1 files changed, 161 insertions, 72 deletions

diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c
index ef8f08c..7903e62 100644
--- a/fs/ceph/dir.c
+++ b/fs/ceph/dir.c
@@ -40,14 +40,6 @@ int ceph_init_dentry(struct dentry *dentry)
 	if (dentry->d_fsdata)
 		return 0;
 
-	if (dentry->d_parent == NULL ||   /* nfs fh_to_dentry */
-	    ceph_snap(dentry->d_parent->d_inode) == CEPH_NOSNAP)
-		d_set_d_op(dentry, &ceph_dentry_ops);
-	else if (ceph_snap(dentry->d_parent->d_inode) == CEPH_SNAPDIR)
-		d_set_d_op(dentry, &ceph_snapdir_dentry_ops);
-	else
-		d_set_d_op(dentry, &ceph_snap_dentry_ops);
-
 	di = kmem_cache_alloc(ceph_dentry_cachep, GFP_NOFS | __GFP_ZERO);
 	if (!di)
 		return -ENOMEM;          /* oh well */
@@ -58,16 +50,42 @@ int ceph_init_dentry(struct dentry *dentry)
 		kmem_cache_free(ceph_dentry_cachep, di);
 		goto out_unlock;
 	}
+
+	if (dentry->d_parent == NULL ||   /* nfs fh_to_dentry */
+	    ceph_snap(dentry->d_parent->d_inode) == CEPH_NOSNAP)
+		d_set_d_op(dentry, &ceph_dentry_ops);
+	else if (ceph_snap(dentry->d_parent->d_inode) == CEPH_SNAPDIR)
+		d_set_d_op(dentry, &ceph_snapdir_dentry_ops);
+	else
+		d_set_d_op(dentry, &ceph_snap_dentry_ops);
+
 	di->dentry = dentry;
 	di->lease_session = NULL;
-	dentry->d_fsdata = di;
 	dentry->d_time = jiffies;
+	/* avoid reordering d_fsdata setup so that the check above is safe */
+	smp_mb();
+	dentry->d_fsdata = di;
 	ceph_dentry_lru_add(dentry);
 out_unlock:
 	spin_unlock(&dentry->d_lock);
 	return 0;
 }
 
+struct inode *ceph_get_dentry_parent_inode(struct dentry *dentry)
+{
+	struct inode *inode = NULL;
+
+	if (!dentry)
+		return NULL;
+
+	spin_lock(&dentry->d_lock);
+	if (dentry->d_parent) {
+		inode = dentry->d_parent->d_inode;
+		ihold(inode);
+	}
+	spin_unlock(&dentry->d_lock);
+	return inode;
+}
 
 
 /*
@@ -86,11 +104,11 @@ static unsigned fpos_off(loff_t p)
 /*
  * When possible, we try to satisfy a readdir by peeking at the
  * dcache.  We make this work by carefully ordering dentries on
- * d_u.d_child when we initially get results back from the MDS, and
+ * d_child when we initially get results back from the MDS, and
  * falling back to a "normal" sync readdir if any dentries in the dir
  * are dropped.
  *
- * I_COMPLETE tells indicates we have all dentries in the dir.  It is
+ * D_COMPLETE tells indicates we have all dentries in the dir.  It is
  * defined IFF we hold CEPH_CAP_FILE_SHARED (which will be revoked by
  * the MDS if/when the directory is modified).
  */
@@ -122,18 +140,18 @@ static int __dcache_readdir(struct file *filp,
 		p = parent->d_subdirs.prev;
 		dout(" initial p %p/%p\n", p->prev, p->next);
 	} else {
-		p = last->d_u.d_child.prev;
+		p = last->d_child.prev;
 	}
 
