[AUFS PATCH v2.6.26-rc2-mm1 15/39] aufs whiteout

[hooanon05]

From: Junjiro Okajima <hooanon05@xxxxxxxxxxx>

initial commit
whiteout for logical deletion and opaque directory

Signed-off-by: Junjiro Okajima <hooanon05@xxxxxxxxxxx>
---
 fs/aufs/whout.c | 1026 +++++++++++++++++++++++++++++++++++++++++++++++++++++++
 fs/aufs/whout.h |  138 ++++++++
 2 files changed, 1164 insertions(+), 0 deletions(-)
 create mode 100644 fs/aufs/whout.c
 create mode 100644 fs/aufs/whout.h

diff --git a/fs/aufs/whout.c b/fs/aufs/whout.c
new file mode 100644
index 0000000..4b18a3f
--- /dev/null
+++ b/fs/aufs/whout.c
@@ -0,0 +1,1026 @@
+/*
+ * Copyright (C) 2005-2008 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
+ */
+
+/*
+ * whiteout for logical deletion and opaque directory
+ */
+
+#include <linux/fs.h>
+#include <linux/namei.h>
+#include "aufs.h"
+
+#define WH_MASK			S_IRUGO
+
+/* If a directory contains this file, then it is opaque.  We start with the
+ * .wh. flag so that it is blocked by lookup.
+ */
+static struct qstr diropq_name = {
+	.name = AUFS_WH_DIROPQ,
+	.len = sizeof(AUFS_WH_DIROPQ) - 1
+};
+
+/*
+ * generate whiteout name, which is NOT terminated by NULL.
+ * @name: original d_name.name
+ * @len: original d_name.len
+ * @wh: whiteout qstr
+ * returns zero when succeeds, otherwise error.
+ * succeeded value as wh->name should be freed by au_wh_name_free().
+ */
+int au_wh_name_alloc(const char *name, int len, struct qstr *wh)
+{
+	char *p;
+
+	AuDebugOn(!name || !len || !wh);
+
+	if (unlikely(len > PATH_MAX - AUFS_WH_PFX_LEN))
+		return -ENAMETOOLONG;
+
+	wh->len = len + AUFS_WH_PFX_LEN;
+	p = kmalloc(wh->len, GFP_KERNEL);
+	wh->name = p;
+	if (p) {
+		memcpy(p, AUFS_WH_PFX, AUFS_WH_PFX_LEN);
+		memcpy(p + AUFS_WH_PFX_LEN, name, len);
+		/* smp_mb(); */
+		return 0;
+	}
+	return -ENOMEM;
+}
+
+void au_wh_name_free(struct qstr *wh)
+{
+	AuDebugOn(!wh || !wh->name);
+	kfree(wh->name);
+}
+
+/* ---------------------------------------------------------------------- */
+
+/*
+ * test if the @wh_name exists under @h_parent.
+ * @try_sio specifies the necessary of super-io.
+ */
+int au_wh_test(struct dentry *h_parent, struct qstr *wh_name, int try_sio,
+	       struct au_ndx *ndx)
+{
+	int err;
+	struct dentry *wh_dentry;
+	struct inode *h_dir;
+	unsigned int flags;
+
+	LKTRTrace("%.*s/%.*s, ndx{%p, 0x%x}\n", AuDLNPair(h_parent),
+		  wh_name->len, wh_name->name, ndx->nfsmnt, ndx->flags);
+	h_dir = h_parent->d_inode;
+	AuDebugOn(!S_ISDIR(h_dir->i_mode));
+
+	flags = 0;
+	if (ndx && ndx->nd) {
+		flags = ndx->nd->flags;
+		ndx->nd->flags &= ~(LOOKUP_OPEN | LOOKUP_CREATE);
+	}
+
+	if (!try_sio)
+		wh_dentry = au_lkup_one(wh_name->name, h_parent,
+					wh_name->len, ndx);
+	else
+		wh_dentry = au_sio_lkup_one(wh_name->name, h_parent,
+					    wh_name->len, ndx);
+	if (flags)
+		ndx->nd->flags = flags;
+	err = PTR_ERR(wh_dentry);
+	if (IS_ERR(wh_dentry))
+		goto out;
+
+	err = 0;
+	if (!wh_dentry->d_inode)
+		goto out_wh; /* success */
+
+	err = 1;
+	if (S_ISREG(wh_dentry->d_inode->i_mode))
+		goto out_wh; /* success */
+
+	err = -EIO;
+	AuIOErr("%.*s Invalid whiteout entry type 0%o.\n",
+		AuDLNPair(wh_dentry), wh_dentry->d_inode->i_mode);
+
+ out_wh:
+	dput(wh_dentry);
+ out:
+	AuTraceErr(err);
+	return err;
+}
+
+/*
+ * test if the @h_dentry sets opaque or not.
+ */
+int au_diropq_test(struct dentry *h_dentry, struct au_ndx *ndx)
+{
+	int err, try_sio;
+	struct inode *h_dir;
+
+	LKTRTrace("dentry %.*s\n", AuDLNPair(h_dentry));
+	h_dir = h_dentry->d_inode;
+	AuDebugOn(!S_ISDIR(h_dir->i_mode));
+
+	try_sio = au_test_h_perm_sio(h_dir, MAY_EXEC,
+				     au_ftest_ndx(ndx->flags, DLGT));
+	err = au_wh_test(h_dentry, &diropq_name, try_sio, ndx);
+	AuTraceErr(err);
+	return err;
+}
+
+/*
+ * returns a negative dentry whose name is unique and temporary.
+ */
+struct dentry *au_whtmp_lkup(struct dentry *h_parent, struct qstr *prefix,
+			     struct au_ndx *ndx)
+{
+#define HEX_LEN 4
+	struct dentry *dentry;
+	int len, i;
+	char defname[AUFS_WH_PFX_LEN * 2 + DNAME_INLINE_LEN_MIN + 1
+		     + HEX_LEN + 1], *name, *p;
+	static unsigned char cnt;
+
+	LKTRTrace("hp %.*s, prefix %.*s\n",
+		  AuDLNPair(h_parent), prefix->len, prefix->name);
+	AuDebugOn(!h_parent->d_inode);
+
+	name = defname;
+	len = sizeof(defname) - DNAME_INLINE_LEN_MIN + prefix->len - 1;
+	if (unlikely(prefix->len > DNAME_INLINE_LEN_MIN)) {
+		dentry = ERR_PTR(-ENAMETOOLONG);
+		if (unlikely(len >= PATH_MAX))
+			goto out;
+		dentry = ERR_PTR(-ENOMEM);
+		name = kmalloc(len + 1, GFP_KERNEL);
+		if (unlikely(!name))
+			goto out;
+	}
+
+	/* doubly whiteout-ed */
+	memcpy(name, AUFS_WH_PFX AUFS_WH_PFX, AUFS_WH_PFX_LEN * 2);
+	p = name + AUFS_WH_PFX_LEN * 2;
+	memcpy(p, prefix->name, prefix->len);
+	p += prefix->len;
+	*p++ = '.';
+	AuDebugOn(name + len + 1 - p <= HEX_LEN);
+
+	for (i = 0; i < 3; i++) {
+		sprintf(p, "%.*d", HEX_LEN, cnt++);
+		dentry = au_sio_lkup_one(name, h_parent, len, ndx);
+		if (IS_ERR(dentry) || !dentry->d_inode)
+			goto out_name;
+		dput(dentry);
+	}
+	/* AuWarn("could not get random name\n"); */
+	dentry = ERR_PTR(-EEXIST);
+	AuDbg("%.*s\n", len, name);
+	BUG();
+
+ out_name:
+	if (unlikely(name != defname))
+		kfree(name);
+ out:
+	AuTraceErrPtr(dentry);
+	return dentry;
+#undef HEX_LEN
+}
+
+/*
+ * rename the @dentry of @bindex to the whiteouted temporary name.
+ */
+int au_whtmp_ren(struct inode *dir, struct dentry *dentry, aufs_bindex_t bindex,
+		 int noself)
+{
+	int err, dlgt;
+	struct inode *h_dir;
+	struct dentry *h_dentry, *h_parent, *tmp_dentry;
+	struct super_block *sb;
+	unsigned int mnt_flags;
+	struct au_hin_ignore ign;
+	struct vfsub_args vargs;
+	struct au_ndx ndx = {
+		.flags	= 0,
+		.nd	= NULL,
+		/* .br	= NULL */
+	};
+
+	LKTRTrace("%.*s, b%d\n", AuDLNPair(dentry), bindex);
+	h_dentry = au_h_dptr(dentry, bindex);
+	AuDebugOn(!h_dentry || !h_dentry->d_inode);
+	h_parent = h_dentry->d_parent; /* dir inode is locked */
+	h_dir = h_parent->d_inode;
+	IMustLock(h_dir);
+
+	sb = dentry->d_sb;
+	mnt_flags = au_mntflags(sb);
+	dlgt = !!au_opt_test_dlgt(mnt_flags);
+	if (unlikely(dlgt))
+		au_fset_ndx(ndx.flags, DLGT);
+	ndx.nfsmnt = au_nfsmnt(sb, bindex);
+	tmp_dentry = au_whtmp_lkup(h_parent, &h_dentry->d_name, &ndx);
+	err = PTR_ERR(tmp_dentry);
+	if (!IS_ERR(tmp_dentry)) {
+		/* under the same dir, no need to lock_rename() */
+		vfsub_args_init(&vargs, &ign, dlgt, 0);
+		AuDebugOn(!S_ISDIR(dentry->d_inode->i_mode));
+		if (unlikely(au_opt_test(mnt_flags, UDBA_INOTIFY) && !noself))
+			vfsub_ign_hinode(&vargs, IN_MOVE_SELF,
+					 au_hi(dentry->d_inode, bindex));
+		err = vfsub_rename(h_dir, h_dentry, h_dir, tmp_dentry, &vargs);
+		AuTraceErr(err);
+		dput(tmp_dentry);
+	}
+
+	AuTraceErr(err);
+	return err;
+}
+
+/* ---------------------------------------------------------------------- */
+
+static int do_unlink_wh(struct inode *h_dir, struct dentry *wh_dentry,
+			struct inode *dir, int dlgt)
+{
+	struct vfsub_args vargs;
+
+	LKTRTrace("hi%lu, wh %.*s\n", h_dir->i_ino, AuDLNPair(wh_dentry));
+	AuDebugOn(!wh_dentry->d_inode || !S_ISREG(wh_dentry->d_inode->i_mode));
+
+	/*
+	 * forces superio when the dir has a sticky bit.
+	 * this may be a violation of unix fs semantics.
