From 44cdc1d952e3f7aa9944c1bbf38fc23f49885017 Mon Sep 17 00:00:00 2001
From: Al Viro <viro@zeniv.linux.org.uk>
Date: Sun, 27 Sep 2020 11:18:30 -0400
Subject: convert ->f_ep_links/->fllink to hlist

we don't care about the order of elements there

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 include/linux/fs.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'include/linux/fs.h')

diff --git a/include/linux/fs.h b/include/linux/fs.h
index 0bd126418bb6..b484ba0e474a 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -946,7 +946,7 @@ struct file {
 
 #ifdef CONFIG_EPOLL
 	/* Used by fs/eventpoll.c to link all the hooks to this file */
-	struct list_head	f_ep_links;
+	struct hlist_head	f_ep_links;
 	struct list_head	f_tfile_llink;
 #endif /* #ifdef CONFIG_EPOLL */
 	struct address_space	*f_mapping;
-- 
cgit v1.3-14-g43fede


From 319c15174757aaedacc89a6e55c965416f130e64 Mon Sep 17 00:00:00 2001
From: Al Viro <viro@zeniv.linux.org.uk>
Date: Thu, 1 Oct 2020 20:45:51 -0400
Subject: epoll: take epitem list out of struct file

Move the head of epitem list out of struct file; for epoll ones it's
moved into struct eventpoll (->refs there), for non-epoll - into
the new object (struct epitem_head).  In place of ->f_ep_links we
leave a pointer to the list head (->f_ep).

->f_ep is protected by ->f_lock and it's zeroed as soon as the list
of epitems becomes empty (that can happen only in ep_remove() by
now).

The list of files for reverse path check is *not* going through
struct file now - it's a single-linked list going through epitem_head
instances.  It's terminated by ERR_PTR(-1) (== EP_UNACTIVE_POINTER),
so the elements of list can be distinguished by head->next != NULL.

epitem_head instances are allocated at ep_insert() time (by
attach_epitem()) and freed either by ep_remove() (if it empties
the set of epitems *and* epitem_head does not belong to the
reverse path check list) or by clear_tfile_check_list() when
the list is emptied (if the set of epitems is empty by that
point).  Allocations are done from a separate slab - minimal kmalloc()
size is too large on some architectures.

As the result, we trim struct file _and_ get rid of the games with
temporary file references.

Locking and barriers are interesting (aren't they always); see unlist_file()
and ep_remove() for details.  The non-obvious part is that ep_remove() needs
to decide if it will be the one to free the damn thing *before* actually
storing NULL to head->epitems.first - that's what smp_load_acquire is for
in there.  unlist_file() lockless path is safe, since we hit it only if
we observe NULL in head->epitems.first and whoever had done that store is
guaranteed to have observed non-NULL in head->next.  IOW, their last access
had been the store of NULL into ->epitems.first and we can safely free
the sucker.  OTOH, we are under rcu_read_lock() and both epitem and
epitem->file have their freeing RCU-delayed.  So if we see non-NULL
->epitems.first, we can grab ->f_lock (all epitems in there share the
same struct file) and safely recheck the emptiness of ->epitems; again,
->next is still non-NULL, so ep_remove() couldn't have freed head yet.
->f_lock serializes us wrt ep_remove(); the rest is trivial.

Note that once head->epitems becomes NULL, nothing can get inserted into
it - the only remaining reference to head after that point is from the
reverse path check list.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/eventpoll.c            | 168 ++++++++++++++++++++++++++++++++++------------
 fs/file_table.c           |   1 -
 include/linux/eventpoll.h |  11 +--
 include/linux/fs.h        |   5 +-
 4 files changed, 129 insertions(+), 56 deletions(-)

(limited to 'include/linux/fs.h')

diff --git a/fs/eventpoll.c b/fs/eventpoll.c
index 8f68095c5ce9..d89aead356df 100644
--- a/fs/eventpoll.c
+++ b/fs/eventpoll.c
@@ -215,6 +215,7 @@ struct eventpoll {
 
 	/* used to optimize loop detection check */
 	u64 gen;
+	struct hlist_head refs;
 
 #ifdef CONFIG_NET_RX_BUSY_POLL
 	/* used to track busy poll napi_id */
@@ -259,7 +260,45 @@ static struct kmem_cache *pwq_cache __read_mostly;
  * List of files with newly added links, where we may need to limit the number
  * of emanating paths. Protected by the epmutex.
  */
-static LIST_HEAD(tfile_check_list);
+struct epitems_head {
+	struct hlist_head epitems;
+	struct epitems_head *next;
+};
+static struct epitems_head *tfile_check_list = EP_UNACTIVE_PTR;
+
+static struct kmem_cache *ephead_cache __read_mostly;
+
+static inline void free_ephead(struct epitems_head *head)
+{
+	if (head)
+		kmem_cache_free(ephead_cache, head);
+}
+
+static void list_file(struct file *file)
+{
+	struct epitems_head *head;
+
+	head = container_of(file->f_ep, struct epitems_head, epitems);
+	if (!head->next) {
+		head->next = tfile_check_list;
+		tfile_check_list = head;
+	}
+}
+
+static void unlist_file(struct epitems_head *head)
+{
+	struct epitems_head *to_free = head;
+	struct hlist_node *p = rcu_dereference(hlist_first_rcu(&head->epitems));
+	if (p) {
+		struct epitem *epi= container_of(p, struct epitem, fllink);
+		spin_lock(&epi->ffd.file->f_lock);
+		if (!hlist_empty(&head->epitems))
+			to_free = NULL;
+		head->next = NULL;
+		spin_unlock(&epi->ffd.file->f_lock);
+	}
+	free_ephead(to_free);
+}
 
