[PATCH] FUSE - read-write operations

This patch adds the write filesystem operations of FUSE.

The following operations are added:

 o setattr
 o symlink
 o mknod
 o mkdir
 o create
 o unlink
 o rmdir
 o rename
 o link

Signed-off-by: Miklos Szeredi <miklos@szeredi.hu>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

3 files changed, 378 insertions, 17 deletions

diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c
index a89730e..92c7188 100644
--- a/fs/fuse/dir.c
+++ b/fs/fuse/dir.c
@@ -42,7 +42,6 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
 		return 0;
 	else if (time_after(jiffies, entry->d_time)) {
 		int err;
-		int version;
 		struct fuse_entry_out outarg;
 		struct inode *inode = entry->d_inode;
 		struct fuse_inode *fi = get_fuse_inode(inode);
@@ -53,15 +52,19 @@ static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
 
 		fuse_lookup_init(req, entry->d_parent->d_inode, entry, &outarg);
 		request_send_nonint(fc, req);
-		version = req->out.h.unique;
 		err = req->out.h.error;
+		if (!err) {
+			if (outarg.nodeid != get_node_id(inode)) {
+				fuse_send_forget(fc, req, outarg.nodeid, 1);
+				return 0;
+			}
+			fi->nlookup ++;
+		}
 		fuse_put_request(fc, req);
-		if (err || outarg.nodeid != get_node_id(inode) ||
-		    (outarg.attr.mode ^ inode->i_mode) & S_IFMT)
+		if (err || (outarg.attr.mode ^ inode->i_mode) & S_IFMT)
 			return 0;
 
 		fuse_change_attributes(inode, &outarg.attr);
-		inode->i_version = version;
 		entry->d_time = time_to_jiffies(outarg.entry_valid,
 						outarg.entry_valid_nsec);
 		fi->i_time = time_to_jiffies(outarg.attr_valid,
@@ -78,7 +81,6 @@ static int fuse_lookup_iget(struct inode *dir, struct dentry *entry,
 			    struct inode **inodep)
 {
 	int err;
-	int version;
 	struct fuse_entry_out outarg;
 	struct inode *inode = NULL;
 	struct fuse_conn *fc = get_fuse_conn(dir);
@@ -93,13 +95,12 @@ static int fuse_lookup_iget(struct inode *dir, struct dentry *entry,
 
 	fuse_lookup_init(req, dir, entry, &outarg);
 	request_send(fc, req);
-	version = req->out.h.unique;
 	err = req->out.h.error;
 	if (!err) {
 		inode = fuse_iget(dir->i_sb, outarg.nodeid, outarg.generation,
-				  &outarg.attr, version);
+				  &outarg.attr);
 		if (!inode) {
-			fuse_send_forget(fc, req, outarg.nodeid, version);
+			fuse_send_forget(fc, req, outarg.nodeid, 1);
 			return -ENOMEM;
 		}
 	}
@@ -120,6 +121,264 @@ static int fuse_lookup_iget(struct inode *dir, struct dentry *entry,
 	return 0;
 }
 
+void fuse_invalidate_attr(struct inode *inode)
+{
+	get_fuse_inode(inode)->i_time = jiffies - 1;
+}
+
+static void fuse_invalidate_entry(struct dentry *entry)
+{
+	d_invalidate(entry);
+	entry->d_time = jiffies - 1;
+}
+
+static int create_new_entry(struct fuse_conn *fc, struct fuse_req *req,
+			    struct inode *dir, struct dentry *entry,
+			    int mode)
+{
+	struct fuse_entry_out outarg;
+	struct inode *inode;
+	struct fuse_inode *fi;
+	int err;
+
+	req->in.h.nodeid = get_node_id(dir);
+	req->inode = dir;
+	req->out.numargs = 1;
+	req->out.args[0].size = sizeof(outarg);
+	req->out.args[0].value = &outarg;
+	request_send(fc, req);
+	err = req->out.h.error;
+	if (err) {
+		fuse_put_request(fc, req);
+		return err;
+	}
+	inode = fuse_iget(dir->i_sb, outarg.nodeid, outarg.generation,
+			  &outarg.attr);
+	if (!inode) {
+		fuse_send_forget(fc, req, outarg.nodeid, 1);
+		return -ENOMEM;
+	}
+	fuse_put_request(fc, req);
+
+	/* Don't allow userspace to do really stupid things... */
+	if ((inode->i_mode ^ mode) & S_IFMT) {
+		iput(inode);
+		return -EIO;
+	}
+
+	entry->d_time = time_to_jiffies(outarg.entry_valid,
+					outarg.entry_valid_nsec);
+
+	fi = get_fuse_inode(inode);
+	fi->i_time = time_to_jiffies(outarg.attr_valid,
+				     outarg.attr_valid_nsec);
+
+	d_instantiate(entry, inode);
+	fuse_invalidate_attr(dir);
+	return 0;
+}
+
+static int fuse_mknod(struct inode *dir, struct dentry *entry, int mode,
+		      dev_t rdev)
+{
+	struct fuse_mknod_in inarg;
+	struct fuse_conn *fc = get_fuse_conn(dir);
+	struct fuse_req *req = fuse_get_request(fc);
+	if (!req)
+		return -ERESTARTNOINTR;
+
+	memset(&inarg, 0, sizeof(inarg));
+	inarg.mode = mode;
+	inarg.rdev = new_encode_dev(rdev);
+	req->in.h.opcode = FUSE_MKNOD;
+	req->in.numargs = 2;
+	req->in.args[0].size = sizeof(inarg);
+	req->in.args[0].value = &inarg;
+	req->in.args[1].size = entry->d_name.len + 1;
+	req->in.args[1].value = entry->d_name.name;
+	return create_new_entry(fc, req, dir, entry, mode);
+}
+
+static int fuse_create(struct inode *dir, struct dentry *entry, int mode,
+		       struct nameidata *nd)
+{
+	return fuse_mknod(dir, entry, mode, 0);
+}
+
+static int fuse_mkdir(struct inode *dir, struct dentry *entry, int mode)
+{
+	struct fuse_mkdir_in inarg;
+	struct fuse_conn *fc = get_fuse_conn(dir);
+	struct fuse_req *req = fuse_get_request(fc);
+	if (!req)
+		return -ERESTARTNOINTR;
+
+	memset(&inarg, 0, sizeof(inarg));
+	inarg.mode = mode;
+	req->in.h.opcode = FUSE_MKDIR;
+	req->in.numargs = 2;
+	req->in.args[0].size = sizeof(inarg);
+	req->in.args[0].value = &inarg;
+	req->in.args[1].size = entry->d_name.len + 1;
+	req->in.args[1].value = entry->d_name.name;
+	return create_new_entry(fc, req, dir, entry, S_IFDIR);
+}
+
+static int fuse_symlink(struct inode *dir, struct dentry *entry,
+			const char *link)
+{
+	struct fuse_conn *fc = get_fuse_conn(dir);
+	unsigned len = strlen(link) + 1;
+	struct fuse_req *req;
+
+	if (len > FUSE_SYMLINK_MAX)
+		return -ENAMETOOLONG;
+
+	req = fuse_get_request(fc);
+	if (!req)
+		return -ERESTARTNOINTR;
+
+	req->in.h.opcode = FUSE_SYMLINK;
+	req->in.numargs = 2;
+	req->in.args[0].size = entry->d_name.len + 1;
+	req->in.args[0].value = entry->d_name.name;
+	req->in.args[1].size = len;
+	req->in.args[1].value = link;
+	return create_new_entry(fc, req, dir, entry, S_IFLNK);
+}
+
+static int fuse_unlink(struct inode *dir, struct dentry *entry)
+{
+	int err;
+	struct fuse_conn *fc = get_fuse_conn(dir);
+	struct fuse_req *req = fuse_get_request(fc);
+	if (!req)
+		return -ERESTARTNOINTR;
+
+	req->in.h.opcode = FUSE_UNLINK;
+	req->in.h.nodeid = get_node_id(dir);
+	req->inode = dir;
+	req->in.numargs = 1;
+	req->in.args[0].size = entry->d_name.len + 1;
+	req->in.args[0].value = entry->d_name.name;
+	request_send(fc, req);
+	err = req->out.h.error;
+	fuse_put_request(fc, req);
+	if (!err) {
+		struct inode *inode = entry->d_inode;
+
+		/* Set nlink to zero so the inode can be cleared, if
+                   the inode does have more links this will be
+                   discovered at the next lookup/getattr */
+		inode->i_nlink = 0;
+		fuse_invalidate_attr(inode);
+		fuse_invalidate_attr(dir);
+	} else if (err == -EINTR)
+		fuse_invalidate_entry(entry);
+	return err;
+}
+
+static int fuse_rmdir(struct inode *dir, struct dentry *entry)
+{
+	int err;
+	struct fuse_conn *fc = get_fuse_conn(dir);
+	struct fuse_req *req = fuse_get_request(fc);
+	if (!req)
+		return -ERESTARTNOINTR;
+
+	req->in.h.opcode = FUSE_RMDIR;
+	req->in.h.nodeid = get_node_id(dir);
+	req->inode = dir;
+	req->in.numargs = 1;
+	req->in.args[0].size = entry->d_name.len + 1;
+	req->in.args[0].value = entry->d_name.name;
+	request_send(fc, req);
+	err = req->out.h.error;
+	fuse_put_request(fc, req);
+	if (!err) {
+		entry->d_inode->i_nlink = 0;
+		fuse_invalidate_attr(dir);
+	} else if (err == -EINTR)
+		fuse_invalidate_entry(entry);
+	return err;
+}
+
+static int fuse_rename(struct inode *olddir, struct dentry *oldent,
+		       struct inode *newdir, struct dentry *newent)
+{
+	int err;
+	struct fuse_rename_in inarg;
+	struct fuse_conn *fc = get_fuse_conn(olddir);
+	struct fuse_req *req = fuse_get_request(fc);
+	if (!req)
+		return -ERESTARTNOINTR;
+
+	memset(&inarg, 0, sizeof(inarg));
+	inarg.newdir = get_node_id(newdir);
+	req->in.h.opcode = FUSE_RENAME;
+	req->in.h.nodeid = get_node_id(olddir);
+	req->inode = olddir;
+	req->inode2 = newdir;
+	req->in.numargs = 3;
+	req->in.args[0].size = sizeof(inarg);
+	req->in.args[0].value = &inarg;
+	req->in.args[1].size = oldent->d_name.len + 1;
+	req->in.args[1].value = oldent->d_name.name;
+	req->in.args[2].size = newent->d_name.len + 1;
+	req->in.args[2].value = newent->d_name.name;
+	request_send(fc, req);
+	err = req->out.h.error;
+	fuse_put_request(fc, req);
+	if (!err) {
+		fuse_invalidate_attr(olddir);
+		if (olddir != newdir)
+			fuse_invalidate_attr(newdir);
+	} else if (err == -EINTR) {
+		/* If request was interrupted, DEITY only knows if the
+		   rename actually took place.  If the invalidation
+		   fails (e.g. some process has CWD under the renamed
+		   directory), then there can be inconsistency between
+		   the dcache and the real filesystem.  Tough luck. */
+		fuse_invalidate_entry(oldent);
+		if (newent->d_inode)
+			fuse_invalidate_entry(newent);
+	}
+
+	return err;
+}
+
+static int fuse_link(struct dentry *entry, struct inode *newdir,
+		     struct dentry *newent)
+{
+	int err;
+	struct fuse_link_in inarg;
+	struct inode *inode = entry->d_inode;
+	struct fuse_conn *fc = get_fuse_conn(inode);
+	struct fuse_req *req = fuse_get_request(fc);
+	if (!req)
+		return -ERESTARTNOINTR;
+
+	memset(&inarg, 0, sizeof(inarg));
+	inarg.oldnodeid = get_node_id(inode);
+	req->in.h.opcode = FUSE_LINK;
+	req->inode2 = inode;
+	req->in.numargs = 2;
+	req->in.args[0].size = sizeof(inarg);
+	req->in.args[0].value = &inarg;
+	req->in.args[1].size = newent->d_name.len + 1;
+	req->in.args[1].value = newent->d_name.name;
+	err = create_new_entry(fc, req, newdir, newent, inode->i_mode);
+	/* Contrary to "normal" filesystems it can happen that link
+	   makes two "logical" inodes point to the same "physical"
+	   inode.  We invalidate the attributes of the old one, so it
+	   will reflect changes in the backing inode (link count,
+	   etc.)
+	*/
+	if (!err || err == -EINTR)
+		fuse_invalidate_attr(inode);
+	return err;
+}
+
 int fuse_do_getattr(struct inode *inode)
 {
 	int err;
@@ -341,6 +600,91 @@ static int fuse_dir_release(struct inode *inode, struct file *file)
 	return 0;
 }
 
+static unsigned iattr_to_fattr(struct iattr *iattr, struct fuse_attr *fattr)
+{
+	unsigned ivalid = iattr->ia_valid;
+	unsigned fvalid = 0;
+
+	memset(fattr, 0, sizeof(*fattr));
+
+	if (ivalid & ATTR_MODE)
+		fvalid |= FATTR_MODE,   fattr->mode = iattr->ia_mode;
+	if (ivalid & ATTR_UID)
+		fvalid |= FATTR_UID,    fattr->uid = iattr->ia_uid;
+	if (ivalid & ATTR_GID)
+		fvalid |= FATTR_GID,    fattr->gid = iattr->ia_gid;
+	if (ivalid & ATTR_SIZE)
+		fvalid |= FATTR_SIZE,   fattr->size = iattr->ia_size;
+	/* You can only _set_ these together (they may change by themselves) */
+	if ((ivalid & (ATTR_ATIME | ATTR_MTIME)) == (ATTR_ATIME | ATTR_MTIME)) {
+		fvalid |= FATTR_ATIME | FATTR_MTIME;
+		fattr->atime = iattr->ia_atime.tv_sec;
+		fattr->mtime = iattr->ia_mtime.tv_sec;
+	}
+
+	return fvalid;
+}
+
+static int fuse_setattr(struct dentry *entry, struct iattr *attr)
+{
+	struct inode *inode = entry->d_inode;
+	struct fuse_conn *fc = get_fuse_conn(inode);
+	struct fuse_inode *fi = get_fuse_inode(inode);
+	struct fuse_req *req;
+	struct fuse_setattr_in inarg;
+	struct fuse_attr_out outarg;
+	int err;
+	int is_truncate = 0;
+
+	if (attr->ia_valid & ATTR_SIZE) {
+		unsigned long limit;
+		is_truncate = 1;
+		limit = current->signal->rlim[RLIMIT_FSIZE].rlim_cur;
+		if (limit != RLIM_INFINITY && attr->ia_size > (loff_t) limit) {
+			send_sig(SIGXFSZ, current, 0);
+			return -EFBIG;
+		}
+	}
+
+	req = fuse_get_request(fc);
+	if (!req)
+		return -ERESTARTNOINTR;
+
+	memset(&inarg, 0, sizeof(inarg));
+	inarg.valid = iattr_to_fattr(attr, &inarg.attr);
+	req->in.h.opcode = FUSE_SETATTR;
+	req->in.h.nodeid = get_node_id(inode);
+	req->inode = inode;
+	req->in.numargs = 1;
+	req->in.args[0].size = sizeof(inarg);
+	req->in.args[0].value = &inarg;
+	req->out.numargs = 1;
+	req->out.args[0].size = sizeof(outarg);
+	req->out.args[0].value = &outarg;
+	request_send(fc, req);
+	err = req->out.h.error;
+	fuse_put_request(fc, req);
+	if (!err) {
+		if ((inode->i_mode ^ outarg.attr.mode) & S_IFMT) {
+			make_bad_inode(inode);
+			err = -EIO;
+		} else {
+			if (is_truncate) {
+				loff_t origsize = i_size_read(inode);
+				i_size_write(inode, outarg.attr.size);
+				if (origsize > outarg.attr.size)
+					vmtruncate(inode, outarg.attr.size);
+			}
+			fuse_change_attributes(inode, &outarg.attr);
+			fi->i_time = time_to_jiffies(outarg.attr_valid,
+						     outarg.attr_valid_nsec);
+		}
+	} else if (err == -EINTR)
+		fuse_invalidate_attr(inode);
+
+	return err;
+}
+
 static int fuse_getattr(struct vfsmount *mnt, struct dentry *entry,
 			struct kstat *stat)
 {
@@ -373,6 +717,15 @@ static struct dentry *fuse_lookup(struct inode *dir, struct dentry *entry,
 
 static struct inode_operations fuse_dir_inode_operations = {
 	.lookup		= fuse_lookup,
+	.mkdir		= fuse_mkdir,
+	.symlink	= fuse_symlink,
+	.unlink		= fuse_unlink,
+	.rmdir		= fuse_rmdir,
+	.rename		= fuse_rename,
+	.link		= fuse_link,
+	.setattr	= fuse_setattr,
+	.create		= fuse_create,
+	.mknod		= fuse_mknod,
 	.permission	= fuse_permission,
 	.getattr	= fuse_getattr,
 };
@@ -385,11 +738,13 @@ static struct file_operations fuse_dir_operations = {
 };
 
 static struct inode_operations fuse_common_inode_operations = {
+	.setattr	= fuse_setattr,
 	.permission	= fuse_permission,
 	.getattr	= fuse_getattr,
 };
 
 static struct inode_operations fuse_symlink_inode_operations = {
+	.setattr	= fuse_setattr,
 	.follow_link	= fuse_follow_link,
 	.put_link	= fuse_put_link,
 	.readlink	= generic_readlink,
diff --git a/fs/fuse/fuse_i.h b/fs/fuse/fuse_i.h
index 8d91e14..87d25b8 100644
--- a/fs/fuse/fuse_i.h
+++ b/fs/fuse/fuse_i.h
@@ -30,6 +30,9 @@ struct fuse_inode {
 	 * and kernel */
 	u64 nodeid;
 
+	/** Number of lookups on this inode */
+	u64 nlookup;
+
 	/** The request used for sending the FORGET message */
 	struct fuse_req *forget_req;
 
@@ -252,13 +255,13 @@ extern spinlock_t fuse_lock;
  * Get a filled in inode
  */
 struct inode *fuse_iget(struct super_block *sb, unsigned long nodeid,
-			int generation, struct fuse_attr *attr, int version);
+			int generation, struct fuse_attr *attr);
 
 /**
  * Send FORGET command
  */
 void fuse_send_forget(struct fuse_conn *fc, struct fuse_req *req,
-		      unsigned long nodeid, int version);
+		      unsigned long nodeid, u64 nlookup);
 
 /**
  * Initialise inode operations on regular files and special files
diff --git a/fs/fuse/inode.c b/fs/fuse/inode.c
index 41498a1..fa03f80 100644
--- a/fs/fuse/inode.c
+++ b/fs/fuse/inode.c
@@ -51,6 +51,7 @@ static struct inode *fuse_alloc_inode(struct super_block *sb)
 	fi = get_fuse_inode(inode);
 	fi->i_time = jiffies - 1;
 	fi->nodeid = 0;
+	fi->nlookup = 0;
 	fi->forget_req = fuse_request_alloc();
 	if (!fi->forget_req) {
 		kmem_cache_free(fuse_inode_cachep, inode);
@@ -74,10 +75,10 @@ static void fuse_read_inode(struct inode *inode)
 }
 
 void fuse_send_forget(struct fuse_conn *fc, struct fuse_req *req,
-		      unsigned long nodeid, int version)
+		      unsigned long nodeid, u64 nlookup)
 {
 	struct fuse_forget_in *inarg = &req->misc.forget_in;
-	inarg->version = version;
+	inarg->nlookup = nlookup;
 	req->in.h.opcode = FUSE_FORGET;
 	req->in.h.nodeid = nodeid;
 	req->in.numargs = 1;
@@ -91,7 +92,7 @@ static void fuse_clear_inode(struct inode *inode)
 	struct fuse_conn *fc = get_fuse_conn(inode);
 	if (fc) {
 		struct fuse_inode *fi = get_fuse_inode(inode);
-		fuse_send_forget(fc, fi->forget_req, fi->nodeid, inode->i_version);
+		fuse_send_forget(fc, fi->forget_req, fi->nodeid, fi->nlookup);
 		fi->forget_req = NULL;
 	}
 }
@@ -156,9 +157,10 @@ static int fuse_inode_set(struct inode *inode, void *_nodeidp)
 }
 
 struct inode *fuse_iget(struct super_block *sb, unsigned long nodeid,
-			int generation, struct fuse_attr *attr, int version)
+			int generation, struct fuse_attr *attr)
 {
 	struct inode *inode;
+	struct fuse_inode *fi;
 	struct fuse_conn *fc = get_fuse_conn_super(sb);
 	int retried = 0;
 
@@ -181,8 +183,9 @@ struct inode *fuse_iget(struct super_block *sb, unsigned long nodeid,
 		goto retry;
 	}
 
+	fi = get_fuse_inode(inode);
+	fi->nlookup ++;
 	fuse_change_attributes(inode, attr);
-	inode->i_version = version;
 	return inode;
 }
 
@@ -389,7 +392,7 @@ static struct inode *get_root_inode(struct super_block *sb, unsigned mode)
 
 	attr.mode = mode;
 	attr.ino = FUSE_ROOT_ID;
-	return fuse_iget(sb, 1, 0, &attr, 0);
+	return fuse_iget(sb, 1, 0, &attr);
 }
 
 static struct super_operations fuse_super_operations = {