1. Background
Previously, if f2fs tries to move data blocks of an *evicting* inode during the
cleaning process, it stops the process incompletely and then restarts the whole
process, since it needs a locked inode to grab victim data pages in its address
space. In order to get a locked inode, iget_locked() by f2fs_iget() is normally
used, but, it waits if the inode is on freeing.
So, here is a deadlock scenario.
1. f2fs_evict_inode() <- inode "A"
2. f2fs_balance_fs()
3. f2fs_gc()
4. gc_data_segment()
5. f2fs_iget() <- inode "A" too!
If step #1 and #5 treat a same inode "A", step #5 would fall into deadlock since
the inode "A" is on freeing. In order to resolve this, f2fs_iget_nowait() which
skips __wait_on_freeing_inode() was introduced in step #5, and stops f2fs_gc()
to complete f2fs_evict_inode().
1. f2fs_evict_inode() <- inode "A"
2. f2fs_balance_fs()
3. f2fs_gc()
4. gc_data_segment()
5. f2fs_iget_nowait() <- inode "A", then stop f2fs_gc() w/ -ENOENT
2. Problem and Solution
In the above scenario, however, f2fs cannot finish f2fs_evict_inode() only if:
o there are not enough free sections, and
o f2fs_gc() tries to move data blocks of the *evicting* inode repeatedly.
So, the final solution is to use f2fs_iget() and remove f2fs_balance_fs() in
f2fs_evict_inode().
The f2fs_evict_inode() actually truncates all the data and node blocks, which
means that it doesn't produce any dirty node pages accordingly.
So, we don't need to do f2fs_balance_fs() in practical.
Signed-off-by: Jaegeuk Kim <jaegeuk.kim@xxxxxxxxxxx>
---
fs/f2fs/f2fs.h | 1 -
fs/f2fs/file.c | 3 +--
fs/f2fs/gc.c | 2 +-
fs/f2fs/inode.c | 33 ---------------------------------
fs/f2fs/recovery.c | 2 +-
5 files changed, 3 insertions(+), 38 deletions(-)
diff --git a/fs/f2fs/f2fs.h b/fs/f2fs/f2fs.h
index 975cb44..5022a7d 100644
--- a/fs/f2fs/f2fs.h
+++ b/fs/f2fs/f2fs.h
@@ -847,7 +847,6 @@ long f2fs_ioctl(struct file *, unsigned int, unsigned long);
* inode.c
*/
void f2fs_set_inode_flags(struct inode *);
-struct inode *f2fs_iget_nowait(struct super_block *, unsigned long);
struct inode *f2fs_iget(struct super_block *, unsigned long);
void update_inode(struct inode *, struct page *);
int f2fs_write_inode(struct inode *, struct writeback_control *);
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c
index 6cdab2c..3efa473 100644
--- a/fs/f2fs/file.c
+++ b/fs/f2fs/file.c
@@ -298,8 +298,6 @@ void f2fs_truncate(struct inode *inode)
inode->i_mtime = inode->i_ctime = CURRENT_TIME;
mark_inode_dirty(inode);
}
-
- f2fs_balance_fs(F2FS_SB(inode->i_sb));
}
static int f2fs_getattr(struct vfsmount *mnt,
@@ -356,6 +354,7 @@ int f2fs_setattr(struct dentry *dentry, struct iattr *attr)
attr->ia_size != i_size_read(inode)) {
truncate_setsize(inode, attr->ia_size);
f2fs_truncate(inode);
+ f2fs_balance_fs(F2FS_SB(inode->i_sb));
}
__setattr_copy(inode, attr);
diff --git a/fs/f2fs/gc.c b/fs/f2fs/gc.c
index fb03be6..375e69e 100644
--- a/fs/f2fs/gc.c
+++ b/fs/f2fs/gc.c
@@ -579,7 +579,7 @@ next_step:
ofs_in_node = le16_to_cpu(entry->ofs_in_node);
if (phase == 2) {
- inode = f2fs_iget_nowait(sb, dni.ino);
+ inode = f2fs_iget(sb, dni.ino);
if (IS_ERR(inode))
continue;
diff --git a/fs/f2fs/inode.c b/fs/f2fs/inode.c
index 62433c6..ddae412 100644
--- a/fs/f2fs/inode.c
+++ b/fs/f2fs/inode.c
@@ -16,11 +16,6 @@
#include "f2fs.h"
#include "node.h"
-struct f2fs_iget_args {
- u64 ino;
- int on_free;
-};
-
void f2fs_set_inode_flags(struct inode *inode)
{
unsigned int flags = F2FS_I(inode)->i_flags;
@@ -40,34 +35,6 @@ void f2fs_set_inode_flags(struct inode *inode)
inode->i_flags |= S_DIRSYNC;
}
-static int f2fs_iget_test(struct inode *inode, void *data)
-{
- struct f2fs_iget_args *args = data;
-
- if (inode->i_ino != args->ino)
- return 0;
- if (inode->i_state & (I_FREEING | I_WILL_FREE)) {
- args->on_free = 1;
- return 0;
- }
- return 1;
-}
-
-struct inode *f2fs_iget_nowait(struct super_block *sb, unsigned long ino)
-{
- struct f2fs_iget_args args = {
- .ino = ino,
- .on_free = 0
- };
- struct inode *inode = ilookup5(sb, ino, f2fs_iget_test, &args);
-
- if (inode)
- return inode;
- if (!args.on_free)
- return f2fs_iget(sb, ino);
- return ERR_PTR(-ENOENT);
-}
-
static int do_read_inode(struct inode *inode)
{
struct f2fs_sb_info *sbi = F2FS_SB(inode->i_sb);
diff --git a/fs/f2fs/recovery.c b/fs/f2fs/recovery.c
index f42e406..e2a3e1a 100644
--- a/fs/f2fs/recovery.c
+++ b/fs/f2fs/recovery.c
@@ -226,7 +226,7 @@ static void check_index_in_prev_nodes(struct f2fs_sb_info *sbi,
f2fs_put_page(node_page, 1);
/* Deallocate previous index in the node page */
- inode = f2fs_iget_nowait(sbi->sb, ino);
+ inode = f2fs_iget(sbi->sb, ino);
if (IS_ERR(inode))
return;
--
1.8.0.1.250.gb7973fb
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