 #ifdef CONFIG_SYSCTL
 
@@ -632,6 +671,8 @@ static void epi_rcu_free(struct rcu_head *head)
 static int ep_remove(struct eventpoll *ep, struct epitem *epi)
 {
 	struct file *file = epi->ffd.file;
+	struct epitems_head *to_free;
+	struct hlist_head *head;
 
 	lockdep_assert_irqs_enabled();
 
@@ -642,8 +683,20 @@ static int ep_remove(struct eventpoll *ep, struct epitem *epi)
 
 	/* Remove the current item from the list of epoll hooks */
 	spin_lock(&file->f_lock);
+	to_free = NULL;
+	head = file->f_ep;
+	if (head->first == &epi->fllink && !epi->fllink.next) {
+		file->f_ep = NULL;
+		if (!is_file_epoll(file)) {
+			struct epitems_head *v;
+			v = container_of(head, struct epitems_head, epitems);
+			if (!smp_load_acquire(&v->next))
+				to_free = v;
+		}
+	}
 	hlist_del_rcu(&epi->fllink);
 	spin_unlock(&file->f_lock);
+	free_ephead(to_free);
 
 	rb_erase_cached(&epi->rbn, &ep->rbr);
 
@@ -852,7 +905,11 @@ void eventpoll_release_file(struct file *file)
 	 * Besides, ep_remove() acquires the lock, so we can't hold it here.
 	 */
 	mutex_lock(&epmutex);
-	hlist_for_each_entry_safe(epi, next, &file->f_ep_links, fllink) {
+	if (unlikely(!file->f_ep)) {
+		mutex_unlock(&epmutex);
+		return;
+	}
+	hlist_for_each_entry_safe(epi, next, file->f_ep, fllink) {
 		ep = epi->ep;
 		mutex_lock_nested(&ep->mtx, 0);
 		ep_remove(ep, epi);
@@ -1248,7 +1305,7 @@ static void path_count_init(void)
 		path_count[i] = 0;
 }
 
-static int reverse_path_check_proc(struct file *file, int depth)
+static int reverse_path_check_proc(struct hlist_head *refs, int depth)
 {
 	int error = 0;
 	struct epitem *epi;
@@ -1257,14 +1314,12 @@ static int reverse_path_check_proc(struct file *file, int depth)
 		return -1;
 
 	/* CTL_DEL can remove links here, but that can't increase our count */
-	hlist_for_each_entry_rcu(epi, &file->f_ep_links, fllink) {
-		struct file *recepient = epi->ep->file;
-		if (WARN_ON(!is_file_epoll(recepient)))
-			continue;
-		if (hlist_empty(&recepient->f_ep_links))
+	hlist_for_each_entry_rcu(epi, refs, fllink) {
+		struct hlist_head *refs = &epi->ep->refs;
+		if (hlist_empty(refs))
 			error = path_count_inc(depth);
 		else
-			error = reverse_path_check_proc(recepient, depth + 1);
+			error = reverse_path_check_proc(refs, depth + 1);
 		if (error != 0)
 			break;
 	}
@@ -1272,7 +1327,7 @@ static int reverse_path_check_proc(struct file *file, int depth)
 }
 
 /**
- * reverse_path_check - The tfile_check_list is list of file *, which have
+ * reverse_path_check - The tfile_check_list is list of epitem_head, which have
  *                      links that are proposed to be newly added. We need to
  *                      make sure that those added links don't add too many
  *                      paths such that we will spend all our time waking up
@@ -1283,19 +1338,18 @@ static int reverse_path_check_proc(struct file *file, int depth)
  */
 static int reverse_path_check(void)
 {
-	int error = 0;
-	struct file *current_file;
+	struct epitems_head *p;
 
-	/* let's call this for all tfiles */
-	list_for_each_entry(current_file, &tfile_check_list, f_tfile_llink) {
+	for (p = tfile_check_list; p != EP_UNACTIVE_PTR; p = p->next) {
+		int error;
 		path_count_init();
 		rcu_read_lock();
-		error = reverse_path_check_proc(current_file, 0);
+		error = reverse_path_check_proc(&p->epitems, 0);
 		rcu_read_unlock();
 		if (error)
-			break;
+			return error;
 	}
-	return error;
+	return 0;
 }
 
 static int ep_create_wakeup_source(struct epitem *epi)
@@ -1336,6 +1390,39 @@ static noinline void ep_destroy_wakeup_source(struct epitem *epi)
 	wakeup_source_unregister(ws);
 }
 
