[PATCH 5/5] fscrypt: remove unnecessary calls to fscrypt_require_key()

[Eric Biggers]

From: Eric Biggers <ebiggers at google.com>

In an encrypted directory, a regular dentry (one that doesn't have the
no-key name flag) can only be created if the directory's encryption key
is available.

Therefore the calls to fscrypt_require_key() in __fscrypt_prepare_link()
and __fscrypt_prepare_rename() are unnecessary, as these functions
already check that the dentries they're given aren't no-key names.

Remove these unnecessary calls to fscrypt_require_key().

Signed-off-by: Eric Biggers <ebiggers at google.com>
---
 fs/crypto/hooks.c       | 26 ++++++++------------------
 include/linux/fscrypt.h |  3 +--
 2 files changed, 9 insertions(+), 20 deletions(-)

diff --git a/fs/crypto/hooks.c b/fs/crypto/hooks.c
index 061418be4b08..c582e2ddb39c 100644
--- a/fs/crypto/hooks.c
+++ b/fs/crypto/hooks.c
@@ -54,15 +54,12 @@ EXPORT_SYMBOL_GPL(fscrypt_file_open);
 int __fscrypt_prepare_link(struct inode *inode, struct inode *dir,
 			   struct dentry *dentry)
 {
-	int err;
-
-	err = fscrypt_require_key(dir);
-	if (err)
-		return err;
-
-	/* ... in case we looked up no-key name before key was added */
 	if (fscrypt_is_nokey_name(dentry))
 		return -ENOKEY;
+	/*
+	 * We don't need to separately check that the directory inode's key is
+	 * available, as it's implied by the dentry not being a no-key name.
+	 */
 
 	if (!fscrypt_has_permitted_context(dir, inode))
 		return -EXDEV;
@@ -75,20 +72,13 @@ int __fscrypt_prepare_rename(struct inode *old_dir, struct dentry *old_dentry,
 			     struct inode *new_dir, struct dentry *new_dentry,
 			     unsigned int flags)
 {
-	int err;
-
-	err = fscrypt_require_key(old_dir);
-	if (err)
-		return err;
-
-	err = fscrypt_require_key(new_dir);
-	if (err)
-		return err;
-
-	/* ... in case we looked up no-key name(s) before key was added */
 	if (fscrypt_is_nokey_name(old_dentry) ||
 	    fscrypt_is_nokey_name(new_dentry))
 		return -ENOKEY;
+	/*
+	 * We don't need to separately check that the directory inodes' keys are
+	 * available, as it's implied by the dentries not being no-key names.
+	 */
 
 	if (old_dir != new_dir) {
 		if (IS_ENCRYPTED(new_dir) &&
diff --git a/include/linux/fscrypt.h b/include/linux/fscrypt.h
index 8e1d31c959bf..0c9e64969b73 100644
--- a/include/linux/fscrypt.h
+++ b/include/linux/fscrypt.h
@@ -710,8 +710,7 @@ static inline int fscrypt_require_key(struct inode *inode)
  *
  * A new link can only be added to an encrypted directory if the directory's
  * encryption key is available --- since otherwise we'd have no way to encrypt
- * the filename.  Therefore, we first set up the directory's encryption key (if
- * not already done) and return an error if it's unavailable.
+ * the filename.
  *
  * We also verify that the link will not violate the constraint that all files
  * in an encrypted directory tree use the same encryption policy.
-- 
2.29.2