+	 */
+	vfsub_args_init(&vargs, NULL, dlgt,
+			(h_dir->i_mode & S_ISVTX)
+			&& wh_dentry->d_inode->i_uid != current->fsuid);
+	return vfsub_unlink(h_dir, wh_dentry, &vargs);
+}
+
+int au_wh_unlink_dentry(struct inode *h_dir, struct dentry *wh_dentry,
+			struct dentry *dentry, struct inode *dir, int dlgt)
+{
+	int err;
+
+	LKTRTrace("hi%lu, wh %.*s, d %p\n", h_dir->i_ino,
+		  AuDLNPair(wh_dentry), dentry);
+	AuDebugOn((dentry && au_dbwh(dentry) < 0)
+		  || !wh_dentry->d_inode
+		  || !S_ISREG(wh_dentry->d_inode->i_mode));
+
+	err = do_unlink_wh(h_dir, wh_dentry, dir, dlgt);
+	if (!err && dentry)
+		au_set_dbwh(dentry, -1);
+
+	AuTraceErr(err);
+	return err;
+}
+
+static int unlink_wh_name(struct dentry *h_parent, struct qstr *wh,
+			  struct inode *dir, struct au_ndx *ndx)
+{
+	int err;
+	struct inode *h_dir;
+	struct dentry *h_dentry;
+
+	LKTRTrace("%.*s/%.*s\n", AuDLNPair(h_parent), AuLNPair(wh));
+	h_dir = h_parent->d_inode;
+
+	/* au_test_h_perm() is already done */
+	h_dentry = au_lkup_one(wh->name, h_parent, wh->len, ndx);
+	if (!IS_ERR(h_dentry)) {
+		err = 0;
+		if (h_dentry->d_inode && S_ISREG(h_dentry->d_inode->i_mode))
+			err = do_unlink_wh(h_dir, h_dentry, dir,
+					   au_ftest_ndx(ndx->flags, DLGT));
+		dput(h_dentry);
+	} else
+		err = PTR_ERR(h_dentry);
+
+	AuTraceErr(err);
+	return err;
+}
+
+/* ---------------------------------------------------------------------- */
+
+static void clean_wh(struct inode *h_dir, struct dentry *wh)
+{
+	int err;
+	struct vfsub_args vargs;
+
+	AuTraceEnter();
+
+	if (wh->d_inode) {
+		vfsub_args_init(&vargs, NULL, 0, 0);
+		err = vfsub_unlink(h_dir, wh, &vargs);
+		if (unlikely(err))
+			AuWarn("failed unlink %.*s (%d), ignored.\n",
+			       AuDLNPair(wh), err);
+	}
+}
+
+static void clean_plink(struct inode *h_dir, struct dentry *plink)
+{
+	int err;
+	struct vfsub_args vargs;
+
+	AuTraceEnter();
+
+	if (plink->d_inode) {
+		vfsub_args_init(&vargs, NULL, 0, 0);
+		err = vfsub_rmdir(h_dir, plink, &vargs);
+		if (unlikely(err))
+			AuWarn("failed rmdir %.*s (%d), ignored.\n",
+			       AuDLNPair(plink), err);
+	}
+}
+
+static int test_linkable(struct inode *h_dir)
+{
+	if (h_dir->i_op && h_dir->i_op->link)
+		return 0;
+	return -ENOSYS;
+}
+
+static int plink_dir(struct inode *h_dir, struct dentry *plink)
+{
+	int err;
+
+	err = -EEXIST;
+	if (!plink->d_inode) {
+		int mode = S_IRWXU;
+		if (unlikely(au_test_nfs(plink->d_sb)))
+			mode |= S_IXUGO;
+		err = vfsub_mkdir(h_dir, plink, mode, /*dlgt*/0);
+	} else if (S_ISDIR(plink->d_inode->i_mode))
+		err = 0;
+	else
+		AuErr("unknown %.*s exists\n", AuDLNPair(plink));
+
+	return err;
+}
+
+/*
+ * initialize the whiteout base file/dir for @br.
+ */
+int au_wh_init(struct dentry *h_root, struct au_branch *br,
+	       struct vfsmount *nfsmnt, struct super_block *sb)
+{
+	int err;
+	struct dentry *wh, *plink;
+	struct inode *h_dir;
+	static struct qstr base_name[] = {
+		{
+			.name	= AUFS_WH_BASENAME,
+			.len	= sizeof(AUFS_WH_BASENAME) - 1
+		},
+		{
+			.name	= AUFS_WH_PLINKDIR,
+			.len	= sizeof(AUFS_WH_PLINKDIR) - 1
+		}
+	};
+	struct au_ndx ndx = {
+		.nfsmnt	= nfsmnt,
+		.flags	= 0, /* always no dlgt */
+		.nd	= NULL,
+		/* .br	= NULL */
+	};
+	const int do_plink = au_opt_test(au_mntflags(sb), PLINK);
+
+	LKTRTrace("nfsmnt %p\n", nfsmnt);
+	BrWhMustWriteLock(br);
+	SiMustWriteLock(sb);
+	h_dir = h_root->d_inode;
+
+	/* doubly whiteouted */
+	wh = au_wh_lkup(h_root, base_name + 0, &ndx);
+	err = PTR_ERR(wh);
+	if (IS_ERR(wh))
+		goto out;
+	AuDebugOn(br->br_wh && br->br_wh != wh);
+
+	plink = au_wh_lkup(h_root, base_name + 1, &ndx);
+	err = PTR_ERR(plink);
+	if (IS_ERR(plink))
+		goto out_dput_wh;
+	AuDebugOn(br->br_plink && br->br_plink != plink);
+
+	dput(br->br_wh);
+	dput(br->br_plink);
+	br->br_wh = NULL;
+	br->br_plink = NULL;
+
+	err = 0;
+	switch (br->br_perm) {
+	case AuBr_RR:
+	case AuBr_RO:
+	case AuBr_RRWH:
+	case AuBr_ROWH:
+		clean_wh(h_dir, wh);
+		clean_plink(h_dir, plink);
+		break;
+
+	case AuBr_RWNoLinkWH:
+		clean_wh(h_dir, wh);
+		if (do_plink) {
+			err = test_linkable(h_dir);
+			if (unlikely(err))
+				goto out_nolink;
+
+			err = plink_dir(h_dir, plink);
+			if (unlikely(err))
+				goto out_err;
+			br->br_plink = dget(plink);
+		} else
+			clean_plink(h_dir, plink);
+		break;
+
+	case AuBr_RW:
+		/*
+		 * for the moment, aufs supports the branch filesystem
+		 * which does not support link(2).
