Commit 58c90473 authored by Dave Chinner's avatar Dave Chinner Committed by Dave Chinner
Browse files

xfs: inodes are new until the dentry cache is set up 



Al Viro noticed a generic set of issues to do with filehandle lookup
racing with dentry cache setup. They involve a filehandle lookup
occurring while an inode is being created and the filehandle lookup
racing with the dentry creation for the real file. This can lead to
multiple dentries for the one path being instantiated. There are a
host of other issues around this same set of paths.

The underlying cause is that file handle lookup only waits on inode
cache instantiation rather than full dentry cache instantiation. XFS
is mostly immune to the problems discovered due to it's own internal
inode cache, but there are a couple of corner cases where races can
happen.

We currently clear the XFS_INEW flag when the inode is fully set up
after insertion into the cache. Newly allocated inodes are inserted
locked and so aren't usable until the allocation transaction
commits. This, however, occurs before the dentry and security
information is fully initialised and hence the inode is unlocked and
available for lookups to find too early.

To solve the problem, only clear the XFS_INEW flag for newly created
inodes once the dentry is fully instantiated. This means lookups
will retry until the XFS_INEW flag is removed from the inode and
hence avoids the race conditions in questions.

THis also means that xfs_create(), xfs_create_tmpfile() and
xfs_symlink() need to finish the setup of the inode in their error
paths if we had allocated the inode but failed later in the creation
process. xfs_symlink(), in particular, needed a lot of help to make
it's error handling match that of xfs_create().

Signed-off-by: default avatarDave Chinner <dchinner@redhat.com>
Reviewed-by: default avatarBrian Foster <bfoster@redhat.com>
Signed-off-by: default avatarDave Chinner <david@fromorbit.com>
parent c517d838

fs/xfs/xfs_icache.c

+2 −2
Original line number Diff line number Diff line
@@ -439,11 +439,11 @@ again: 
	*ipp = ip;



	/*

	 * If we have a real type for an on-disk inode, we can set ops(&unlock)

	 * If we have a real type for an on-disk inode, we can setup the inode

	 * now.	 If it's a new inode being created, xfs_ialloc will handle it.

	 */

	if (xfs_iflags_test(ip, XFS_INEW) && ip->i_d.di_mode != 0)

		xfs_setup_inode(ip);

		xfs_setup_existing_inode(ip);

	return 0;



out_error_or_again:

fs/xfs/xfs_inode.c

+13 −9
Original line number Diff line number Diff line
@@ -818,7 +818,7 @@ xfs_ialloc( 
	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);

	xfs_trans_log_inode(tp, ip, flags);



	/* now that we have an i_mode we can setup inode ops and unlock */

	/* now that we have an i_mode we can setup the inode structure */

	xfs_setup_inode(ip);



	*ipp = ip;
@@ -1235,12 +1235,14 @@ xfs_create( 
	xfs_trans_cancel(tp, cancel_flags);

 out_release_inode:

	/*

	 * Wait until after the current transaction is aborted to

	 * release the inode.  This prevents recursive transactions

	 * and deadlocks from xfs_inactive.

	 * Wait until after the current transaction is aborted to finish the

	 * setup of the inode and release the inode.  This prevents recursive

	 * transactions and deadlocks from xfs_inactive.

	 */

	if (ip)

	if (ip) {

		xfs_finish_inode_setup(ip);

		IRELE(ip);

	}



	xfs_qm_dqrele(udqp);

	xfs_qm_dqrele(gdqp);
@@ -1345,12 +1347,14 @@ xfs_create_tmpfile( 
	xfs_trans_cancel(tp, cancel_flags);

 out_release_inode:

	/*

	 * Wait until after the current transaction is aborted to

	 * release the inode.  This prevents recursive transactions

	 * and deadlocks from xfs_inactive.

	 * Wait until after the current transaction is aborted to finish the

	 * setup of the inode and release the inode.  This prevents recursive

	 * transactions and deadlocks from xfs_inactive.

	 */

	if (ip)

	if (ip) {

		xfs_finish_inode_setup(ip);

		IRELE(ip);

	}



	xfs_qm_dqrele(udqp);

	xfs_qm_dqrele(gdqp);

fs/xfs/xfs_inode.h

+22 −0
Original line number Diff line number Diff line
@@ -390,6 +390,28 @@ int xfs_zero_eof(struct xfs_inode *, xfs_off_t, xfs_fsize_t); 
int		xfs_iozero(struct xfs_inode *, loff_t, size_t);





/* from xfs_iops.c */

/*

 * When setting up a newly allocated inode, we need to call

 * xfs_finish_inode_setup() once the inode is fully instantiated at

 * the VFS level to prevent the rest of the world seeing the inode

 * before we've completed instantiation. Otherwise we can do it

 * the moment the inode lookup is complete.

 */

extern void xfs_setup_inode(struct xfs_inode *ip);

static inline void xfs_finish_inode_setup(struct xfs_inode *ip)

{

	xfs_iflags_clear(ip, XFS_INEW);

	barrier();

	unlock_new_inode(VFS_I(ip));

}



static inline void xfs_setup_existing_inode(struct xfs_inode *ip)

{

	xfs_setup_inode(ip);

	xfs_finish_inode_setup(ip);

}



#define IHOLD(ip) \

do { \

	ASSERT(atomic_read(&VFS_I(ip)->i_count) > 0) ; \

fs/xfs/xfs_iops.c

+10 −14
Original line number Diff line number Diff line
@@ -187,6 +187,8 @@ xfs_generic_create( 
	else

		d_instantiate(dentry, inode);



	xfs_finish_inode_setup(ip);



 out_free_acl:

	if (default_acl)

		posix_acl_release(default_acl);
@@ -195,6 +197,7 @@ xfs_generic_create( 
	return error;



 out_cleanup_inode:

	xfs_finish_inode_setup(ip);

	if (!tmpfile)

		xfs_cleanup_inode(dir, inode, dentry);

	iput(inode);
@@ -367,9 +370,11 @@ xfs_vn_symlink( 
		goto out_cleanup_inode;



	d_instantiate(dentry, inode);

	xfs_finish_inode_setup(cip);

	return 0;



 out_cleanup_inode:

	xfs_finish_inode_setup(cip);

	xfs_cleanup_inode(dir, inode, dentry);

	iput(inode);

 out:
@@ -1236,16 +1241,12 @@ xfs_diflags_to_iflags( 
}



/*

 * Initialize the Linux inode, set up the operation vectors and

 * unlock the inode.

 *

 * When reading existing inodes from disk this is called directly

 * from xfs_iget, when creating a new inode it is called from

 * xfs_ialloc after setting up the inode.

 * Initialize the Linux inode and set up the operation vectors.

 *

 * We are always called with an uninitialised linux inode here.

 * We need to initialise the necessary fields and take a reference

 * on it.

 * When reading existing inodes from disk this is called directly from xfs_iget,

 * when creating a new inode it is called from xfs_ialloc after setting up the

 * inode. These callers have different criteria for clearing XFS_INEW, so leave

 * it up to the caller to deal with unlocking the inode appropriately.

 */

void

xfs_setup_inode(
@@ -1332,9 +1333,4 @@ xfs_setup_inode( 
		inode_has_no_xattr(inode);

		cache_no_acl(inode);

	}



	xfs_iflags_clear(ip, XFS_INEW);

	barrier();



	unlock_new_inode(inode);

}

fs/xfs/xfs_iops.h

+0 −2
Original line number Diff line number Diff line
@@ -25,8 +25,6 @@ extern const struct file_operations xfs_dir_file_operations; 


extern ssize_t xfs_vn_listxattr(struct dentry *, char *data, size_t size);



extern void xfs_setup_inode(struct xfs_inode *);



/*

 * Internal setattr interfaces.

 */

CRA Git | Maintained and supported by SUSTech CRA and CCSE