struct ocfs2_alloc_context *meta_ac;
struct ocfs2_alloc_context *data_ac;
struct ocfs2_cached_dealloc_ctxt dealloc;
+ int set_abort;
};
#define OCFS2_XATTR_ROOT_SIZE (sizeof(struct ocfs2_xattr_def_value_root))
.xv.xr_list.l_count = cpu_to_le16(1),
};
-struct xattr_handler *ocfs2_xattr_handlers[] = {
+const struct xattr_handler *ocfs2_xattr_handlers[] = {
&ocfs2_xattr_user_handler,
&ocfs2_xattr_acl_access_handler,
&ocfs2_xattr_acl_default_handler,
NULL
};
-static struct xattr_handler *ocfs2_xattr_handler_map[OCFS2_XATTR_MAX] = {
+static const struct xattr_handler *ocfs2_xattr_handler_map[OCFS2_XATTR_MAX] = {
[OCFS2_XATTR_INDEX_USER] = &ocfs2_xattr_user_handler,
[OCFS2_XATTR_INDEX_POSIX_ACL_ACCESS]
= &ocfs2_xattr_acl_access_handler,
static inline const char *ocfs2_xattr_prefix(int name_index)
{
- struct xattr_handler *handler = NULL;
+ const struct xattr_handler *handler = NULL;
if (name_index > 0 && name_index < OCFS2_XATTR_MAX)
handler = ocfs2_xattr_handler_map[name_index];
goto leave;
}
- status = ocfs2_journal_dirty(handle, vb->vb_bh);
- if (status < 0) {
- mlog_errno(status);
- goto leave;
- }
+ ocfs2_journal_dirty(handle, vb->vb_bh);
clusters_to_add -= le32_to_cpu(vb->vb_xv->xr_clusters) - prev_clusters;
}
le32_add_cpu(&vb->vb_xv->xr_clusters, -len);
-
- ret = ocfs2_journal_dirty(handle, vb->vb_bh);
- if (ret) {
- mlog_errno(ret);
- goto out;
- }
+ ocfs2_journal_dirty(handle, vb->vb_bh);
if (ext_flags & OCFS2_EXT_REFCOUNTED)
ret = ocfs2_decrease_refcount(inode, handle,
memset(bh->b_data + cp_len, 0,
blocksize - cp_len);
- ret = ocfs2_journal_dirty(handle, bh);
- if (ret < 0) {
- mlog_errno(ret);
- goto out;
- }
+ ocfs2_journal_dirty(handle, bh);
brelse(bh);
bh = NULL;
/* Now tell xh->xh_entries about it */
for (i = 0; i < count; i++) {
offset = le16_to_cpu(xh->xh_entries[i].xe_name_offset);
- if (offset < namevalue_offset)
+ if (offset <= namevalue_offset)
le16_add_cpu(&xh->xh_entries[i].xe_name_offset,
namevalue_size);
}
.xlo_fill_value_buf = ocfs2_xa_bucket_fill_value_buf,
};
+static unsigned int ocfs2_xa_value_clusters(struct ocfs2_xa_loc *loc)
+{
+ struct ocfs2_xattr_value_buf vb;
+
+ if (ocfs2_xattr_is_local(loc->xl_entry))
+ return 0;
+
+ ocfs2_xa_fill_value_buf(loc, &vb);
+ return le32_to_cpu(vb.vb_xv->xr_clusters);
+}
+
static int ocfs2_xa_value_truncate(struct ocfs2_xa_loc *loc, u64 bytes,
struct ocfs2_xattr_set_ctxt *ctxt)
{
}
}
+/*
+ * If we have a problem adjusting the size of an external value during
+ * ocfs2_xa_prepare_entry() or ocfs2_xa_remove(), we may have an xattr
+ * in an intermediate state. For example, the value may be partially
+ * truncated.
+ *
+ * If the value tree hasn't changed, the extend/truncate went nowhere.
+ * We have nothing to do. The caller can treat it as a straight error.
+ *
+ * If the value tree got partially truncated, we now have a corrupted
+ * extended attribute. We're going to wipe its entry and leak the
+ * clusters. Better to leak some storage than leave a corrupt entry.
