vfs_intent-2.6-fc5.patch

lustre 1.6.5 source code

Index: linux-2.6.15-fc5/fs/inode.c
===================================================================
--- linux-2.6.15-fc5.orig/fs/inode.c	2006-05-31 04:08:19.000000000 +0800
+++ linux-2.6.15-fc5/fs/inode.c	2006-05-31 17:24:35.000000000 +0800
@@ -236,6 +236,7 @@
 	inodes_stat.nr_unused--;
 }
 
+EXPORT_SYMBOL(__iget);
 /**
  * clear_inode - clear an inode
  * @inode: inode to clear
Index: linux-2.6.15-fc5/fs/open.c
===================================================================
--- linux-2.6.15-fc5.orig/fs/open.c	2006-05-31 04:08:39.000000000 +0800
+++ linux-2.6.15-fc5/fs/open.c	2006-05-31 17:24:35.000000000 +0800
@@ -224,12 +224,12 @@
 	struct nameidata nd;
 	struct inode * inode;
 	int error;
-
+	intent_init(&nd.intent, IT_GETATTR);
 	error = -EINVAL;
 	if (length < 0)	/* sorry, but loff_t says... */
 		goto out;
 
-	error = user_path_walk(path, &nd);
+	error = user_path_walk_it(path, &nd);
 	if (error)
 		goto out;
 	inode = nd.dentry->d_inode;
@@ -488,6 +488,7 @@
 	int old_fsuid, old_fsgid;
 	kernel_cap_t old_cap;
 	int res;
+	intent_init(&nd.intent, IT_GETATTR);
 
 	if (mode & ~S_IRWXO)	/* where's F_OK, X_OK, W_OK, R_OK? */
 		return -EINVAL;
@@ -512,7 +513,7 @@
 	else
 		current->cap_effective = current->cap_permitted;
 
-	res = __user_walk_fd(dfd, filename, LOOKUP_FOLLOW|LOOKUP_ACCESS, &nd);
+	res = __user_walk_fd_it(dfd, filename, LOOKUP_FOLLOW|LOOKUP_ACCESS, &nd);
 	if (!res) {
 		res = vfs_permission(&nd, mode);
 		/* SuS v2 requires we report a read only fs too */
@@ -538,8 +539,9 @@
 {
 	struct nameidata nd;
 	int error;
+	intent_init(&nd.intent, IT_GETATTR);
 
-	error = __user_walk(filename, LOOKUP_FOLLOW|LOOKUP_DIRECTORY, &nd);
+	error = __user_walk_it(filename, LOOKUP_FOLLOW|LOOKUP_DIRECTORY, &nd);
 	if (error)
 		goto out;
 
@@ -591,8 +593,9 @@
 {
 	struct nameidata nd;
 	int error;
+	intent_init(&nd.intent, IT_GETATTR);
 
-	error = __user_walk(filename, LOOKUP_FOLLOW | LOOKUP_DIRECTORY | LOOKUP_NOALT, &nd);
+	error = __user_walk_it(filename, LOOKUP_FOLLOW | LOOKUP_DIRECTORY | LOOKUP_NOALT, &nd);
 	if (error)
 		goto out;
 
@@ -815,6 +818,7 @@
 		error = open(inode, f);
 		if (error)
 			goto cleanup_all;
+		intent_release(f->f_it);
 	}
 
 	f->f_flags &= ~(O_CREAT | O_EXCL | O_NOCTTY | O_TRUNC);
@@ -841,6 +845,7 @@
 	f->f_dentry = NULL;
 	f->f_vfsmnt = NULL;
 cleanup_file:
+	intent_release(f->f_it);
 	put_filp(f);
 	dput(dentry);
 	mntput(mnt);
@@ -866,6 +871,7 @@
 {
 	int namei_flags, error;
 	struct nameidata nd;
+	intent_init(&nd.intent, IT_OPEN);
 
 	namei_flags = flags;
 	if ((namei_flags+1) & O_ACCMODE)
@@ -902,19 +908,19 @@
 struct file *lookup_instantiate_filp(struct nameidata *nd, struct dentry *dentry,
 		int (*open)(struct inode *, struct file *))
 {
-	if (IS_ERR(nd->intent.open.file))
+	if (IS_ERR(nd->intent.file))
 		goto out;
 	if (IS_ERR(dentry))
 		goto out_err;
-	nd->intent.open.file = __dentry_open(dget(dentry), mntget(nd->mnt),
-					     nd->intent.open.flags - 1,
-					     nd->intent.open.file,
+	nd->intent.file = __dentry_open(dget(dentry), mntget(nd->mnt),
+					     nd->intent.flags - 1,
+					     nd->intent.file,
 					     open);
 out:
-	return nd->intent.open.file;
+	return nd->intent.file;
 out_err:
 	release_open_intent(nd);
-	nd->intent.open.file = (struct file *)dentry;
+	nd->intent.file = (struct file *)dentry;
 	goto out;
 }
 EXPORT_SYMBOL_GPL(lookup_instantiate_filp);
@@ -931,7 +937,8 @@
 	struct file *filp;
 
