Skip to contentSkip to conversation

smb/server: fix posix state check for directory rename

PATCH In review ChenXiaoSong 2 participants 5 messages Aug 2 – Aug 3

Subsystems (2)
lore
Share & tools
lore

ChenXiaoSong

author reviewer
#0 lore raw

cover smb/server: fix posix state check for directory rename

From: ChenXiaoSong <[email protected]>

ChenXiaoSong (2):
smb/server: pass source ksmbd_file to rename helpers smb/server: fix posix state check for directory rename

 fs/smb/server/smb2pdu.c   | 2 +-
 fs/smb/server/vfs.c       | 6 +++---
 fs/smb/server/vfs.h       | 4 ++--
 fs/smb/server/vfs_cache.c | 5 ++++-
 fs/smb/server/vfs_cache.h | 2 +-
 5 files changed, 11 insertions(+), 8 deletions(-)
Signature · 1 line
2.54.0
author reviewer
#1 lore raw

PATCH 1/2 smb/server: pass source ksmbd_file to rename helpers

From: ChenXiaoSong <[email protected]>

Pass the source ksmbd_file to ksmbd_vfs_rename() and ksmbd_has_open_files().

This is a preparatory change for the next patch. Keeping the interface change separate makes the functional fix easier to review.

No functional change.

---
 fs/smb/server/smb2pdu.c   | 2 +-
 fs/smb/server/vfs.c       | 5 +++--
 fs/smb/server/vfs.h       | 4 ++--
 fs/smb/server/vfs_cache.c | 3 ++-
 fs/smb/server/vfs_cache.h | 2 +-
 5 files changed, 9 insertions(+), 7 deletions(-)
fs/smb/server/smb2pdu.c +1 -1
index 2039a44d4b17..cb6c222b937b 100644
@@ -7690,7 +7690,7 @@ static int smb2_rename(struct ksmbd_work *work,
 		goto out;
 
 	smb_break_all_levII_oplock_rename(work, fp);
-	rc = ksmbd_vfs_rename(work, &fp->filp->f_path, new_name, flags);
+	rc = ksmbd_vfs_rename(work, fp, new_name, flags);
 out:
 	kfree(new_name);
 	return rc;
fs/smb/server/vfs.c +3 -2
index 8a80d6d8e02d..34975e18528c 100644
@@ -681,9 +681,10 @@ int ksmbd_vfs_check_rename_share(struct ksmbd_work *work,
 	return err;
 }
 
-int ksmbd_vfs_rename(struct ksmbd_work *work, const struct path *old_path,
+int ksmbd_vfs_rename(struct ksmbd_work *work, struct ksmbd_file *old_fp,
 			     char *newname, int flags)
 {
+	const struct path *old_path = &old_fp->filp->f_path;
 	struct dentry *old_child = old_path->dentry;
 	struct path new_path;
 	struct qstr new_last;
@@ -721,7 +722,7 @@ int ksmbd_vfs_rename(struct ksmbd_work *work, const struct path *old_path,
 		goto out_drop_write;
 
 	if (!work->tcon->posix_extensions && d_is_dir(old_child) &&
-	    ksmbd_has_open_files(old_child)) {
+	    ksmbd_has_open_files(old_fp)) {
 		err = -EACCES;
 		goto out3;
 	}
fs/smb/server/vfs.h +2 -2
index 1818b3f1971c..f922cd15ce3e 100644
@@ -88,8 +88,8 @@ int ksmbd_vfs_remove_file(struct ksmbd_work *work, const struct path *path);
 int ksmbd_vfs_link(struct ksmbd_work *work,
 		   const char *oldname, const char *newname);
 int ksmbd_vfs_getattr(const struct path *path, struct kstat *stat);
-int ksmbd_vfs_rename(struct ksmbd_work *work, const struct path *old_path,
-			     char *newname, int flags);
+int ksmbd_vfs_rename(struct ksmbd_work *work, struct ksmbd_file *old_fp,
+		     char *newname, int flags);
 int ksmbd_vfs_check_rename_share(struct ksmbd_work *work,
 				 const struct path *old_path);
 int ksmbd_vfs_truncate(struct ksmbd_work *work,
fs/smb/server/vfs_cache.c +2 -1
index a35df2ab59c9..c66584ed23ab 100644
@@ -1137,8 +1137,9 @@ struct ksmbd_file *ksmbd_lookup_fd_inode(struct dentry *dentry)
 	return NULL;
 }
 