+ *
+ * If the value tree grew, it obviously didn't grow enough for the
+ * new entry. We're not going to try and reclaim those clusters either.
+ * If there was already an external value there (orig_clusters != 0),
+ * the new clusters are attached safely and we can just leave the old
+ * value in place. If there was no external value there, we remove
+ * the entry.
+ *
+ * This way, the xattr block we store in the journal will be consistent.
+ * If the size change broke because of the journal, no changes will hit
+ * disk anyway.
+ */
+static void ocfs2_xa_cleanup_value_truncate(struct ocfs2_xa_loc *loc,
+ const char *what,
+ unsigned int orig_clusters)
+{
+ unsigned int new_clusters = ocfs2_xa_value_clusters(loc);
+ char *nameval_buf = ocfs2_xa_offset_pointer(loc,
+ le16_to_cpu(loc->xl_entry->xe_name_offset));
+
+ if (new_clusters < orig_clusters) {
+ mlog(ML_ERROR,
+ "Partial truncate while %s xattr %.*s. Leaking "
+ "%u clusters and removing the entry\n",
+ what, loc->xl_entry->xe_name_len, nameval_buf,
+ orig_clusters - new_clusters);
+ ocfs2_xa_remove_entry(loc);
+ } else if (!orig_clusters) {
+ mlog(ML_ERROR,
+ "Unable to allocate an external value for xattr "
+ "%.*s safely. Leaking %u clusters and removing the "
+ "entry\n",
+ loc->xl_entry->xe_name_len, nameval_buf,
+ new_clusters - orig_clusters);
+ ocfs2_xa_remove_entry(loc);
+ } else if (new_clusters > orig_clusters)
+ mlog(ML_ERROR,
+ "Unable to grow xattr %.*s safely. %u new clusters "
+ "have been added, but the value will not be "
+ "modified\n",
+ loc->xl_entry->xe_name_len, nameval_buf,
+ new_clusters - orig_clusters);
+}
+
static int ocfs2_xa_remove(struct ocfs2_xa_loc *loc,
struct ocfs2_xattr_set_ctxt *ctxt)
{
int rc = 0;
+ unsigned int orig_clusters;
if (!ocfs2_xattr_is_local(loc->xl_entry)) {
+ orig_clusters = ocfs2_xa_value_clusters(loc);
rc = ocfs2_xa_value_truncate(loc, 0, ctxt);
if (rc) {
mlog_errno(rc);
- goto out;
+ /*
+ * Since this is remove, we can return 0 if
+ * ocfs2_xa_cleanup_value_truncate() is going to
+ * wipe the entry anyway. So we check the
+ * cluster count as well.
+ */
+ if (orig_clusters != ocfs2_xa_value_clusters(loc))
+ rc = 0;
+ ocfs2_xa_cleanup_value_truncate(loc, "removing",
+ orig_clusters);
+ if (rc)
+ goto out;
}
}
{
int rc = 0;
int name_size = OCFS2_XATTR_SIZE(xi->xi_name_len);
+ unsigned int orig_clusters;
char *nameval_buf;
int xe_local = ocfs2_xattr_is_local(loc->xl_entry);
int xi_local = xi->xi_value_len <= OCFS2_XATTR_INLINE_SIZE;
if (!xi_local)
ocfs2_xa_install_value_root(loc);
} else {
+ orig_clusters = ocfs2_xa_value_clusters(loc);
if (xi_local) {
rc = ocfs2_xa_value_truncate(loc, 0, ctxt);
- if (rc < 0) {
+ if (rc < 0)
mlog_errno(rc);
- goto out;
- }
- memset(nameval_buf + name_size, 0,
- namevalue_size_xe(loc->xl_entry) -
- name_size);
+ else
+ memset(nameval_buf + name_size, 0,