+static int attach_epitem(struct file *file, struct epitem *epi)
+{
+	struct epitems_head *to_free = NULL;
+	struct hlist_head *head = NULL;
+	struct eventpoll *ep = NULL;
+
+	if (is_file_epoll(file))
+		ep = file->private_data;
+
+	if (ep) {
+		head = &ep->refs;
+	} else if (!READ_ONCE(file->f_ep)) {
+allocate:
+		to_free = kmem_cache_zalloc(ephead_cache, GFP_KERNEL);
+		if (!to_free)
+			return -ENOMEM;
+		head = &to_free->epitems;
+	}
+	spin_lock(&file->f_lock);
+	if (!file->f_ep) {
+		if (unlikely(!head)) {
+			spin_unlock(&file->f_lock);
+			goto allocate;
+		}
+		file->f_ep = head;
+		to_free = NULL;
+	}
+	hlist_add_head_rcu(&epi->fllink, file->f_ep);
+	spin_unlock(&file->f_lock);
+	free_ephead(to_free);
+	return 0;
+}
+
 /*
  * Must be called with "mtx" held.
  */
@@ -1367,19 +1454,21 @@ static int ep_insert(struct eventpoll *ep, const struct epoll_event *event,
 	epi->event = *event;
 	epi->next = EP_UNACTIVE_PTR;
 
-	atomic_long_inc(&ep->user->epoll_watches);
-
 	if (tep)
 		mutex_lock_nested(&tep->mtx, 1);
 	/* Add the current item to the list of active epoll hook for this file */
-	spin_lock(&tfile->f_lock);
-	hlist_add_head_rcu(&epi->fllink, &tfile->f_ep_links);
-	spin_unlock(&tfile->f_lock);
-	if (full_check && !tep) {
-		get_file(tfile);
-		list_add(&tfile->f_tfile_llink, &tfile_check_list);
+	if (unlikely(attach_epitem(tfile, epi) < 0)) {
+		kmem_cache_free(epi_cache, epi);
+		if (tep)
+			mutex_unlock(&tep->mtx);
+		return -ENOMEM;
 	}
 
+	if (full_check && !tep)
+		list_file(tfile);
+
+	atomic_long_inc(&ep->user->epoll_watches);
+
 	/*
 	 * Add the current item to the RB tree. All RB tree operations are
 	 * protected by "mtx", and ep_insert() is called with "mtx" held.
@@ -1813,11 +1902,7 @@ static int ep_loop_check_proc(struct eventpoll *ep, int depth)
 			 * not already there, and calling reverse_path_check()
 			 * during ep_insert().
 			 */
-			if (list_empty(&epi->ffd.file->f_tfile_llink)) {
-				if (get_file_rcu(epi->ffd.file))
-					list_add(&epi->ffd.file->f_tfile_llink,
-						 &tfile_check_list);
-			}
+			list_file(epi->ffd.file);
 		}
 	}
 	mutex_unlock(&ep->mtx);
@@ -1844,16 +1929,13 @@ static int ep_loop_check(struct eventpoll *ep, struct eventpoll *to)
 
 static void clear_tfile_check_list(void)
 {
-	struct file *file;
-
-	/* first clear the tfile_check_list */
-	while (!list_empty(&tfile_check_list)) {
-		file = list_first_entry(&tfile_check_list, struct file,
-					f_tfile_llink);
-		list_del_init(&file->f_tfile_llink);
-		fput(file);
+	rcu_read_lock();
+	while (tfile_check_list != EP_UNACTIVE_PTR) {
+		struct epitems_head *head = tfile_check_list;
+		tfile_check_list = head->next;
+		unlist_file(head);
 	}
-	INIT_LIST_HEAD(&tfile_check_list);
+	rcu_read_unlock();
 }
 
 /*
@@ -2003,9 +2085,8 @@ int do_epoll_ctl(int epfd, int op, int fd, struct epoll_event *epds,
 	if (error)
 		goto error_tgt_fput;
 	if (op == EPOLL_CTL_ADD) {
-		if (!hlist_empty(&f.file->f_ep_links) ||
-				ep->gen == loop_check_gen ||
-						is_file_epoll(tf.file)) {
+		if (READ_ONCE(f.file->f_ep) || ep->gen == loop_check_gen ||
+		    is_file_epoll(tf.file)) {
 			mutex_unlock(&ep->mtx);
 			error = epoll_mutex_lock(&epmutex, 0, nonblock);
 			if (error)
@@ -2216,6 +2297,9 @@ static int __init eventpoll_init(void)
 	pwq_cache = kmem_cache_create("eventpoll_pwq",
 		sizeof(struct eppoll_entry), 0, SLAB_PANIC|SLAB_ACCOUNT, NULL);
 
+	ephead_cache = kmem_cache_create("ep_head",
+		sizeof(struct epitems_head), 0, SLAB_PANIC|SLAB_ACCOUNT, NULL);
+
 	return 0;
 }
 fs_initcall(eventpoll_init);
diff --git a/fs/file_table.c b/fs/file_table.c
index 709ada3151da..45437f8e1003 100644
--- a/fs/file_table.c
+++ b/fs/file_table.c
@@ -113,7 +113,6 @@ static struct file *__alloc_file(int flags, const struct cred *cred)
 	rwlock_init(&f->f_owner.lock);
 	spin_lock_init(&f->f_lock);
 	mutex_init(&f->f_pos_lock);
-	eventpoll_init_file(f);
 	f->f_flags = flags;
 	f->f_mode = OPEN_FMODE(flags);
 	/* f->f_version: 0 */
diff --git a/include/linux/eventpoll.h b/include/linux/eventpoll.h
index 4e215ccfa792..0350393465d4 100644
--- a/include/linux/eventpoll.h
+++ b/include/linux/eventpoll.h
@@ -22,14 +22,6 @@ struct file;
 struct file *get_epoll_tfile_raw_ptr(struct file *file, int tfd, unsigned long toff);
 #endif
 
-/* Used to initialize the epoll bits inside the "struct file" */
-static inline void eventpoll_init_file(struct file *file)
-{
-	INIT_HLIST_HEAD(&file->f_ep_links);
-	INIT_LIST_HEAD(&file->f_tfile_llink);
-}
-
-
 /* Used to release the epoll bits inside the "struct file" */
 void eventpoll_release_file(struct file *file);
 
@@ -50,7 +42,7 @@ static inline void eventpoll_release(struct file *file)
 	 * because the file in on the way to be removed and nobody ( but
 	 * eventpoll ) has still a reference to this file.
 	 */
-	if (likely(hlist_empty(&file->f_ep_links)))
+	if (likely(!file->f_ep))
 		return;
 
 	/*
@@ -72,7 +64,6 @@ static inline int ep_op_has_event(int op)
 
 #else
 
-static inline void eventpoll_init_file(struct file *file) {}
 static inline void eventpoll_release(struct file *file) {}
 
 #endif
diff --git a/include/linux/fs.h b/include/linux/fs.h
index b484ba0e474a..993c540e3b77 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -923,7 +923,7 @@ struct file {
 	const struct file_operations	*f_op;
 
 	/*
-	 * Protects f_ep_links, f_flags.
+	 * Protects f_ep, f_flags.
 	 * Must not be taken from IRQ context.
 	 */
 	spinlock_t		f_lock;
@@ -946,8 +946,7 @@ struct file {
 
 #ifdef CONFIG_EPOLL
 	/* Used by fs/eventpoll.c to link all the hooks to this file */
-	struct hlist_head	f_ep_links;
-	struct list_head	f_tfile_llink;
+	struct hlist_head	*f_ep;
 #endif /* #ifdef CONFIG_EPOLL */
 	struct address_space	*f_mapping;
 	errseq_t		f_wb_err;
-- 
cgit v1.3-14-g43fede


From 8a3c84b649b033024d2349f96234b26cbd6083a6 Mon Sep 17 00:00:00 2001
From: "Darrick J. Wong" <darrick.wong@oracle.com>
Date: Tue, 10 Nov 2020 16:50:21 -0800
Subject: vfs: separate __sb_start_write into blocking and non-blocking helpers

Break this function into two helpers so that it's obvious that the
trylock versions return a value that must be checked, and the blocking
versions don't require that.  While we're at it, clean up the return
type mismatch.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Christoph Hellwig <hch@lst.de>
---
 fs/aio.c           |  2 +-
 fs/io_uring.c      |  3 +--
 fs/super.c         | 18 ++++++++++++------
 include/linux/fs.h | 21 +++++++++++----------
 4 files changed, 25 insertions(+), 19 deletions(-)

(limited to 'include/linux/fs.h')

