Add a lock structure to vnode structure. Previously it was either allocated

separately (nfs, cd9660 etc) or keept as a first element of structure
referenced by v_data pointer(ffs). Such organization leads to known problems
with stacked filesystems.

From this point vop_no*lock*() functions maintain only interlock lock.
vop_std*lock*() functions maintain built-in v_lock structure using lockmgr().
vop_sharedlock() is compatible with vop_stdunlock(), but maintains a shared
lock on vnode.

If filesystem wishes to export lockmgr compatible lock, it can put an address
of this lock to v_vnlock field. This indicates that the upper filesystem
can take advantage of it and use single lock structure for entire (or part)
of stack of vnodes. This field shouldn't be examined or modified by VFS code
except for initialization purposes.

Reviewed in general by:	mckusick

13 files changed, 47 insertions, 68 deletions

diff --git a/sys/gnu/ext2fs/ext2_ihash.c b/sys/gnu/ext2fs/ext2_ihash.c
index 0a7c7d5..36176f0 100644
--- a/sys/gnu/ext2fs/ext2_ihash.c
+++ b/sys/gnu/ext2fs/ext2_ihash.c
@@ -130,7 +130,7 @@ ufs_ihashins(ip)
 	struct ihashhead *ipp;
 
 	/* lock the inode, then put it on the appropriate hash list */
-	lockmgr(&ip->i_lock, LK_EXCLUSIVE, (struct simplelock *)0, p);
+	lockmgr(&ip->i_vnode->v_lock, LK_EXCLUSIVE, (struct simplelock *)0, p);
 
 	simple_lock(&ufs_ihash_slock);
 	ipp = INOHASH(ip->i_dev, ip->i_number);
diff --git a/sys/gnu/ext2fs/inode.h b/sys/gnu/ext2fs/inode.h
index 6417a10..79e79b1 100644
--- a/sys/gnu/ext2fs/inode.h
+++ b/sys/gnu/ext2fs/inode.h
@@ -66,7 +66,6 @@ typedef long ufs_lbn_t;
  * active, and is put back when the file is no longer being used.
  */
 struct inode {
-	struct	 lock i_lock;	/* Inode lock. >Keep this first< */
 	LIST_ENTRY(inode) i_hash;/* Hash chain. */
 	struct	vnode  *i_vnode;/* Vnode associated with this inode. */
 	struct	vnode  *i_devvp;/* Vnode for block I/O. */
diff --git a/sys/gnu/fs/ext2fs/inode.h b/sys/gnu/fs/ext2fs/inode.h
index 6417a10..79e79b1 100644
--- a/sys/gnu/fs/ext2fs/inode.h
+++ b/sys/gnu/fs/ext2fs/inode.h
@@ -66,7 +66,6 @@ typedef long ufs_lbn_t;
  * active, and is put back when the file is no longer being used.
  */
 struct inode {
-	struct	 lock i_lock;	/* Inode lock. >Keep this first< */
 	LIST_ENTRY(inode) i_hash;/* Hash chain. */
 	struct	vnode  *i_vnode;/* Vnode associated with this inode. */
 	struct	vnode  *i_devvp;/* Vnode for block I/O. */
diff --git a/sys/kern/vfs_default.c b/sys/kern/vfs_default.c
index 2b651c8..01418f0 100644
--- a/sys/kern/vfs_default.c
+++ b/sys/kern/vfs_default.c
@@ -231,19 +231,13 @@ vop_stdlock(ap)
 		struct proc *a_p;
 	} */ *ap;
 {               
-	struct lock *l;
-
-	if ((l = (struct lock *)ap->a_vp->v_data) == NULL) {
-		if (ap->a_flags & LK_INTERLOCK)
-			simple_unlock(&ap->a_vp->v_interlock);
-		return 0;
-	}
+	struct vnode *vp = ap->a_vp;
 
 #ifndef	DEBUG_LOCKS
-	return (lockmgr(l, ap->a_flags, &ap->a_vp->v_interlock, ap->a_p));
+	return (lockmgr(&vp->v_lock, ap->a_flags, &vp->v_interlock, ap->a_p));
 #else
-	return (debuglockmgr(l, ap->a_flags, &ap->a_vp->v_interlock, ap->a_p,
-	    "vop_stdlock", ap->a_vp->filename, ap->a_vp->line));
+	return (debuglockmgr(&vp->v_lock, ap->a_flags, &vp->v_interlock,
+	    ap->a_p, "vop_stdlock", vp->filename, vp->line));
 #endif
 }
 
@@ -255,15 +249,9 @@ vop_stdunlock(ap)
 		struct proc *a_p;
 	} */ *ap;
 {
-	struct lock *l;
-
-	if ((l = (struct lock *)ap->a_vp->v_data) == NULL) {
-		if (ap->a_flags & LK_INTERLOCK)
-			simple_unlock(&ap->a_vp->v_interlock);
-		return 0;
-	}
+	struct vnode *vp = ap->a_vp;
 
-	return (lockmgr(l, ap->a_flags | LK_RELEASE, &ap->a_vp->v_interlock, 
+	return (lockmgr(&vp->v_lock, ap->a_flags | LK_RELEASE, &vp->v_interlock, 
 	    ap->a_p));
 }
 
@@ -274,12 +262,8 @@ vop_stdislocked(ap)
 		struct proc *a_p;
 	} */ *ap;
 {
-	struct lock *l;
 
-	if ((l = (struct lock *)ap->a_vp->v_data) == NULL)
-		return 0;
-
-	return (lockstatus(l, ap->a_p));
+	return (lockstatus(&ap->a_vp->v_lock, ap->a_p));
 }
 
 int
@@ -374,13 +358,6 @@ vop_sharedlock(ap)
 	struct vnode *vp = ap->a_vp;
 	int vnflags, flags = ap->a_flags;
 
-	if (vp->v_vnlock == NULL) {
-		if ((flags & LK_TYPE_MASK) == LK_DRAIN)
-			return (0);
-		MALLOC(vp->v_vnlock, struct lock *, sizeof(struct lock),
-		    M_VNODE, M_WAITOK);
-		lockinit(vp->v_vnlock, PVFS, "vnlock", 0, LK_NOPAUSE);
-	}
 	switch (flags & LK_TYPE_MASK) {
 	case LK_DRAIN:
 		vnflags = LK_DRAIN;
@@ -408,9 +385,9 @@ vop_sharedlock(ap)
 	if (flags & LK_INTERLOCK)
 		vnflags |= LK_INTERLOCK;
 #ifndef	DEBUG_LOCKS
-	return (lockmgr(vp->v_vnlock, vnflags, &vp->v_interlock, ap->a_p));
+	return (lockmgr(&vp->v_lock, vnflags, &vp->v_interlock, ap->a_p));
 #else
-	return (debuglockmgr(vp->v_vnlock, vnflags, &vp->v_interlock, ap->a_p,
+	return (debuglockmgr(&vp->v_lock, vnflags, &vp->v_interlock, ap->a_p,
 	    "vop_sharedlock", vp->filename, vp->line));
 #endif
 }
@@ -447,13 +424,6 @@ vop_nolock(ap)
 	struct vnode *vp = ap->a_vp;
 	int vnflags, flags = ap->a_flags;
 
-	if (vp->v_vnlock == NULL) {
-		if ((flags & LK_TYPE_MASK) == LK_DRAIN)
-			return (0);
-		MALLOC(vp->v_vnlock, struct lock *, sizeof(struct lock),
-		    M_VNODE, M_WAITOK);
-		lockinit(vp->v_vnlock, PVFS, "vnlock", 0, LK_NOPAUSE);
-	}
 	switch (flags & LK_TYPE_MASK) {
 	case LK_DRAIN:
 		vnflags = LK_DRAIN;
@@ -472,7 +442,7 @@ vop_nolock(ap)
 	}
 	if (flags & LK_INTERLOCK)
 		vnflags |= LK_INTERLOCK;
-	return(lockmgr(vp->v_vnlock, vnflags, &vp->v_interlock, ap->a_p));
+	return(lockmgr(&vp->v_lock, vnflags, &vp->v_interlock, ap->a_p));
 #else /* for now */
 	/*
 	 * Since we are not using the lock manager, we must clear
@@ -495,15 +465,14 @@ vop_nounlock(ap)
 		struct proc *a_p;
 	} */ *ap;
 {
-	struct vnode *vp = ap->a_vp;
 
-	if (vp->v_vnlock == NULL) {
-		if (ap->a_flags & LK_INTERLOCK)
-			simple_unlock(&ap->a_vp->v_interlock);
-		return (0);
-	}
-	return (lockmgr(vp->v_vnlock, LK_RELEASE | ap->a_flags,
-		&ap->a_vp->v_interlock, ap->a_p));
+	/*
+	 * Since we are not using the lock manager, we must clear
+	 * the interlock here.
+	 */
+	if (ap->a_flags & LK_INTERLOCK)
+		simple_unlock(&ap->a_vp->v_interlock);
+	return (0);
 }
 
