[people/ms/linux.git] / fs / attr.c

// SPDX-License-Identifier: GPL-2.0
/*
 *  linux/fs/attr.c
 *
 *  Copyright (C) 1991, 1992  Linus Torvalds
 *  changes by Thomas Schoebel-Theuer
 */

#include <linux/export.h>
#include <linux/time.h>
#include <linux/mm.h>
#include <linux/string.h>
#include <linux/sched/signal.h>
#include <linux/capability.h>
#include <linux/fsnotify.h>
#include <linux/fcntl.h>
#include <linux/security.h>
#include <linux/evm.h>
#include <linux/ima.h>

/**
 * chown_ok - verify permissions to chown inode
 * @mnt_userns:	user namespace of the mount @inode was found from
 * @inode:	inode to check permissions on
 * @ia_vfsuid:	uid to chown @inode to
 *
 * If the inode has been found through an idmapped mount the user namespace of
 * the vfsmount must be passed through @mnt_userns. This function will then
 * take care to map the inode according to @mnt_userns before checking
 * permissions. On non-idmapped mounts or if permission checking is to be
 * performed on the raw inode simply passs init_user_ns.
 */
static bool chown_ok(struct user_namespace *mnt_userns,
		     const struct inode *inode, vfsuid_t ia_vfsuid)
{
	vfsuid_t vfsuid = i_uid_into_vfsuid(mnt_userns, inode);
	if (vfsuid_eq_kuid(vfsuid, current_fsuid()) &&
	    vfsuid_eq(ia_vfsuid, vfsuid))
		return true;
	if (capable_wrt_inode_uidgid(mnt_userns, inode, CAP_CHOWN))
		return true;
	if (!vfsuid_valid(vfsuid) &&
	    ns_capable(inode->i_sb->s_user_ns, CAP_CHOWN))
		return true;
	return false;
}

/**
 * chgrp_ok - verify permissions to chgrp inode
 * @mnt_userns:	user namespace of the mount @inode was found from
 * @inode:	inode to check permissions on
 * @ia_vfsgid:	gid to chown @inode to
 *
 * If the inode has been found through an idmapped mount the user namespace of
 * the vfsmount must be passed through @mnt_userns. This function will then
 * take care to map the inode according to @mnt_userns before checking
 * permissions. On non-idmapped mounts or if permission checking is to be
 * performed on the raw inode simply passs init_user_ns.
 */
static bool chgrp_ok(struct user_namespace *mnt_userns,
		     const struct inode *inode, vfsgid_t ia_vfsgid)
{
	vfsgid_t vfsgid = i_gid_into_vfsgid(mnt_userns, inode);
	vfsuid_t vfsuid = i_uid_into_vfsuid(mnt_userns, inode);
	if (vfsuid_eq_kuid(vfsuid, current_fsuid())) {
		if (vfsgid_eq(ia_vfsgid, vfsgid))
			return true;
		if (vfsgid_in_group_p(ia_vfsgid))
			return true;
	}
	if (capable_wrt_inode_uidgid(mnt_userns, inode, CAP_CHOWN))
		return true;
	if (!vfsgid_valid(vfsgid) &&
	    ns_capable(inode->i_sb->s_user_ns, CAP_CHOWN))
		return true;
	return false;
}

/**
 * setattr_prepare - check if attribute changes to a dentry are allowed
 * @mnt_userns:	user namespace of the mount the inode was found from
 * @dentry:	dentry to check
 * @attr:	attributes to change
 *
 * Check if we are allowed to change the attributes contained in @attr
 * in the given dentry.  This includes the normal unix access permission
 * checks, as well as checks for rlimits and others. The function also clears
 * SGID bit from mode if user is not allowed to set it. Also file capabilities
 * and IMA extended attributes are cleared if ATTR_KILL_PRIV is set.
 *
 * If the inode has been found through an idmapped mount the user namespace of
 * the vfsmount must be passed through @mnt_userns. This function will then
 * take care to map the inode according to @mnt_userns before checking
 * permissions. On non-idmapped mounts or if permission checking is to be
 * performed on the raw inode simply passs init_user_ns.
 *
 * Should be called as the first thing in ->setattr implementations,
 * possibly after taking additional locks.
 */
int setattr_prepare(struct user_namespace *mnt_userns, struct dentry *dentry,
		    struct iattr *attr)
{
	struct inode *inode = d_inode(dentry);
	unsigned int ia_valid = attr->ia_valid;

	/*
	 * First check size constraints.  These can't be overriden using
	 * ATTR_FORCE.
	 */
	if (ia_valid & ATTR_SIZE) {
		int error = inode_newsize_ok(inode, attr->ia_size);
		if (error)
			return error;
	}

	/* If force is set do it anyway. */
	if (ia_valid & ATTR_FORCE)
		goto kill_priv;

	/* Make sure a caller can chown. */
	if ((ia_valid & ATTR_UID) &&
	    !chown_ok(mnt_userns, inode, attr->ia_vfsuid))
		return -EPERM;

	/* Make sure caller can chgrp. */
	if ((ia_valid & ATTR_GID) &&
	    !chgrp_ok(mnt_userns, inode, attr->ia_vfsgid))
		return -EPERM;

	/* Make sure a caller can chmod. */
	if (ia_valid & ATTR_MODE) {
		vfsgid_t vfsgid;

		if (!inode_owner_or_capable(mnt_userns, inode))
			return -EPERM;

		if (ia_valid & ATTR_GID)
			vfsgid = attr->ia_vfsgid;
		else
			vfsgid = i_gid_into_vfsgid(mnt_userns, inode);

		/* Also check the setgid bit! */
		if (!vfsgid_in_group_p(vfsgid) &&
		    !capable_wrt_inode_uidgid(mnt_userns, inode, CAP_FSETID))
			attr->ia_mode &= ~S_ISGID;
	}

	/* Check for setting the inode time. */
	if (ia_valid & (ATTR_MTIME_SET | ATTR_ATIME_SET | ATTR_TIMES_SET)) {
		if (!inode_owner_or_capable(mnt_userns, inode))
			return -EPERM;
	}

kill_priv:
	/* User has permission for the change */
	if (ia_valid & ATTR_KILL_PRIV) {
		int error;

		error = security_inode_killpriv(mnt_userns, dentry);
		if (error)
			return error;
	}

	return 0;
}
EXPORT_SYMBOL(setattr_prepare);

/**
 * inode_newsize_ok - may this inode be truncated to a given size
 * @inode:	the inode to be truncated
 * @offset:	the new size to assign to the inode
 *
 * inode_newsize_ok must be called with i_mutex held.
 *
 * inode_newsize_ok will check filesystem limits and ulimits to check that the
 * new inode size is within limits. inode_newsize_ok will also send SIGXFSZ
 * when necessary. Caller must not proceed with inode size change if failure is
 * returned. @inode must be a file (not directory), with appropriate
 * permissions to allow truncate (inode_newsize_ok does NOT check these
 * conditions).
 *
 * Return: 0 on success, -ve errno on failure
 */
int inode_newsize_ok(const struct inode *inode, loff_t offset)
{
	if (offset < 0)
		return -EINVAL;
	if (inode->i_size < offset) {
		unsigned long limit;

		limit = rlimit(RLIMIT_FSIZE);
		if (limit != RLIM_INFINITY && offset > limit)
			goto out_sig;
		if (offset > inode->i_sb->s_maxbytes)
			goto out_big;
	} else {
		/*
		 * truncation of in-use swapfiles is disallowed - it would
		 * cause subsequent swapout to scribble on the now-freed
		 * blocks.
		 */
		if (IS_SWAPFILE(inode))
			return -ETXTBSY;
	}

	return 0;
out_sig:
	send_sig(SIGXFSZ, current, 0);
out_big:
	return -EFBIG;
}
EXPORT_SYMBOL(inode_newsize_ok);

/**
 * setattr_copy - copy simple metadata updates into the generic inode
 * @mnt_userns:	user namespace of the mount the inode was found from
 * @inode:	the inode to be updated
 * @attr:	the new attributes
 *
 * setattr_copy must be called with i_mutex held.
 *
 * setattr_copy updates the inode's metadata with that specified
 * in attr on idmapped mounts. Necessary permission checks to determine
 * whether or not the S_ISGID property needs to be removed are performed with
 * the correct idmapped mount permission helpers.
 * Noticeably missing is inode size update, which is more complex
 * as it requires pagecache updates.
 *
 * If the inode has been found through an idmapped mount the user namespace of
 * the vfsmount must be passed through @mnt_userns. This function will then
 * take care to map the inode according to @mnt_userns before checking
 * permissions. On non-idmapped mounts or if permission checking is to be
 * performed on the raw inode simply passs init_user_ns.
 *
 * The inode is not marked as dirty after this operation. The rationale is
 * that for "simple" filesystems, the struct inode is the inode storage.
 * The caller is free to mark the inode dirty afterwards if needed.
 */
void setattr_copy(struct user_namespace *mnt_userns, struct inode *inode,
		  const struct iattr *attr)
{
	unsigned int ia_valid = attr->ia_valid;

	i_uid_update(mnt_userns, attr, inode);
	i_gid_update(mnt_userns, attr, inode);
	if (ia_valid & ATTR_ATIME)
		inode->i_atime = attr->ia_atime;
	if (ia_valid & ATTR_MTIME)
		inode->i_mtime = attr->ia_mtime;
	if (ia_valid & ATTR_CTIME)
		inode->i_ctime = attr->ia_ctime;
	if (ia_valid & ATTR_MODE) {
		umode_t mode = attr->ia_mode;
		vfsgid_t vfsgid = i_gid_into_vfsgid(mnt_userns, inode);
		if (!vfsgid_in_group_p(vfsgid) &&
		    !capable_wrt_inode_uidgid(mnt_userns, inode, CAP_FSETID))
			mode &= ~S_ISGID;
		inode->i_mode = mode;
	}
}
EXPORT_SYMBOL(setattr_copy);

int may_setattr(struct user_namespace *mnt_userns, struct inode *inode,
		unsigned int ia_valid)
{
	int error;

	if (ia_valid & (ATTR_MODE | ATTR_UID | ATTR_GID | ATTR_TIMES_SET)) {
		if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
			return -EPERM;
	}