diff --git a/fs/aio.c b/fs/aio.c
index c45c20d87538..6a21d8919409 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -1572,7 +1572,7 @@ static int aio_write(struct kiocb *req, const struct iocb *iocb,
 		 * we return to userspace.
 		 */
 		if (S_ISREG(file_inode(file)->i_mode)) {
-			__sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
+			sb_start_write(file_inode(file)->i_sb);
 			__sb_writers_release(file_inode(file)->i_sb, SB_FREEZE_WRITE);
 		}
 		req->ki_flags |= IOCB_WRITE;
diff --git a/fs/io_uring.c b/fs/io_uring.c
index b42dfa0243bf..4cbaddfe3d80 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -3532,8 +3532,7 @@ static int io_write(struct io_kiocb *req, bool force_nonblock,
 	 * we return to userspace.
 	 */
 	if (req->flags & REQ_F_ISREG) {
-		__sb_start_write(file_inode(req->file)->i_sb,
-					SB_FREEZE_WRITE, true);
+		sb_start_write(file_inode(req->file)->i_sb);
 		__sb_writers_release(file_inode(req->file)->i_sb,
 					SB_FREEZE_WRITE);
 	}
diff --git a/fs/super.c b/fs/super.c
index e1fd667454d4..59aa59279133 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -1645,16 +1645,22 @@ EXPORT_SYMBOL(__sb_end_write);
  * This is an internal function, please use sb_start_{write,pagefault,intwrite}
  * instead.
  */
-int __sb_start_write(struct super_block *sb, int level, bool wait)
+void __sb_start_write(struct super_block *sb, int level)
 {
-	if (!wait)
-		return percpu_down_read_trylock(sb->s_writers.rw_sem + level-1);
-
-	percpu_down_read(sb->s_writers.rw_sem + level-1);
-	return 1;
+	percpu_down_read(sb->s_writers.rw_sem + level - 1);
 }
 EXPORT_SYMBOL(__sb_start_write);
 
+/*
+ * This is an internal function, please use sb_start_{write,pagefault,intwrite}
+ * instead.
+ */
+bool __sb_start_write_trylock(struct super_block *sb, int level)
+{
+	return percpu_down_read_trylock(sb->s_writers.rw_sem + level - 1);
+}
+EXPORT_SYMBOL_GPL(__sb_start_write_trylock);
+
 /**
  * sb_wait_write - wait until all writers to given file system finish
  * @sb: the super for which we wait
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 0bd126418bb6..305989afd49c 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -1581,7 +1581,8 @@ extern struct timespec64 current_time(struct inode *inode);
  */
 
 void __sb_end_write(struct super_block *sb, int level);
-int __sb_start_write(struct super_block *sb, int level, bool wait);
+void __sb_start_write(struct super_block *sb, int level);
+bool __sb_start_write_trylock(struct super_block *sb, int level);
 
 #define __sb_writers_acquired(sb, lev)	\
 	percpu_rwsem_acquire(&(sb)->s_writers.rw_sem[(lev)-1], 1, _THIS_IP_)
@@ -1645,12 +1646,12 @@ static inline void sb_end_intwrite(struct super_block *sb)
  */
 static inline void sb_start_write(struct super_block *sb)
 {
-	__sb_start_write(sb, SB_FREEZE_WRITE, true);
+	__sb_start_write(sb, SB_FREEZE_WRITE);
 }
 
-static inline int sb_start_write_trylock(struct super_block *sb)
+static inline bool sb_start_write_trylock(struct super_block *sb)
 {
-	return __sb_start_write(sb, SB_FREEZE_WRITE, false);
+	return __sb_start_write_trylock(sb, SB_FREEZE_WRITE);
 }
 
 /**
@@ -1674,7 +1675,7 @@ static inline int sb_start_write_trylock(struct super_block *sb)
  */
 static inline void sb_start_pagefault(struct super_block *sb)
 {
-	__sb_start_write(sb, SB_FREEZE_PAGEFAULT, true);
+	__sb_start_write(sb, SB_FREEZE_PAGEFAULT);
 }
 
 /*
@@ -1692,12 +1693,12 @@ static inline void sb_start_pagefault(struct super_block *sb)
  */
 static inline void sb_start_intwrite(struct super_block *sb)
 {
-	__sb_start_write(sb, SB_FREEZE_FS, true);
+	__sb_start_write(sb, SB_FREEZE_FS);
 }
 
-static inline int sb_start_intwrite_trylock(struct super_block *sb)
+static inline bool sb_start_intwrite_trylock(struct super_block *sb)
 {
-	return __sb_start_write(sb, SB_FREEZE_FS, false);
+	return __sb_start_write_trylock(sb, SB_FREEZE_FS);
 }
 
 
@@ -2756,14 +2757,14 @@ static inline void file_start_write(struct file *file)
 {
 	if (!S_ISREG(file_inode(file)->i_mode))
 		return;
-	__sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, true);
+	sb_start_write(file_inode(file)->i_sb);
 }
 
 static inline bool file_start_write_trylock(struct file *file)
 {
 	if (!S_ISREG(file_inode(file)->i_mode))
 		return true;
-	return __sb_start_write(file_inode(file)->i_sb, SB_FREEZE_WRITE, false);
+	return sb_start_write_trylock(file_inode(file)->i_sb);
 }
 
 static inline void file_end_write(struct file *file)
-- 
cgit v1.3-14-g43fede


From 9b8523423b23ee3dfd88e32f5b7207be56a4e782 Mon Sep 17 00:00:00 2001
From: "Darrick J. Wong" <darrick.wong@oracle.com>
Date: Tue, 10 Nov 2020 16:50:21 -0800
Subject: vfs: move __sb_{start,end}_write* to fs.h

