigned int bytes = PAGE_SIZE;
- if (map_data)
- bytes = 1U << (PAGE_SHIFT + map_data->page_order);
- else
- bytes = PAGE_SIZE;
+ bytes -= offset;
if (bytes > len)
bytes = len;
if (map_data) {
- if (i == map_data->nr_entries) {
+ if (i == map_data->nr_entries * nr_pages) {
ret = -ENOMEM;
break;
}
- page = map_data->pages[i++];
- } else
+
+ page = map_data->pages[i / nr_pages];
+ page += (i % nr_pages);
+
+ i++;
+ } else {
page = alloc_page(q->bounce_gfp | gfp_mask);
- if (!page) {
- ret = -ENOMEM;
- break;
+ if (!page) {
+ ret = -ENOMEM;
+ break;
+ }
}
- if (bio_add_pc_page(q, bio, page, bytes, 0) < bytes)
+ if (bio_add_pc_page(q, bio, page, bytes, offset) < bytes)
break;
len -= bytes;
+ offset = 0;
}
if (ret)
/*
* success
*/
- if (!write_to_vm) {
+ if (!write_to_vm && (!map_data || !map_data->null_mapped)) {
ret = __bio_copy_iov(bio, bio->bi_io_vec, iov, iov_count, 0, 0);
if (ret)
goto cleanup;
}
/*
- * split a bio - only worry about a bio with a single page
- * in it's iovec
+ * split a bio - only worry about a bio with a single page in its iovec
*/
struct bio_pair *bio_split(struct bio *bi, int first_sectors)
{
if (bs->bio_pool)
mempool_destroy(bs->bio_pool);
- bioset_integrity_free(bs);
biovec_free_pools(bs);
+ bio_put_slab(bs);
kfree(bs);
}
-struct bio_set *bioset_create(int bio_pool_size, int bvec_pool_size)
+/**
+ * bioset_create - Create a bio_set
+ * @pool_size: Number of bio and bio_vecs to cache in the mempool
+ * @front_pad: Number of bytes to allocate in front of the returned bio
+ *
+ * Description:
+ * Set up a bio_set to be used with @bio_alloc_bioset. Allows the caller
+ * to ask for a number of bytes to be allocated in front of the bio.
+ * Front pad allocation is useful for embedding the bio inside
+ * another structure, to avoid allocating extra data to go with the bio.
+ * Note that the bio must be embedded at the END of that structure always,
+ * or things will break badly.
+ */
+struct bio_set *bioset_create(unsigned int pool_size, unsigned int front_pad)
{
- struct bio_set *bs = kzalloc(sizeof(*bs), GFP_KERNEL);
+ unsigned int back_pad = BIO_INLINE_VECS * sizeof(struct bio_vec);
+ struct bio_set *bs;
+ bs = kzalloc(sizeof(*bs), GFP_KERNEL);
if (!bs)
return NULL;
- bs->bio_pool = mempool_create_slab_pool(bio_pool_size, bio_slab);
- if (!bs->bio_pool)
- goto bad;
+ bs->front_pad = front_pad;
+
+ bs->bio_slab = bio_find_or_create_slab(front_pad + back_pad);
+ if (!bs->bio_slab) {
+ kfree(bs);
+ return NULL;
+ }
- if (bioset_integrity_create(bs, bio_pool_size))
+ bs->bio_pool = mempool_create_slab_pool(pool_size, bs->bio_slab);
+ if (!bs->bio_pool)
goto bad;
- if (!biovec_create_pools(bs, bvec_pool_size))
+ if (!biovec_create_pools(bs, pool_size))
return bs;
bad:
int size;
struct biovec_slab *bvs = bvec_slabs + i;
+#ifndef CONFIG_BLK_DEV_INTEGRITY
+ if (bvs->nr_vecs <= BIO_INLINE_VECS) {
+ bvs->slab = NULL;
+ continue;
+ }
+#endif
+
size = bvs->nr_vecs * sizeof(struct bio_vec);
bvs->slab = kmem_cache_create(bvs->name, size, 0,
SLAB_HWCACHE_ALIGN|SLAB_PANIC, NULL);
static int __init init_bio(void)
{
- bio_slab = KMEM_CACHE(bio, SLAB_HWCACHE_ALIGN|SLAB_PANIC);
+ bio_slab_max = 2;
+ bio_slab_nr = 0;
+ bio_slabs = kzalloc(bio_slab_max * sizeof(struct bio_slab), GFP_KERNEL);
+ if (!bio_slabs)
+ panic("bio: can't allocate bios\n");
- bio_integrity_init_slab();
biovec_init_slabs();
- fs_bio_set = bioset_create(BIO_POOL_SIZE, 2);
+ fs_bio_set = bioset_create(BIO_POOL_SIZE, 0);
if (!fs_bio_set)
panic("bio: can't allocate bios\n");
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff