Message-Id: <12109484023300-git-send-email-hooanon05@yahoo.co.jp>
Date:	Fri, 16 May 2008 23:33:09 +0900
From:	hooanon05@...oo.co.jp
To:	linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org
Cc:	Junjiro Okajima <hooanon05@...oo.co.jp>
Subject: [PATCH 55/67] aufs export via nfs

From: Junjiro Okajima <hooanon05@...oo.co.jp>

	initial commit
	aufs export via nfs

Signed-off-by: Junjiro Okajima <hooanon05@...oo.co.jp>
---
 fs/aufs/export.c |  597 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 1 files changed, 597 insertions(+), 0 deletions(-)

diff --git a/fs/aufs/export.c b/fs/aufs/export.c
new file mode 100644
index 0000000..edcbb6f
--- /dev/null
+++ b/fs/aufs/export.c
@@ -0,0 +1,597 @@
+/*
+ * Copyright (C) 2007-2008 Junjiro Okajima
+ *
+ * This program, aufs is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+/*
+ * export via nfs
+ *
+ * $Id: export.c,v 1.4 2008/05/04 23:52:08 sfjro Exp $
+ */
+
+#include <linux/exportfs.h>
+#include "aufs.h"
+
+union conv {
+#if BITS_PER_LONG == 32
+	__u32 a[1];
+#else
+	__u32 a[2];
+#endif
+	ino_t ino;
+};
+
+static ino_t decode_ino(__u32 *a)
+{
+	union conv u;
+
+	u.a[0] = a[0];
+#if BITS_PER_LONG == 64
+	u.a[1] = a[1];
+#endif
+	return u.ino;
+}
+
+static void encode_ino(__u32 *a, ino_t ino)
+{
+	union conv u;
+
+	u.ino = ino;
+	a[0] = u.a[0];
+#if BITS_PER_LONG == 64
+	a[1] = u.a[1];
+#endif
+}
+
+/* NFS file handle */
+enum {
+	Fh_br_id,
+	Fh_sigen,
+#if BITS_PER_LONG == 64
+	/* support 64bit inode number */
+	Fh_ino1,
+	Fh_ino2,
+	Fh_dir_ino1,
+	Fh_dir_ino2,
+	Fh_h_ino1,
+	Fh_h_ino2,
+#else
+	Fh_ino1,
+	Fh_dir_ino1,
+	Fh_h_ino1,
+#endif
+	Fh_h_igen,
+	Fh_h_type,
+	Fh_tail,
+
+	Fh_ino = Fh_ino1,
+	Fh_dir_ino = Fh_dir_ino1,
+	Fh_h_ino = Fh_h_ino1,
+};
+
+static int au_test_anon(struct dentry *dentry)
+{
+	return !!(dentry->d_flags & DCACHE_DISCONNECTED);
+}
+
+/* ---------------------------------------------------------------------- */
+
+static struct dentry *decode_by_ino(struct super_block *sb, ino_t ino,
+				    ino_t dir_ino)
+{
+	struct dentry *dentry, *parent;
+	struct inode *inode;
+
+	LKTRTrace("i%lu, diri%lu\n", ino, dir_ino);
+
+	dentry = NULL;
+	inode = ilookup(sb, ino);
+	if (unlikely(!inode))
+		goto out;
+
+	dentry = ERR_PTR(-ESTALE);
+	if (unlikely(is_bad_inode(inode)))
+		goto out_iput;
+
+	dentry = NULL;
+	if (!S_ISDIR(inode->i_mode)) {
+		struct dentry *d;
+		spin_lock(&dcache_lock);
+		list_for_each_entry(d, &inode->i_dentry, d_alias)
+			if (!au_test_anon(d)
+			    && d->d_parent->d_inode->i_ino == dir_ino) {
+				dentry = dget_locked(d);
+				break;
+			}
+		spin_unlock(&dcache_lock);
+	} else {
+		dentry = d_find_alias(inode);
+		if (dentry && !au_test_anon(dentry)) {
+			int same_ino;
+			parent = dget_parent(dentry);
+			same_ino = (parent->d_inode->i_ino == dir_ino);
+			dput(parent);
+			if (same_ino)
+				goto out_iput; /* success */
+		}
+
+		dput(dentry);
+		dentry = NULL;
+	}
+
+ out_iput:
+	iput(inode);
+ out:
+	AuTraceErrPtr(dentry);
+	return dentry;
+}
+
+/* ---------------------------------------------------------------------- */
+
+struct find_name_by_ino {
+	int called, found;
+	ino_t ino;
+	char *name;
+	int namelen;
+};
+
+static int
+find_name_by_ino(void *arg, const char *name, int namelen, loff_t offset,
+		 u64 ino, unsigned int d_type)
+{
+	struct find_name_by_ino *a = arg;
+
+	a->called++;
+	if (a->ino != ino)
+		return 0;
+
+	memcpy(a->name, name, namelen);
+	a->namelen = namelen;
+	a->found = 1;
+	return 1;
+}
+
+static struct dentry *decode_by_dir_ino(struct super_block *sb, ino_t ino,
+					ino_t dir_ino)
+{
+	struct dentry *dentry, *parent;
+	struct inode *dir;
+	struct find_name_by_ino arg;
+	struct file *file;
+	int err;
+
+	LKTRTrace("i%lu, diri%lu\n", ino, dir_ino);
+
+	dentry = NULL;
+	dir = ilookup(sb, dir_ino);
+	if (unlikely(!dir))
+		goto out;
+
+	dentry = ERR_PTR(-ESTALE);
+	if (unlikely(is_bad_inode(dir)))
+		goto out_iput;
+
+	dentry = NULL;
+	parent = d_find_alias(dir);
+	if (parent) {
+		if (unlikely(au_test_anon(parent))) {
+			dput(parent);
+			goto out_iput;
+		}
+	} else
+		goto out_iput;
+
+	file = dentry_open(parent, au_mntcache_get(sb), au_dir_roflags);
+	dentry = (void *)file;
+	if (IS_ERR(file))
+		goto out_iput;
+
+	dentry = ERR_PTR(-ENOMEM);
+	arg.name = __getname();
+	if (unlikely(!arg.name))
+		goto out_fput;
+	arg.ino = ino;
+	arg.found = 0;
+
+	do {
+		arg.called = 0;
+		//smp_mb();
+		err = vfsub_readdir(file, find_name_by_ino, &arg, /*dlgt*/0);
+	} while (!err && !arg.found && arg.called);
+	dentry = ERR_PTR(err);
+	if (arg.found) {
+		/* do not call au_lkup_one(), nor dlgt */
+		mutex_lock(&dir->i_mutex);
+		dentry = vfsub_lookup_one_len(arg.name, parent, arg.namelen);
+		mutex_unlock(&dir->i_mutex);
+		AuTraceErrPtr(dentry);
+	}
+
+	//out_putname:
+	__putname(arg.name);
+ out_fput:
+	fput(file);
+ out_iput:
+	iput(dir);
+ out:
+	AuTraceErrPtr(dentry);
+	return dentry;
+}
+
+/* ---------------------------------------------------------------------- */
+
+struct append_name {
+	int found, called, len;
+	char *h_path;
+	ino_t h_ino;
+};
+
+static int append_name(void *arg, const char *name, int len, loff_t pos,
+		       u64 ino, unsigned int d_type)
+{
+	struct append_name *a = arg;
+	char *p;
+
+	a->called++;
+	if (ino != a->h_ino)
+		return 0;
+
+	AuDebugOn(len == 1 && *name == '.');
+	AuDebugOn(len == 2 && name[0] == '.' && name[1] == '.');
+	a->len = strlen(a->h_path);
+	memmove(a->h_path - len - 1, a->h_path, a->len);
+	a->h_path -= len + 1;
+	p = a->h_path + a->len;
+	*p++ = '/';
+	memcpy(p, name, len);
+	a->len += 1 + len;
+	a->found++;
+	return 1;
+}
+
+static int h_acceptable(void *expv, struct dentry *dentry)
+{
+	return 1;
+}
+
+static char *au_build_path(struct super_block *sb, __u32 *fh, char *path,
+			     struct vfsmount *h_mnt, struct dentry *h_root,
+			     struct dentry *h_parent)
+{
+	char *ret;
+	int err, len;
+	struct file *h_file;
+	struct append_name arg;
+	struct path dm_path = {
+		.mnt	= h_mnt,
+		.dentry	= h_root
+	};
+
+	AuTraceEnter();
+
+	arg.h_path = d_path(&dm_path, path, PATH_MAX);
+	ret = arg.h_path;
+	if (unlikely(!arg.h_path || IS_ERR(arg.h_path)))
+		goto out;
+
+	len = strlen(arg.h_path);
+	dm_path.dentry = h_parent;
+	arg.h_path = d_path(&dm_path, path, PATH_MAX);
+	ret = arg.h_path;
+	if (unlikely(!arg.h_path || IS_ERR(arg.h_path)))
+		goto out;
+	LKTRTrace("%s\n", arg.h_path);
+	if (len != 1)
+		arg.h_path += len;
+	LKTRTrace("%p, %s, %ld\n",
+		  arg.h_path, arg.h_path, (long)(arg.h_path - path));
+
+	/* cf. fs/exportfs/expfs.c */
+	h_file = dentry_open(dget(h_parent), mntget(h_mnt), au_dir_roflags);
+	ret = (void *)h_file;
+	if (IS_ERR(h_file))
+		goto out;
+
+	arg.found = 0;
+	arg.h_ino = decode_ino(fh + Fh_h_ino);
+	do {
+		arg.called = 0;
+		err = vfsub_readdir(h_file, append_name, &arg, /*dlgt*/0);
+	} while (!err && !arg.found && arg.called);
+	LKTRTrace("%p, %s, %d\n", arg.h_path, arg.h_path, arg.len);
+	fput(h_file);
+	ret = ERR_PTR(err);
+	if (unlikely(err))
+		goto out;
+
+	dm_path.mnt = au_mntcache_get(sb);
+	dm_path.dentry = sb->s_root;
+	ret = d_path(&dm_path, path, PATH_MAX - arg.len);
+	mntput(dm_path.mnt);
+	if (unlikely(!ret || IS_ERR(ret)))
+		goto out;
+	ret[strlen(ret)] = '/';
+	LKTRTrace("%s\n", ret);
+
+ out:
+	AuTraceErrPtr(ret);
+	return ret;
+}
+
+static struct dentry *
+decode_by_path(struct super_block *sb, aufs_bindex_t bindex, __u32 *fh,
+	       int fh_len, void *context)
+{
+	struct dentry *dentry, *h_parent, *root, *h_root;
+	struct super_block *h_sb;
+	char *path, *p;
+	struct vfsmount *h_mnt;
+	int err;
+	struct nameidata nd;
+	struct au_branch *br;
+
+	LKTRTrace("b%d\n", bindex);
+	SiMustAnyLock(sb);
+
+	br = au_sbr(sb, bindex);
+	/* au_br_get(br); */
+	h_mnt = br->br_mnt;
+	h_sb = h_mnt->mnt_sb;
+	LKTRTrace("%s, h_decode_fh\n", au_sbtype(h_sb));
+	/* in linux-2.6.24, it takes struct fid * as file handle */
+	//todo: call lower fh_to_dentry()? fh_to_parent()?