+		 * testing on FAT which does not support i_op->setattr() fully
+		 * either, copyup failed.
+		 * finally, such filesystem will not be used as the writable
+		 * branch.
+		 */
+		err = test_linkable(h_dir);
+		if (unlikely(err))
+			goto out_nolink;
+
+		err = -EEXIST;
+		if (!wh->d_inode)
+			err = au_h_create(h_dir, wh, WH_MASK, /*dlgt*/0,
+					  /*nd*/NULL, nfsmnt);
+		else if (S_ISREG(wh->d_inode->i_mode))
+			err = 0;
+		else
+			AuErr("unknown %.*s/%.*s exists\n",
+			      AuDLNPair(h_root), AuDLNPair(wh));
+		if (unlikely(err))
+			goto out_err;
+
+		if (do_plink) {
+			err = plink_dir(h_dir, plink);
+			if (unlikely(err))
+				goto out_err;
+			br->br_plink = dget(plink);
+		} else
+			clean_plink(h_dir, plink);
+		br->br_wh = dget(wh);
+		break;
+
+	default:
+		BUG();
+	}
+
+ out_dput:
+	dput(plink);
+ out_dput_wh:
+	dput(wh);
+ out:
+	AuTraceErr(err);
+	return err;
+ out_nolink:
+	AuErr("%.*s doesn't support link(2), use noplink and rw+nolwh\n",
+	      AuDLNPair(h_root));
+	goto out_dput;
+ out_err:
+	AuErr("an error(%d) on the writable branch %.*s(%s)\n",
+	      err, AuDLNPair(h_root), au_sbtype(h_root->d_sb));
+	goto out_dput;
+}
+
+struct reinit_br_wh {
+	struct super_block *sb;
+	struct au_branch *br;
+};
+
+static void reinit_br_wh(void *arg)
+{
+	int err;
+	struct reinit_br_wh *a = arg;
+	struct inode *h_dir, *dir;
+	struct dentry *h_root;
+	aufs_bindex_t bindex;
+	struct vfsub_args vargs;
+
+	AuTraceEnter();
+	AuDebugOn(!a->br->br_wh || !a->br->br_wh->d_inode || current->fsuid);
+
+	err = 0;
+	/* big aufs lock */
+	si_write_lock(a->sb);
+	if (unlikely(!au_br_writable(a->br->br_perm)))
+		goto out;
+	bindex = au_br_index(a->sb, a->br->br_id);
+	if (unlikely(bindex < 0))
+		goto out;
+
+	dir = a->sb->s_root->d_inode;
+	h_root = dget_parent(a->br->br_wh);
+	h_dir = h_root->d_inode;
+	AuDebugOn(!h_dir->i_op || !h_dir->i_op->link);
+	vfsub_args_init(&vargs, NULL, /*dlgt*/0, 0);
+	au_hdir_lock(h_dir, dir, bindex);
+	/* todo: revalidate h_wh? */
+	br_wh_write_lock(a->br);
+	err = vfsub_unlink(h_dir, a->br->br_wh, &vargs);
+	dput(a->br->br_wh);
+	a->br->br_wh = NULL;
+	if (!err)
+		err = au_wh_init(h_root, a->br, au_do_nfsmnt(a->br->br_mnt),
+				 a->sb);
+	br_wh_write_unlock(a->br);
+	au_hdir_unlock(h_dir, dir, bindex);
+	dput(h_root);
+
+ out:
+	atomic_dec_return(&a->br->br_wh_running);
+	au_br_put(a->br);
+	si_write_unlock(a->sb);
+	kfree(arg);
+	if (unlikely(err))
+		AuIOErr("err %d\n", err);
+}
+
+static void kick_reinit_br_wh(struct super_block *sb, struct au_branch *br)
+{
+	int do_dec, wkq_err;
+	struct reinit_br_wh *arg;
+
+	do_dec = 1;
+	if (atomic_inc_return(&br->br_wh_running) != 1)
+		goto out;
+
+	/* ignore ENOMEM */
+	arg = kmalloc(sizeof(*arg), GFP_TEMPORARY);
+	if (arg) {
+		/*
+		 * dec(wh_running), kfree(arg) and au_br_put()
+		 * in reinit function
+		 */
+		arg->sb = sb;
+		arg->br = br;
+		au_br_get(br);
+		wkq_err = au_wkq_nowait(reinit_br_wh, arg, sb, /*dlgt*/0);
+		if (unlikely(wkq_err)) {
+			atomic_dec_return(&br->br_wh_running);
+			au_br_put(br);
+			kfree(arg);
+		}
+		do_dec = 0;
+	}
+
+ out:
+	if (do_dec)
+		atomic_dec_return(&br->br_wh_running);
+}
+
+/*
+ * create the whiteout @wh.