	/*
	 * If utimes(2) and friends are called with times == NULL (or both
	 * times are UTIME_NOW), then we need to check for write permission
	 */
	if (ia_valid & ATTR_TOUCH) {
		if (IS_IMMUTABLE(inode))
			return -EPERM;

		if (!inode_owner_or_capable(mnt_userns, inode)) {
			error = inode_permission(mnt_userns, inode, MAY_WRITE);
			if (error)
				return error;
		}
	}
	return 0;
}
EXPORT_SYMBOL(may_setattr);

/**
 * notify_change - modify attributes of a filesytem object
 * @mnt_userns:	user namespace of the mount the inode was found from
 * @dentry:	object affected
 * @attr:	new attributes
 * @delegated_inode: returns inode, if the inode is delegated
 *
 * The caller must hold the i_mutex on the affected object.
 *
 * If notify_change discovers a delegation in need of breaking,
 * it will return -EWOULDBLOCK and return a reference to the inode in
 * delegated_inode.  The caller should then break the delegation and
 * retry.  Because breaking a delegation may take a long time, the
 * caller should drop the i_mutex before doing so.
 *
 * Alternatively, a caller may pass NULL for delegated_inode.  This may
 * be appropriate for callers that expect the underlying filesystem not
 * to be NFS exported.  Also, passing NULL is fine for callers holding
 * the file open for write, as there can be no conflicting delegation in
 * that case.
 *
 * If the inode has been found through an idmapped mount the user namespace of
 * the vfsmount must be passed through @mnt_userns. This function will then
 * take care to map the inode according to @mnt_userns before checking
 * permissions. On non-idmapped mounts or if permission checking is to be
 * performed on the raw inode simply passs init_user_ns.
 */
int notify_change(struct user_namespace *mnt_userns, struct dentry *dentry,
		  struct iattr *attr, struct inode **delegated_inode)
{
	struct inode *inode = dentry->d_inode;
	umode_t mode = inode->i_mode;
	int error;
	struct timespec64 now;
	unsigned int ia_valid = attr->ia_valid;

	WARN_ON_ONCE(!inode_is_locked(inode));

	error = may_setattr(mnt_userns, inode, ia_valid);
	if (error)
		return error;

	if ((ia_valid & ATTR_MODE)) {
		umode_t amode = attr->ia_mode;
		/* Flag setting protected by i_mutex */
		if (is_sxid(amode))
			inode->i_flags &= ~S_NOSEC;
	}

	now = current_time(inode);

	attr->ia_ctime = now;
	if (!(ia_valid & ATTR_ATIME_SET))
		attr->ia_atime = now;
	else
		attr->ia_atime = timestamp_truncate(attr->ia_atime, inode);
	if (!(ia_valid & ATTR_MTIME_SET))
		attr->ia_mtime = now;
	else
		attr->ia_mtime = timestamp_truncate(attr->ia_mtime, inode);

	if (ia_valid & ATTR_KILL_PRIV) {
		error = security_inode_need_killpriv(dentry);
		if (error < 0)
			return error;
		if (error == 0)
			ia_valid = attr->ia_valid &= ~ATTR_KILL_PRIV;
	}

	/*
	 * We now pass ATTR_KILL_S*ID to the lower level setattr function so
	 * that the function has the ability to reinterpret a mode change
	 * that's due to these bits. This adds an implicit restriction that
	 * no function will ever call notify_change with both ATTR_MODE and
	 * ATTR_KILL_S*ID set.
	 */
	if ((ia_valid & (ATTR_KILL_SUID|ATTR_KILL_SGID)) &&
	    (ia_valid & ATTR_MODE))
		BUG();

	if (ia_valid & ATTR_KILL_SUID) {
		if (mode & S_ISUID) {
			ia_valid = attr->ia_valid |= ATTR_MODE;
			attr->ia_mode = (inode->i_mode & ~S_ISUID);
		}
	}
	if (ia_valid & ATTR_KILL_SGID) {
		if ((mode & (S_ISGID | S_IXGRP)) == (S_ISGID | S_IXGRP)) {
			if (!(ia_valid & ATTR_MODE)) {
				ia_valid = attr->ia_valid |= ATTR_MODE;
				attr->ia_mode = inode->i_mode;
			}
			attr->ia_mode &= ~S_ISGID;
		}
	}
	if (!(attr->ia_valid & ~(ATTR_KILL_SUID | ATTR_KILL_SGID)))
		return 0;

	/*
	 * Verify that uid/gid changes are valid in the target
	 * namespace of the superblock.
	 */
	if (ia_valid & ATTR_UID &&
	    !vfsuid_has_fsmapping(mnt_userns, inode->i_sb->s_user_ns,
				  attr->ia_vfsuid))
		return -EOVERFLOW;
	if (ia_valid & ATTR_GID &&
	    !vfsgid_has_fsmapping(mnt_userns, inode->i_sb->s_user_ns,
				  attr->ia_vfsgid))
		return -EOVERFLOW;

	/* Don't allow modifications of files with invalid uids or
	 * gids unless those uids & gids are being made valid.
	 */
	if (!(ia_valid & ATTR_UID) &&
	    !vfsuid_valid(i_uid_into_vfsuid(mnt_userns, inode)))
		return -EOVERFLOW;
	if (!(ia_valid & ATTR_GID) &&
	    !vfsgid_valid(i_gid_into_vfsgid(mnt_userns, inode)))
		return -EOVERFLOW;

	error = security_inode_setattr(mnt_userns, dentry, attr);
	if (error)
		return error;
	error = try_break_deleg(inode, delegated_inode);
	if (error)
		return error;

	if (inode->i_op->setattr)
		error = inode->i_op->setattr(mnt_userns, dentry, attr);
	else
		error = simple_setattr(mnt_userns, dentry, attr);

	if (!error) {
		fsnotify_change(dentry, ia_valid);
		ima_inode_post_setattr(mnt_userns, dentry);
		evm_inode_post_setattr(dentry, ia_valid);
	}

