From: Al Viro <viro@zeniv.linux.org.uk>
To: Mateusz Guzik <mjguzik@gmail.com>
Cc: brauner@kernel.org, jack@suse.cz, linux-kernel@vger.kernel.org,
linux-fsdevel@vger.kernel.org
Subject: [PATCH 2/3] lift grabbing path into caller of do_dentry_open()
Date: Thu, 22 Aug 2024 01:41:01 +0100 [thread overview]
Message-ID: <20240822004101.GQ504335@ZenIV> (raw)
In-Reply-To: <20240822003359.GO504335@ZenIV>
... and do that after the call. Legitimate, since no ->open()
instance tries to modify ->f_path. Never had been promised to
work, never had been done by any such instance, now it's clearly
forbidden.
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
Documentation/filesystems/porting.rst | 9 +++++++++
fs/open.c | 10 +++++++---
2 files changed, 16 insertions(+), 3 deletions(-)
diff --git a/Documentation/filesystems/porting.rst b/Documentation/filesystems/porting.rst
index 92bffcc6747a..34f83613ad6f 100644
--- a/Documentation/filesystems/porting.rst
+++ b/Documentation/filesystems/porting.rst
@@ -1141,3 +1141,12 @@ pointer are gone.
set_blocksize() takes opened struct file instead of struct block_device now
and it *must* be opened exclusive.
+
+
+---
+
+**mandatory**
+
+do not even think of modifying ->f_path in ->open() instance; it never had
+been expected to work and nobody had been insane enough to try it. Now
+it is explicitly forbidden.
diff --git a/fs/open.c b/fs/open.c
index 2bda3aadfa24..0ec2e9a33856 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -912,7 +912,6 @@ static int do_dentry_open(struct file *f,
struct inode *inode = f->f_path.dentry->d_inode;
int error;
- path_get(&f->f_path);
f->f_inode = inode;
f->f_mapping = inode->i_mapping;
f->f_wb_err = filemap_sample_wb_err(f->f_mapping);
@@ -1015,7 +1014,6 @@ static int do_dentry_open(struct file *f,
fops_put(f->f_op);
put_file_access(f);
cleanup_file:
- path_put(&f->f_path);
f->f_path.mnt = NULL;
f->f_path.dentry = NULL;
f->f_inode = NULL;
@@ -1042,10 +1040,14 @@ static int do_dentry_open(struct file *f,
int finish_open(struct file *file, struct dentry *dentry,
int (*open)(struct inode *, struct file *))
{
+ int err;
BUG_ON(file->f_mode & FMODE_OPENED); /* once it's opened, it's opened */
file->f_path.dentry = dentry;
- return do_dentry_open(file, open);
+ err = do_dentry_open(file, open);
+ if (file->f_mode & FMODE_OPENED)
+ path_get(&file->f_path);
+ return err;
}
EXPORT_SYMBOL(finish_open);
@@ -1095,6 +1097,8 @@ int vfs_open(const struct path *path, struct file *file)
*/
fsnotify_open(file);
}
+ if (file->f_mode & FMODE_OPENED)
+ path_get(&file->f_path);
return ret;
}
--
2.39.2
next prev parent reply other threads:[~2024-08-22 0:41 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-08-06 14:46 [PATCH] vfs: avoid spurious dentry ref/unref cycle on open Mateusz Guzik
2024-08-06 15:53 ` Al Viro
2024-08-06 16:09 ` Mateusz Guzik
2024-08-06 16:14 ` Mateusz Guzik
2024-08-07 3:38 ` Al Viro
2024-08-07 3:57 ` Mateusz Guzik
2024-08-07 5:32 ` Al Viro
2024-08-07 5:46 ` Mateusz Guzik
2024-08-07 6:23 ` Al Viro
2024-08-07 6:33 ` Al Viro
2024-08-07 6:40 ` Mateusz Guzik
2024-08-07 7:05 ` Al Viro
2024-08-07 7:22 ` Mateusz Guzik
2024-08-07 7:52 ` Al Viro
2024-08-07 7:59 ` Mateusz Guzik
2024-08-07 9:50 ` Mateusz Guzik
2024-08-07 12:43 ` Al Viro
2024-08-07 20:38 ` Al Viro
2024-08-20 11:38 ` Mateusz Guzik
2024-08-22 0:33 ` Al Viro
2024-08-22 0:34 ` [PATCH 1/3] don't duplicate vfs_open() in kernel_file_open() Al Viro
2024-08-22 7:53 ` Christian Brauner
2024-08-22 0:41 ` Al Viro [this message]
2024-08-22 7:54 ` [PATCH 2/3] lift grabbing path into caller of do_dentry_open() Christian Brauner
2024-08-22 0:41 ` [PATCH 3/3] avoid extra path_get/path_put cycle in path_openat() Al Viro
2024-08-22 9:31 ` Christian Brauner
2024-08-22 10:21 ` Mateusz Guzik
2025-02-17 8:03 ` [PATCH] vfs: avoid spurious dentry ref/unref cycle on open Mateusz Guzik
2024-08-08 6:26 ` Mateusz Guzik
2024-08-06 22:51 ` Dave Chinner
2024-08-06 22:55 ` Mateusz Guzik
2024-08-07 2:56 ` Dave Chinner
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=20240822004101.GQ504335@ZenIV \
--to=viro@zeniv.linux.org.uk \
--cc=brauner@kernel.org \
--cc=jack@suse.cz \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=mjguzik@gmail.com \
/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).