+ */
+static int link_or_create_wh(struct dentry *wh, struct super_block *sb,
+			     aufs_bindex_t bindex, struct inode *dir)
+{
+	int err, dlgt;
+	struct au_branch *br;
+	struct dentry *h_parent;
+	struct inode *h_dir;
+
+	LKTRTrace("%.*s\n", AuDLNPair(wh));
+	SiMustReadLock(sb);
+	h_parent = wh->d_parent; /* dir inode is locked */
+	h_dir = h_parent->d_inode;
+	IMustLock(h_dir);
+
+	dlgt = !!au_opt_test_dlgt(au_mntflags(sb));
+	br = au_sbr(sb, bindex);
+	br_wh_read_lock(br);
+	if (br->br_wh) {
+		err = vfsub_link(br->br_wh, h_dir, wh, dlgt);
+		if (!err || err != -EMLINK)
+			goto out;
+
+		/* link count full. re-initialize br_wh. */
+		kick_reinit_br_wh(sb, br);
+	}
+
+	/* return this error in this context */
+	err = au_h_create(h_dir, wh, WH_MASK, dlgt, /*nd*/NULL,
+			  au_do_nfsmnt(br->br_mnt));
+
+ out:
+	br_wh_read_unlock(br);
+	AuTraceErr(err);
+	return err;
+}
+
+/* ---------------------------------------------------------------------- */
+
+/*
+ * create or remove the diropq.
+ */
+static struct dentry *do_diropq(struct dentry *dentry, aufs_bindex_t bindex,
+				unsigned int flags)
+{
+	struct dentry *opq_dentry, *h_dentry;
+	struct inode *h_dir;
+	int err, dlgt;
+	struct super_block *sb;
+	struct au_ndx ndx = {
+		.flags	= 0,
+		.nd	= NULL,
+		/* .br	= NULL */
+	};
+
+	LKTRTrace("%.*s, bindex %d, flags 0x%x\n",
+		  AuDLNPair(dentry), bindex, flags);
+	h_dentry = au_h_dptr(dentry, bindex);
+	AuDebugOn(!h_dentry);
+	h_dir = h_dentry->d_inode;
+	AuDebugOn(!h_dir || !S_ISDIR(h_dir->i_mode));
+
+	/* already checked by au_test_h_perm(). */
+	sb = dentry->d_sb;
+	ndx.nfsmnt = au_nfsmnt(sb, bindex);
+	dlgt = 0;
+	if (unlikely(au_ftest_diropq(flags, DLGT))) {
+		dlgt = 1;
+		au_fset_ndx(ndx.flags, DLGT);
+	}
+	opq_dentry = au_lkup_one(diropq_name.name, h_dentry, diropq_name.len,
+				 &ndx);
+	if (IS_ERR(opq_dentry))
+		goto out;
+
+	if (au_ftest_diropq(flags, CREATE)) {
+		AuDebugOn(opq_dentry->d_inode);
+		err = link_or_create_wh(opq_dentry, sb, bindex,
+					dentry->d_inode);
+		if (!err) {
+			au_set_dbdiropq(dentry, bindex);
+			goto out; /* success */
+		}
+	} else {
+		AuDebugOn(/* !S_ISDIR(dentry->d_inode->i_mode)
+			   * ||  */!opq_dentry->d_inode);
+		err = do_unlink_wh(h_dir, opq_dentry, dentry->d_inode, dlgt);
+		if (!err)
+			au_set_dbdiropq(dentry, -1);
+	}
+	dput(opq_dentry);
+	opq_dentry = ERR_PTR(err);
+
+ out:
+	AuTraceErrPtr(opq_dentry);
+	return opq_dentry;
+}
+
+struct do_diropq_args {
+	struct dentry **errp;
+	struct dentry *dentry;
+	aufs_bindex_t bindex;
+	unsigned int flags;
+};
+
+static void call_do_diropq(void *args)
+{
+	struct do_diropq_args *a = args;
+	*a->errp = do_diropq(a->dentry, a->bindex, a->flags);
+}
+
+struct dentry *au_diropq_sio(struct dentry *dentry, aufs_bindex_t bindex,
+			     unsigned int flags)
+{
+	struct dentry *diropq, *h_dentry;
+
+	LKTRTrace("%.*s, bindex %d, flags 0x%x\n",
+		  AuDLNPair(dentry), bindex, flags);
+
+	h_dentry = au_h_dptr(dentry, bindex);
+	if (!au_test_h_perm_sio(h_dentry->d_inode, MAY_EXEC | MAY_WRITE,
+				au_ftest_diropq(flags, DLGT)))
+		diropq = do_diropq(dentry, bindex, flags);
+	else {
+		int wkq_err;
+		struct do_diropq_args args = {
+			.errp		= &diropq,
+			.dentry		= dentry,
+			.bindex		= bindex,
+			.flags		= flags
+		};
+		wkq_err = au_wkq_wait(call_do_diropq, &args, /*dlgt*/0);
+		if (unlikely(wkq_err))
+			diropq = ERR_PTR(wkq_err);
+	}
+
+	AuTraceErrPtr(diropq);
+	return diropq;
+}
+
+/* ---------------------------------------------------------------------- */
+
+/*
+ * lookup whiteout dentry.
+ * @h_parent: hidden parent dentry which must exist and be locked
+ * @base_name: name of dentry which will be whiteouted
+ * returns dentry for whiteout.
+ */
+struct dentry *au_wh_lkup(struct dentry *h_parent, struct qstr *base_name,
+			  struct au_ndx *ndx)
+{
+	int err;
+	struct qstr wh_name;
+	struct dentry *wh_dentry;
+
+	LKTRTrace("%.*s/%.*s\n", AuDLNPair(h_parent), AuLNPair(base_name));
+
+	err = au_wh_name_alloc(base_name->name, base_name->len, &wh_name);
+	wh_dentry = ERR_PTR(err);
+	if (!err) {
+		/* do not superio. */
+		wh_dentry = au_lkup_one(wh_name.name, h_parent,
+					wh_name.len, ndx);
+		au_wh_name_free(&wh_name);
+	}
+	AuTraceErrPtr(wh_dentry);
+	return wh_dentry;
+}
+
+/*
+ * link/create a whiteout for @dentry on @bindex.
+ */
+struct dentry *au_wh_create(struct inode *dir, struct dentry *dentry,
+			    aufs_bindex_t bindex, struct dentry *h_parent,
+			    struct au_ndx *ndx)
+{
+	struct dentry *wh_dentry;
+	int err;
+	struct super_block *sb;
+
+	LKTRTrace("%.*s/%.*s on b%d\n", AuDLNPair(h_parent),
+		  AuDLNPair(dentry), bindex);
+
+	sb = dentry->d_sb;
+	wh_dentry = au_wh_lkup(h_parent, &dentry->d_name, ndx);
+	if (!IS_ERR(wh_dentry) && !wh_dentry->d_inode) {
+		err = link_or_create_wh(wh_dentry, sb, bindex, dir);
+		if (!err)
+			au_set_dbwh(dentry, bindex);
+		else {
+			dput(wh_dentry);
+			wh_dentry = ERR_PTR(err);
+		}
+	}
+
+	AuTraceErrPtr(wh_dentry);
+	return wh_dentry;
+}
+
+/* ---------------------------------------------------------------------- */
+
+/* Delete all whiteouts in this directory on branch bindex. */
+static int del_wh_children(struct au_nhash *whlist, struct dentry *h_parent,
+			   aufs_bindex_t bindex, struct inode *inode,
+			   struct au_ndx *ndx)
+{
+	int err, i;
+	struct qstr wh_name;
+	char *p;
+	struct inode *h_dir;
+	struct hlist_head *head;
+	struct au_vdir_wh *tpos;
+	struct hlist_node *pos;
+	struct au_vdir_destr *str;
+
+	LKTRTrace("%.*s\n", AuDLNPair(h_parent));
+	h_dir = h_parent->d_inode;
+	AuDebugOn(IS_RDONLY(h_dir));
+
+	err = -ENOMEM;
+	p = __getname();
+	wh_name.name = p;
+	if (unlikely(!wh_name.name))
+		goto out;
+	memcpy(p, AUFS_WH_PFX, AUFS_WH_PFX_LEN);
+	p += AUFS_WH_PFX_LEN;
+
+	/* already checked by au_test_h_perm(). */
+	err = 0;
+	for (i = 0; !err && i < AuSize_NHASH; i++) {
+		head = whlist->heads + i;
+		hlist_for_each_entry(tpos, pos, head, wh_hash) {
+			if (tpos->wh_bindex != bindex)
+				continue;
+			str = &tpos->wh_str;
+			if (str->len + AUFS_WH_PFX_LEN <= PATH_MAX) {
+				memcpy(p, str->name, str->len);
+				wh_name.len = AUFS_WH_PFX_LEN + str->len;
+				err = unlink_wh_name(h_parent, &wh_name, inode,
+						     ndx);
+				if (!err)
+					continue;
+				break;
+			}
+			AuIOErr("whiteout name too long %.*s\n",
+				str->len, str->name);
+			err = -EIO;
+			break;
+		}
+	}
+	__putname(wh_name.name);
+
+ out:
+	AuTraceErr(err);
+	return err;
+}
+
+struct del_wh_children_args {
+	int *errp;
+	struct au_nhash *whlist;
+	struct dentry *h_parent;
+	aufs_bindex_t bindex;
+	struct inode *inode;
+	struct au_ndx *ndx;
+};
+
+static void call_del_wh_children(void *args)
+{
+	struct del_wh_children_args *a = args;
+	*a->errp = del_wh_children(a->whlist, a->h_parent, a->bindex,
+				   a->inode, a->ndx);
+}
+
+/* ---------------------------------------------------------------------- */
+
+/*
+ * rmdir the whiteouted temporary named dir @h_dentry.
+ * @whlist: whiteouted children.
+ */
+int au_whtmp_rmdir(struct dentry *h_dentry, struct au_nhash *whlist,
+		   aufs_bindex_t bindex, struct inode *dir, struct inode *inode,
+		   int noself)
+{
+	int err, dlgt;
+	struct inode *h_inode, *h_dir;
+	struct super_block *sb;
+	unsigned int mnt_flags;
+	struct au_hin_ignore ign;
+	struct vfsub_args vargs;
+	struct au_ndx ndx = {
+		.flags	= 0,
+		.nd	= NULL,
+		/* .br	= NULL */
+	};
+
+	LKTRTrace("hd %.*s, b%d, i%lu\n",
+		  AuDLNPair(h_dentry), bindex, dir->i_ino);
+	IMustLock(dir);
+	IiMustAnyLock(dir);
+	h_dir = h_dentry->d_parent->d_inode; /* dir inode is locked */
+	IMustLock(h_dir);
+
+	sb = inode->i_sb;
+	mnt_flags = au_mntflags(sb);
+	dlgt = !!au_opt_test_dlgt(mnt_flags);
+	if (unlikely(dlgt))
+		au_fset_ndx(ndx.flags, DLGT);
+	ndx.nfsmnt = au_nfsmnt(sb, bindex);
+	h_inode = h_dentry->d_inode;
+	AuDebugOn(h_inode != au_h_iptr(inode, bindex));
+	au_hdir2_lock(h_inode, inode, bindex);
+
+	/*
+	 * someone else might change some whiteouts while we were sleeping.
+	 * it means this whlist may have an obsoleted entry.
+	 */
+	if (!au_test_h_perm_sio(h_inode, MAY_EXEC | MAY_WRITE, dlgt))
+		err = del_wh_children(whlist, h_dentry, bindex, inode, &ndx);
+	else {
+		int wkq_err;
+		/* ugly */
+		unsigned int flags = ndx.flags;
+		struct del_wh_children_args args = {
+			.errp		= &err,
+			.whlist		= whlist,
+			.h_parent	= h_dentry,
+			.bindex		= bindex,
+			.inode		= inode,
+			.ndx		= &ndx
+		};
+
+		ndx.flags = 0;
+		wkq_err = au_wkq_wait(call_del_wh_children, &args, /*dlgt*/0);
+		if (unlikely(wkq_err))
+			err = wkq_err;
+		ndx.flags = flags;
+	}
+	au_hdir_unlock(h_inode, inode, bindex);
+
+	if (!err) {
+		vfsub_args_init(&vargs, &ign, dlgt, 0);
+		if (unlikely(au_opt_test(mnt_flags, UDBA_INOTIFY) && !noself))
+			vfsub_ign_hinode(&vargs, IN_DELETE_SELF,
+					 au_hi(inode, bindex));
+		err = vfsub_rmdir(h_dir, h_dentry, &vargs);
+		/* d_drop(h_dentry); */
+	}
+
+	if (!err) {
+		if (au_ibstart(dir) == bindex) {
+			au_cpup_attr_timesizes(dir);
+			/* au_cpup_attr_nlink(dir); */
+			drop_nlink(dir);
+		}
+		return 0; /* success */
+	}
+
+	AuWarn("failed removing %.*s(%d), ignored\n", AuDLNPair(h_dentry), err);
+	return err;
+}
+
+static void au_whtmp_rmdir_free_args(struct au_whtmp_rmdir_args *args)
+{
+	dput(args->h_dentry);
+	au_nhash_fin(&args->whlist);
+	iput(args->inode);
+	mutex_unlock(&args->dir->i_mutex);
+	iput(args->dir);
+	kfree(args);
+}
+
+static void do_rmdir_whtmp(void *args)
+{
+	int err;
+	struct au_whtmp_rmdir_args *a = args;
+	struct super_block *sb;
+
+	LKTRTrace("%.*s, b%d, dir i%lu\n",
+		  AuDLNPair(a->h_dentry), a->bindex, a->dir->i_ino);
+
+	mutex_lock(&a->dir->i_mutex);
+	sb = a->dir->i_sb;
+	si_noflush_read_lock(sb);
+	err = au_test_ro(sb, a->bindex, NULL);
+	if (!err) {
+		struct dentry *h_parent = dget_parent(a->h_dentry);
+		struct inode *h_dir = h_parent->d_inode;
+
+		ii_write_lock_child(a->inode);
+		ii_write_lock_parent(a->dir);
+		au_hdir_lock(h_dir, a->dir, a->bindex);
+		/* todo: revalidate h_dentry? */
+		err = au_whtmp_rmdir(a->h_dentry, &a->whlist, a->bindex,
+				     a->dir, a->inode, a->noself);
+		au_hdir_unlock(h_dir, a->dir, a->bindex);
+		ii_write_unlock(a->dir);
+		ii_write_unlock(a->inode);
+		dput(h_parent);
+	}
+	si_read_unlock(sb);
+	au_whtmp_rmdir_free_args(a);
+	if (unlikely(err))
+		AuIOErr("err %d\n", err);
+}
+
+void au_whtmp_kick_rmdir(struct dentry *h_dentry, struct au_nhash *whlist,
+			 aufs_bindex_t bindex, struct inode *dir,
+			 struct inode *inode, int noself,
+			 struct au_whtmp_rmdir_args *args)
+{
+	int wkq_err;
+
+	LKTRTrace("%.*s\n", AuDLNPair(h_dentry));
+	IMustLock(dir);
+
+	/* all post-process will be done in do_rmdir_whtmp(). */
+	args->h_dentry = dget(h_dentry);
+	au_nhash_init(&args->whlist);
+	au_nhash_move(&args->whlist, whlist);
+	args->bindex = bindex;
+	args->dir = igrab(dir);
+	args->inode = igrab(inode);
+	args->noself = noself;
+	wkq_err = au_wkq_nowait(do_rmdir_whtmp, args, dir->i_sb, /*dlgt*/0);
+	if (unlikely(wkq_err)) {
+		AuWarn("rmdir error %.*s (%d), ignored\n",
+		       AuDLNPair(h_dentry), wkq_err);
+		au_whtmp_rmdir_free_args(args);
+	}
+}
diff --git a/fs/aufs/whout.h b/fs/aufs/whout.h
new file mode 100644
index 0000000..45333cb
--- /dev/null
+++ b/fs/aufs/whout.h
@@ -0,0 +1,138 @@
+/*
+ * Copyright (C) 2005-2008 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
+ */
+
+/*
+ * whiteout for logical deletion and opaque directory
+ */
+
+#ifndef __AUFS_WHOUT_H__
+#define __AUFS_WHOUT_H__
+
+#ifdef __KERNEL__
+
+#include <linux/fs.h>
+#include <linux/aufs_type.h>
+#include "dir.h"
+#include "opts.h"
+#include "super.h"
+
+int au_wh_name_alloc(const char *name, int len, struct qstr *wh);
+void au_wh_name_free(struct qstr *wh);
+
+struct au_ndx;
+int au_wh_test(struct dentry *h_parent, struct qstr *wh_name, int try_sio,
+	       struct au_ndx *ndx);
+int au_diropq_test(struct dentry *h_dentry, struct au_ndx *ndx);
+
+struct dentry *au_whtmp_lkup(struct dentry *h_parent, struct qstr *prefix,
+			     struct au_ndx *ndx);
+int au_whtmp_ren(struct inode *dir, struct dentry *dentry, aufs_bindex_t bindex,
+		 int noself);
+int au_wh_unlink_dentry(struct inode *h_dir, struct dentry *wh_dentry,
+			struct dentry *dentry, struct inode *dir, int dlgt);
+
+struct au_branch;
+int au_wh_init(struct dentry *h_parent, struct au_branch *br,
+	       struct vfsmount *nfsmnt, struct super_block *sb);
+
+/* diropq flags */
+#define AuDiropq_CREATE	1
+#define AuDiropq_DLGT	(1 << 1)
+#define au_ftest_diropq(flags, name)	((flags) & AuDiropq_##name)
+#define au_fset_diropq(flags, name)	{ (flags) |= AuDiropq_##name; }
+#define au_fclr_diropq(flags, name)	{ (flags) &= ~AuDiropq_##name; }
+#ifndef CONFIG_AUFS_DLGT
+#undef AuDiropq_DLGT
+#define AuDiropq_DLGT	0
+#endif
+
+struct dentry *au_diropq_sio(struct dentry *dentry, aufs_bindex_t bindex,
+			     unsigned int flags);
+
+struct dentry *au_wh_lkup(struct dentry *h_parent, struct qstr *base_name,
+			  struct au_ndx *ndx);
+struct dentry *au_wh_create(struct inode *dir, struct dentry *dentry,
+			    aufs_bindex_t bindex, struct dentry *h_parent,
+			    struct au_ndx *ndx);
+
+/* real rmdir the whiteout-ed dir */
+struct au_whtmp_rmdir_args {
+	struct dentry *h_dentry;
+	struct au_nhash whlist;
+	aufs_bindex_t bindex;
+	struct inode *dir, *inode;
+	int noself;
+};
+
+struct au_nhash;
+int au_whtmp_rmdir(struct dentry *h_dentry, struct au_nhash *whlist,
+		   aufs_bindex_t bindex, struct inode *dir, struct inode *inode,
+		   int noself);
+void au_whtmp_kick_rmdir(struct dentry *h_dentry, struct au_nhash *whlist,
+			 aufs_bindex_t bindex, struct inode *dir,
+			 struct inode *inode, int noself,
+			 struct au_whtmp_rmdir_args *args);
+
+/* ---------------------------------------------------------------------- */
+
+static inline
+struct dentry *au_diropq_create(struct dentry *dentry, aufs_bindex_t bindex,
+				int dlgt)
+{
+	unsigned int flags = AuDiropq_CREATE;
+	if (unlikely(dlgt))
+		au_fset_diropq(flags, DLGT);
+	return au_diropq_sio(dentry, bindex, flags);
+}
+
+static inline
+int au_diropq_remove(struct dentry *dentry, aufs_bindex_t bindex, int dlgt)
+{
+	unsigned int flags = !AuDiropq_CREATE;
+	if (unlikely(dlgt))
+		au_fset_diropq(flags, DLGT);
+	return PTR_ERR(au_diropq_sio(dentry, bindex, flags));
+}
+
+/* ---------------------------------------------------------------------- */
+
+#ifdef CONFIG_AUFS_ROBR
+/* robr.c */
+int au_test_robr_wh(struct qstr *name, struct dentry *h_parent,
+		    struct qstr *wh_name, int try_sio, struct au_ndx *ndx);
+int au_test_robr_shwh(struct super_block *sb, const struct qstr *name);
+#else
+static inline
+int au_test_robr_wh(struct qstr *name, struct dentry *h_parent,
+		    struct qstr *wh_name, int try_sio, struct au_ndx *ndx)
+{
+	return au_wh_test(h_parent, wh_name, try_sio, ndx);
+}
+
+static inline
+int au_test_robr_shwh(struct super_block *sb, const struct qstr *name)
+{
+	if (unlikely(!au_opt_test(au_mntflags(sb), SHWH)
+		     && !strncmp(name->name, AUFS_WH_PFX, AUFS_WH_PFX_LEN)))
+		return -EPERM;
+	return 0;
+}
+#endif /* CONFIG_AUFS_ROBR */
+
+#endif /* __KERNEL__ */
+#endif /* __AUFS_WHOUT_H__ */
-- 
1.5.5.1.308.g1fbb5.dirty

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