+	h_parent = exportfs_decode_fh(h_mnt, (void *)(fh + Fh_tail),
+				      fh_len - Fh_tail, fh[Fh_h_type],
+				      h_acceptable, /*context*/NULL);
+	dentry = h_parent;
+	if (unlikely(!h_parent || IS_ERR(h_parent))) {
+		AuWarn1("%s decode_fh failed\n", au_sbtype(h_sb));
+		goto out;
+	}
+	dentry = NULL;
+	if (unlikely(au_test_anon(h_parent))) {
+		AuWarn1("%s decode_fh returned a disconnected dentry\n",
+			au_sbtype(h_sb));
+		dput(h_parent);
+		goto out;
+	}
+
+	dentry = ERR_PTR(-ENOMEM);
+	path = __getname();
+	if (unlikely(!path)) {
+		dput(h_parent);
+		goto out;
+	}
+
+	root = sb->s_root;
+	di_read_lock_parent(root, !AuLock_IR);
+	h_root = au_h_dptr(root, bindex);
+	di_read_unlock(root, !AuLock_IR);
+	p = au_build_path(sb, fh, path, h_mnt, h_root, h_parent);
+	dput(h_parent);
+	if (IS_ERR(p))
+		goto out_putname;
+
+	err = vfsub_path_lookup(p, LOOKUP_FOLLOW, &nd);
+	dentry = ERR_PTR(err);
+	if (!err) {
+		dentry = dget(nd.path.dentry);
+		if (unlikely(au_test_anon(dentry))) {
+			dput(dentry);
+			dentry = ERR_PTR(-ESTALE);
+		}
+		path_put(&nd.path);
+	}
+
+ out_putname:
+	__putname(path);
+ out:
+	/* au_br_put(br); */
+	AuTraceErrPtr(dentry);
+	return dentry;
+}
+
+/* ---------------------------------------------------------------------- */
+
+static noinline_for_stack struct dentry *
+aufs_decode_fh(struct super_block *sb, __u32 *fh, int fh_len, int fh_type,
+	       int (*acceptable)(void *context, struct dentry *de),
+	       void *context)
+{
+	struct dentry *dentry;
+	ino_t ino, dir_ino;
+	aufs_bindex_t bindex, br_id;
+	struct inode *inode, *h_inode;
+	au_gen_t sigen;
+
+	//au_debug_on();
+	LKTRTrace("%d, fh{i%u, br_id %u, sigen %u, hi%u}\n",
+		  fh_type, fh[Fh_ino], fh[Fh_br_id], fh[Fh_sigen],
+		  fh[Fh_h_ino]);
+	AuDebugOn(fh_len < Fh_tail);
+
+	si_read_lock(sb, AuLock_FLUSH);
+	lockdep_off();
+
+	/* branch id may be wrapped around */
+	dentry = ERR_PTR(-ESTALE);
+	br_id = fh[Fh_br_id];
+	sigen = fh[Fh_sigen];
+	bindex = au_br_index(sb, br_id);
+	if (unlikely(bindex < 0 || sigen + AUFS_BRANCH_MAX <= au_sigen(sb)))
+		goto out;
+
+	/* is this inode still cached? */
+	ino = decode_ino(fh + Fh_ino);
+	dir_ino = decode_ino(fh + Fh_dir_ino);
+	dentry = decode_by_ino(sb, ino, dir_ino);
+	if (IS_ERR(dentry))
+		goto out;
+	if (dentry)
+		goto accept;
+
+	/* is the parent dir cached? */
+	dentry = decode_by_dir_ino(sb, ino, dir_ino);
+	if (IS_ERR(dentry))
+		goto out;
+	if (dentry)
+		goto accept;
+
+	/* lookup path */
+	dentry = decode_by_path(sb, bindex, fh, fh_len, context);
+	if (IS_ERR(dentry))
+		goto out;
+	if (unlikely(!dentry))
+		goto out_stale;
+	if (unlikely(dentry->d_inode->i_ino != ino))
+		goto out_dput;
+
+ accept:
+	inode = dentry->d_inode;
+	h_inode = NULL;
+	ii_read_lock_child(inode);
+	if (au_ibstart(inode) <= bindex && bindex <= au_ibend(inode))
+		h_inode = au_h_iptr(inode, bindex);
+	ii_read_unlock(inode);
+	if (h_inode
+	    && h_inode->i_generation == fh[Fh_h_igen]
+	    && acceptable(context, dentry))
+		goto out; /* success */
+ out_dput:
+	dput(dentry);
+ out_stale:
+	dentry = ERR_PTR(-ESTALE);
+ out:
+	lockdep_on();
+	si_read_unlock(sb);
+	AuTraceErrPtr(dentry);
+	//au_debug_off();
+	return dentry;
+}
+