 more:
-	dentry = list_entry(p, struct dentry, d_u.d_child);
+	dentry = list_entry(p, struct dentry, d_child);
 	di = ceph_dentry(dentry);
 	while (1) {
 		dout(" p %p/%p %s d_subdirs %p/%p\n", p->prev, p->next,
 		     d_unhashed(dentry) ? "!hashed" : "hashed",
 		     parent->d_subdirs.prev, parent->d_subdirs.next);
 		if (p == &parent->d_subdirs) {
-			fi->at_end = 1;
+			fi->flags |= CEPH_F_ATEND;
 			goto out_unlock;
 		}
 		spin_lock_nested(&dentry->d_lock, DENTRY_D_LOCK_NESTED);
@@ -148,7 +166,7 @@ more:
 		     !dentry->d_inode ? " null" : "");
 		spin_unlock(&dentry->d_lock);
 		p = p->prev;
-		dentry = list_entry(p, struct dentry, d_u.d_child);
+		dentry = list_entry(p, struct dentry, d_child);
 		di = ceph_dentry(dentry);
 	}
 
@@ -181,8 +199,8 @@ more:
 	filp->f_pos++;
 
 	/* make sure a dentry wasn't dropped while we didn't have parent lock */
-	if (!ceph_i_test(dir, CEPH_I_COMPLETE)) {
-		dout(" lost I_COMPLETE on %p; falling back to mds\n", dir);
+	if (!ceph_dir_test_complete(dir)) {
+		dout(" lost D_COMPLETE on %p; falling back to mds\n", dir);
 		err = -EAGAIN;
 		goto out;
 	}
@@ -234,7 +252,7 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir)
 	const int max_bytes = fsc->mount_options->max_readdir_bytes;
 
 	dout("readdir %p filp %p frag %u off %u\n", inode, filp, frag, off);
-	if (fi->at_end)
+	if (fi->flags & CEPH_F_ATEND)
 		return 0;
 
 	/* always start with . and .. */
@@ -252,7 +270,7 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir)
 		off = 1;
 	}
 	if (filp->f_pos == 1) {
-		ino_t ino = filp->f_dentry->d_parent->d_inode->i_ino;
+		ino_t ino = parent_ino(filp->f_dentry);
 		dout("readdir off 1 -> '..'\n");
 		if (filldir(dirent, "..", 2, ceph_make_fpos(0, 1),
 			    ceph_translate_ino(inode->i_sb, ino),
@@ -263,18 +281,18 @@ static int ceph_readdir(struct file *filp, void *dirent, filldir_t filldir)
 	}
 
 	/* can we use the dcache? */
-	spin_lock(&inode->i_lock);
+	spin_lock(&ci->i_ceph_lock);
 	if ((filp->f_pos == 2 || fi->dentry) &&
 	    !ceph_test_mount_opt(fsc, NOASYNCREADDIR) &&
 	    ceph_snap(inode) != CEPH_SNAPDIR &&
-	    (ci->i_ceph_flags & CEPH_I_COMPLETE) &&
+	    ceph_dir_test_complete(inode) &&
 	    __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1)) {
-		spin_unlock(&inode->i_lock);
+		spin_unlock(&ci->i_ceph_lock);
 		err = __dcache_readdir(filp, dirent, filldir);
 		if (err != -EAGAIN)
 			return err;
 	} else {
-		spin_unlock(&inode->i_lock);
+		spin_unlock(&ci->i_ceph_lock);
 	}
 	if (fi->dentry) {
 		err = note_last_dentry(fi, fi->dentry->d_name.name,
@@ -333,7 +351,7 @@ more:
 
 		if (!req->r_did_prepopulate) {
 			dout("readdir !did_prepopulate");
-			fi->dir_release_count--;    /* preclude I_COMPLETE */
+			fi->dir_release_count--;    /* preclude D_COMPLETE */
 		}
 
 		/* note next offset and last dentry name */
@@ -403,20 +421,19 @@ more:
 		dout("readdir next frag is %x\n", frag);
 		goto more;
 	}
-	fi->at_end = 1;
+	fi->flags |= CEPH_F_ATEND;
 
