[v4,21/40] open: handle idmapped mounts in do_truncate()

Commit Message

Christian Brauner Dec. 3, 2020, 11:57 p.m. UTC

When truncating files the vfs will verify that the caller is privileged
over the inode. Extend it to handle idmapped mounts. If the inode is
accessed through an idmapped mount it is mapped according to the mount's
user namespace. Afterwards the permissions checks are identical to
non-idmapped mounts. If the initial user namespace is passed nothing
changes so non-idmapped mounts will see identical behavior as before.

Cc: Christoph Hellwig <[e-mail]>
Cc: David Howells <[e-mail]>
Cc: Al Viro <[e-mail]>
Cc: [e-mail]
Signed-off-by: Christian Brauner <[e-mail]>
---
/* v2 */
unchanged

/* v3 */
unchanged

/* v4 */
- Serge Hallyn <[e-mail]>:
  - Use "mnt_userns" to refer to a vfsmount's userns everywhere to make
    terminology consistent.
---
 fs/coredump.c      | 14 ++++++++++----
 fs/inode.c         | 13 +++++++++----
 fs/namei.c         |  6 +++---
 fs/open.c          | 21 +++++++++++++--------
 include/linux/fs.h |  4 ++--
 5 files changed, 37 insertions(+), 21 deletions(-)

Patch

diff --git a/fs/coredump.c b/fs/coredump.c
index 0cd9056d79cc..478add03c554 100644
--- a/fs/coredump.c
+++ b/fs/coredump.c
@@ -703,6 +703,7 @@  void do_coredump(const kernel_siginfo_t *siginfo)
 			goto close_fail;
 		}
 	} else {
+		struct user_namespace *mnt_userns;
 		struct inode *inode;
 		int open_flags = O_CREAT | O_RDWR | O_NOFOLLOW |
 				 O_LARGEFILE | O_EXCL;
@@ -780,13 +781,14 @@  void do_coredump(const kernel_siginfo_t *siginfo)
 		 * a process dumps core while its cwd is e.g. on a vfat
 		 * filesystem.
 		 */
-		if (!uid_eq(inode->i_uid, current_fsuid()))
+		mnt_userns = mnt_user_ns(cprm.file->f_path.mnt);
+		if (!uid_eq(i_uid_into_mnt(mnt_userns, inode), current_fsuid()))
 			goto close_fail;
 		if ((inode->i_mode & 0677) != 0600)
 			goto close_fail;
 		if (!(cprm.file->f_mode & FMODE_CAN_WRITE))
 			goto close_fail;
-		if (do_truncate(cprm.file->f_path.dentry, 0, 0, cprm.file))
+		if (do_truncate(mnt_userns, cprm.file->f_path.dentry, 0, 0, cprm.file))
 			goto close_fail;
 	}
 
@@ -931,8 +933,12 @@  void dump_truncate(struct coredump_params *cprm)
 
 	if (file->f_op->llseek && file->f_op->llseek != no_llseek) {
 		offset = file->f_op->llseek(file, 0, SEEK_CUR);
-		if (i_size_read(file->f_mapping->host) < offset)
-			do_truncate(file->f_path.dentry, offset, 0, file);
+		if (i_size_read(file->f_mapping->host) < offset) {
+			struct user_namespace *mnt_userns;
+
+			mnt_userns = mnt_user_ns(file->f_path.mnt);
+			do_truncate(mnt_userns, file->f_path.dentry, offset, 0, file);
+		}
 	}
 }
 EXPORT_SYMBOL(dump_truncate);
diff --git a/fs/inode.c b/fs/inode.c
index 159b55538347..d7a6842cf2ad 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -1904,7 +1904,8 @@  int dentry_needs_remove_privs(struct dentry *dentry)
 	return mask;
 }
 
-static int __remove_privs(struct dentry *dentry, int kill)
+static int __remove_privs(struct user_namespace *mnt_userns, struct dentry *dentry,
+			  int kill)
 {
 	struct iattr newattrs;
 
@@ -1913,7 +1914,7 @@  static int __remove_privs(struct dentry *dentry, int kill)
 	 * Note we call this on write, so notify_change will not
 	 * encounter any conflicting delegations:
 	 */
-	return notify_change(&init_user_ns, dentry, &newattrs, NULL);
+	return notify_change(mnt_userns, dentry, &newattrs, NULL);
 }
 