+static struct dentry *
+aufs_fh_to_dentry(struct super_block *sb, struct fid *fid, int fh_len,
+		  int fh_type)
+{
+	return aufs_decode_fh(sb, fid->raw, fh_len, fh_type, h_acceptable,
+			      /*context*/NULL);
+}
+
+#if 0
+/* support subtreecheck option */
+static struct dentry *au_fh_to_parent(struct super_block *sb, struct fid *fid,
+				      int fh_len, int fh_type)
+{
+}
+#endif
+
+/* ---------------------------------------------------------------------- */
+
+static int aufs_encode_fh(struct dentry *dentry, __u32 *fh, int *max_len,
+			  int connectable)
+{
+	int err;
+	struct super_block *sb, *h_sb;
+	struct inode *inode, *h_inode, *dir;
+	aufs_bindex_t bindex;
+	union conv u;
+	struct dentry *parent, *h_parent;
+
+	//au_debug_on();
+	BUILD_BUG_ON(sizeof(u.ino) != sizeof(u.a));
+	LKTRTrace("%.*s, max %d, conn %d\n",
+		  AuDLNPair(dentry), *max_len, connectable);
+	AuDebugOn(au_test_anon(dentry));
+	inode = dentry->d_inode;
+	AuDebugOn(!inode);
+	parent = dget_parent(dentry);
+	AuDebugOn(au_test_anon(parent));
+
+	err = -ENOSPC;
+	if (unlikely(*max_len <= Fh_tail)) {
+		AuWarn1("NFSv2 client (max_len %d)?\n", *max_len);
+		goto out;
+	}
+
+	sb = dentry->d_sb;
+	si_read_lock(sb, AuLock_FLUSH);
+	di_read_lock_child(dentry, AuLock_IR);
+	di_read_lock_parent(parent, AuLock_IR);
+#ifdef CONFIG_AUFS_DEBUG
+	if (unlikely(!au_opt_test(au_mntflags(sb), XINO)))
+		AuWarn1("NFS-exporting requires xino\n");
+#if 0 // temp
+	if (unlikely(au_opt_test(au_mntflags(sb), UDBA_INOTIFY)))
+		AuWarn1("udba=inotify is not recommended when exporting\n");
+#endif
+#endif
+
+	err = -EPERM;
+	bindex = au_dbstart(dentry);
+	h_sb = au_sbr_sb(sb, bindex);
+	if (unlikely(!h_sb->s_export_op)) {
+		AuErr1("%s branch is not exportable\n", au_sbtype(h_sb));
+		goto out_unlock;
+	}
+
+#if 0 //def CONFIG_AUFS_ROBR
+	if (unlikely(au_test_aufs(h_sb))) {
+		AuErr1("aufs branch is not supported\n");
+		goto out_unlock;
+	}
+#endif
+
+	fh[Fh_br_id] = au_sbr_id(sb, bindex);
+	fh[Fh_sigen] = au_sigen(sb);
+	encode_ino(fh + Fh_ino, inode->i_ino);
+	dir = parent->d_inode;
+	encode_ino(fh + Fh_dir_ino, dir->i_ino);
+	h_inode = au_h_dptr(dentry, bindex)->d_inode;
+	encode_ino(fh + Fh_h_ino, h_inode->i_ino);
+	fh[Fh_h_igen] = h_inode->i_generation;
+
+	*max_len -= Fh_tail;
+	//LKTRTrace("Fh_tail %d, max_len %d\n", Fh_tail, *max_len);
+	h_parent = au_h_dptr(parent, bindex);
+	AuDebugOn(au_test_anon(h_parent));
+	/* in linux-2.6.24, it takes struct fid * as file handle */
+	fh[Fh_h_type] = exportfs_encode_fh(h_parent, (void *)(fh + Fh_tail),
+					   max_len, connectable);
+	err = fh[Fh_h_type];
+	*max_len += Fh_tail;
+	if (err != 255)
+		err = 2; //??
+	else
+		AuWarn1("%s encode_fh failed\n", au_sbtype(h_sb));
+
+ out_unlock:
+	di_read_unlock(parent, AuLock_IR);
+	aufs_read_unlock(dentry, AuLock_IR);
+ out:
+	dput(parent);
+	AuTraceErr(err);
+	//au_debug_off();
+	if (unlikely(err < 0))
+		err = 255;
+	return err;
+}
+
+/* ---------------------------------------------------------------------- */
+
+struct export_operations aufs_export_op = {
+	.fh_to_dentry	= aufs_fh_to_dentry,
+	.encode_fh	= aufs_encode_fh
+};
-- 
1.4.4.4

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists