1 files changed, 200 insertions, 52 deletions

diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 8d501093660f..193d6fb363b7 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -58,7 +58,7 @@ static void nfs_readdir_clear_array(struct page*);
 const struct file_operations nfs_dir_operations = {
 	.llseek		= nfs_llseek_dir,
 	.read		= generic_read_dir,
-	.iterate	= nfs_readdir,
+	.iterate_shared	= nfs_readdir,
 	.open		= nfs_opendir,
 	.release	= nfs_closedir,
 	.fsync		= nfs_fsync_dir,
@@ -155,6 +155,7 @@ typedef struct {
 	loff_t		current_index;
 	decode_dirent_t	decode;
 
+	unsigned long	dir_verifier;
 	unsigned long	timestamp;
 	unsigned long	gencount;
 	unsigned int	cache_entry_index;
@@ -162,6 +163,17 @@ typedef struct {
 	bool eof;
 } nfs_readdir_descriptor_t;
 
+static
+void nfs_readdir_init_array(struct page *page)
+{
+	struct nfs_cache_array *array;
+
+	array = kmap_atomic(page);
+	memset(array, 0, sizeof(struct nfs_cache_array));
+	array->eof_index = -1;
+	kunmap_atomic(array);
+}
+
 /*
  * we are freeing strings created by nfs_add_to_readdir_array()
  */
@@ -174,6 +186,7 @@ void nfs_readdir_clear_array(struct page *page)
 	array = kmap_atomic(page);
 	for (i = 0; i < array->size; i++)
 		kfree(array->array[i].string.name);
+	array->size = 0;
 	kunmap_atomic(array);
 }
 
@@ -186,7 +199,7 @@ static
 int nfs_readdir_make_qstr(struct qstr *string, const char *name, unsigned int len)
 {
 	string->len = len;
-	string->name = kmemdup(name, len, GFP_KERNEL);
+	string->name = kmemdup_nul(name, len, GFP_KERNEL);
 	if (string->name == NULL)
 		return -ENOMEM;
 	/*
@@ -341,6 +354,7 @@ int nfs_readdir_xdr_filler(struct page **pages, nfs_readdir_descriptor_t *desc,
  again:
 	timestamp = jiffies;
 	gencount = nfs_inc_attr_generation_counter();
+	desc->dir_verifier = nfs_save_change_attribute(inode);
 	error = NFS_PROTO(inode)->readdir(file_dentry(file), cred, entry->cookie, pages,
 					  NFS_SERVER(inode)->dtsize, desc->plus);
 	if (error < 0) {
@@ -437,18 +451,19 @@ void nfs_force_use_readdirplus(struct inode *dir)
 	if (nfs_server_capable(dir, NFS_CAP_READDIRPLUS) &&
 	    !list_empty(&nfsi->open_files)) {
 		set_bit(NFS_INO_ADVISE_RDPLUS, &nfsi->flags);
-		invalidate_mapping_pages(dir->i_mapping, 0, -1);
+		invalidate_mapping_pages(dir->i_mapping,
+			nfsi->page_index + 1, -1);
 	}
 }
 
 static
-void nfs_prime_dcache(struct dentry *parent, struct nfs_entry *entry)
+void nfs_prime_dcache(struct dentry *parent, struct nfs_entry *entry,
+		unsigned long dir_verifier)
 {
 	struct qstr filename = QSTR_INIT(entry->name, entry->len);
 	DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq);
 	struct dentry *dentry;
 	struct dentry *alias;
-	struct inode *dir = d_inode(parent);
 	struct inode *inode;
 	int status;
 
@@ -487,7 +502,7 @@ again:
 		if (nfs_same_file(dentry, entry)) {
 			if (!entry->fh->size)
 				goto out;
-			nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
+			nfs_set_verifier(dentry, dir_verifier);
 			status = nfs_refresh_inode(d_inode(dentry), entry->fattr);
 			if (!status)
 				nfs_setsecurity(d_inode(dentry), entry->fattr, entry->label);
@@ -513,7 +528,7 @@ again:
 		dput(dentry);
 		dentry = alias;
 	}
-	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
+	nfs_set_verifier(dentry, dir_verifier);
 out:
 	dput(dentry);
 }
@@ -551,7 +566,8 @@ int nfs_readdir_page_filler(nfs_readdir_descriptor_t *desc, struct nfs_entry *en
 		count++;
 
 		if (desc->plus)
-			nfs_prime_dcache(file_dentry(desc->file), entry);
+			nfs_prime_dcache(file_dentry(desc->file), entry,
+					desc->dir_verifier);
 
 		status = nfs_readdir_add_to_array(entry, page);
 		if (status != 0)
@@ -610,6 +626,8 @@ int nfs_readdir_xdr_to_array(nfs_readdir_descriptor_t *desc, struct page *page,
 	int status = -ENOMEM;
 	unsigned int array_size = ARRAY_SIZE(pages);
 
