Btrfs: add flushoncommit mount option

The 'flushoncommit' mount option forces any data dirtied by a write in a
prior transaction to commit as part of the current commit.  This makes
the committed state a fully consistent view of the file system from the
application's perspective (i.e., it includes all completed file system
operations).  This was previously the behavior only when a snapshot is
created.

This is used by Ceph to ensure that completed writes make it to the
platter along with the metadata operations they are bound to (by
BTRFS_IOC_TRANS_{START,END}).

Signed-off-by: Sage Weil <sage@newdream.net>
Signed-off-by: Chris Mason <chris.mason@oracle.com>
This commit is contained in:
Sage Weil 2009-04-02 16:59:01 -04:00 committed by Chris Mason
parent 3a5e14048a
commit dccae99995
3 changed files with 15 additions and 5 deletions

View File

@ -1037,6 +1037,7 @@ struct btrfs_root {
#define BTRFS_MOUNT_DEGRADED (1 << 4) #define BTRFS_MOUNT_DEGRADED (1 << 4)
#define BTRFS_MOUNT_COMPRESS (1 << 5) #define BTRFS_MOUNT_COMPRESS (1 << 5)
#define BTRFS_MOUNT_NOTREELOG (1 << 6) #define BTRFS_MOUNT_NOTREELOG (1 << 6)
#define BTRFS_MOUNT_FLUSHONCOMMIT (1 << 7)
#define btrfs_clear_opt(o, opt) ((o) &= ~BTRFS_MOUNT_##opt) #define btrfs_clear_opt(o, opt) ((o) &= ~BTRFS_MOUNT_##opt)
#define btrfs_set_opt(o, opt) ((o) |= BTRFS_MOUNT_##opt) #define btrfs_set_opt(o, opt) ((o) |= BTRFS_MOUNT_##opt)

View File

@ -68,7 +68,7 @@ enum {
Opt_degraded, Opt_subvol, Opt_device, Opt_nodatasum, Opt_nodatacow, Opt_degraded, Opt_subvol, Opt_device, Opt_nodatasum, Opt_nodatacow,
Opt_max_extent, Opt_max_inline, Opt_alloc_start, Opt_nobarrier, Opt_max_extent, Opt_max_inline, Opt_alloc_start, Opt_nobarrier,
Opt_ssd, Opt_thread_pool, Opt_noacl, Opt_compress, Opt_notreelog, Opt_ssd, Opt_thread_pool, Opt_noacl, Opt_compress, Opt_notreelog,
Opt_err, Opt_flushoncommit, Opt_err,
}; };
static match_table_t tokens = { static match_table_t tokens = {
@ -86,6 +86,7 @@ static match_table_t tokens = {
{Opt_ssd, "ssd"}, {Opt_ssd, "ssd"},
{Opt_noacl, "noacl"}, {Opt_noacl, "noacl"},
{Opt_notreelog, "notreelog"}, {Opt_notreelog, "notreelog"},
{Opt_flushoncommit, "flushoncommit"},
{Opt_err, NULL}, {Opt_err, NULL},
}; };
@ -229,6 +230,10 @@ int btrfs_parse_options(struct btrfs_root *root, char *options)
printk(KERN_INFO "btrfs: disabling tree log\n"); printk(KERN_INFO "btrfs: disabling tree log\n");
btrfs_set_opt(info->mount_opt, NOTREELOG); btrfs_set_opt(info->mount_opt, NOTREELOG);
break; break;
case Opt_flushoncommit:
printk(KERN_INFO "btrfs: turning on flush-on-commit\n");
btrfs_set_opt(info->mount_opt, FLUSHONCOMMIT);
break;
default: default:
break; break;
} }
@ -370,9 +375,8 @@ fail_close:
int btrfs_sync_fs(struct super_block *sb, int wait) int btrfs_sync_fs(struct super_block *sb, int wait)
{ {
struct btrfs_trans_handle *trans; struct btrfs_trans_handle *trans;
struct btrfs_root *root; struct btrfs_root *root = btrfs_sb(sb);
int ret; int ret;
root = btrfs_sb(sb);
if (sb->s_flags & MS_RDONLY) if (sb->s_flags & MS_RDONLY)
return 0; return 0;
@ -419,7 +423,9 @@ static int btrfs_show_options(struct seq_file *seq, struct vfsmount *vfs)
if (btrfs_test_opt(root, SSD)) if (btrfs_test_opt(root, SSD))
seq_puts(seq, ",ssd"); seq_puts(seq, ",ssd");
if (btrfs_test_opt(root, NOTREELOG)) if (btrfs_test_opt(root, NOTREELOG))
seq_puts(seq, ",notreelog"); seq_puts(seq, ",no-treelog");
if (btrfs_test_opt(root, FLUSHONCOMMIT))
seq_puts(seq, ",flush-on-commit");
if (!(root->fs_info->sb->s_flags & MS_POSIXACL)) if (!(root->fs_info->sb->s_flags & MS_POSIXACL))
seq_puts(seq, ",noacl"); seq_puts(seq, ",noacl");
return 0; return 0;

View File

@ -972,6 +972,7 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
int ret; int ret;
int should_grow = 0; int should_grow = 0;
unsigned long now = get_seconds(); unsigned long now = get_seconds();
int flush_on_commit = btrfs_test_opt(root, FLUSHONCOMMIT);
btrfs_run_ordered_operations(root, 0); btrfs_run_ordered_operations(root, 0);
@ -1051,7 +1052,9 @@ int btrfs_commit_transaction(struct btrfs_trans_handle *trans,
mutex_unlock(&root->fs_info->trans_mutex); mutex_unlock(&root->fs_info->trans_mutex);
if (snap_pending) { if (flush_on_commit || snap_pending) {
if (flush_on_commit)
btrfs_start_delalloc_inodes(root);
ret = btrfs_wait_ordered_extents(root, 1); ret = btrfs_wait_ordered_extents(root, 1);
BUG_ON(ret); BUG_ON(ret);
} }