	return error;
}
EXPORT_SYMBOL(notify_change);
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
1da177e4 LT	2	/*
	3	* linux/fs/attr.c
	4	*
	5	* Copyright (C) 1991, 1992 Linus Torvalds
	6	* changes by Thomas Schoebel-Theuer
	7	*/
	8
630d9c47	9	#include <linux/export.h>
1da177e4 LT	10	#include <linux/time.h>
	11	#include <linux/mm.h>
	12	#include <linux/string.h>
3f07c014	13	#include <linux/sched/signal.h>
16f7e0fe	14	#include <linux/capability.h>
0eeca283	15	#include <linux/fsnotify.h>
1da177e4	16	#include <linux/fcntl.h>
1da177e4	17	#include <linux/security.h>
975d2943	18	#include <linux/evm.h>
9957a504	19	#include <linux/ima.h>
1da177e4	20
2f221d6f CB	21	/**
	22	* chown_ok - verify permissions to chown inode
	23	* @mnt_userns: user namespace of the mount @inode was found from
	24	* @inode: inode to check permissions on
81a1807d	25	* @ia_vfsuid: uid to chown @inode to
2f221d6f CB	26	*
	27	* If the inode has been found through an idmapped mount the user namespace of
	28	* the vfsmount must be passed through @mnt_userns. This function will then
	29	* take care to map the inode according to @mnt_userns before checking
	30	* permissions. On non-idmapped mounts or if permission checking is to be
	31	* performed on the raw inode simply passs init_user_ns.
	32	*/
	33	static bool chown_ok(struct user_namespace *mnt_userns,
b27c82e1	34	const struct inode *inode, vfsuid_t ia_vfsuid)
0031181c	35	{
b27c82e1 CB	36	vfsuid_t vfsuid = i_uid_into_vfsuid(mnt_userns, inode);
	37	if (vfsuid_eq_kuid(vfsuid, current_fsuid()) &&
	38	vfsuid_eq(ia_vfsuid, vfsuid))
0031181c	39	return true;
2f221d6f	40	if (capable_wrt_inode_uidgid(mnt_userns, inode, CAP_CHOWN))
0031181c	41	return true;
b27c82e1	42	if (!vfsuid_valid(vfsuid) &&
0031181c EB	43	ns_capable(inode->i_sb->s_user_ns, CAP_CHOWN))
	44	return true;
	45	return false;
	46	}
	47
2f221d6f CB	48	/**
	49	* chgrp_ok - verify permissions to chgrp inode
	50	* @mnt_userns: user namespace of the mount @inode was found from
	51	* @inode: inode to check permissions on
81a1807d	52	* @ia_vfsgid: gid to chown @inode to
2f221d6f CB	53	*
	54	* If the inode has been found through an idmapped mount the user namespace of
	55	* the vfsmount must be passed through @mnt_userns. This function will then
	56	* take care to map the inode according to @mnt_userns before checking
	57	* permissions. On non-idmapped mounts or if permission checking is to be
	58	* performed on the raw inode simply passs init_user_ns.
	59	*/
	60	static bool chgrp_ok(struct user_namespace *mnt_userns,
b27c82e1	61	const struct inode *inode, vfsgid_t ia_vfsgid)
0031181c	62	{
b27c82e1 CB	63	vfsgid_t vfsgid = i_gid_into_vfsgid(mnt_userns, inode);
	64	vfsuid_t vfsuid = i_uid_into_vfsuid(mnt_userns, inode);
	65	if (vfsuid_eq_kuid(vfsuid, current_fsuid())) {
	66	if (vfsgid_eq(ia_vfsgid, vfsgid))
168f9128	67	return true;
b27c82e1	68	if (vfsgid_in_group_p(ia_vfsgid))
168f9128 CB	69	return true;
168f9128 CB	70	}
2f221d6f	71	if (capable_wrt_inode_uidgid(mnt_userns, inode, CAP_CHOWN))
0031181c	72	return true;
b27c82e1	73	if (!vfsgid_valid(vfsgid) &&
0031181c EB	74	ns_capable(inode->i_sb->s_user_ns, CAP_CHOWN))
	75	return true;
	76	return false;
	77	}
	78
2c27c65e	79	/**
31051c85	80	* setattr_prepare - check if attribute changes to a dentry are allowed
2f221d6f	81	* @mnt_userns: user namespace of the mount the inode was found from
31051c85	82	* @dentry: dentry to check
2c27c65e CH	83	* @attr: attributes to change
	84	*
	85	* Check if we are allowed to change the attributes contained in @attr
31051c85 JK	86	* in the given dentry. This includes the normal unix access permission
	87	* checks, as well as checks for rlimits and others. The function also clears
	88	* SGID bit from mode if user is not allowed to set it. Also file capabilities
	89	* and IMA extended attributes are cleared if ATTR_KILL_PRIV is set.
2c27c65e	90	*
2f221d6f CB	91	* If the inode has been found through an idmapped mount the user namespace of
	92	* the vfsmount must be passed through @mnt_userns. This function will then
	93	* take care to map the inode according to @mnt_userns before checking
	94	* permissions. On non-idmapped mounts or if permission checking is to be
	95	* performed on the raw inode simply passs init_user_ns.
	96	*
