fscrypt: remove fscrypt_operations::max_namelen

The max_namelen field is unnecessary, as it is set to 255 (NAME_MAX) on
all filesystems that support fscrypt (or plan to support fscrypt).  For
simplicity, just use NAME_MAX directly instead.

Link: https://lore.kernel.org/r/20210909184513.139281-1-ebiggers@kernel.org
Signed-off-by: Eric Biggers <ebiggers@google.com>

fs/crypto/fname.c

@@ -429,8 +429,7 @@ int fscrypt_setup_filename(struct inode *dir, const struct qstr *iname,
 
 	if (fscrypt_has_encryption_key(dir)) {
 		if (!fscrypt_fname_encrypted_size(&dir->i_crypt_info->ci_policy,
-						  iname->len,
-						  dir->i_sb->s_cop->max_namelen,
+						  iname->len, NAME_MAX,
 						  &fname->crypto_buf.len))
 			return -ENAMETOOLONG;
 		fname->crypto_buf.name = kmalloc(fname->crypto_buf.len,

fs/ext4/super.c

@@ -1566,7 +1566,6 @@ static const struct fscrypt_operations ext4_cryptops = {
 	.set_context		= ext4_set_context,
 	.get_dummy_policy	= ext4_get_dummy_policy,
 	.empty_dir		= ext4_empty_dir,
-	.max_namelen		= EXT4_NAME_LEN,
 	.has_stable_inodes	= ext4_has_stable_inodes,
 	.get_ino_and_lblk_bits	= ext4_get_ino_and_lblk_bits,
 };

fs/f2fs/super.c

@@ -2976,7 +2976,6 @@ static const struct fscrypt_operations f2fs_cryptops = {
 	.set_context		= f2fs_set_context,
 	.get_dummy_policy	= f2fs_get_dummy_policy,
 	.empty_dir		= f2fs_empty_dir,
-	.max_namelen		= F2FS_NAME_LEN,
 	.has_stable_inodes	= f2fs_has_stable_inodes,
 	.get_ino_and_lblk_bits	= f2fs_get_ino_and_lblk_bits,
 	.get_num_devices	= f2fs_get_num_devices,

fs/ubifs/crypto.c

@@ -82,5 +82,4 @@ const struct fscrypt_operations ubifs_crypt_operations = {
 	.get_context		= ubifs_crypt_get_context,
 	.set_context		= ubifs_crypt_set_context,
 	.empty_dir		= ubifs_crypt_empty_dir,
-	.max_namelen		= UBIFS_MAX_NLEN,
 };

include/linux/fscrypt.h

@@ -118,9 +118,6 @@ struct fscrypt_operations {
 	 */
 	bool (*empty_dir)(struct inode *inode);
 
-	/* The filesystem's maximum ciphertext filename length, in bytes */
-	unsigned int max_namelen;
-
 	/*
 	 * Check whether the filesystem's inode numbers and UUID are stable,
 	 * meaning that they will never be changed even by offline operations