+	nfs_readdir_init_array(page);
+
 	entry.prev_cookie = 0;
 	entry.cookie = desc->last_cookie;
 	entry.eof = 0;
@@ -626,8 +644,6 @@ int nfs_readdir_xdr_to_array(nfs_readdir_descriptor_t *desc, struct page *page,
 	}
 
 	array = kmap(page);
-	memset(array, 0, sizeof(struct nfs_cache_array));
-	array->eof_index = -1;
 
 	status = nfs_readdir_alloc_pages(pages, array_size);
 	if (status < 0)
@@ -682,6 +698,7 @@ int nfs_readdir_filler(void *data, struct page* page)
 	unlock_page(page);
 	return 0;
  error:
+	nfs_readdir_clear_array(page);
 	unlock_page(page);
 	return ret;
 }
@@ -689,8 +706,6 @@ int nfs_readdir_filler(void *data, struct page* page)
 static
 void cache_page_release(nfs_readdir_descriptor_t *desc)
 {
-	if (!desc->page->mapping)
-		nfs_readdir_clear_array(desc->page);
 	put_page(desc->page);
 	desc->page = NULL;
 }
@@ -704,19 +719,32 @@ struct page *get_cache_page(nfs_readdir_descriptor_t *desc)
 
 /*
  * Returns 0 if desc->dir_cookie was found on page desc->page_index
+ * and locks the page to prevent removal from the page cache.
  */
 static
-int find_cache_page(nfs_readdir_descriptor_t *desc)
+int find_and_lock_cache_page(nfs_readdir_descriptor_t *desc)
 {
+	struct inode *inode = file_inode(desc->file);
+	struct nfs_inode *nfsi = NFS_I(inode);
 	int res;
 
 	desc->page = get_cache_page(desc);
 	if (IS_ERR(desc->page))
 		return PTR_ERR(desc->page);
-
-	res = nfs_readdir_search_array(desc);
+	res = lock_page_killable(desc->page);
 	if (res != 0)
-		cache_page_release(desc);
+		goto error;
+	res = -EAGAIN;
+	if (desc->page->mapping != NULL) {
+		res = nfs_readdir_search_array(desc);
+		if (res == 0) {
+			nfsi->page_index = desc->page_index;
+			return 0;
+		}
+	}
+	unlock_page(desc->page);
+error:
+	cache_page_release(desc);
 	return res;
 }
 
@@ -731,7 +759,7 @@ int readdir_search_pagecache(nfs_readdir_descriptor_t *desc)
 		desc->last_cookie = 0;
 	}
 	do {
-		res = find_cache_page(desc);
+		res = find_and_lock_cache_page(desc);
 	} while (res == -EAGAIN);
 	return res;
 }
@@ -770,7 +798,6 @@ int nfs_do_filldir(nfs_readdir_descriptor_t *desc)
 		desc->eof = true;
 
 	kunmap(desc->page);
-	cache_page_release(desc);
 	dfprintk(DIRCACHE, "NFS: nfs_do_filldir() filling ended @ cookie %Lu; returning = %d\n",
 			(unsigned long long)*desc->dir_cookie, res);
 	return res;
@@ -816,13 +843,13 @@ int uncached_readdir(nfs_readdir_descriptor_t *desc)
 
 	status = nfs_do_filldir(desc);
 
+ out_release:
+	nfs_readdir_clear_array(desc->page);
+	cache_page_release(desc);
  out:
 	dfprintk(DIRCACHE, "NFS: %s: returns %d\n",
 			__func__, status);
 	return status;
- out_release:
-	cache_page_release(desc);
-	goto out;
 }
 
 /* The file offset position represents the dirent entry number.  A
@@ -887,6 +914,8 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)
 			break;
 
 		res = nfs_do_filldir(desc);
+		unlock_page(desc->page);
+		cache_page_release(desc);
 		if (res < 0)
 			break;
 	} while (!desc->eof);
@@ -957,14 +986,113 @@ static int nfs_fsync_dir(struct file *filp, loff_t start, loff_t end,
  * full lookup on all child dentries of 'dir' whenever a change occurs
  * on the server that might have invalidated our dcache.
  *
+ * Note that we reserve bit '0' as a tag to let us know when a dentry
+ * was revalidated while holding a delegation on its inode.
+ *
  * The caller should be holding dir->i_lock
  */
 void nfs_force_lookup_revalidate(struct inode *dir)
 {
-	NFS_I(dir)->cache_change_attribute++;
+	NFS_I(dir)->cache_change_attribute += 2;
 }
 EXPORT_SYMBOL_GPL(nfs_force_lookup_revalidate);
 
+/**
+ * nfs_verify_change_attribute - Detects NFS remote directory changes
+ * @dir: pointer to parent directory inode
+ * @verf: previously saved change attribute
+ *
+ * Return "false" if the verifiers doesn't match the change attribute.
+ * This would usually indicate that the directory contents have changed on
+ * the server, and that any dentries need revalidating.
+ */
+static bool nfs_verify_change_attribute(struct inode *dir, unsigned long verf)
+{
+	return (verf & ~1UL) == nfs_save_change_attribute(dir);
+}
+
+static void nfs_set_verifier_delegated(unsigned long *verf)
+{
+	*verf |= 1UL;
+}
+
+#if IS_ENABLED(CONFIG_NFS_V4)
+static void nfs_unset_verifier_delegated(unsigned long *verf)
+{
+	*verf &= ~1UL;
+}
+#endif /* IS_ENABLED(CONFIG_NFS_V4) */
+
+static bool nfs_test_verifier_delegated(unsigned long verf)
+{
+	return verf & 1;
+}
+
+static bool nfs_verifier_is_delegated(struct dentry *dentry)
+{
+	return nfs_test_verifier_delegated(dentry->d_time);
+}
+
+static void nfs_set_verifier_locked(struct dentry *dentry, unsigned long verf)
+{
+	struct inode *inode = d_inode(dentry);
+
+	if (!nfs_verifier_is_delegated(dentry) &&
+	    !nfs_verify_change_attribute(d_inode(dentry->d_parent), verf))
+		goto out;
+	if (inode && NFS_PROTO(inode)->have_delegation(inode, FMODE_READ))
+		nfs_set_verifier_delegated(&verf);
+out:
+	dentry->d_time = verf;
+}
+
+/**
+ * nfs_set_verifier - save a parent directory verifier in the dentry
+ * @dentry: pointer to dentry
+ * @verf: verifier to save
+ *
+ * Saves the parent directory verifier in @dentry. If the inode has
+ * a delegation, we also tag the dentry as having been revalidated
+ * while holding a delegation so that we know we don't have to
+ * look it up again after a directory change.
+ */
+void nfs_set_verifier(struct dentry *dentry, unsigned long verf)
+{
+
+	spin_lock(&dentry->d_lock);
+	nfs_set_verifier_locked(dentry, verf);
+	spin_unlock(&dentry->d_lock);
+}
+EXPORT_SYMBOL_GPL(nfs_set_verifier);
+
+#if IS_ENABLED(CONFIG_NFS_V4)
+/**
+ * nfs_clear_verifier_delegated - clear the dir verifier delegation tag
+ * @inode: pointer to inode
+ *
+ * Iterates through the dentries in the inode alias list and clears
+ * the tag used to indicate that the dentry has been revalidated
+ * while holding a delegation.
+ * This function is intended for use when the delegation is being
+ * returned or revoked.
+ */
+void nfs_clear_verifier_delegated(struct inode *inode)
+{
+	struct dentry *alias;
+
+	if (!inode)
+		return;
+	spin_lock(&inode->i_lock);
+	hlist_for_each_entry(alias, &inode->i_dentry, d_u.d_alias) {
+		spin_lock(&alias->d_lock);
+		nfs_unset_verifier_delegated(&alias->d_time);
+		spin_unlock(&alias->d_lock);
+	}
+	spin_unlock(&inode->i_lock);
+}
+EXPORT_SYMBOL_GPL(nfs_clear_verifier_delegated);
+#endif /* IS_ENABLED(CONFIG_NFS_V4) */
+
 /*
  * A check for whether or not the parent directory has changed.
  * In the case it has, we assume that the dentries are untrustworthy
@@ -1133,6 +1261,7 @@ nfs_lookup_revalidate_dentry(struct inode *dir, struct dentry *dentry,
 	struct nfs_fh *fhandle;
 	struct nfs_fattr *fattr;
 	struct nfs4_label *label;
+	unsigned long dir_verifier;
 	int ret;
 