 	/*
 	 * if dir_release_count still matches the dir, no dentries
 	 * were released during the whole readdir, and we should have
 	 * the complete dir contents in our cache.
 	 */
-	spin_lock(&inode->i_lock);
+	spin_lock(&ci->i_ceph_lock);
 	if (ci->i_release_count == fi->dir_release_count) {
-		dout(" marking %p complete\n", inode);
-		/* ci->i_ceph_flags |= CEPH_I_COMPLETE; */
+		ceph_dir_set_complete(inode);
 		ci->i_max_offset = filp->f_pos;
 	}
-	spin_unlock(&inode->i_lock);
+	spin_unlock(&ci->i_ceph_lock);
 
 	dout("readdir %p filp %p done.\n", inode, filp);
 	return 0;
@@ -435,7 +452,7 @@ static void reset_readdir(struct ceph_file_info *fi)
 		dput(fi->dentry);
 		fi->dentry = NULL;
 	}
-	fi->at_end = 0;
+	fi->flags &= ~CEPH_F_ATEND;
 }
 
 static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int origin)
@@ -446,19 +463,24 @@ static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int origin)
 	loff_t retval;
 
 	mutex_lock(&inode->i_mutex);
+	retval = -EINVAL;
 	switch (origin) {
 	case SEEK_END:
 		offset += inode->i_size + 2;   /* FIXME */
 		break;
 	case SEEK_CUR:
 		offset += file->f_pos;
+	case SEEK_SET:
+		break;
+	default:
+		goto out;
 	}
-	retval = -EINVAL;
+
 	if (offset >= 0 && offset <= inode->i_sb->s_maxbytes) {
 		if (offset != file->f_pos) {
 			file->f_pos = offset;
 			file->f_version = 0;
-			fi->at_end = 0;
+			fi->flags &= ~CEPH_F_ATEND;
 		}
 		retval = offset;
 
@@ -477,26 +499,19 @@ static loff_t ceph_dir_llseek(struct file *file, loff_t offset, int origin)
 		if (offset > old_offset)
 			fi->dir_release_count--;
 	}
+out:
 	mutex_unlock(&inode->i_mutex);
 	return retval;
 }
 
 /*
- * Process result of a lookup/open request.
- *
- * Mainly, make sure we return the final req->r_dentry (if it already
- * existed) in place of the original VFS-provided dentry when they
- * differ.
- *
- * Gracefully handle the case where the MDS replies with -ENOENT and
- * no trace (which it may do, at its discretion, e.g., if it doesn't
- * care to issue a lease on the negative dentry).
+ * Handle lookups for the hidden .snap directory.
  */
-struct dentry *ceph_finish_lookup(struct ceph_mds_request *req,
-				  struct dentry *dentry, int err)
+int ceph_handle_snapdir(struct ceph_mds_request *req,
+			struct dentry *dentry, int err)
 {
 	struct ceph_fs_client *fsc = ceph_sb_to_client(dentry->d_sb);
-	struct inode *parent = dentry->d_parent->d_inode;
+	struct inode *parent = dentry->d_parent->d_inode; /* we hold i_mutex */
 
 	/* .snap dir? */
 	if (err == -ENOENT &&
@@ -510,7 +525,23 @@ struct dentry *ceph_finish_lookup(struct ceph_mds_request *req,
 		d_add(dentry, inode);
 		err = 0;
 	}
+	return err;
+}
 
