aufs: hnotify 3/3, callers

In order to prevent firing the notify event from aufs itself, hnotify
feature is suspend/resume-able. They are combined with mutex lock/unlock
for the parent dir.

See also previous commits.
Signed-off-by: J. R. Okajima <hooanon05g@gmail.com>

fs/aufs/branch.c

@@ -19,6 +19,7 @@ static void au_br_do_free(struct au_branch *br)
 	struct au_wbr *wbr;
 	struct au_dykey **key;
 
+	au_hnotify_fin_br(br);
 	au_xino_put(br);
 
 	AuLCntZero(au_lcnt_read(&br->br_nfiles, /*do_rev*/0));
@@ -120,13 +121,16 @@ static struct au_branch *au_br_alloc(struct super_block *sb, int new_nbranch,
 	add_branch->br_xino = au_xino_alloc(/*nfile*/1);
 	if (unlikely(!add_branch->br_xino))
 		goto out_br;
+	err = au_hnotify_init_br(add_branch, perm);
+	if (unlikely(err))
+		goto out_xino;
 
 	if (au_br_writable(perm)) {
 		/* may be freed separately at changing the branch permission */
 		add_branch->br_wbr = kzalloc(sizeof(*add_branch->br_wbr),
 					     GFP_NOFS);
 		if (unlikely(!add_branch->br_wbr))
-			goto out_xino;
+			goto out_hnotify;
 	}
 
 	root = sb->s_root;
@@ -143,6 +147,8 @@ static struct au_branch *au_br_alloc(struct super_block *sb, int new_nbranch,
 
 	au_kfree_rcu(add_branch->br_wbr);
 
+out_hnotify:
+	au_hnotify_fin_br(add_branch);
 out_xino:
 	au_xino_put(add_branch);
 out_br:
@@ -275,7 +281,7 @@ static int au_br_init_wh(struct super_block *sb, struct au_branch *br,
 	bindex = au_br_index(sb, br->br_id);
 	if (0 <= bindex) {
 		hdir = au_hi(d_inode(sb->s_root), bindex);
-		inode_lock_nested(hdir->hi_inode, AuLsc_I_PARENT);
+		au_hn_inode_lock_nested(hdir, AuLsc_I_PARENT);
 	} else {
 		h_dentry = au_br_dentry(br);
 		h_inode = d_inode(h_dentry);
@@ -289,7 +295,7 @@ static int au_br_init_wh(struct super_block *sb, struct au_branch *br,
 		wbr_wh_write_unlock(wbr);
 	}
 	if (hdir)
-		inode_unlock(hdir->hi_inode);
+		au_hn_inode_unlock(hdir);
 	else
 		inode_unlock(h_inode);
 	vfsub_mnt_drop_write(au_br_mnt(br));

fs/aufs/debug.c

@@ -55,7 +55,7 @@ char *au_plevel = KERN_DEBUG;
 
 /* ---------------------------------------------------------------------- */
 
-static int do_pri_inode(aufs_bindex_t bindex, struct inode *inode,
+static int do_pri_inode(aufs_bindex_t bindex, struct inode *inode, int hn,
 			struct dentry *wh)
 {
 	char *n = NULL;
@@ -75,12 +75,12 @@ static int do_pri_inode(aufs_bindex_t bindex, struct inode *inode,
 	}
 
 	dpri("i%d: %p, i%lu, %s, cnt %d, nl %u, 0%o, sz %llu, blk %llu,"
-	     " ct %lld, np %lu, st 0x%lx, f 0x%x, v %llu, g %x%s%.*s\n",
+	     " hn %d, ct %lld, np %lu, st 0x%lx, f 0x%x, v %llu, g %x%s%.*s\n",
 	     bindex, inode,
 	     inode->i_ino, inode->i_sb ? au_sbtype(inode->i_sb) : "??",
 	     atomic_read(&inode->i_count), inode->i_nlink, inode->i_mode,
 	     i_size_read(inode), (unsigned long long)inode->i_blocks,
-	     (long long)timespec64_to_ns(&inode->i_ctime) & 0x0ffff,
+	     hn, (long long)timespec64_to_ns(&inode->i_ctime) & 0x0ffff,
 	     inode->i_mapping ? inode->i_mapping->nrpages : 0,
 	     inode->i_state, inode->i_flags, inode_peek_iversion(inode),
 	     inode->i_generation,
@@ -91,10 +91,11 @@ static int do_pri_inode(aufs_bindex_t bindex, struct inode *inode,
 void au_dpri_inode(struct inode *inode)
 {
 	struct au_iinfo *iinfo;
+	struct au_hinode *hi;
 	aufs_bindex_t bindex;
-	int err;
+	int err, hn;
 