 	/* Pick up the filp from the open intent */
-	filp = nd->intent.open.file;
+	filp = nd->intent.file;
+	filp->f_it = &nd->intent; 
 	/* Has the filesystem initialised the file for us? */
 	if (filp->f_dentry == NULL)
 		filp = __dentry_open(nd->dentry, nd->mnt, flags, filp, NULL);
Index: linux-2.6.15-fc5/fs/nfsctl.c
===================================================================
--- linux-2.6.15-fc5.orig/fs/nfsctl.c	2006-05-31 04:08:20.000000000 +0800
+++ linux-2.6.15-fc5/fs/nfsctl.c	2006-05-31 17:24:35.000000000 +0800
@@ -26,6 +26,7 @@
 	struct nameidata nd;
 	int error;
 
+	intent_init(&nd.intent, IT_OPEN);
 	nd.mnt = do_kern_mount("nfsd", 0, "nfsd", NULL);
 
 	if (IS_ERR(nd.mnt))
Index: linux-2.6.15-fc5/fs/namei.c
===================================================================
--- linux-2.6.15-fc5.orig/fs/namei.c	2006-05-31 04:08:33.000000000 +0800
+++ linux-2.6.15-fc5/fs/namei.c	2006-05-31 17:24:35.000000000 +0800
@@ -337,8 +337,19 @@
 	return 0;
 }
 
+void intent_release(struct lookup_intent *it)
+{
+	if (!it)
+		return;
+	if (it->it_magic != INTENT_MAGIC)
+		return;
+	if (it->it_op_release)
+		it->it_op_release(it);
+}
+
 void path_release(struct nameidata *nd)
 {
+	intent_release(&nd->intent);
 	dput(nd->dentry);
 	mntput(nd->mnt);
 }
@@ -359,10 +370,10 @@
  */
 void release_open_intent(struct nameidata *nd)
 {
-	if (nd->intent.open.file->f_dentry == NULL)
-		put_filp(nd->intent.open.file);
+	if (nd->intent.file->f_dentry == NULL)
+		put_filp(nd->intent.file);
 	else
-		fput(nd->intent.open.file);
+		fput(nd->intent.file);
 }
 
 /*
@@ -440,8 +451,12 @@
 {
 	struct dentry * result;
 	struct inode *dir = parent->d_inode;
-
+	int counter = 0;
+	
 	mutex_lock(&dir->i_mutex);
+again:
+	counter++;
+
 	/*
 	 * First re-do the cached lookup just in case it was created
 	 * while we waited for the directory semaphore..
@@ -475,13 +490,16 @@
 	 * Uhhuh! Nasty case: the cache was re-populated while
 	 * we waited on the semaphore. Need to revalidate.
 	 */
-	mutex_unlock(&dir->i_mutex);
 	if (result->d_op && result->d_op->d_revalidate) {
 		if (!result->d_op->d_revalidate(result, nd) && !d_invalidate(result)) {
 			dput(result);
-			result = ERR_PTR(-ENOENT);
+			if (counter > 10)
+				result = ERR_PTR(-ESTALE);
+			if (!IS_ERR(result))
+				goto again;
 		}
 	}
+	mutex_unlock(&dir->i_mutex);
 	return result;
 }
 
@@ -509,7 +527,9 @@
 static __always_inline int __vfs_follow_link(struct nameidata *nd, const char *link)
 {
 	int res = 0;
+	struct lookup_intent it = nd->intent;
 	char *name;
+
 	if (IS_ERR(link))
 		goto fail;
 
@@ -519,6 +539,10 @@
 			/* weird __emul_prefix() stuff did it */
 			goto out;
 	}
+	intent_init(&nd->intent, it.it_op);
+	nd->intent.it_flags = it.it_flags;
+	nd->intent.it_create_mode = it.it_create_mode;
+	nd->intent.file = it.file;
 	res = link_path_walk(link, nd);
 out:
 	if (nd->depth || res || nd->last_type!=LAST_NORM)
@@ -772,6 +796,33 @@
 	return PTR_ERR(dentry);
 }
 
+static int revalidate_special(struct nameidata *nd)
+{
+	struct dentry *dentry = nd->dentry;
+	int err, counter = 0;
+
+ revalidate_again:
+	if (!dentry->d_op || !dentry->d_op->d_revalidate)
+		return 0;
+	if (!dentry->d_op->d_revalidate(dentry, nd)) {
+		struct dentry *new;
+		if ((err = permission(dentry->d_parent->d_inode, MAY_EXEC, nd)))
+			return err;
+		new = real_lookup(dentry->d_parent, &dentry->d_name, nd);
+		if (IS_ERR(new))
+			return PTR_ERR(new);
+		d_invalidate(dentry);
+		dput(dentry);
+		nd->dentry = dentry = new;
+		counter++;
+		if (counter < 10)
+			goto revalidate_again;
+		//printk("excessive revalidate_it loops\n");
+		return -ESTALE;
+	}
+	return 0;
+}
+
 /*
  * Name resolution.
  * This is the basic name resolution function, turning a pathname into
@@ -870,7 +921,11 @@
 			goto out_dput;
 