Now that we've straightened out the callers, move these three functions
to fs.h since they're fairly trivial.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Jan Kara <jack@suse.cz>
---
 fs/super.c         | 30 ------------------------------
 include/linux/fs.h | 21 ++++++++++++++++++---
 2 files changed, 18 insertions(+), 33 deletions(-)

(limited to 'include/linux/fs.h')

diff --git a/fs/super.c b/fs/super.c
index 59aa59279133..98bb0629ee10 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -1631,36 +1631,6 @@ int super_setup_bdi(struct super_block *sb)
 }
 EXPORT_SYMBOL(super_setup_bdi);
 
-/*
- * This is an internal function, please use sb_end_{write,pagefault,intwrite}
- * instead.
- */
-void __sb_end_write(struct super_block *sb, int level)
-{
-	percpu_up_read(sb->s_writers.rw_sem + level-1);
-}
-EXPORT_SYMBOL(__sb_end_write);
-
-/*
- * This is an internal function, please use sb_start_{write,pagefault,intwrite}
- * instead.
- */
-void __sb_start_write(struct super_block *sb, int level)
-{
-	percpu_down_read(sb->s_writers.rw_sem + level - 1);
-}
-EXPORT_SYMBOL(__sb_start_write);
-
-/*
- * This is an internal function, please use sb_start_{write,pagefault,intwrite}
- * instead.
- */
-bool __sb_start_write_trylock(struct super_block *sb, int level)
-{
-	return percpu_down_read_trylock(sb->s_writers.rw_sem + level - 1);
-}
-EXPORT_SYMBOL_GPL(__sb_start_write_trylock);
-
 /**
  * sb_wait_write - wait until all writers to given file system finish
  * @sb: the super for which we wait
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 305989afd49c..6dabd019cab0 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -1580,9 +1580,24 @@ extern struct timespec64 current_time(struct inode *inode);
  * Snapshotting support.
  */
 
-void __sb_end_write(struct super_block *sb, int level);
-void __sb_start_write(struct super_block *sb, int level);
-bool __sb_start_write_trylock(struct super_block *sb, int level);
+/*
+ * These are internal functions, please use sb_start_{write,pagefault,intwrite}
+ * instead.
+ */
+static inline void __sb_end_write(struct super_block *sb, int level)
+{
+	percpu_up_read(sb->s_writers.rw_sem + level-1);
+}
+
+static inline void __sb_start_write(struct super_block *sb, int level)
+{
+	percpu_down_read(sb->s_writers.rw_sem + level - 1);
+}
+
+static inline bool __sb_start_write_trylock(struct super_block *sb, int level)
+{
+	return percpu_down_read_trylock(sb->s_writers.rw_sem + level - 1);
+}
 
 #define __sb_writers_acquired(sb, lev)	\
 	percpu_rwsem_acquire(&(sb)->s_writers.rw_sem[(lev)-1], 1, _THIS_IP_)
-- 
cgit v1.3-14-g43fede


From 60b498852bf219c0bf2b0864c69972840978ca43 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Mon, 16 Nov 2020 15:21:18 +0100
Subject: fs: remove get_super_thawed and get_super_exclusive_thawed

Just open code the wait in the only caller of both functions.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/internal.h      |  2 ++
 fs/quota/quota.c   | 31 ++++++++++++++++++++++++-------
 fs/super.c         | 51 ++-------------------------------------------------
 include/linux/fs.h |  4 +---
 4 files changed, 29 insertions(+), 59 deletions(-)

(limited to 'include/linux/fs.h')

diff --git a/fs/internal.h b/fs/internal.h
index a7cd0f64faa4..47be21dfeebe 100644
--- a/fs/internal.h
+++ b/fs/internal.h
@@ -114,7 +114,9 @@ extern struct file *alloc_empty_file_noaccount(int, const struct cred *);
  */
 extern int reconfigure_super(struct fs_context *);
 extern bool trylock_super(struct super_block *sb);
+struct super_block *__get_super(struct block_device *bdev, bool excl);
 extern struct super_block *user_get_super(dev_t);
+void put_super(struct super_block *sb);
 extern bool mount_capable(struct fs_context *);
 