-	err = do_pri_inode(-1, inode, NULL);
+	err = do_pri_inode(-1, inode, -1, NULL);
 	if (err || !au_test_aufs(inode->i_sb) || au_is_bad_inode(inode))
 		return;
 
@@ -103,9 +104,12 @@ void au_dpri_inode(struct inode *inode)
 	     iinfo->ii_btop, iinfo->ii_bbot, au_iigen(inode));
 	if (iinfo->ii_btop < 0)
 		return;
-	for (bindex = iinfo->ii_btop; bindex <= iinfo->ii_bbot; bindex++)
-		do_pri_inode(bindex, iinfo->ii_hinode[0 + bindex].hi_inode,
-			     iinfo->ii_hinode[0 + bindex].hi_whdentry);
+	hn = 0;
+	for (bindex = iinfo->ii_btop; bindex <= iinfo->ii_bbot; bindex++) {
+		hi = au_hinode(iinfo, bindex);
+		hn = !!au_hn(hi);
+		do_pri_inode(bindex, hi->hi_inode, hn, hi->hi_whdentry);
+	}
 }
 
 void au_dpri_dalias(struct inode *inode)
@@ -121,6 +125,7 @@ void au_dpri_dalias(struct inode *inode)
 static int do_pri_dentry(aufs_bindex_t bindex, struct dentry *dentry)
 {
 	struct dentry *wh = NULL;
+	int hn;
 	struct inode *inode;
 	struct au_iinfo *iinfo;
 	struct au_hinode *hi;
@@ -136,6 +141,7 @@ static int do_pri_dentry(aufs_bindex_t bindex, struct dentry *dentry)
 	     dentry->d_sb ? au_sbtype(dentry->d_sb) : "??",
 	     au_dcount(dentry), dentry->d_flags,
 	     d_unhashed(dentry) ? "un" : "");
+	hn = -1;
 	inode = NULL;
 	if (d_is_positive(dentry))
 		inode = d_inode(dentry);
@@ -145,9 +151,10 @@ static int do_pri_dentry(aufs_bindex_t bindex, struct dentry *dentry)
 	    && !au_is_bad_inode(inode)) {
 		iinfo = au_ii(inode);
 		hi = au_hinode(iinfo, bindex);
+		hn = !!au_hn(hi);
 		wh = hi->hi_whdentry;
 	}
-	do_pri_inode(bindex, inode, wh);
+	do_pri_inode(bindex, inode, hn, wh);
 	return 0;
 }
 

fs/aufs/dir.c

@@ -85,14 +85,14 @@ static void au_do_dir_ts(void *arg)
 	if (err)
 		goto out_unlock;
 	hdir = au_hi(dir, btop);
-	inode_lock_nested(hdir->hi_inode, AuLsc_I_PARENT);
+	au_hn_inode_lock_nested(hdir, AuLsc_I_PARENT);
 	h_dir = au_h_iptr(dir, btop);
 	if (h_dir->i_nlink
 	    && timespec64_compare(&h_dir->i_mtime, &dt.dt_mtime) < 0) {
 		dt.dt_h_path = h_path;
 		au_dtime_revert(&dt);
 	}
-	inode_unlock(hdir->hi_inode);
+	au_hn_inode_unlock(hdir);
 	vfsub_mnt_drop_write(h_path.mnt);
 	au_cpup_attr_timesizes(dir);
 

fs/aufs/i_op.c

@@ -12,7 +12,7 @@
 void au_pin_hdir_unlock(struct au_pin *p)
 {
 	if (p->hdir)
-		inode_unlock(p->hdir->hi_inode);
+		au_hn_inode_unlock(p->hdir);
 }
 
 int au_pin_hdir_lock(struct au_pin *p)
@@ -24,7 +24,7 @@ int au_pin_hdir_lock(struct au_pin *p)
 		goto out;
 
 	/* even if an error happens later, keep this lock */
-	inode_lock_nested(p->hdir->hi_inode, p->lsc_hi);
+	au_hn_inode_lock_nested(p->hdir, p->lsc_hi);
 
 	err = -EBUSY;
 	if (unlikely(p->hdir->hi_inode != d_inode(p->h_parent)))

fs/aufs/iinfo.c

@@ -25,6 +25,7 @@ struct inode *au_h_iptr(struct inode *inode, aufs_bindex_t bindex)
 /* todo: hard/soft set? */
 void au_hiput(struct au_hinode *hinode)
 {
+	au_hn_free(hinode);
 	dput(hinode->hi_whdentry);
 	iput(hinode->hi_inode);
 }
@@ -37,6 +38,8 @@ unsigned int au_hi_flags(struct inode *inode, int isdir)
 	flags = 0;
 	if (au_opt_test(mnt_flags, XINO))
 		au_fset_hi(flags, XINO);
+	if (isdir && au_opt_test(mnt_flags, UDBA_HNOTIFY))
+		au_fset_hi(flags, HNOTIFY);
 	return flags;
 }
 
@@ -74,6 +77,13 @@ void au_set_h_iptr(struct inode *inode, aufs_bindex_t bindex,
 			if (unlikely(err))
 				AuIOErr1("failed au_xino_write() %d\n", err);
 		}
+
+		if (au_ftest_hi(flags, HNOTIFY)
+		    && au_br_hnotifyable(br->br_perm)) {
+			err = au_hn_alloc(hinode, inode);
+			if (unlikely(err))
+				AuIOErr1("au_hn_alloc() %d\n", err);
+		}
 	}
 }
 
@@ -159,6 +169,7 @@ void au_hinode_init(struct au_hinode *hinode)
 {
 	hinode->hi_inode = NULL;
 	hinode->hi_id = -1;
+	au_hn_init(hinode);
 	hinode->hi_whdentry = NULL;
 }
 

fs/aufs/module.c

@@ -81,7 +81,10 @@ static void au_cache_fin(void)
 	 * destroy cache.
 	 */
 	rcu_barrier();
-	for (i = 0; i < AuCache_Last; i++) {
+
+	/* excluding AuCache_HNOTIFY */
+	BUILD_BUG_ON(AuCache_HNOTIFY + 1 != AuCache_Last);
+	for (i = 0; i < AuCache_HNOTIFY; i++) {
 		kmem_cache_destroy(au_cache[i]);
 		au_cache[i] = NULL;
 	}
@@ -161,7 +164,7 @@ static int __init aufs_init(void)
 
 	au_dir_roflags = au_file_roflags(O_DIRECTORY | O_LARGEFILE);
 
-	memset(au_cache, 0, sizeof(au_cache));
+	memset(au_cache, 0, sizeof(au_cache));	/* including hnotify */
 
 	au_sbilist_init();
 	sysaufs_brs_init();
@@ -175,14 +178,19 @@ static int __init aufs_init(void)
 	err = au_wkq_init();
 	if (unlikely(err))
 		goto out_procfs;
-	err = au_cache_init();
+	err = au_hnotify_init();
 	if (unlikely(err))
 		goto out_wkq;
+	err = au_cache_init();
+	if (unlikely(err))
+		goto out_hin;
 
 	/* since we define pr_fmt, call printk directly */
 	printk(KERN_INFO AUFS_NAME " " AUFS_VERSION "\n");
 	goto out; /* success */
 
+out_hin:
+	au_hnotify_fin();
 out_wkq:
 	au_wkq_fin();
 out_procfs:

fs/aufs/opts.c

@@ -22,6 +22,7 @@ enum {
 	Opt_trunc_xino_path, Opt_itrunc_xino,
 	Opt_trunc_xib, Opt_notrunc_xib,
 	Opt_plink, Opt_noplink, Opt_list_plink,
+	Opt_udba,
 	Opt_dio, Opt_nodio,
 	Opt_wbr_copyup, Opt_wbr_create,
 	Opt_tail, Opt_ignore, Opt_ignore_silent, Opt_err
@@ -54,6 +55,8 @@ static match_table_t options = {
 	{Opt_list_plink, "list_plink"},
 #endif
 
