Thread (16 messages) 16 messages, 1 author, 2026-05-02
COLD34d REVIEWED: 10 (9M)

[PATCH v13 12/15] isofs: Implement fileattr_get for case sensitivity

From: Chuck Lever <cel@kernel.org>
Date: 2026-05-02 14:22:24
Also in: linux-cifs, linux-ext4, linux-f2fs-devel, linux-fsdevel, linux-nfs, linux-xfs
Subsystem: filesystems (vfs and infrastructure), isofs filesystem, the rest · Maintainers: Alexander Viro, Christian Brauner, Jan Kara, Linus Torvalds

From: Chuck Lever <chuck.lever@oracle.com>

Upper layers such as NFSD need a way to query whether a
filesystem handles filenames in a case-sensitive manner so
they can provide correct semantics to remote clients. Without
this information, NFS exports of ISO 9660 filesystems cannot
advertise their filename case behavior.

Implement isofs_fileattr_get() to report ISO 9660 case handling
behavior. The 'check=r' (relaxed) mount option enables
case-insensitive lookups and is reported via FS_XFLAG_CASEFOLD.
By default, Joliet extensions operate in relaxed mode while
plain ISO 9660 uses strict (case-sensitive) mode.

Plain ISO 9660 names on the medium are uppercase. When neither
Rock Ridge nor Joliet is in effect, the default 'map=n' option
(and 'map=a') routes lookup and readdir through
isofs_name_translate(), which forces A-Z to a-z. The names
visible to userspace then differ in case from the on-disc form,
so report FS_XFLAG_CASENONPRESERVING in that configuration. Rock
Ridge and Joliet both deliver names as authored, and 'map=o'
emits the raw on-disc name unchanged, so those configurations
remain case-preserving.

Casefolding is a directory property, and the in-tree consumers
(NFSD, ksmbd) issue the query against a directory: NFSD walks
to the parent for non-directory dentries before calling
vfs_fileattr_get(), and ksmbd reports per-share attributes from
the share root. Wire .fileattr_get only on
isofs_dir_inode_operations. The CASEFOLD flag is set in both
fa->fsx_xflags and fa->flags so FS_IOC_FSGETXATTR and
FS_IOC_GETFLAGS agree.

Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Roland Mainz <redacted>
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
---
 fs/isofs/dir.c   | 16 ++++++++++++++++
 fs/isofs/isofs.h |  3 +++
 2 files changed, 19 insertions(+)
diff --git a/fs/isofs/dir.c b/fs/isofs/dir.c
index 2fd9948d606e..55385a72a4ce 100644
--- a/fs/isofs/dir.c
+++ b/fs/isofs/dir.c
@@ -14,6 +14,7 @@
 #include <linux/gfp.h>
 #include <linux/filelock.h>
 #include "isofs.h"
+#include <linux/fileattr.h>
 
 int isofs_name_translate(struct iso_directory_record *de, char *new, struct inode *inode)
 {
@@ -267,6 +268,20 @@ static int isofs_readdir(struct file *file, struct dir_context *ctx)
 	return result;
 }
 
+int isofs_fileattr_get(struct dentry *dentry, struct file_kattr *fa)
+{
+	struct isofs_sb_info *sbi = ISOFS_SB(dentry->d_sb);
+
+	if (sbi->s_check == 'r') {
+		fa->fsx_xflags |= FS_XFLAG_CASEFOLD;
+		fa->flags |= FS_CASEFOLD_FL;
+	}
+	if (!sbi->s_joliet_level && !sbi->s_rock &&
+	    (sbi->s_mapping == 'n' || sbi->s_mapping == 'a'))
+		fa->fsx_xflags |= FS_XFLAG_CASENONPRESERVING;
+	return 0;
+}
+
 const struct file_operations isofs_dir_operations =
 {
 	.llseek = generic_file_llseek,
@@ -281,6 +296,7 @@ const struct file_operations isofs_dir_operations =
 const struct inode_operations isofs_dir_inode_operations =
 {
 	.lookup = isofs_lookup,
+	.fileattr_get = isofs_fileattr_get,
 };
 
 
diff --git a/fs/isofs/isofs.h b/fs/isofs/isofs.h
index 506555837533..0ec8b24a42ed 100644
--- a/fs/isofs/isofs.h
+++ b/fs/isofs/isofs.h
@@ -197,6 +197,9 @@ isofs_normalize_block_and_offset(struct iso_directory_record* de,
 	}
 }
 
+struct file_kattr;
+int isofs_fileattr_get(struct dentry *dentry, struct file_kattr *fa);
+
 extern const struct inode_operations isofs_dir_inode_operations;
 extern const struct file_operations isofs_dir_operations;
 extern const struct address_space_operations isofs_symlink_aops;
-- 
2.53.0
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help