-bool ksmbd_has_open_files(struct dentry *dentry)
+bool ksmbd_has_open_files(struct ksmbd_file *old_fp)
 {
+	struct dentry *dentry = old_fp->filp->f_path.dentry;
 	struct ksmbd_file *fp;
 	unsigned int id;
 	bool ret = false;
fs/smb/server/vfs_cache.h +1 -1
index d80f379d4e12..127ea4987e3f 100644
@@ -212,7 +212,7 @@ bool ksmbd_has_stream_without_delete_share(struct ksmbd_file *fp);
 int ksmbd_close_fd_app_instance_id(char *app_instance_id);
 struct ksmbd_file *ksmbd_lookup_fd_cguid(char *cguid);
 struct ksmbd_file *ksmbd_lookup_fd_inode(struct dentry *dentry);
-bool ksmbd_has_open_files(struct dentry *dentry);
+bool ksmbd_has_open_files(struct ksmbd_file *old_fp);
 unsigned int ksmbd_open_durable_fd(struct ksmbd_file *fp);
 struct ksmbd_file *ksmbd_open_fd(struct ksmbd_work *work, struct file *filp);
 void ksmbd_launch_ksmbd_durable_scavenger(void);
Signature · 1 line
2.54.0
author reviewer
#2 lore raw

PATCH 2/2 smb/server: fix posix state check for directory rename

In reply to ChenXiaoSong

From: ChenXiaoSong <[email protected]>

Reproducer:

1. server: systemctl start ksmbd
2. client: mount -t cifs //${server_ip}/export /mnt # without posix option
3. client: mkdir /mnt/dir1/; touch /mnt/dir1/file
4. client: tail -f /mnt/dir1/file # open file
5. client: mv /mnt/dir1 /mnt/dir2
rename succeeded, but it was expected to fail with "Permission denied"

See POSIX-FSA 2.1.3.1.

work->tcon->posix_extensions only records whether POSIX extensions were negotiated on the connection. It does not indicate that the handles were opened with POSIX create contexts.

Fixes c841bd3d8dec ("ksmbd: deny renaming directory with open children")
---
 fs/smb/server/vfs.c       | 3 +--
 fs/smb/server/vfs_cache.c | 2 ++
 2 files changed, 3 insertions(+), 2 deletions(-)
fs/smb/server/vfs.c +1 -2
index 34975e18528c..28940c7b5f83 100644
@@ -721,8 +721,7 @@ int ksmbd_vfs_rename(struct ksmbd_work *work, struct ksmbd_file *old_fp,
 	if (err)
 		goto out_drop_write;
 
-	if (!work->tcon->posix_extensions && d_is_dir(old_child) &&
-	    ksmbd_has_open_files(old_fp)) {
+	if (d_is_dir(old_child) && ksmbd_has_open_files(old_fp)) {
 		err = -EACCES;
 		goto out3;
 	}
fs/smb/server/vfs_cache.c +2 -0
index c66584ed23ab..eac9886eb6e3 100644
@@ -1152,6 +1152,8 @@ bool ksmbd_has_open_files(struct ksmbd_file *old_fp)
 			continue;
 		if (fp_dentry == dentry)
 			continue;
+		if (old_fp->is_posix_ctxt && fp->is_posix_ctxt)
+			continue;
 		if (is_subdir(fp_dentry, dentry)) {
 			ret = true;
 			break;
Signature · 1 line
2.54.0
maintainer
#3 lore raw
In reply to ChenXiaoSong
Quoted 6 lines · ChenXiaoSong

From: ChenXiaoSong <[email protected]>

ChenXiaoSong (2):
smb/server: pass source ksmbd_file to rename helpers smb/server: fix posix state check for directory rename

The argument-only change in 0002 patch has no standalone functional purpose, so there is no reason to keep it as a separate patch. Additionally, 0002 patch must be applied first, and applying only 0001 patch can result in a build failure. This is another source of error caused by unnecessarily splitting the change. Therefore, I squashed the two patches into a single patch and applied it to #ksmbd-for-next-next.
Thanks!

author reviewer
#4 lore raw

Agreed. Squashing them into a single patch is fine.

Quoted 8 lines · Namjae Jeon

The argument-only change in 0002 patch has no standalone functional purpose, so there is no reason to keep it as a separate patch. Additionally, 0002 patch must be applied first, and applying only 0001 patch can result in a build failure. This is another source of error caused by unnecessarily splitting the change. Therefore, I squashed the two patches into a single patch and applied it to #ksmbd-for-next-next.
Thanks!

Signature · 3 lines
ChenXiaoSong <[email protected]>
Chinese Homepage: https://chenxiaosong.com
English Homepage: https://chenxiaosong.com/en