From: Mateusz Guzik <mjguzik@gmail.com>
To: brauner@kernel.org
Cc: viro@zeniv.linux.org.uk, jack@suse.cz,
linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org,
josef@toxicpanda.com, kernel-team@fb.com, amir73il@gmail.com,
linux-btrfs@vger.kernel.org, linux-ext4@vger.kernel.org,
linux-xfs@vger.kernel.org, ceph-devel@vger.kernel.org,
linux-unionfs@vger.kernel.org, Mateusz Guzik <mjguzik@gmail.com>
Subject: [PATCH v4 12/12] fs: make plain ->i_state access fail to compile
Date: Tue, 16 Sep 2025 15:59:00 +0200 [thread overview]
Message-ID: <20250916135900.2170346-13-mjguzik@gmail.com> (raw)
In-Reply-To: <20250916135900.2170346-1-mjguzik@gmail.com>
... to make sure all accesses are properly validated.
Merely renaming the var to __i_state still lets the compiler make the
following suggestion:
error: 'struct inode' has no member named 'i_state'; did you mean '__i_state'?
Unfortunately some people will add the __'s and call it a day.
In order to make it harder to mess up in this way, hide it behind a
struct. The resulting error message should be convincing in terms of
checking what to do:
error: invalid operands to binary & (have 'struct inode_state_flags' and 'int')
Of course people determined to do a plain access can still do it, but
nothing can be done for that case.
Signed-off-by: Mateusz Guzik <mjguzik@gmail.com>
---
include/linux/fs.h | 19 +++++++++++++------
1 file changed, 13 insertions(+), 6 deletions(-)
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 11eef4ef5ace..80c53af7bc5a 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -782,6 +782,13 @@ enum inode_state_flags_enum {
#define I_DIRTY (I_DIRTY_INODE | I_DIRTY_PAGES)
#define I_DIRTY_ALL (I_DIRTY | I_DIRTY_TIME)
+/*
+ * Use inode_state_read() & friends to access.
+ */
+struct inode_state_flags {
+ enum inode_state_flags_enum __state;
+};
+
/*
* Keep mostly read-only and often accessed (especially for
* the RCU path lookup and 'stat' data) fields at the beginning
@@ -840,7 +847,7 @@ struct inode {
#endif
/* Misc */
- enum inode_state_flags_enum i_state;
+ struct inode_state_flags i_state;
/* 32-bit hole */
struct rw_semaphore i_rwsem;
@@ -906,19 +913,19 @@ struct inode {
*/
static inline enum inode_state_flags_enum inode_state_read_once(struct inode *inode)
{
- return READ_ONCE(inode->i_state);
+ return READ_ONCE(inode->i_state.__state);
}
static inline enum inode_state_flags_enum inode_state_read(struct inode *inode)
{
lockdep_assert_held(&inode->i_lock);
- return inode->i_state;
+ return inode->i_state.__state;
}
static inline void inode_state_add_raw(struct inode *inode,
enum inode_state_flags_enum addflags)
{
- WRITE_ONCE(inode->i_state, inode->i_state | addflags);
+ WRITE_ONCE(inode->i_state.__state, inode->i_state.__state | addflags);
}
static inline void inode_state_add(struct inode *inode,
@@ -931,7 +938,7 @@ static inline void inode_state_add(struct inode *inode,
static inline void inode_state_del_raw(struct inode *inode,
enum inode_state_flags_enum delflags)
{
- WRITE_ONCE(inode->i_state, inode->i_state & ~delflags);
+ WRITE_ONCE(inode->i_state.__state, inode->i_state.__state & ~delflags);
}
static inline void inode_state_del(struct inode *inode,
@@ -944,7 +951,7 @@ static inline void inode_state_del(struct inode *inode,
static inline void inode_state_set_raw(struct inode *inode,
enum inode_state_flags_enum setflags)
{
- WRITE_ONCE(inode->i_state, setflags);
+ WRITE_ONCE(inode->i_state.__state, setflags);
}
static inline void inode_state_set(struct inode *inode,
--
2.43.0
next prev parent reply other threads:[~2025-09-16 14:00 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-09-16 13:58 [PATCH v4 00/12] hide ->i_state behind accessors Mateusz Guzik
2025-09-16 13:58 ` [PATCH v4 01/12] fs: provide accessors for ->i_state Mateusz Guzik
2025-09-16 13:58 ` [PATCH v4 02/12] fs: use ->i_state accessors in core kernel Mateusz Guzik
2025-09-16 13:58 ` [PATCH v4 03/12] fs: mechanically convert most filesystems to use ->i_state accessors Mateusz Guzik
2025-09-16 13:58 ` [PATCH v4 04/12] btrfs: use the new " Mateusz Guzik
2025-09-16 13:58 ` [PATCH v4 05/12] netfs: " Mateusz Guzik
2025-09-16 13:58 ` [PATCH v4 06/12] nilfs2: " Mateusz Guzik
2025-09-16 13:58 ` [PATCH v4 07/12] xfs: " Mateusz Guzik
2025-09-16 13:58 ` [PATCH v4 08/12] ext4: " Mateusz Guzik
2025-09-16 13:58 ` [PATCH v4 09/12] f2fs: " Mateusz Guzik
2025-09-16 13:58 ` [PATCH v4 10/12] ceph: " Mateusz Guzik
2025-09-16 17:36 ` Viacheslav Dubeyko
2025-09-16 17:38 ` Mateusz Guzik
2025-09-16 13:58 ` [PATCH v4 11/12] overlayfs: " Mateusz Guzik
2025-09-16 13:59 ` Mateusz Guzik [this message]
2025-09-19 12:19 ` [PATCH v4 00/12] hide ->i_state behind accessors Christian Brauner
2025-09-19 13:09 ` Mateusz Guzik
2025-09-19 13:39 ` Mateusz Guzik
2025-09-19 15:51 ` Mateusz Guzik
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20250916135900.2170346-13-mjguzik@gmail.com \
--to=mjguzik@gmail.com \
--cc=amir73il@gmail.com \
--cc=brauner@kernel.org \
--cc=ceph-devel@vger.kernel.org \
--cc=jack@suse.cz \
--cc=josef@toxicpanda.com \
--cc=kernel-team@fb.com \
--cc=linux-btrfs@vger.kernel.org \
--cc=linux-ext4@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-unionfs@vger.kernel.org \
--cc=linux-xfs@vger.kernel.org \
--cc=viro@zeniv.linux.org.uk \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).