+	{Opt_udba, "udba=%s"},
+
 	{Opt_dio, "dio"},
 	{Opt_nodio, "nodio"},
 
@@ -259,10 +262,15 @@ void au_optstr_br_perm(au_br_perm_str_t *str, int perm)
 static match_table_t udbalevel = {
 	{AuOpt_UDBA_REVAL, "reval"},
 	{AuOpt_UDBA_NONE, "none"},
+#ifdef CONFIG_AUFS_HNOTIFY
+	{AuOpt_UDBA_HNOTIFY, "notify"}, /* abstraction */
+#ifdef CONFIG_AUFS_HFSNOTIFY
+	{AuOpt_UDBA_HNOTIFY, "fsnotify"},
+#endif
+#endif
 	{-1, NULL}
 };
 
-/* re-commit later */ __maybe_unused
 static int noinline_for_stack udba_val(char *str)
 {
 	substring_t args[MAX_OPT_ARGS];
@@ -461,6 +469,10 @@ static void dump_opts(struct au_opts *opts)
 		case Opt_list_plink:
 			AuLabel(list_plink);
 			break;
+		case Opt_udba:
+			AuDbg("udba %d, %s\n",
+				  opt->udba, au_optstr_udba(opt->udba));
+			break;
 		case Opt_dio:
 			AuLabel(dio);
 			break;
@@ -730,6 +742,15 @@ int au_opts_parse(struct super_block *sb, char *str, struct au_opts *opts)
 			opt->type = token;
 			break;
 
+		case Opt_udba:
+			opt->udba = udba_val(a->args[0].from);
+			if (opt->udba >= 0) {
+				err = 0;
+				opt->type = token;
+			} else
+				pr_err("wrong value, %s\n", opt_str);
+			break;
+
 		case Opt_wbr_create:
 			u.create = &opt->wbr_create;
 			u.create->wbr_create
@@ -839,6 +860,12 @@ static int au_opt_simple(struct super_block *sb, struct au_opt *opt,
 	err = 1; /* handled */
 	sbinfo = au_sbi(sb);
 	switch (opt->type) {
+	case Opt_udba:
+		sbinfo->si_mntflags &= ~AuOptMask_UDBA;
+		sbinfo->si_mntflags |= opt->udba;
+		opts->given_udba |= opt->udba;
+		break;
+
 	case Opt_plink:
 		au_opt_set(sbinfo->si_mntflags, PLINK);
 		break;
@@ -964,6 +991,7 @@ int au_opts_verify(struct super_block *sb, unsigned long sb_flags,
 	SiMustAnyLock(sb);
 
 	sbinfo = au_sbi(sb);
+	AuDebugOn(!(sbinfo->si_mntflags & AuOptMask_UDBA));
 
 	if (!(sb_flags & SB_RDONLY)) {
 		if (unlikely(!au_br_writable(au_sbr_perm(sb, 0))))
@@ -1017,13 +1045,13 @@ int au_opts_verify(struct super_block *sb, unsigned long sb_flags,
 			continue;
 
 		hdir = au_hi(dir, bindex);
-		inode_lock_nested(hdir->hi_inode, AuLsc_I_PARENT);
+		au_hn_inode_lock_nested(hdir, AuLsc_I_PARENT);
 		if (wbr)
 			wbr_wh_write_lock(wbr);
 		err = au_wh_init(br, sb);
 		if (wbr)
 			wbr_wh_write_unlock(wbr);
-		inode_unlock(hdir->hi_inode);
+		au_hn_inode_unlock(hdir);
 
 		if (!err && do_free) {
 			au_kfree_rcu(wbr);
@@ -1038,10 +1066,12 @@ int au_opts_mount(struct super_block *sb, struct au_opts *opts)
 {
 	int err;
 	unsigned int tmp;
-	aufs_bindex_t bbot;
+	aufs_bindex_t bindex, bbot;
 	struct au_opt *opt;
 	struct au_opt_xino *opt_xino, xino;
 	struct au_sbinfo *sbinfo;
+	struct au_branch *br;
+	struct inode *dir;
 
 	SiMustWriteLock(sb);
 
@@ -1055,10 +1085,11 @@ int au_opts_mount(struct super_block *sb, struct au_opts *opts)
 	else if (unlikely(err < 0))
 		goto out;
 
-	/* disable xino temporary */
+	/* disable xino and udba temporary */
 	sbinfo = au_sbi(sb);
 	tmp = sbinfo->si_mntflags;
 	au_opt_clr(sbinfo->si_mntflags, XINO);
+	au_opt_set_udba(sbinfo->si_mntflags, UDBA_REVAL);
 
 	opt = opts->opt;
 	while (err >= 0 && opt->type != Opt_tail)
@@ -1100,7 +1131,23 @@ int au_opts_mount(struct super_block *sb, struct au_opts *opts)
 			goto out;
 	}
 