 /*
@@ -516,11 +485,8 @@ vop_noislocked(ap)
 		struct proc *a_p;
 	} */ *ap;
 {
-	struct vnode *vp = ap->a_vp;
 
-	if (vp->v_vnlock == NULL)
-		return (0);
-	return (lockstatus(vp->v_vnlock, ap->a_p));
+	return (0);
 }
 
 /*
diff --git a/sys/kern/vfs_export.c b/sys/kern/vfs_export.c
index e0a758c..6d60d44 100644
--- a/sys/kern/vfs_export.c
+++ b/sys/kern/vfs_export.c
@@ -618,6 +618,7 @@ getnewvnode(tag, mp, vops, vpp)
 	vp->v_type = VNON;
 	vp->v_tag = tag;
 	vp->v_op = vops;
+	lockinit(&vp->v_lock, PVFS, "vnlock", 0, LK_NOPAUSE);
 	insmntque(vp, mp);
 	*vpp = vp;
 	vp->v_usecount = 1;
@@ -1373,6 +1374,9 @@ addaliasu(nvp, nvp_rdev)
 	ops = nvp->v_op;
 	nvp->v_op = ovp->v_op;
 	ovp->v_op = ops;
+	lockinit(&ovp->v_lock, PVFS, "vnlock", 0, LK_NOPAUSE);
+	if (nvp->v_vnlock)
+		ovp->v_vnlock = &ovp->v_lock;
 	insmntque(ovp, nvp->v_mount);
 	vrele(nvp);
 	vgone(nvp);
@@ -1773,10 +1777,7 @@ vclean(vp, flags, p)
 	}
 
 	cache_purge(vp);
-	if (vp->v_vnlock) {
-		FREE(vp->v_vnlock, M_VNODE);
-		vp->v_vnlock = NULL;
-	}
+	vp->v_vnlock = NULL;
 
 	if (VSHOULDFREE(vp))
 		vfree(vp);
diff --git a/sys/kern/vfs_subr.c b/sys/kern/vfs_subr.c
index e0a758c..6d60d44 100644
--- a/sys/kern/vfs_subr.c
+++ b/sys/kern/vfs_subr.c
@@ -618,6 +618,7 @@ getnewvnode(tag, mp, vops, vpp)
 	vp->v_type = VNON;
 	vp->v_tag = tag;
 	vp->v_op = vops;
+	lockinit(&vp->v_lock, PVFS, "vnlock", 0, LK_NOPAUSE);
 	insmntque(vp, mp);
 	*vpp = vp;
 	vp->v_usecount = 1;
@@ -1373,6 +1374,9 @@ addaliasu(nvp, nvp_rdev)
 	ops = nvp->v_op;
 	nvp->v_op = ovp->v_op;
 	ovp->v_op = ops;
+	lockinit(&ovp->v_lock, PVFS, "vnlock", 0, LK_NOPAUSE);
+	if (nvp->v_vnlock)
+		ovp->v_vnlock = &ovp->v_lock;
 	insmntque(ovp, nvp->v_mount);
 	vrele(nvp);
 	vgone(nvp);
@@ -1773,10 +1777,7 @@ vclean(vp, flags, p)
 	}
 
 	cache_purge(vp);
-	if (vp->v_vnlock) {
-		FREE(vp->v_vnlock, M_VNODE);
-		vp->v_vnlock = NULL;
-	}
+	vp->v_vnlock = NULL;
 