 		if (inode->i_op->follow_link) {
+			int save_flags = nd->flags;
+			nd->flags |= LOOKUP_LINK_NOTLAST;
 			err = do_follow_link(&next, nd);
+			if (!(save_flags & LOOKUP_LINK_NOTLAST))
+				nd->flags &= ~LOOKUP_LINK_NOTLAST;
 			if (err)
 				goto return_err;
 			err = -ENOENT;
@@ -905,6 +960,23 @@
 				inode = nd->dentry->d_inode;
 				/* fallthrough */
 			case 1:
+				nd->flags |= LOOKUP_LAST;
+				err = revalidate_special(nd);
+				nd->flags &= ~LOOKUP_LAST;
+				if (!nd->dentry->d_inode)
+					err = -ENOENT;
+				if (err) {
+					path_release(nd);
+					goto return_err;
+				}
+				if (lookup_flags & LOOKUP_DIRECTORY) {
+					err = -ENOTDIR;
+					if(!nd->dentry->d_inode->i_op ||
+					  !nd->dentry->d_inode->i_op->lookup) {
+						path_release(nd);
+						goto return_err;
+					}
+				}
 				goto return_reval;
 		}
 		if (nd->dentry->d_op && nd->dentry->d_op->d_hash) {
@@ -912,7 +984,9 @@
 			if (err < 0)
 				break;
 		}
+		nd->flags |= LOOKUP_LAST;
 		err = do_lookup(nd, &this, &next, atomic);
+		nd->flags &= ~LOOKUP_LAST;
 		if (err)
 			break;
 		inode = next.dentry->d_inode;
@@ -1154,13 +1228,13 @@
 
 	if (filp == NULL)
 		return -ENFILE;
-	nd->intent.open.file = filp;
-	nd->intent.open.flags = open_flags;
-	nd->intent.open.create_mode = create_mode;
+	nd->intent.file = filp;
+	nd->intent.flags = open_flags;
+	nd->intent.create_mode = create_mode;
 	err = do_path_lookup(dfd, name, lookup_flags|LOOKUP_OPEN, nd);
-	if (IS_ERR(nd->intent.open.file)) {
+	if (IS_ERR(nd->intent.file)) {
 		if (err == 0) {
-			err = PTR_ERR(nd->intent.open.file);
+			err = PTR_ERR(nd->intent.file);
 			path_release(nd);
 		}
 	} else if (err != 0)
@@ -1263,7 +1337,7 @@
 }
 
 /* SMP-safe */
-struct dentry * lookup_one_len(const char * name, struct dentry * base, int len)
+struct dentry * lookup_one_len_it(const char * name, struct dentry * base, int len, struct nameidata *nd)
 {
 	unsigned long hash;
 	struct qstr this;
@@ -1283,11 +1357,17 @@
 	}
 	this.hash = end_name_hash(hash);
 
-	return __lookup_hash(&this, base, NULL);
+	return __lookup_hash(&this, base, nd);
 access:
 	return ERR_PTR(-EACCES);
 }
 
+struct dentry * lookup_one_len(const char * name, struct dentry * base, int len)
+{
+ 	return lookup_one_len_it(name, base, len, NULL);
+}
+
+
 /*
  *	namei()
  *
@@ -1299,8 +1379,9 @@
  * that namei follows links, while lnamei does not.
  * SMP-safe
  */
-int fastcall __user_walk_fd(int dfd, const char __user *name, unsigned flags,
-			    struct nameidata *nd)
+
+int fastcall __user_walk_fd_it(int dfd, const char __user *name, unsigned flags,
+			       struct nameidata *nd)
 {
 	char *tmp = getname(name);
 	int err = PTR_ERR(tmp);
@@ -1312,9 +1393,22 @@
 	return err;
 }
 
+int fastcall __user_walk_fd(int dfd, const char __user *name, unsigned flags,
+			    struct nameidata *nd)
+{
+	intent_init(&nd->intent, IT_LOOKUP);
+	return __user_walk_fd_it(dfd, name, flags, nd);	
+}
+
+int fastcall __user_walk_it(const char __user *name, unsigned flags, struct nameidata *nd)
+{
+	return __user_walk_fd_it(AT_FDCWD, name, flags, nd);
+}
+
 int fastcall __user_walk(const char __user *name, unsigned flags, struct nameidata *nd)
 {
-	return __user_walk_fd(AT_FDCWD, name, flags, nd);
+	intent_init(&nd->intent, IT_LOOKUP);
+	return __user_walk_it(name, flags, nd);
 }
 
 /*
@@ -1596,6 +1690,8 @@
 	if (flag & O_APPEND)
 		acc_mode |= MAY_APPEND;
 
+ 	nd->intent.it_flags = flag;
+ 	nd->intent.it_create_mode = mode;
 	/*
 	 * The simplest case - just a plain lookup.
 	 */
@@ -1610,6 +1706,7 @@
 	/*
 	 * Create - we need to know the parent.
 	 */
+ 	nd->intent.it_op |= IT_CREAT;
 	error = path_lookup_create(dfd,pathname,LOOKUP_PARENT,nd,flag,mode);
 	if (error)
 		return error;