+/*
+ * Figure out final result of a lookup/open request.
+ *
+ * Mainly, make sure we return the final req->r_dentry (if it already
+ * existed) in place of the original VFS-provided dentry when they
+ * differ.
+ *
+ * Gracefully handle the case where the MDS replies with -ENOENT and
+ * no trace (which it may do, at its discretion, e.g., if it doesn't
+ * care to issue a lease on the negative dentry).
+ */
+struct dentry *ceph_finish_lookup(struct ceph_mds_request *req,
+				  struct dentry *dentry, int err)
+{
 	if (err == -ENOENT) {
 		/* no trace? */
 		err = 0;
@@ -566,7 +597,6 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry,
 	/* open (but not create!) intent? */
 	if (nd &&
 	    (nd->flags & LOOKUP_OPEN) &&
-	    (nd->flags & LOOKUP_CONTINUE) == 0 && /* only open last component */
 	    !(nd->intent.open.flags & O_CREAT)) {
 		int mode = nd->intent.open.create_mode & ~current->fs->umask;
 		return ceph_lookup_open(dir, dentry, nd, mode, 1);
@@ -577,21 +607,21 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry,
 		struct ceph_inode_info *ci = ceph_inode(dir);
 		struct ceph_dentry_info *di = ceph_dentry(dentry);
 
-		spin_lock(&dir->i_lock);
+		spin_lock(&ci->i_ceph_lock);
 		dout(" dir %p flags are %d\n", dir, ci->i_ceph_flags);
 		if (strncmp(dentry->d_name.name,
 			    fsc->mount_options->snapdir_name,
 			    dentry->d_name.len) &&
 		    !is_root_ceph_dentry(dir, dentry) &&
-		    (ci->i_ceph_flags & CEPH_I_COMPLETE) &&
+		    ceph_dir_test_complete(dir) &&
 		    (__ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1))) {
-			spin_unlock(&dir->i_lock);
+			spin_unlock(&ci->i_ceph_lock);
 			dout(" dir %p complete, -ENOENT\n", dir);
 			d_add(dentry, NULL);
 			di->lease_shared_gen = ci->i_shared_gen;
 			return NULL;
 		}
-		spin_unlock(&dir->i_lock);
+		spin_unlock(&ci->i_ceph_lock);
 	}
 
 	op = ceph_snap(dir) == CEPH_SNAPDIR ?
@@ -605,6 +635,7 @@ static struct dentry *ceph_lookup(struct inode *dir, struct dentry *dentry,
 	req->r_args.getattr.mask = cpu_to_le32(CEPH_STAT_CAP_INODE);
 	req->r_locked_dir = dir;
 	err = ceph_mdsc_do_request(mdsc, NULL, req);
+	err = ceph_handle_snapdir(req, dentry, err);
 	dentry = ceph_finish_lookup(req, dentry, err);
 	ceph_mdsc_put_request(req);  /* will dput(dentry) */
 	dout("lookup result=%p\n", dentry);
@@ -784,6 +815,7 @@ static int ceph_link(struct dentry *old_dentry, struct inode *dir,
 	req->r_dentry = dget(dentry);
 	req->r_num_caps = 2;
 	req->r_old_dentry = dget(old_dentry); /* or inode? hrm. */
+	req->r_old_dentry_dir = ceph_get_dentry_parent_inode(old_dentry);
 	req->r_locked_dir = dir;
 	req->r_dentry_drop = CEPH_CAP_FILE_SHARED;
 	req->r_dentry_unless = CEPH_CAP_FILE_EXCL;
@@ -809,12 +841,12 @@ static int drop_caps_for_unlink(struct inode *inode)
 	struct ceph_inode_info *ci = ceph_inode(inode);
 	int drop = CEPH_CAP_LINK_SHARED | CEPH_CAP_LINK_EXCL;
 
-	spin_lock(&inode->i_lock);
+	spin_lock(&ci->i_ceph_lock);
 	if (inode->i_nlink == 1) {
 		drop |= ~(__ceph_caps_wanted(ci) | CEPH_CAP_PIN);
 		ci->i_ceph_flags |= CEPH_I_NODELAY;
 	}
-	spin_unlock(&inode->i_lock);
+	spin_unlock(&ci->i_ceph_lock);
 	return drop;
 }
 
@@ -882,6 +914,7 @@ static int ceph_rename(struct inode *old_dir, struct dentry *old_dentry,
 	req->r_dentry = dget(new_dentry);
 	req->r_num_caps = 2;
 	req->r_old_dentry = dget(old_dentry);
+	req->r_old_dentry_dir = ceph_get_dentry_parent_inode(old_dentry);
 	req->r_locked_dir = new_dir;
 	req->r_old_dentry_drop = CEPH_CAP_FILE_SHARED;
 	req->r_old_dentry_unless = CEPH_CAP_FILE_EXCL;
@@ -900,7 +933,7 @@ static int ceph_rename(struct inode *old_dir, struct dentry *old_dentry,
 		 */
 
 		/* d_move screws up d_subdirs order */
-		ceph_i_clear(new_dir, CEPH_I_COMPLETE);
+		ceph_dir_clear_complete(new_dir);
 
 		d_move(old_dentry, new_dentry);
 
@@ -982,10 +1015,10 @@ static int dir_lease_is_valid(struct inode *dir, struct dentry *dentry)
 	struct ceph_dentry_info *di = ceph_dentry(dentry);
 	int valid = 0;
 
-	spin_lock(&dir->i_lock);
+	spin_lock(&ci->i_ceph_lock);
 	if (ci->i_shared_gen == di->lease_shared_gen)
 		valid = __ceph_caps_issued_mask(ci, CEPH_CAP_FILE_SHARED, 1);
-	spin_unlock(&dir->i_lock);
+	spin_unlock(&ci->i_ceph_lock);
 	dout("dir_lease_is_valid dir %p v%u dentry %p v%u = %d\n",
 	     dir, (unsigned)ci->i_shared_gen, dentry,
 	     (unsigned)di->lease_shared_gen, valid);
@@ -997,36 +1030,38 @@ static int dir_lease_is_valid(struct inode *dir, struct dentry *dentry)
  */
 static int ceph_d_revalidate(struct dentry *dentry, struct nameidata *nd)
 {
+	int valid = 0;
 	struct inode *dir;
 
 	if (nd && nd->flags & LOOKUP_RCU)
 		return -ECHILD;
 
-	dir = dentry->d_parent->d_inode;
-
 	dout("d_revalidate %p '%.*s' inode %p offset %lld\n", dentry,
 	     dentry->d_name.len, dentry->d_name.name, dentry->d_inode,
 	     ceph_dentry(dentry)->offset);
 
+	dir = ceph_get_dentry_parent_inode(dentry);
+
 	/* always trust cached snapped dentries, snapdir dentry */
 	if (ceph_snap(dir) != CEPH_NOSNAP) {
 		dout("d_revalidate %p '%.*s' inode %p is SNAPPED\n", dentry,
 		     dentry->d_name.len, dentry->d_name.name, dentry->d_inode);
-		goto out_touch;
+		valid = 1;
+	} else if (dentry->d_inode &&
+		   ceph_snap(dentry->d_inode) == CEPH_SNAPDIR) {
+		valid = 1;
+	} else if (dentry_lease_is_valid(dentry) ||
+		   dir_lease_is_valid(dir, dentry)) {
+		valid = 1;
 	}
-	if (dentry->d_inode && ceph_snap(dentry->d_inode) == CEPH_SNAPDIR)
-		goto out_touch;
 
-	if (dentry_lease_is_valid(dentry) ||
-	    dir_lease_is_valid(dir, dentry))
-		goto out_touch;
-
-	dout("d_revalidate %p invalid\n", dentry);
-	d_drop(dentry);
-	return 0;
-out_touch:
-	ceph_dentry_lru_touch(dentry);
-	return 1;
+	dout("d_revalidate %p %s\n", dentry, valid ? "valid" : "invalid");
+	if (valid)
+		ceph_dentry_lru_touch(dentry);
+	else
+		d_drop(dentry);
+	iput(dir);
+	return valid;
 }
 
 /*
@@ -1056,7 +1091,52 @@ static int ceph_snapdir_d_revalidate(struct dentry *dentry,
 	return 1;
 }
 
+/*
+ * Set/clear/test dir complete flag on the dir's dentry.
+ */
+void ceph_dir_set_complete(struct inode *inode)
+{
+	/* not yet implemented */
+}
+
+void ceph_dir_clear_complete(struct inode *inode)
+{
+	/* not yet implemented */
+}
 
+bool ceph_dir_test_complete(struct inode *inode)
+{
+	/* not yet implemented */
+	return false;
+}
+
+/*
+ * When the VFS prunes a dentry from the cache, we need to clear the
+ * complete flag on the parent directory.
+ *
+ * Called under dentry->d_lock.
+ */
+static void ceph_d_prune(struct dentry *dentry)
+{
+	struct ceph_dentry_info *di;
+
+	dout("ceph_d_prune %p\n", dentry);
+
+	/* do we have a valid parent? */
+	if (!dentry->d_parent || IS_ROOT(dentry))
+		return;
+
+	/* if we are not hashed, we don't affect D_COMPLETE */
+	if (d_unhashed(dentry))
+		return;
+
+	/*
+	 * we hold d_lock, so d_parent is stable, and d_fsdata is never
+	 * cleared until d_release
+	 */
+	di = ceph_dentry(dentry->d_parent);
+	clear_bit(CEPH_D_COMPLETE, &di->flags);
+}
 
 /*
  * read() on a dir.  This weird interface hack only works if mounted
@@ -1113,7 +1193,8 @@ static ssize_t ceph_read_dir(struct file *file, char __user *buf, size_t size,
  * an fsync() on a dir will wait for any uncommitted directory
  * operations to commit.
  */
-static int ceph_dir_fsync(struct file *file, int datasync)
+static int ceph_dir_fsync(struct file *file, loff_t start, loff_t end,
+			  int datasync)
 {
 	struct inode *inode = file->f_path.dentry->d_inode;
 	struct ceph_inode_info *ci = ceph_inode(inode);
@@ -1123,6 +1204,11 @@ static int ceph_dir_fsync(struct file *file, int datasync)
 	int ret = 0;
 
 	dout("dir_fsync %p\n", inode);
+	ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
+	if (ret)
+		return ret;
+	mutex_lock(&inode->i_mutex);
+
 	spin_lock(&ci->i_unsafe_lock);
 	if (list_empty(head))
 		goto out;
@@ -1156,6 +1242,8 @@ static int ceph_dir_fsync(struct file *file, int datasync)
 	} while (req->r_tid < last_tid);
 out:
 	spin_unlock(&ci->i_unsafe_lock);
+	mutex_unlock(&inode->i_mutex);
+
 	return ret;
 }
 
@@ -1215,9 +1303,8 @@ void ceph_dentry_lru_del(struct dentry *dn)
  * Return name hash for a given dentry.  This is dependent on
  * the parent directory's hash function.
  */
-unsigned ceph_dentry_hash(struct dentry *dn)
+unsigned ceph_dentry_hash(struct inode *dir, struct dentry *dn)
 {
-	struct inode *dir = dn->d_parent->d_inode;
 	struct ceph_inode_info *dci = ceph_inode(dir);
 
 	switch (dci->i_dir_layout.dl_dir_hash) {
@@ -1263,6 +1350,7 @@ const struct inode_operations ceph_dir_iops = {
 const struct dentry_operations ceph_dentry_ops = {
 	.d_revalidate = ceph_d_revalidate,
 	.d_release = ceph_d_release,
+	.d_prune = ceph_d_prune,
 };
 
 const struct dentry_operations ceph_snapdir_dentry_ops = {
@@ -1272,4 +1360,5 @@ const struct dentry_operations ceph_snapdir_dentry_ops = {
 
 const struct dentry_operations ceph_snap_dentry_ops = {
 	.d_release = ceph_d_release,
+	.d_prune = ceph_d_prune,
 };