2c27c65e CH	97	* Should be called as the first thing in ->setattr implementations,
	98	* possibly after taking additional locks.
	99	*/
2f221d6f CB	100	int setattr_prepare(struct user_namespace mnt_userns, struct dentry dentry,
2f221d6f CB	101	struct iattr *attr)
1da177e4	102	{
31051c85	103	struct inode *inode = d_inode(dentry);
1da177e4 LT	104	unsigned int ia_valid = attr->ia_valid;
1da177e4 LT	105
2c27c65e CH	106	/*
	107	* First check size constraints. These can't be overriden using
	108	* ATTR_FORCE.
	109	*/
	110	if (ia_valid & ATTR_SIZE) {
	111	int error = inode_newsize_ok(inode, attr->ia_size);
	112	if (error)
	113	return error;
	114	}
	115
1da177e4 LT	116	/* If force is set do it anyway. */
1da177e4 LT	117	if (ia_valid & ATTR_FORCE)
030b533c	118	goto kill_priv;
1da177e4 LT	119
1da177e4 LT	120	/* Make sure a caller can chown. */
b27c82e1 CB	121	if ((ia_valid & ATTR_UID) &&
b27c82e1 CB	122	!chown_ok(mnt_userns, inode, attr->ia_vfsuid))
2c27c65e	123	return -EPERM;
1da177e4 LT	124
1da177e4 LT	125	/* Make sure caller can chgrp. */
b27c82e1 CB	126	if ((ia_valid & ATTR_GID) &&
b27c82e1 CB	127	!chgrp_ok(mnt_userns, inode, attr->ia_vfsgid))
2c27c65e	128	return -EPERM;
1da177e4 LT	129
	130	/* Make sure a caller can chmod. */
	131	if (ia_valid & ATTR_MODE) {
b27c82e1	132	vfsgid_t vfsgid;
168f9128	133
2f221d6f	134	if (!inode_owner_or_capable(mnt_userns, inode))
2c27c65e	135	return -EPERM;
168f9128 CB	136
168f9128 CB	137	if (ia_valid & ATTR_GID)
b27c82e1	138	vfsgid = attr->ia_vfsgid;
168f9128	139	else
b27c82e1	140	vfsgid = i_gid_into_vfsgid(mnt_userns, inode);
168f9128	141
1da177e4	142	/* Also check the setgid bit! */
b27c82e1	143	if (!vfsgid_in_group_p(vfsgid) &&
168f9128	144	!capable_wrt_inode_uidgid(mnt_userns, inode, CAP_FSETID))
1da177e4 LT	145	attr->ia_mode &= ~S_ISGID;
	146	}
	147
	148	/* Check for setting the inode time. */
9767d749	149	if (ia_valid & (ATTR_MTIME_SET \| ATTR_ATIME_SET \| ATTR_TIMES_SET)) {
2f221d6f	150	if (!inode_owner_or_capable(mnt_userns, inode))
2c27c65e	151	return -EPERM;
1da177e4	152	}
2c27c65e	153
030b533c JK	154	kill_priv:
	155	/* User has permission for the change */
	156	if (ia_valid & ATTR_KILL_PRIV) {
	157	int error;
	158
71bc356f	159	error = security_inode_killpriv(mnt_userns, dentry);
030b533c JK	160	if (error)
	161	return error;
	162	}
	163
2c27c65e	164	return 0;
1da177e4	165	}
31051c85	166	EXPORT_SYMBOL(setattr_prepare);
1da177e4	167
25d9e2d1 NP	168	/**
	169	* inode_newsize_ok - may this inode be truncated to a given size
	170	* @inode: the inode to be truncated
	171	* @offset: the new size to assign to the inode
25d9e2d1	172	*
7bb46a67 NP	173	* inode_newsize_ok must be called with i_mutex held.
7bb46a67 NP	174	*
25d9e2d1 NP	175	* inode_newsize_ok will check filesystem limits and ulimits to check that the
	176	* new inode size is within limits. inode_newsize_ok will also send SIGXFSZ
	177	* when necessary. Caller must not proceed with inode size change if failure is
	178	* returned. @inode must be a file (not directory), with appropriate
	179	* permissions to allow truncate (inode_newsize_ok does NOT check these
	180	* conditions).
3fae1746 MW	181	*
3fae1746 MW	182	* Return: 0 on success, -ve errno on failure
25d9e2d1 NP	183	*/
	184	int inode_newsize_ok(const struct inode *inode, loff_t offset)
	185	{
e2ebff9c DH	186	if (offset < 0)
e2ebff9c DH	187	return -EINVAL;
25d9e2d1 NP	188	if (inode->i_size < offset) {
	189	unsigned long limit;
	190
d554ed89	191	limit = rlimit(RLIMIT_FSIZE);
25d9e2d1 NP	192	if (limit != RLIM_INFINITY && offset > limit)
	193	goto out_sig;
	194	if (offset > inode->i_sb->s_maxbytes)
	195	goto out_big;
	196	} else {
	197	/*
	198	* truncation of in-use swapfiles is disallowed - it would
	199	* cause subsequent swapout to scribble on the now-freed
	200	* blocks.
	201	*/
	202	if (IS_SWAPFILE(inode))
	203	return -ETXTBSY;
	204	}
	205
	206	return 0;
	207	out_sig:
	208	send_sig(SIGXFSZ, current, 0);
	209	out_big:
	210	return -EFBIG;
	211	}
	212	EXPORT_SYMBOL(inode_newsize_ok);
	213
7bb46a67	214	/**
6a1a90ad	215	* setattr_copy - copy simple metadata updates into the generic inode
2f221d6f	216	* @mnt_userns: user namespace of the mount the inode was found from
7bb46a67 NP	217	* @inode: the inode to be updated
	218	* @attr: the new attributes
	219	*
6a1a90ad	220	* setattr_copy must be called with i_mutex held.
7bb46a67	221	*
6a1a90ad	222	* setattr_copy updates the inode's metadata with that specified
b27c82e1	223	* in attr on idmapped mounts. Necessary permission checks to determine
2f221d6f CB	224	* whether or not the S_ISGID property needs to be removed are performed with
	225	* the correct idmapped mount permission helpers.
	226	* Noticeably missing is inode size update, which is more complex
2c27c65e	227	* as it requires pagecache updates.
7bb46a67	228	*
2f221d6f CB	229	* If the inode has been found through an idmapped mount the user namespace of
	230	* the vfsmount must be passed through @mnt_userns. This function will then
	231	* take care to map the inode according to @mnt_userns before checking
	232	* permissions. On non-idmapped mounts or if permission checking is to be
	233	* performed on the raw inode simply passs init_user_ns.
	234	*
7bb46a67 NP	235	* The inode is not marked as dirty after this operation. The rationale is
	236	* that for "simple" filesystems, the struct inode is the inode storage.
	237	* The caller is free to mark the inode dirty afterwards if needed.
	238	*/
2f221d6f CB	239	void setattr_copy(struct user_namespace mnt_userns, struct inode inode,
2f221d6f CB	240	const struct iattr *attr)
1da177e4 LT	241	{
1da177e4 LT	242	unsigned int ia_valid = attr->ia_valid;
4a30131e	243
b27c82e1 CB	244	i_uid_update(mnt_userns, attr, inode);
b27c82e1 CB	245	i_gid_update(mnt_userns, attr, inode);
eb31e2f6 AG	246	if (ia_valid & ATTR_ATIME)
	247	inode->i_atime = attr->ia_atime;
	248	if (ia_valid & ATTR_MTIME)
	249	inode->i_mtime = attr->ia_mtime;
	250	if (ia_valid & ATTR_CTIME)
	251	inode->i_ctime = attr->ia_ctime;
1da177e4 LT	252	if (ia_valid & ATTR_MODE) {
1da177e4 LT	253	umode_t mode = attr->ia_mode;
b27c82e1 CB	254	vfsgid_t vfsgid = i_gid_into_vfsgid(mnt_userns, inode);
b27c82e1 CB	255	if (!vfsgid_in_group_p(vfsgid) &&
2f221d6f	256	!capable_wrt_inode_uidgid(mnt_userns, inode, CAP_FSETID))
1da177e4 LT	257	mode &= ~S_ISGID;
	258	inode->i_mode = mode;
	259	}
7bb46a67	260	}
6a1a90ad	261	EXPORT_SYMBOL(setattr_copy);
7bb46a67	262
7bb698f0 AG	263	int may_setattr(struct user_namespace mnt_userns, struct inode inode,
	264	unsigned int ia_valid)
	265	{
	266	int error;
	267
	268	if (ia_valid & (ATTR_MODE \| ATTR_UID \| ATTR_GID \| ATTR_TIMES_SET)) {
	269	if (IS_IMMUTABLE(inode) \|\| IS_APPEND(inode))
	270	return -EPERM;
	271	}
	272
	273	/*
	274	* If utimes(2) and friends are called with times == NULL (or both
	275	* times are UTIME_NOW), then we need to check for write permission
	276	*/
	277	if (ia_valid & ATTR_TOUCH) {
	278	if (IS_IMMUTABLE(inode))
	279	return -EPERM;
	280
	281	if (!inode_owner_or_capable(mnt_userns, inode)) {
	282	error = inode_permission(mnt_userns, inode, MAY_WRITE);
	283	if (error)
	284	return error;
	285	}
	286	}
	287	return 0;
	288	}
	289	EXPORT_SYMBOL(may_setattr);
	290
27ac0ffe BF	291	/**
27ac0ffe BF	292	* notify_change - modify attributes of a filesytem object
2f221d6f	293	* @mnt_userns: user namespace of the mount the inode was found from
27ac0ffe	294	* @dentry: object affected
3fae1746	295	* @attr: new attributes
27ac0ffe BF	296	* @delegated_inode: returns inode, if the inode is delegated
	297	*
	298	* The caller must hold the i_mutex on the affected object.
	299	*
	300	* If notify_change discovers a delegation in need of breaking,
	301	* it will return -EWOULDBLOCK and return a reference to the inode in
	302	* delegated_inode. The caller should then break the delegation and
	303	* retry. Because breaking a delegation may take a long time, the
	304	* caller should drop the i_mutex before doing so.
	305	*
	306	* Alternatively, a caller may pass NULL for delegated_inode. This may
	307	* be appropriate for callers that expect the underlying filesystem not
	308	* to be NFS exported. Also, passing NULL is fine for callers holding
	309	* the file open for write, as there can be no conflicting delegation in
	310	* that case.
2f221d6f CB	311	*
	312	* If the inode has been found through an idmapped mount the user namespace of
	313	* the vfsmount must be passed through @mnt_userns. This function will then
	314	* take care to map the inode according to @mnt_userns before checking
	315	* permissions. On non-idmapped mounts or if permission checking is to be
