aboutsummaryrefslogtreecommitdiffstats
path: root/fs/btrfs/volumes.c
diff options
context:
space:
mode:
Diffstat (limited to 'fs/btrfs/volumes.c')
-rw-r--r--fs/btrfs/volumes.c115
1 files changed, 78 insertions, 37 deletions
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index a6d35b0054ca..5dbefd11b4af 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -161,8 +161,10 @@ static noinline int run_scheduled_bios(struct btrfs_device *device)
int again = 0;
unsigned long num_run;
unsigned long num_sync_run;
+ unsigned long batch_run = 0;
unsigned long limit;
unsigned long last_waited = 0;
+ int force_reg = 0;
bdi = blk_get_backing_dev_info(device->bdev);
fs_info = device->dev_root->fs_info;
@@ -176,19 +178,22 @@ static noinline int run_scheduled_bios(struct btrfs_device *device)
loop:
spin_lock(&device->io_lock);
- num_run = 0;
loop_lock:
+ num_run = 0;
/* take all the bios off the list at once and process them
* later on (without the lock held). But, remember the
* tail and other pointers so the bios can be properly reinserted
* into the list if we hit congestion
*/
- if (device->pending_sync_bios.head)
+ if (!force_reg && device->pending_sync_bios.head) {
pending_bios = &device->pending_sync_bios;
- else
+ force_reg = 1;
+ } else {
pending_bios = &device->pending_bios;
+ force_reg = 0;
+ }
pending = pending_bios->head;
tail = pending_bios->tail;
@@ -228,10 +233,14 @@ loop_lock:
while (pending) {
rmb();
- if (pending_bios != &device->pending_sync_bios &&
- device->pending_sync_bios.head &&
- num_run > 16) {
- cond_resched();
+ /* we want to work on both lists, but do more bios on the
+ * sync list than the regular list
+ */
+ if ((num_run > 32 &&
+ pending_bios != &device->pending_sync_bios &&
+ device->pending_sync_bios.head) ||
+ (num_run > 64 && pending_bios == &device->pending_sync_bios &&
+ device->pending_bios.head)) {
spin_lock(&device->io_lock);
requeue_list(pending_bios, pending, tail);
goto loop_lock;
@@ -249,6 +258,8 @@ loop_lock:
BUG_ON(atomic_read(&cur->bi_cnt) == 0);
submit_bio(cur->bi_rw, cur);
num_run++;
+ batch_run++;
+
if (bio_sync(cur))
num_sync_run++;
@@ -265,7 +276,7 @@ loop_lock:
* is now congested. Back off and let other work structs
* run instead
*/
- if (pending && bdi_write_congested(bdi) && num_run > 16 &&
+ if (pending && bdi_write_congested(bdi) && batch_run > 32 &&
fs_info->fs_devices->open_devices > 1) {
struct io_context *ioc;
@@ -366,6 +377,7 @@ static noinline int device_list_add(const char *path,
memcpy(fs_devices->fsid, disk_super->fsid, BTRFS_FSID_SIZE);
fs_devices->latest_devid = devid;
fs_devices->latest_trans = found_transid;
+ mutex_init(&fs_devices->device_list_mutex);
device = NULL;
} else {
device = __find_device(&fs_devices->devices, devid,
@@ -392,7 +404,11 @@ static noinline int device_list_add(const char *path,
return -ENOMEM;
}
INIT_LIST_HEAD(&device->dev_alloc_list);
+
+ mutex_lock(&fs_devices->device_list_mutex);
list_add(&device->dev_list, &fs_devices->devices);
+ mutex_unlock(&fs_devices->device_list_mutex);
+
device->fs_devices = fs_devices;
fs_devices->num_devices++;
}
@@ -418,10 +434,12 @@ static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig)
INIT_LIST_HEAD(&fs_devices->devices);
INIT_LIST_HEAD(&fs_devices->alloc_list);
INIT_LIST_HEAD(&fs_devices->list);
+ mutex_init(&fs_devices->device_list_mutex);
fs_devices->latest_devid = orig->latest_devid;
fs_devices->latest_trans = orig->latest_trans;
memcpy(fs_devices->fsid, orig->fsid, sizeof(fs_devices->fsid));
+ mutex_lock(&orig->device_list_mutex);
list_for_each_entry(orig_dev, &orig->devices, dev_list) {
device = kzalloc(sizeof(*device), GFP_NOFS);
if (!device)
@@ -443,8 +461,10 @@ static struct btrfs_fs_devices *clone_fs_devices(struct btrfs_fs_devices *orig)
device->fs_devices = fs_devices;
fs_devices->num_devices++;
}
+ mutex_unlock(&orig->device_list_mutex);
return fs_devices;
error:
+ mutex_unlock(&orig->device_list_mutex);
free_fs_devices(fs_devices);
return ERR_PTR(-ENOMEM);
}
@@ -455,6 +475,7 @@ int btrfs_close_extra_devices(struct btrfs_fs_devices *fs_devices)
mutex_lock(&uuid_mutex);
again:
+ mutex_lock(&fs_devices->device_list_mutex);
list_for_each_entry_safe(device, next, &fs_devices->devices, dev_list) {
if (device->in_fs_metadata)
continue;
@@ -474,6 +495,7 @@ again:
kfree(device->name);
kfree(device);
}
+ mutex_unlock(&fs_devices->device_list_mutex);
if (fs_devices->seed) {
fs_devices = fs_devices->seed;
@@ -594,6 +616,9 @@ static int __btrfs_open_devices(struct btrfs_fs_devices *fs_devices,
device->in_fs_metadata = 0;
device->mode = flags;
+ if (!blk_queue_nonrot(bdev_get_queue(bdev)))
+ fs_devices->rotating = 1;
+
fs_devices->open_devices++;
if (device->writeable) {
fs_devices->rw_devices++;
@@ -696,7 +721,8 @@ error:
*/
static noinline int find_free_dev_extent(struct btrfs_trans_handle *trans,
struct btrfs_device *device,
- u64 num_bytes, u64 *start)
+ u64 num_bytes, u64 *start,
+ u64 *max_avail)
{
struct btrfs_key key;
struct btrfs_root *root = device->dev_root;
@@ -733,9 +759,13 @@ static noinline int find_free_dev_extent(struct btrfs_trans_handle *trans,
ret = btrfs_search_slot(trans, root, &key, path, 0, 0);
if (ret < 0)
goto error;
- ret = btrfs_previous_item(root, path, 0, key.type);
- if (ret < 0)
- goto error;
+ if (ret > 0) {
+ ret = btrfs_previous_item(root, path, key.objectid, key.type);
+ if (ret < 0)
+ goto error;
+ if (ret > 0)
+ start_found = 1;
+ }
l = path->nodes[0];
btrfs_item_key_to_cpu(l, &key, path->slots[0]);
while (1) {
@@ -778,6 +808,10 @@ no_more_items:
if (last_byte < search_start)
last_byte = search_start;
hole_size = key.offset - last_byte;
+
+ if (hole_size > *max_avail)
+ *max_avail = hole_size;
+
if (key.offset > last_byte &&
hole_size >= num_bytes) {
*start = last_byte;
@@ -1121,12 +1155,14 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
device = NULL;
devices = &root->fs_info->fs_devices->devices;
+ mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
list_for_each_entry(tmp, devices, dev_list) {
if (tmp->in_fs_metadata && !tmp->bdev) {
device = tmp;
break;
}
}
+ mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
bdev = NULL;
bh = NULL;
disk_super = NULL;
@@ -1181,7 +1217,16 @@ int btrfs_rm_device(struct btrfs_root *root, char *device_path)
goto error_brelse;
device->in_fs_metadata = 0;
+
+ /*
+ * the device list mutex makes sure that we don't change
+ * the device list while someone else is writing out all
+ * the device supers.
+ */
+ mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
list_del_init(&device->dev_list);
+ mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
+
device->fs_devices->num_devices--;
next_device = list_entry(root->fs_info->fs_devices->devices.next,
@@ -1275,6 +1320,7 @@ static int btrfs_prepare_sprout(struct btrfs_trans_handle *trans,
seed_devices->opened = 1;
INIT_LIST_HEAD(&seed_devices->devices);
INIT_LIST_HEAD(&seed_devices->alloc_list);
+ mutex_init(&seed_devices->device_list_mutex);
list_splice_init(&fs_devices->devices, &seed_devices->devices);
list_splice_init(&fs_devices->alloc_list, &seed_devices->alloc_list);
list_for_each_entry(device, &seed_devices->devices, dev_list) {
@@ -1400,6 +1446,10 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
mutex_lock(&root->fs_info->volume_mutex);
devices = &root->fs_info->fs_devices->devices;
+ /*
+ * we have the volume lock, so we don't need the extra
+ * device list mutex while reading the list here.
+ */
list_for_each_entry(device, devices, dev_list) {
if (device->bdev == bdev) {
ret = -EEXIST;
@@ -1454,6 +1504,12 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
}
device->fs_devices = root->fs_info->fs_devices;
+
+ /*
+ * we don't want write_supers to jump in here with our device
+ * half setup
+ */
+ mutex_lock(&root->fs_info->fs_devices->device_list_mutex);
list_add(&device->dev_list, &root->fs_info->fs_devices->devices);
list_add(&device->dev_alloc_list,
&root->fs_info->fs_devices->alloc_list);
@@ -1462,6 +1518,9 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
root->fs_info->fs_devices->rw_devices++;
root->fs_info->fs_devices->total_rw_bytes += device->total_bytes;
+ if (!blk_queue_nonrot(bdev_get_queue(bdev)))
+ root->fs_info->fs_devices->rotating = 1;
+
total_bytes = btrfs_super_total_bytes(&root->fs_info->super_copy);
btrfs_set_super_total_bytes(&root->fs_info->super_copy,
total_bytes + device->total_bytes);
@@ -1469,6 +1528,7 @@ int btrfs_init_new_device(struct btrfs_root *root, char *device_path)
total_bytes = btrfs_super_num_devices(&root->fs_info->super_copy);
btrfs_set_super_num_devices(&root->fs_info->super_copy,
total_bytes + 1);
+ mutex_unlock(&root->fs_info->fs_devices->device_list_mutex);
if (seeding_dev) {
ret = init_first_rw_device(trans, root, device);
@@ -1570,6 +1630,7 @@ static int __btrfs_grow_device(struct btrfs_trans_handle *trans,
device->fs_devices->total_rw_bytes += diff;
device->total_bytes = new_size;
+ device->disk_total_bytes = new_size;
btrfs_clear_space_info_full(device->dev_root->fs_info);
return btrfs_update_device(trans, device);
@@ -1671,8 +1732,6 @@ static int btrfs_relocate_chunk(struct btrfs_root *root,
int ret;
int i;
- printk(KERN_INFO "btrfs relocating chunk %llu\n",
- (unsigned long long)chunk_offset);
root = root->fs_info->chunk_root;
extent_root = root->fs_info->extent_root;
em_tree = &root->fs_info->mapping_tree.map_tree;
@@ -1958,7 +2017,7 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
goto done;
if (ret) {
ret = 0;
- goto done;
+ break;
}
l = path->nodes[0];
@@ -1966,7 +2025,7 @@ int btrfs_shrink_device(struct btrfs_device *device, u64 new_size)
btrfs_item_key_to_cpu(l, &key, path->slots[0]);
if (key.objectid != device->devid)
- goto done;
+ break;
dev_extent = btrfs_item_ptr(l, slot, struct btrfs_dev_extent);
length = btrfs_dev_extent_length(l, dev_extent);
@@ -2122,6 +2181,7 @@ static int __btrfs_alloc_chunk(struct btrfs_trans_handle *trans,
max_chunk_size);
again:
+ max_avail = 0;
if (!map || map->num_stripes != num_stripes) {
kfree(map);
map = kmalloc(map_lookup_size(num_stripes), GFP_NOFS);
@@ -2170,7 +2230,8 @@ again:
if (device->in_fs_metadata && avail >= min_free) {
ret = find_free_dev_extent(trans, device,
- min_free, &dev_offset);
+ min_free, &dev_offset,
+ &max_avail);
if (ret == 0) {
list_move_tail(&device->dev_alloc_list,
&private_devs);
@@ -2746,26 +2807,6 @@ int btrfs_rmap_block(struct btrfs_mapping_tree *map_tree,
}
}
- for (i = 0; i > nr; i++) {
- struct btrfs_multi_bio *multi;
- struct btrfs_bio_stripe *stripe;
- int ret;
-
- length = 1;
- ret = btrfs_map_block(map_tree, WRITE, buf[i],
- &length, &multi, 0);
- BUG_ON(ret);
-
- stripe = multi->stripes;
- for (j = 0; j < multi->num_stripes; j++) {
- if (stripe->physical >= physical &&
- physical < stripe->physical + length)
- break;
- }
- BUG_ON(j >= multi->num_stripes);
- kfree(multi);
- }
-
*logical = buf;
*naddrs = nr;
*stripe_len = map->stripe_len;
Copyright © 1996 – 2023 Jason A. Donenfeld. All Rights Reverse Engineered.