+ namevalue_size_xe(loc->xl_entry) -
+ name_size);
} else if (le64_to_cpu(loc->xl_entry->xe_value_size) >
xi->xi_value_len) {
rc = ocfs2_xa_value_truncate(loc, xi->xi_value_len,
ctxt);
- if (rc < 0) {
+ if (rc < 0)
mlog_errno(rc);
- goto out;
- }
+ }
+
+ if (rc) {
+ ocfs2_xa_cleanup_value_truncate(loc, "reusing",
+ orig_clusters);
+ goto out;
}
}
struct ocfs2_xattr_set_ctxt *ctxt)
{
int rc = 0;
+ unsigned int orig_clusters;
+ __le64 orig_value_size = 0;
rc = ocfs2_xa_check_space(loc, xi);
if (rc)
if (loc->xl_entry) {
if (ocfs2_xa_can_reuse_entry(loc, xi)) {
+ orig_value_size = loc->xl_entry->xe_value_size;
rc = ocfs2_xa_reuse_entry(loc, xi, ctxt);
if (rc)
goto out;
}
if (!ocfs2_xattr_is_local(loc->xl_entry)) {
+ orig_clusters = ocfs2_xa_value_clusters(loc);
rc = ocfs2_xa_value_truncate(loc, 0, ctxt);
if (rc) {
mlog_errno(rc);
+ ocfs2_xa_cleanup_value_truncate(loc,
+ "overwriting",
+ orig_clusters);
goto out;
}
}
alloc_value:
if (xi->xi_value_len > OCFS2_XATTR_INLINE_SIZE) {
+ orig_clusters = ocfs2_xa_value_clusters(loc);
rc = ocfs2_xa_value_truncate(loc, xi->xi_value_len, ctxt);
- if (rc < 0)
+ if (rc < 0) {
+ ctxt->set_abort = 1;
+ ocfs2_xa_cleanup_value_truncate(loc, "growing",
+ orig_clusters);
+ /*
+ * If we were growing an existing value,
+ * ocfs2_xa_cleanup_value_truncate() won't remove
+ * the entry. We need to restore the original value
+ * size.
+ */
+ if (loc->xl_entry) {
+ BUG_ON(!orig_value_size);
+ loc->xl_entry->xe_value_size = orig_value_size;
+ }
mlog_errno(rc);
+ }
}
out:
goto out;
}
+ /*
+ * From here on out, everything is going to modify the buffer a
+ * little. Errors are going to leave the xattr header in a
+ * sane state. Thus, even with errors we dirty the sucker.
+ */
+
/* Don't worry, we are never called with !xi_value and !xl_entry */
if (!xi->xi_value) {
ret = ocfs2_xa_remove(loc, ctxt);
- goto out;
+ goto out_dirty;
}
ret = ocfs2_xa_prepare_entry(loc, xi, name_hash, ctxt);
if (ret) {
if (ret != -ENOSPC)
mlog_errno(ret);
- goto out;
+ goto out_dirty;
}
ret = ocfs2_xa_store_value(loc, xi, ctxt);
- if (ret) {
+ if (ret)
mlog_errno(ret);
- goto out;
- }
+out_dirty:
ocfs2_xa_journal_dirty(ctxt->handle, loc);
out:
xb = (struct ocfs2_xattr_block *)blk_bh->b_data;
blk = le64_to_cpu(xb->xb_blkno);
bit = le16_to_cpu(xb->xb_suballoc_bit);
- bg_blkno = ocfs2_which_suballoc_group(blk, bit);
+ if (xb->xb_suballoc_loc)
+ bg_blkno = le64_to_cpu(xb->xb_suballoc_loc);
+ else
+ bg_blkno = ocfs2_which_suballoc_group(blk, bit);
xb_alloc_inode = ocfs2_get_system_file_inode(osb,
EXTENT_ALLOC_SYSTEM_INODE,
di->i_dyn_features = cpu_to_le16(oi->ip_dyn_features);
spin_unlock(&oi->ip_lock);
- ret = ocfs2_journal_dirty(handle, di_bh);
- if (ret < 0)
- mlog_errno(ret);
+ ocfs2_journal_dirty(handle, di_bh);
out_commit:
ocfs2_commit_trans(OCFS2_SB(inode->i_sb), handle);
out:
di->i_dyn_features = cpu_to_le16(oi->ip_dyn_features);
spin_unlock(&oi->ip_lock);
- ret = ocfs2_journal_dirty(ctxt->handle, di_bh);
- if (ret < 0)
- mlog_errno(ret);
+ ocfs2_journal_dirty(ctxt->handle, di_bh);
out:
return ret;
int ret;
u16 suballoc_bit_start;
u32 num_got;
- u64 first_blkno;
+ u64 suballoc_loc, first_blkno;
struct ocfs2_dinode *di = (struct ocfs2_dinode *)inode_bh->b_data;
- struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct buffer_head *new_bh = NULL;
struct ocfs2_xattr_block *xblk;
goto end;
}
- ret = ocfs2_claim_metadata(osb, ctxt->handle, ctxt->meta_ac, 1,
- &suballoc_bit_start, &num_got,
- &first_blkno);
+ ret = ocfs2_claim_metadata(ctxt->handle, ctxt->meta_ac, 1,
+ &suballoc_loc, &suballoc_bit_start,
+ &num_got, &first_blkno);
if (ret < 0) {
mlog_errno(ret);
goto end;
memset(xblk, 0, inode->i_sb->s_blocksize);
strcpy((void *)xblk, OCFS2_XATTR_BLOCK_SIGNATURE);
xblk->xb_suballoc_slot = cpu_to_le16(ctxt->meta_ac->ac_alloc_slot);
+ xblk->xb_suballoc_loc = cpu_to_le64(suballoc_loc);
xblk->xb_suballoc_bit = cpu_to_le16(suballoc_bit_start);
- xblk->xb_fs_generation = cpu_to_le32(osb->fs_generation);
+ xblk->xb_fs_generation =
+ cpu_to_le32(OCFS2_SB(inode->i_sb)->fs_generation);
xblk->xb_blkno = cpu_to_le64(first_blkno);
if (indexed) {
struct ocfs2_xattr_tree_root *xr = &xblk->xb_attrs.xb_root;
ret = ocfs2_xa_set(&loc, xi, ctxt);
if (!ret)
xs->here = loc.xl_entry;
- else if (ret != -ENOSPC)
+ else if ((ret != -ENOSPC) || ctxt->set_abort)
goto end;
else {
ret = ocfs2_xattr_create_index_block(inode, xs, ctxt);
goto out;
}
- ret = ocfs2_extend_trans(ctxt->handle, credits +
- ctxt->handle->h_buffer_credits);
+ ret = ocfs2_extend_trans(ctxt->handle, credits);
if (ret) {
mlog_errno(ret);
goto out;
}
ret = ocfs2_xattr_block_set(inode, xi, xbs, ctxt);
- } else if (ret == -ENOSPC) {
+ } else if ((ret == -ENOSPC) && !ctxt->set_abort) {
if (di->i_xattr_loc && !xbs->xattr_bh) {
ret = ocfs2_xattr_block_find(inode,
xi->xi_name_index,
goto out;
}
- ret = ocfs2_extend_trans(ctxt->handle, credits +
- ctxt->handle->h_buffer_credits);
+ ret = ocfs2_extend_trans(ctxt->handle, credits);
if (ret) {
mlog_errno(ret);
goto out;
goto out;
}
- ret = ocfs2_extend_trans(ctxt->handle, credits +
- ctxt->handle->h_buffer_credits);
+ ret = ocfs2_extend_trans(ctxt->handle, credits);
if (ret) {
mlog_errno(ret);
goto out;
u32 bit_off, len;
u64 blkno;
handle_t *handle = ctxt->handle;
- struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
struct ocfs2_inode_info *oi = OCFS2_I(inode);
struct buffer_head *xb_bh = xs->xattr_bh;
struct ocfs2_xattr_block *xb =
goto out;
}
- ret = __ocfs2_claim_clusters(osb, handle, ctxt->data_ac,
+ ret = __ocfs2_claim_clusters(handle, ctxt->data_ac,
1, 1, &bit_off, &len);
if (ret) {
mlog_errno(ret);
* We need to update the first bucket of the old extent and all
* the buckets going to the new extent.
