lk/linux-kernel-test
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

ocfs2: Attach xattr clusters to refcount tree.

Browse Source 
In ocfs2, when xattr's value is larger than OCFS2_XATTR_INLINE_SIZE,
it will be kept outside of the blocks we store xattr entry. And they
are stored in a b-tree also. So this patch try to attach all these
clusters to refcount tree also.

Signed-off-by: Tao Ma <tao.ma@oracle.com>
This commit is contained in:
Tao Ma
2009-09-21 13:04:19 +08:00
committed by Joel Becker
parent 47bca4950b
commit 0129241e2b
4 changed files with 329 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files

29
fs/ocfs2/refcounttree.c
View File

@ -3547,7 +3547,8 @@ int ocfs2_add_refcount_flag(struct inode *inode,
struct ocfs2_caching_info *ref_ci,
struct buffer_head *ref_root_bh,
u32 cpos, u32 p_cluster, u32 num_clusters,
struct ocfs2_cached_dealloc_ctxt *dealloc)
struct ocfs2_cached_dealloc_ctxt *dealloc,
struct ocfs2_post_refcount *post)
{
int ret;
handle_t *handle;
@ -3576,6 +3577,9 @@ int ocfs2_add_refcount_flag(struct inode *inode,
}
}
if (post)
credits += post->credits;
handle = ocfs2_start_trans(osb, credits);
if (IS_ERR(handle)) {
ret = PTR_ERR(handle);
@ -3594,8 +3598,16 @@ int ocfs2_add_refcount_flag(struct inode *inode,
ret = __ocfs2_increase_refcount(handle, ref_ci, ref_root_bh,
p_cluster, num_clusters,
meta_ac, dealloc);
if (ret)
if (ret) {
mlog_errno(ret);
goto out_commit;
}
if (post && post->func) {
ret = post->func(inode, handle, post->para);
if (ret)
mlog_errno(ret);
}
out_commit:
ocfs2_commit_trans(osb, handle);
@ -3688,7 +3700,7 @@ static int ocfs2_attach_refcount_tree(struct inode *inode,
&ref_tree->rf_ci,
ref_root_bh, cpos,
p_cluster, num_clusters,
&dealloc);
&dealloc, NULL);
if (ret) {
mlog_errno(ret);
goto unlock;
@ -3699,6 +3711,17 @@ static int ocfs2_attach_refcount_tree(struct inode *inode,
cpos += num_clusters;
}
if (oi->ip_dyn_features & OCFS2_HAS_XATTR_FL) {
ret = ocfs2_xattr_attach_refcount_tree(inode, di_bh,
&ref_tree->rf_ci,
ref_root_bh,
&dealloc);
if (ret) {
mlog_errno(ret);
goto unlock;
}
}
if (data_changed) {
ret = ocfs2_change_ctime(inode, di_bh);
if (ret)