 	if (VSHOULDFREE(vp))
 		vfree(vp);
diff --git a/sys/nfs/nfs_vnops.c b/sys/nfs/nfs_vnops.c
index 91f9435..28fe55b 100644
--- a/sys/nfs/nfs_vnops.c
+++ b/sys/nfs/nfs_vnops.c
@@ -148,6 +148,7 @@ static struct vnodeopv_entry_desc nfsv2_vnodeop_entries[] = {
 	{ &vop_getpages_desc,		(vop_t *) nfs_getpages },
 	{ &vop_putpages_desc,		(vop_t *) nfs_putpages },
 	{ &vop_inactive_desc,		(vop_t *) nfs_inactive },
+	{ &vop_islocked_desc,		(vop_t *) vop_stdislocked },
 	{ &vop_lease_desc,		(vop_t *) vop_null },
 	{ &vop_link_desc,		(vop_t *) nfs_link },
 	{ &vop_lock_desc,		(vop_t *) vop_sharedlock },
@@ -168,6 +169,7 @@ static struct vnodeopv_entry_desc nfsv2_vnodeop_entries[] = {
 	{ &vop_setattr_desc,		(vop_t *) nfs_setattr },
 	{ &vop_strategy_desc,		(vop_t *) nfs_strategy },
 	{ &vop_symlink_desc,		(vop_t *) nfs_symlink },
+	{ &vop_unlock_desc,		(vop_t *) vop_stdunlock },
 	{ &vop_write_desc,		(vop_t *) nfs_write },
 	{ NULL, NULL }
 };
@@ -185,12 +187,14 @@ static struct vnodeopv_entry_desc nfsv2_specop_entries[] = {
 	{ &vop_close_desc,		(vop_t *) nfsspec_close },
 	{ &vop_fsync_desc,		(vop_t *) nfs_fsync },
 	{ &vop_getattr_desc,		(vop_t *) nfs_getattr },
+	{ &vop_islocked_desc,		(vop_t *) vop_stdislocked },
 	{ &vop_inactive_desc,		(vop_t *) nfs_inactive },
 	{ &vop_lock_desc,		(vop_t *) vop_sharedlock },
 	{ &vop_print_desc,		(vop_t *) nfs_print },
 	{ &vop_read_desc,		(vop_t *) nfsspec_read },
 	{ &vop_reclaim_desc,		(vop_t *) nfs_reclaim },
 	{ &vop_setattr_desc,		(vop_t *) nfs_setattr },
+	{ &vop_unlock_desc,		(vop_t *) vop_stdunlock },
 	{ &vop_write_desc,		(vop_t *) nfsspec_write },
 	{ NULL, NULL }
 };
diff --git a/sys/nfsclient/nfs_vnops.c b/sys/nfsclient/nfs_vnops.c
index 91f9435..28fe55b 100644
--- a/sys/nfsclient/nfs_vnops.c
+++ b/sys/nfsclient/nfs_vnops.c
@@ -148,6 +148,7 @@ static struct vnodeopv_entry_desc nfsv2_vnodeop_entries[] = {
 	{ &vop_getpages_desc,		(vop_t *) nfs_getpages },
 	{ &vop_putpages_desc,		(vop_t *) nfs_putpages },
 	{ &vop_inactive_desc,		(vop_t *) nfs_inactive },
+	{ &vop_islocked_desc,		(vop_t *) vop_stdislocked },
 	{ &vop_lease_desc,		(vop_t *) vop_null },
 	{ &vop_link_desc,		(vop_t *) nfs_link },
 	{ &vop_lock_desc,		(vop_t *) vop_sharedlock },
@@ -168,6 +169,7 @@ static struct vnodeopv_entry_desc nfsv2_vnodeop_entries[] = {
 	{ &vop_setattr_desc,		(vop_t *) nfs_setattr },
 	{ &vop_strategy_desc,		(vop_t *) nfs_strategy },
 	{ &vop_symlink_desc,		(vop_t *) nfs_symlink },
+	{ &vop_unlock_desc,		(vop_t *) vop_stdunlock },
 	{ &vop_write_desc,		(vop_t *) nfs_write },
 	{ NULL, NULL }
 };
@@ -185,12 +187,14 @@ static struct vnodeopv_entry_desc nfsv2_specop_entries[] = {
 	{ &vop_close_desc,		(vop_t *) nfsspec_close },
 	{ &vop_fsync_desc,		(vop_t *) nfs_fsync },
 	{ &vop_getattr_desc,		(vop_t *) nfs_getattr },
+	{ &vop_islocked_desc,		(vop_t *) vop_stdislocked },
 	{ &vop_inactive_desc,		(vop_t *) nfs_inactive },
 	{ &vop_lock_desc,		(vop_t *) vop_sharedlock },
 	{ &vop_print_desc,		(vop_t *) nfs_print },
 	{ &vop_read_desc,		(vop_t *) nfsspec_read },
 	{ &vop_reclaim_desc,		(vop_t *) nfs_reclaim },
 	{ &vop_setattr_desc,		(vop_t *) nfs_setattr },
+	{ &vop_unlock_desc,		(vop_t *) vop_stdunlock },
 	{ &vop_write_desc,		(vop_t *) nfsspec_write },
 	{ NULL, NULL }
 };
diff --git a/sys/sys/vnode.h b/sys/sys/vnode.h
index 10b2ad4..707df71 100644
--- a/sys/sys/vnode.h
+++ b/sys/sys/vnode.h
@@ -37,6 +37,7 @@
 #ifndef _SYS_VNODE_H_
 #define	_SYS_VNODE_H_
 