 /*
@@ -1939,8 +1940,12 @@  int file_remove_privs(struct file *file)
 	kill = dentry_needs_remove_privs(dentry);
 	if (kill < 0)
 		return kill;
-	if (kill)
-		error = __remove_privs(dentry, kill);
+	if (kill) {
+		struct user_namespace *mnt_userns;
+
+		mnt_userns = mnt_user_ns(file->f_path.mnt);
+		error = __remove_privs(mnt_userns, dentry, kill);
+	}
 	if (!error)
 		inode_has_no_xattr(inode);
 
diff --git a/fs/namei.c b/fs/namei.c
index 6cc655310bd9..a2c4ecd26966 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -3005,9 +3005,9 @@  static int handle_truncate(struct file *filp)
 	if (!error)
 		error = security_path_truncate(path);
 	if (!error) {
-		error = do_truncate(path->dentry, 0,
-				    ATTR_MTIME|ATTR_CTIME|ATTR_OPEN,
-				    filp);
+		error = do_truncate(mnt_user_ns(filp->f_path.mnt),
+				    path->dentry, 0,
+				    ATTR_MTIME | ATTR_CTIME | ATTR_OPEN, filp);
 	}
 	put_write_access(inode);
 	return error;
diff --git a/fs/open.c b/fs/open.c
index 2dc94689a7dc..648477aa7e0b 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -35,8 +35,8 @@ 
 
 #include "internal.h"
 
-int do_truncate(struct dentry *dentry, loff_t length, unsigned int time_attrs,
-	struct file *filp)
+int do_truncate(struct user_namespace *mnt_userns, struct dentry *dentry,
+		loff_t length, unsigned int time_attrs, struct file *filp)
 {
 	int ret;
 	struct iattr newattrs;
@@ -61,13 +61,14 @@  int do_truncate(struct dentry *dentry, loff_t length, unsigned int time_attrs,
 
 	inode_lock(dentry->d_inode);
 	/* Note any delegations or leases have already been broken: */
-	ret = notify_change(&init_user_ns, dentry, &newattrs, NULL);
+	ret = notify_change(mnt_userns, dentry, &newattrs, NULL);
 	inode_unlock(dentry->d_inode);
 	return ret;
 }
 
 long vfs_truncate(const struct path *path, loff_t length)
 {
+	struct user_namespace *mnt_userns;
 	struct inode *inode;
 	long error;
 
@@ -83,7 +84,8 @@  long vfs_truncate(const struct path *path, loff_t length)
 	if (error)
 		goto out;
 
-	error = inode_permission(&init_user_ns, inode, MAY_WRITE);
+	mnt_userns = mnt_user_ns(path->mnt);
+	error = inode_permission(mnt_userns, inode, MAY_WRITE);
 	if (error)
 		goto mnt_drop_write_and_out;
 
@@ -107,7 +109,7 @@  long vfs_truncate(const struct path *path, loff_t length)
 	if (!error)
 		error = security_path_truncate(path);
 	if (!error)
-		error = do_truncate(path->dentry, length, 0, NULL);
+		error = do_truncate(mnt_userns, path->dentry, length, 0, NULL);
 
 put_write_and_out:
 	put_write_access(inode);
@@ -186,13 +188,16 @@  long do_sys_ftruncate(unsigned int fd, loff_t length, int small)
 	/* Check IS_APPEND on real upper inode */
 	if (IS_APPEND(file_inode(f.file)))
 		goto out_putf;
-
 	sb_start_write(inode->i_sb);
 	error = locks_verify_truncate(inode, f.file, length);
 	if (!error)
 		error = security_path_truncate(&f.file->f_path);
-	if (!error)
-		error = do_truncate(dentry, length, ATTR_MTIME|ATTR_CTIME, f.file);
+	if (!error) {
+		struct user_namespace *mnt_userns;
+
+		mnt_userns = mnt_user_ns(f.file->f_path.mnt);
+		error = do_truncate(mnt_userns, dentry, length, ATTR_MTIME | ATTR_CTIME, f.file);
+	}
 	sb_end_write(inode->i_sb);
 out_putf:
 	fdput(f);
diff --git a/include/linux/fs.h b/include/linux/fs.h
index c88dafc4152f..616aeee07d69 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -2586,8 +2586,8 @@  struct filename {
 static_assert(offsetof(struct filename, iname) % sizeof(long) == 0);
 
 extern long vfs_truncate(const struct path *, loff_t);
-extern int do_truncate(struct dentry *, loff_t start, unsigned int time_attrs,
-		       struct file *filp);
+extern int do_truncate(struct user_namespace *, struct dentry *, loff_t start,
+		       unsigned int time_attrs, struct file *filp);
 extern int vfs_fallocate(struct file *file, int mode, loff_t offset,
 			loff_t len);
 extern long do_sys_open(int dfd, const char __user *filename, int flags,