 	ret = -ENOMEM;
@@ -1142,10 +1271,18 @@ nfs_lookup_revalidate_dentry(struct inode *dir, struct dentry *dentry,
 	if (fhandle == NULL || fattr == NULL || IS_ERR(label))
 		goto out;
 
-	ret = NFS_PROTO(dir)->lookup(dir, &dentry->d_name, fhandle, fattr, label);
+	dir_verifier = nfs_save_change_attribute(dir);
+	ret = NFS_PROTO(dir)->lookup(dir, dentry, fhandle, fattr, label);
 	if (ret < 0) {
-		if (ret == -ESTALE || ret == -ENOENT)
+		switch (ret) {
+		case -ESTALE:
+		case -ENOENT:
 			ret = 0;
+			break;
+		case -ETIMEDOUT:
+			if (NFS_SERVER(inode)->flags & NFS_MOUNT_SOFTREVAL)
+				ret = 1;
+		}
 		goto out;
 	}
 	ret = 0;
@@ -1155,7 +1292,7 @@ nfs_lookup_revalidate_dentry(struct inode *dir, struct dentry *dentry,
 		goto out;
 
 	nfs_setsecurity(inode, fattr, label);
-	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
+	nfs_set_verifier(dentry, dir_verifier);
 
 	/* set a readdirplus hint that we had a cache miss */
 	nfs_force_use_readdirplus(dir);
@@ -1197,7 +1334,7 @@ nfs_do_lookup_revalidate(struct inode *dir, struct dentry *dentry,
 		goto out_bad;
 	}
 
-	if (NFS_PROTO(dir)->have_delegation(inode, FMODE_READ))
+	if (nfs_verifier_is_delegated(dentry))
 		return nfs_lookup_revalidate_delegated(dir, dentry, inode);
 
 	/* Force a full look up iff the parent directory has changed */
@@ -1382,6 +1519,7 @@ struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, unsigned in
 	struct nfs_fh *fhandle = NULL;
 	struct nfs_fattr *fattr = NULL;
 	struct nfs4_label *label = NULL;
+	unsigned long dir_verifier;
 	int error;
 
 	dfprintk(VFS, "NFS: lookup(%pd2)\n", dentry);
@@ -1407,8 +1545,9 @@ struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, unsigned in
 	if (IS_ERR(label))
 		goto out;
 
+	dir_verifier = nfs_save_change_attribute(dir);
 	trace_nfs_lookup_enter(dir, dentry, flags);
-	error = NFS_PROTO(dir)->lookup(dir, &dentry->d_name, fhandle, fattr, label);
+	error = NFS_PROTO(dir)->lookup(dir, dentry, fhandle, fattr, label);
 	if (error == -ENOENT)
 		goto no_entry;
 	if (error < 0) {
@@ -1430,7 +1569,7 @@ no_entry:
 			goto out_label;
 		dentry = res;
 	}
-	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
+	nfs_set_verifier(dentry, dir_verifier);
 out_label:
 	trace_nfs_lookup_exit(dir, dentry, flags, error);
 	nfs4_label_free(label);
@@ -1487,7 +1626,7 @@ static int nfs_finish_open(struct nfs_open_context *ctx,
 	if (S_ISREG(file->f_path.dentry->d_inode->i_mode))
 		nfs_file_set_open_context(file, ctx);
 	else
-		err = -ESTALE;
+		err = -EOPENSTALE;
 out:
 	return err;
 }
@@ -1635,7 +1774,7 @@ nfs4_do_lookup_revalidate(struct inode *dir, struct dentry *dentry,
 	if (inode == NULL)
 		goto full_reval;
 
-	if (NFS_PROTO(dir)->have_delegation(inode, FMODE_READ))
+	if (nfs_verifier_is_delegated(dentry))
 		return nfs_lookup_revalidate_delegated(dir, dentry, inode);
 
 	/* NFS only supports OPEN on regular files */
@@ -1669,10 +1808,8 @@ static int nfs4_lookup_revalidate(struct dentry *dentry, unsigned int flags)
 