	316	* performed on the raw inode simply passs init_user_ns.
27ac0ffe	317	*/
2f221d6f CB	318	int notify_change(struct user_namespace mnt_userns, struct dentry dentry,
2f221d6f CB	319	struct iattr attr, struct inode *delegated_inode)
1da177e4 LT	320	{
1da177e4 LT	321	struct inode *inode = dentry->d_inode;
8d334acd	322	umode_t mode = inode->i_mode;
1da177e4	323	int error;
95582b00	324	struct timespec64 now;
1da177e4 LT	325	unsigned int ia_valid = attr->ia_valid;
1da177e4 LT	326
5955102c	327	WARN_ON_ONCE(!inode_is_locked(inode));
c4107b30	328
7bb698f0 AG	329	error = may_setattr(mnt_userns, inode, ia_valid);
	330	if (error)
	331	return error;
f2b20f6e	332
69b45732	333	if ((ia_valid & ATTR_MODE)) {
8d334acd	334	umode_t amode = attr->ia_mode;
69b45732 AK	335	/* Flag setting protected by i_mutex */
	336	if (is_sxid(amode))
	337	inode->i_flags &= ~S_NOSEC;
	338	}
	339
c2050a45	340	now = current_time(inode);
1da177e4 LT	341
	342	attr->ia_ctime = now;
	343	if (!(ia_valid & ATTR_ATIME_SET))
	344	attr->ia_atime = now;
eb31e2f6 AG	345	else
eb31e2f6 AG	346	attr->ia_atime = timestamp_truncate(attr->ia_atime, inode);
1da177e4 LT	347	if (!(ia_valid & ATTR_MTIME_SET))
1da177e4 LT	348	attr->ia_mtime = now;
eb31e2f6 AG	349	else
	350	attr->ia_mtime = timestamp_truncate(attr->ia_mtime, inode);
	351
b5376771	352	if (ia_valid & ATTR_KILL_PRIV) {
b5376771	353	error = security_inode_need_killpriv(dentry);
030b533c	354	if (error < 0)
b5376771	355	return error;
030b533c JK	356	if (error == 0)
030b533c JK	357	ia_valid = attr->ia_valid &= ~ATTR_KILL_PRIV;
b5376771	358	}
6de0ec00 JL	359
	360	/*
	361	* We now pass ATTR_KILL_S*ID to the lower level setattr function so
	362	* that the function has the ability to reinterpret a mode change
	363	* that's due to these bits. This adds an implicit restriction that
	364	* no function will ever call notify_change with both ATTR_MODE and
	365	* ATTR_KILL_S*ID set.
	366	*/
	367	if ((ia_valid & (ATTR_KILL_SUID\|ATTR_KILL_SGID)) &&
	368	(ia_valid & ATTR_MODE))
	369	BUG();
	370
1da177e4	371	if (ia_valid & ATTR_KILL_SUID) {
1da177e4	372	if (mode & S_ISUID) {
6de0ec00 JL	373	ia_valid = attr->ia_valid \|= ATTR_MODE;
6de0ec00 JL	374	attr->ia_mode = (inode->i_mode & ~S_ISUID);
1da177e4 LT	375	}
	376	}
	377	if (ia_valid & ATTR_KILL_SGID) {
1da177e4 LT	378	if ((mode & (S_ISGID \| S_IXGRP)) == (S_ISGID \| S_IXGRP)) {
	379	if (!(ia_valid & ATTR_MODE)) {
	380	ia_valid = attr->ia_valid \|= ATTR_MODE;
	381	attr->ia_mode = inode->i_mode;
	382	}
	383	attr->ia_mode &= ~S_ISGID;
	384	}
	385	}
6de0ec00	386	if (!(attr->ia_valid & ~(ATTR_KILL_SUID \| ATTR_KILL_SGID)))
1da177e4 LT	387	return 0;
1da177e4 LT	388
a475acf0 SF	389	/*
	390	* Verify that uid/gid changes are valid in the target
	391	* namespace of the superblock.
	392	*/
	393	if (ia_valid & ATTR_UID &&
b27c82e1 CB	394	!vfsuid_has_fsmapping(mnt_userns, inode->i_sb->s_user_ns,
b27c82e1 CB	395	attr->ia_vfsuid))
a475acf0 SF	396	return -EOVERFLOW;
a475acf0 SF	397	if (ia_valid & ATTR_GID &&
b27c82e1 CB	398	!vfsgid_has_fsmapping(mnt_userns, inode->i_sb->s_user_ns,
b27c82e1 CB	399	attr->ia_vfsgid))
a475acf0 SF	400	return -EOVERFLOW;
a475acf0 SF	401
0bd23d09 EB	402	/* Don't allow modifications of files with invalid uids or
	403	* gids unless those uids & gids are being made valid.
	404	*/
2f221d6f	405	if (!(ia_valid & ATTR_UID) &&
b27c82e1	406	!vfsuid_valid(i_uid_into_vfsuid(mnt_userns, inode)))
0bd23d09	407	return -EOVERFLOW;
2f221d6f	408	if (!(ia_valid & ATTR_GID) &&
b27c82e1	409	!vfsgid_valid(i_gid_into_vfsgid(mnt_userns, inode)))
0bd23d09 EB	410	return -EOVERFLOW;
0bd23d09 EB	411
b27c82e1	412	error = security_inode_setattr(mnt_userns, dentry, attr);
27ac0ffe BF	413	if (error)
	414	return error;
	415	error = try_break_deleg(inode, delegated_inode);
a77b72da MS	416	if (error)
	417	return error;
	418
eef2380c	419	if (inode->i_op->setattr)
549c7297	420	error = inode->i_op->setattr(mnt_userns, dentry, attr);
eef2380c	421	else
549c7297	422	error = simple_setattr(mnt_userns, dentry, attr);
1da177e4	423
975d2943	424	if (!error) {
0eeca283	425	fsnotify_change(dentry, ia_valid);
a2d2329e	426	ima_inode_post_setattr(mnt_userns, dentry);
975d2943 MZ	427	evm_inode_post_setattr(dentry, ia_valid);
975d2943 MZ	428	}
0eeca283	429
1da177e4 LT	430	return error;
1da177e4 LT	431	}
1da177e4	432	EXPORT_SYMBOL(notify_change);