*/
- credits = ((num_buckets + 1) * blks_per_bucket) +
- handle->h_buffer_credits;
+ credits = ((num_buckets + 1) * blks_per_bucket);
ret = ocfs2_extend_trans(handle, credits);
if (ret) {
mlog_errno(ret);
u32 *first_hash)
{
u16 blk_per_bucket = ocfs2_blocks_per_xattr_bucket(inode->i_sb);
- int ret, credits = 2 * blk_per_bucket + handle->h_buffer_credits;
+ int ret, credits = 2 * blk_per_bucket;
BUG_ON(OCFS2_XATTR_BUCKET_SIZE < OCFS2_SB(inode->i_sb)->s_clustersize);
goto leave;
}
- ret = __ocfs2_claim_clusters(osb, handle, ctxt->data_ac, 1,
+ ret = __ocfs2_claim_clusters(handle, ctxt->data_ac, 1,
clusters_to_add, &bit_off, &num_bits);
if (ret < 0) {
if (ret != -ENOSPC)
goto leave;
}
- ret = ocfs2_journal_dirty(handle, root_bh);
- if (ret < 0)
- mlog_errno(ret);
+ ocfs2_journal_dirty(handle, root_bh);
leave:
return ret;
* existing bucket. Then we add the last existing bucket, the
* new bucket, and the first bucket (3 * blk_per_bucket).
*/
- credits = (end_blk - target_blk) + (3 * blk_per_bucket) +
- handle->h_buffer_credits;
+ credits = (end_blk - target_blk) + (3 * blk_per_bucket);
ret = ocfs2_extend_trans(handle, credits);
if (ret) {
mlog_errno(ret);
}
le32_add_cpu(&xb->xb_attrs.xb_root.xt_clusters, -len);
-
- ret = ocfs2_journal_dirty(handle, root_bh);
- if (ret) {
- mlog_errno(ret);
- goto out_commit;
- }
+ ocfs2_journal_dirty(handle, root_bh);
ret = ocfs2_truncate_log_append(osb, handle, blkno, len);
if (ret)
int indexed)
{
int ret;
- struct ocfs2_alloc_context *meta_ac;
struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
- struct ocfs2_xattr_set_ctxt ctxt = {
- .meta_ac = meta_ac,
- };
+ struct ocfs2_xattr_set_ctxt ctxt;
- ret = ocfs2_reserve_new_metadata_blocks(osb, 1, &meta_ac);
+ memset(&ctxt, 0, sizeof(ctxt));
+ ret = ocfs2_reserve_new_metadata_blocks(osb, 1, &ctxt.meta_ac);
if (ret < 0) {
mlog_errno(ret);
return ret;
ocfs2_commit_trans(osb, ctxt.handle);
out:
- ocfs2_free_alloc_context(meta_ac);
+ ocfs2_free_alloc_context(ctxt.meta_ac);
return ret;
}
goto out;
}
- ret = ocfs2_claim_clusters(osb, handle, data_ac,
+ ret = ocfs2_claim_clusters(handle, data_ac,
len, &p_cluster, &num_clusters);
if (ret) {
mlog_errno(ret);
xattr_ac, data_ac);
}
-struct xattr_handler ocfs2_xattr_security_handler = {
+const struct xattr_handler ocfs2_xattr_security_handler = {
.prefix = XATTR_SECURITY_PREFIX,
.list = ocfs2_xattr_security_list,
.get = ocfs2_xattr_security_get,
name, value, size, flags);
}
-struct xattr_handler ocfs2_xattr_trusted_handler = {
+const struct xattr_handler ocfs2_xattr_trusted_handler = {
.prefix = XATTR_TRUSTED_PREFIX,
.list = ocfs2_xattr_trusted_list,
.get = ocfs2_xattr_trusted_get,
name, value, size, flags);
}
-struct xattr_handler ocfs2_xattr_user_handler = {
+const struct xattr_handler ocfs2_xattr_user_handler = {
.prefix = XATTR_USER_PREFIX,
.list = ocfs2_xattr_user_list,
.get = ocfs2_xattr_user_get,