 #endif /* CONFIG_NFSV4 */
 
-/*
- * Code common to create, mkdir, and mknod.
- */
-int nfs_instantiate(struct dentry *dentry, struct nfs_fh *fhandle,
+struct dentry *
+nfs_add_or_obtain(struct dentry *dentry, struct nfs_fh *fhandle,
 				struct nfs_fattr *fattr,
 				struct nfs4_label *label)
 {
@@ -1680,15 +1817,12 @@ int nfs_instantiate(struct dentry *dentry, struct nfs_fh *fhandle,
 	struct inode *dir = d_inode(parent);
 	struct inode *inode;
 	struct dentry *d;
-	int error = -EACCES;
+	int error;
 
 	d_drop(dentry);
 
-	/* We may have been initialized further down */
-	if (d_really_is_positive(dentry))
-		goto out;
 	if (fhandle->size == 0) {
-		error = NFS_PROTO(dir)->lookup(dir, &dentry->d_name, fhandle, fattr, NULL);
+		error = NFS_PROTO(dir)->lookup(dir, dentry, fhandle, fattr, NULL);
 		if (error)
 			goto out_error;
 	}
@@ -1702,18 +1836,32 @@ int nfs_instantiate(struct dentry *dentry, struct nfs_fh *fhandle,
 	}
 	inode = nfs_fhget(dentry->d_sb, fhandle, fattr, label);
 	d = d_splice_alias(inode, dentry);
-	if (IS_ERR(d)) {
-		error = PTR_ERR(d);
-		goto out_error;
-	}
-	dput(d);
 out:
 	dput(parent);
-	return 0;
+	return d;
 out_error:
 	nfs_mark_for_revalidate(dir);
-	dput(parent);
-	return error;
+	d = ERR_PTR(error);
+	goto out;
+}
+EXPORT_SYMBOL_GPL(nfs_add_or_obtain);
+
+/*
+ * Code common to create, mkdir, and mknod.
+ */
+int nfs_instantiate(struct dentry *dentry, struct nfs_fh *fhandle,
+				struct nfs_fattr *fattr,
+				struct nfs4_label *label)
+{
+	struct dentry *d;
+
+	d = nfs_add_or_obtain(dentry, fhandle, fattr, label);
+	if (IS_ERR(d))
+		return PTR_ERR(d);
+
+	/* Callers don't care */
+	dput(d);
+	return 0;
 }
 EXPORT_SYMBOL_GPL(nfs_instantiate);
 
@@ -2303,11 +2451,11 @@ static int nfs_access_get_cached(struct inode *inode, const struct cred *cred, s
 		/* Found an entry, is our attribute cache valid? */
 		if (!nfs_check_cache_invalid(inode, NFS_INO_INVALID_ACCESS))
 			break;
+		if (!retry)
+			break;
 		err = -ECHILD;
 		if (!may_block)
 			goto out;
-		if (!retry)
-			goto out_zap;
 		spin_unlock(&inode->i_lock);
 		err = __nfs_revalidate_inode(NFS_SERVER(inode), inode);
 		if (err)
@@ -2344,7 +2492,7 @@ static int nfs_access_get_cached_rcu(struct inode *inode, const struct cred *cre
 	lh = rcu_dereference(nfsi->access_cache_entry_lru.prev);
 	cache = list_entry(lh, struct nfs_access_entry, lru);
 	if (lh == &nfsi->access_cache_entry_lru ||
-	    cred != cache->cred)
+	    cred_fscmp(cred, cache->cred) != 0)
 		cache = NULL;
 	if (cache == NULL)
 		goto out;
@@ -2467,7 +2615,7 @@ static int nfs_do_access(struct inode *inode, const struct cred *cred, int mask)
 {
 	struct nfs_access_entry cache;
 	bool may_block = (mask & MAY_NOT_BLOCK) == 0;
-	int cache_mask;
+	int cache_mask = -1;
 	int status;
 
 	trace_nfs_access_enter(inode);
@@ -2506,7 +2654,7 @@ out_cached:
 	if ((mask & ~cache_mask & (MAY_READ | MAY_WRITE | MAY_EXEC)) != 0)
 		status = -EACCES;
 out:
-	trace_nfs_access_exit(inode, status);
+	trace_nfs_access_exit(inode, mask, cache_mask, status);
 	return status;
 }