+#include <sys/lock.h>
 #include <sys/queue.h>
 #include <sys/select.h>
 #include <sys/uio.h>
@@ -115,7 +116,8 @@ struct vnode {
 	int	v_clen;				/* length of current cluster */
 	struct vm_object *v_object;		/* Place to store VM object */
 	struct	simplelock v_interlock;		/* lock on usecount and flag */
-	struct	lock *v_vnlock;			/* used for non-locking fs's */
+	struct	lock v_lock;			/* used if fs don't have one */
+	struct	lock *v_vnlock;			/* pointer to vnode lock */
 	enum	vtagtype v_tag;			/* type of underlying data */
 	void 	*v_data;			/* private data for fs */
 	LIST_HEAD(, namecache) v_cache_src;	/* Cache entries from us */
diff --git a/sys/ufs/ffs/ffs_vfsops.c b/sys/ufs/ffs/ffs_vfsops.c
index dc04eea..6806734 100644
--- a/sys/ufs/ffs/ffs_vfsops.c
+++ b/sys/ufs/ffs/ffs_vfsops.c
@@ -1056,7 +1056,11 @@ restart:
 		return (error);
 	}
 	bzero((caddr_t)ip, sizeof(struct inode));
-	lockinit(&ip->i_lock, PINOD, "inode", 0, LK_CANRECURSE);
+	/*
+	 * FFS supports lock sharing in the stack of vnodes
+	 */
+	vp->v_vnlock = &vp->v_lock;
+	lockinit(vp->v_vnlock, PINOD, "inode", 0, LK_CANRECURSE);
 	vp->v_data = ip;
 	ip->i_vnode = vp;
 	ip->i_fs = fs = ump->um_fs;
diff --git a/sys/ufs/ufs/inode.h b/sys/ufs/ufs/inode.h
index 6417a10..79e79b1 100644
--- a/sys/ufs/ufs/inode.h
+++ b/sys/ufs/ufs/inode.h
@@ -66,7 +66,6 @@ typedef long ufs_lbn_t;
  * active, and is put back when the file is no longer being used.
  */
 struct inode {
-	struct	 lock i_lock;	/* Inode lock. >Keep this first< */
 	LIST_ENTRY(inode) i_hash;/* Hash chain. */
 	struct	vnode  *i_vnode;/* Vnode associated with this inode. */
 	struct	vnode  *i_devvp;/* Vnode for block I/O. */
diff --git a/sys/ufs/ufs/ufs_ihash.c b/sys/ufs/ufs/ufs_ihash.c
index 0a7c7d5..36176f0 100644
--- a/sys/ufs/ufs/ufs_ihash.c
+++ b/sys/ufs/ufs/ufs_ihash.c
@@ -130,7 +130,7 @@ ufs_ihashins(ip)
 	struct ihashhead *ipp;
 
 	/* lock the inode, then put it on the appropriate hash list */
-	lockmgr(&ip->i_lock, LK_EXCLUSIVE, (struct simplelock *)0, p);
+	lockmgr(&ip->i_vnode->v_lock, LK_EXCLUSIVE, (struct simplelock *)0, p);
 
 	simple_lock(&ufs_ihash_slock);
 	ipp = INOHASH(ip->i_dev, ip->i_number);
diff --git a/sys/ufs/ufs/ufs_vnops.c b/sys/ufs/ufs/ufs_vnops.c
index 2b34d3c..fe8de37 100644
--- a/sys/ufs/ufs/ufs_vnops.c
+++ b/sys/ufs/ufs/ufs_vnops.c
@@ -1782,7 +1782,7 @@ ufs_print(ap)
 	    minor(ip->i_dev));
 	if (vp->v_type == VFIFO)
 		fifo_printinfo(vp);
-	lockmgr_printinfo(&ip->i_lock);
+	lockmgr_printinfo(&vp->v_lock);
 	printf("\n");
 	return (0);
 }