+	/* restore udba */
+	tmp &= AuOptMask_UDBA;
+	sbinfo->si_mntflags &= ~AuOptMask_UDBA;
+	sbinfo->si_mntflags |= tmp;
 	bbot = au_sbbot(sb);
+	for (bindex = 0; bindex <= bbot; bindex++) {
+		br = au_sbr(sb, bindex);
+		err = au_hnotify_reset_br(tmp, br, br->br_perm);
+		if (unlikely(err))
+			AuIOErr("hnotify failed on br %d, %d, ignored\n",
+				bindex, err);
+		/* go on even if err */
+	}
+	if (au_opt_test(tmp, UDBA_HNOTIFY)) {
+		dir = d_inode(sb->s_root);
+		au_hn_reset(dir, au_hi_flags(dir, /*isdir*/1) & ~AuHi_XINO);
+	}
 
 out:
 	return err;

fs/aufs/opts.h

@@ -120,6 +120,7 @@ struct au_opt {
 		struct au_opt_xino	xino;
 		struct au_opt_xino_itrunc xino_itrunc;
 		struct au_opt_add	add;
+		int			udba;
 		struct au_opt_wbr_create wbr_create;
 		int			wbr_copyup;
 		/* add more later */
@@ -139,6 +140,7 @@ struct au_opts {
 	struct au_opt	*opt;
 	int		max_opt;
 
+	unsigned int	given_udba;
 	unsigned int	flags;
 	unsigned long	sb_flags;
 };

fs/aufs/whout.c

@@ -496,7 +496,7 @@ static void reinit_br_wh(void *arg)
 	h_root = au_h_dptr(a->sb->s_root, bindex);
 	AuDebugOn(h_root != au_br_dentry(a->br));
 
-	inode_lock_nested(hdir->hi_inode, AuLsc_I_PARENT);
+	au_hn_inode_lock_nested(hdir, AuLsc_I_PARENT);
 	wbr_wh_write_lock(wbr);
 	err = au_h_verify(wbr->wbr_whbase, au_opt_udba(a->sb), hdir->hi_inode,
 			  h_root, a->br);
@@ -520,7 +520,7 @@ static void reinit_br_wh(void *arg)
 	if (!err)
 		err = au_wh_init(a->br, a->sb);
 	wbr_wh_write_unlock(wbr);
-	inode_unlock(hdir->hi_inode);
+	au_hn_inode_unlock(hdir);
 	di_read_unlock(a->sb->s_root, AuLock_IR);
 
 out:

fs/aufs/wkq.c

@@ -258,6 +258,14 @@ static void au_wkq_comp_free(struct completion *comp __maybe_unused)
 
 static void au_wkq_run(struct au_wkinfo *wkinfo)
 {
+	if (au_ftest_wkq(wkinfo->flags, NEST)) {
+		if (au_wkq_test()) {
+			AuWarn1("wkq from wkq, unless silly-rename on NFS,"
+				" due to a dead dir by UDBA,"
+				" or async xino write?\n");
+			AuDebugOn(au_ftest_wkq(wkinfo->flags, WAIT));
+		}
+	} else
 		au_dbg_verify_kthread();
 
 	if (au_ftest_wkq(wkinfo->flags, WAIT)) {

fs/aufs/xino.c

@@ -124,7 +124,7 @@ static void au_xino_lock_dir(struct super_block *sb, struct path *xipath,
 	if (bindex >= 0) {
 		/* rw branch root */
 		ldir->hdir = au_hi(d_inode(sb->s_root), bindex);
-		inode_lock_nested(ldir->hdir->hi_inode, AuLsc_I_PARENT);
+		au_hn_inode_lock_nested(ldir->hdir, AuLsc_I_PARENT);
 	} else {
 		/* other */
 		ldir->parent = au_dget_parent_lock(xipath->dentry,
@@ -136,7 +136,7 @@ static void au_xino_lock_dir(struct super_block *sb, struct path *xipath,
 static void au_xino_unlock_dir(struct au_xino_lock_dir *ldir)
 {
 	if (ldir->hdir)
-		inode_unlock(ldir->hdir->hi_inode);
+		au_hn_inode_unlock(ldir->hdir);
 	else {
 		inode_unlock(ldir->dir);
 		dput(ldir->parent);