 /*
diff --git a/fs/quota/quota.c b/fs/quota/quota.c
index 9af95c7a0bbe..f3d32b0d9008 100644
--- a/fs/quota/quota.c
+++ b/fs/quota/quota.c
@@ -20,6 +20,7 @@
 #include <linux/writeback.h>
 #include <linux/nospec.h>
 #include "compat.h"
+#include "../internal.h"
 
 static int check_quotactl_permission(struct super_block *sb, int type, int cmd,
 				     qid_t id)
@@ -868,6 +869,7 @@ static struct super_block *quotactl_block(const char __user *special, int cmd)
 	struct block_device *bdev;
 	struct super_block *sb;
 	struct filename *tmp = getname(special);
+	bool excl = false, thawed = false;
 
 	if (IS_ERR(tmp))
 		return ERR_CAST(tmp);
@@ -875,17 +877,32 @@ static struct super_block *quotactl_block(const char __user *special, int cmd)
 	putname(tmp);
 	if (IS_ERR(bdev))
 		return ERR_CAST(bdev);
-	if (quotactl_cmd_onoff(cmd))
-		sb = get_super_exclusive_thawed(bdev);
-	else if (quotactl_cmd_write(cmd))
-		sb = get_super_thawed(bdev);
-	else
-		sb = get_super(bdev);
+
+	if (quotactl_cmd_onoff(cmd)) {
+		excl = true;
+		thawed = true;
+	} else if (quotactl_cmd_write(cmd)) {
+		thawed = true;
+	}
+
+retry:
+	sb = __get_super(bdev, excl);
+	if (thawed && sb && sb->s_writers.frozen != SB_UNFROZEN) {
+		if (excl)
+			up_write(&sb->s_umount);
+		else
+			up_read(&sb->s_umount);
+		wait_event(sb->s_writers.wait_unfrozen,
+			   sb->s_writers.frozen == SB_UNFROZEN);
+		put_super(sb);
+		goto retry;
+	}
+
 	bdput(bdev);
 	if (!sb)
 		return ERR_PTR(-ENODEV);
-
 	return sb;
+
 #else
 	return ERR_PTR(-ENODEV);
 #endif
diff --git a/fs/super.c b/fs/super.c
index 98bb0629ee10..343e5c1e538d 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -307,7 +307,7 @@ static void __put_super(struct super_block *s)
  *	Drops a temporary reference, frees superblock if there's no
  *	references left.
  */
-static void put_super(struct super_block *sb)
+void put_super(struct super_block *sb)
 {
 	spin_lock(&sb_lock);
 	__put_super(sb);
@@ -740,7 +740,7 @@ void iterate_supers_type(struct file_system_type *type,
 
 EXPORT_SYMBOL(iterate_supers_type);
 
-static struct super_block *__get_super(struct block_device *bdev, bool excl)
+struct super_block *__get_super(struct block_device *bdev, bool excl)
 {
 	struct super_block *sb;
 
@@ -789,53 +789,6 @@ struct super_block *get_super(struct block_device *bdev)
 }
 EXPORT_SYMBOL(get_super);
 
-static struct super_block *__get_super_thawed(struct block_device *bdev,
-					      bool excl)
-{
-	while (1) {
-		struct super_block *s = __get_super(bdev, excl);
-		if (!s || s->s_writers.frozen == SB_UNFROZEN)
-			return s;
-		if (!excl)
-			up_read(&s->s_umount);
-		else
-			up_write(&s->s_umount);
-		wait_event(s->s_writers.wait_unfrozen,
-			   s->s_writers.frozen == SB_UNFROZEN);
-		put_super(s);
-	}
-}
-
-/**
- *	get_super_thawed - get thawed superblock of a device
- *	@bdev: device to get the superblock for
- *
- *	Scans the superblock list and finds the superblock of the file system
- *	mounted on the device. The superblock is returned once it is thawed
- *	(or immediately if it was not frozen). %NULL is returned if no match
- *	is found.
- */
-struct super_block *get_super_thawed(struct block_device *bdev)
-{
-	return __get_super_thawed(bdev, false);
-}
-EXPORT_SYMBOL(get_super_thawed);
-
-/**
- *	get_super_exclusive_thawed - get thawed superblock of a device
- *	@bdev: device to get the superblock for
- *
- *	Scans the superblock list and finds the superblock of the file system
- *	mounted on the device. The superblock is returned once it is thawed
- *	(or immediately if it was not frozen) and s_umount semaphore is held
- *	in exclusive mode. %NULL is returned if no match is found.
- */
-struct super_block *get_super_exclusive_thawed(struct block_device *bdev)
-{
-	return __get_super_thawed(bdev, true);
-}
-EXPORT_SYMBOL(get_super_exclusive_thawed);
-
 /**
  * get_active_super - get an active reference to the superblock of a device
  * @bdev: device to get the superblock for
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 8667d0cdc71e..a61df0dd4f19 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -1409,7 +1409,7 @@ enum {
 
 struct sb_writers {
 	int				frozen;		/* Is sb frozen? */
-	wait_queue_head_t		wait_unfrozen;	/* for get_super_thawed() */
+	wait_queue_head_t		wait_unfrozen;	/* wait for thaw */
 	struct percpu_rw_semaphore	rw_sem[SB_FREEZE_LEVELS];
 };
 
@@ -3132,8 +3132,6 @@ extern struct file_system_type *get_filesystem(struct file_system_type *fs);
 extern void put_filesystem(struct file_system_type *fs);
 extern struct file_system_type *get_fs_type(const char *name);
 extern struct super_block *get_super(struct block_device *);
-extern struct super_block *get_super_thawed(struct block_device *);
-extern struct super_block *get_super_exclusive_thawed(struct block_device *bdev);
 extern struct super_block *get_active_super(struct block_device *bdev);
 extern void drop_super(struct super_block *sb);
 extern void drop_super_exclusive(struct super_block *sb);
-- 
cgit v1.3-14-g43fede


From 4e7b5671c6a883d94b5428e1a9c141bbd56cb2a6 Mon Sep 17 00:00:00 2001
From: Christoph Hellwig <hch@lst.de>
Date: Mon, 23 Nov 2020 13:38:40 +0100
Subject: block: remove i_bdev

Switch the block device lookup interfaces to directly work with a dev_t
so that struct block_device references are only acquired by the
blkdev_get variants (and the blk-cgroup special case).  This means that
we now don't need an extra reference in the inode and can generally
simplify handling of struct block_device to keep the lookups contained
in the core block layer code.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Acked-by: Tejun Heo <tj@kernel.org>
Acked-by: Coly Li <colyli@suse.de>		[bcache]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 block/ioctl.c                                |   3 +-
 drivers/block/loop.c                         |   8 +-
 drivers/md/bcache/super.c                    |  20 +--
 drivers/md/dm-table.c                        |   9 +-
 drivers/mtd/mtdsuper.c                       |  17 +--
 drivers/target/target_core_file.c            |   6 +-
 drivers/usb/gadget/function/storage_common.c |   8 +-
 fs/block_dev.c                               | 196 +++++++--------------------
 fs/btrfs/volumes.c                           |  13 +-
 fs/inode.c                                   |   3 -
 fs/internal.h                                |   7 +-
 fs/io_uring.c                                |  10 +-
 fs/pipe.c                                    |   5 +-
 fs/quota/quota.c                             |  19 ++-
 fs/statfs.c                                  |   2 +-
 fs/super.c                                   |  44 +++---
 include/linux/blkdev.h                       |   2 +-
 include/linux/fs.h                           |   1 -
 18 files changed, 121 insertions(+), 252 deletions(-)

(limited to 'include/linux/fs.h')

diff --git a/block/ioctl.c b/block/ioctl.c
index 0c09bb7a6ff3..a6d8171221c7 100644
--- a/block/ioctl.c
+++ b/block/ioctl.c
@@ -590,8 +590,7 @@ long compat_blkdev_ioctl(struct file *file, unsigned cmd, unsigned long arg)
 {
 	int ret;
 	void __user *argp = compat_ptr(arg);
-	struct inode *inode = file->f_mapping->host;
-	struct block_device *bdev = inode->i_bdev;
+	struct block_device *bdev = I_BDEV(file->f_mapping->host);
 	struct gendisk *disk = bdev->bd_disk;
 	fmode_t mode = file->f_mode;
 	loff_t size;
diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index b42c728620c9..26c7aafba7c5 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -675,10 +675,10 @@ static int loop_validate_file(struct file *file, struct block_device *bdev)
 	while (is_loop_device(f)) {
 		struct loop_device *l;
 
-		if (f->f_mapping->host->i_bdev == bdev)
+		if (f->f_mapping->host->i_rdev == bdev->bd_dev)
 			return -EBADF;
 
-		l = f->f_mapping->host->i_bdev->bd_disk->private_data;
+		l = I_BDEV(f->f_mapping->host)->bd_disk->private_data;
 		if (l->lo_state != Lo_bound) {
 			return -EINVAL;
 		}
@@ -885,9 +885,7 @@ static void loop_config_discard(struct loop_device *lo)
 	 * file-backed loop devices: discarded regions read back as zero.
 	 */
 	if (S_ISBLK(inode->i_mode) && !lo->lo_encrypt_key_size) {
-		struct request_queue *backingq;
-
-		backingq = bdev_get_queue(inode->i_bdev);
+		struct request_queue *backingq = bdev_get_queue(I_BDEV(inode));
 
 		max_discard_sectors = backingq->limits.max_write_zeroes_sectors;
 		granularity = backingq->limits.discard_granularity ?:
diff --git a/drivers/md/bcache/super.c b/drivers/md/bcache/super.c
index a6a5e21e4fd1..c55d3c58a7ef 100644
--- a/drivers/md/bcache/super.c
+++ b/drivers/md/bcache/super.c
@@ -2380,38 +2380,38 @@ kobj_attribute_write(register,		register_bcache);
 kobj_attribute_write(register_quiet,	register_bcache);
 kobj_attribute_write(pendings_cleanup,	bch_pending_bdevs_cleanup);
 
-static bool bch_is_open_backing(struct block_device *bdev)
+static bool bch_is_open_backing(dev_t dev)
 {
 	struct cache_set *c, *tc;
 	struct cached_dev *dc, *t;
 
 	list_for_each_entry_safe(c, tc, &bch_cache_sets, list)
 		list_for_each_entry_safe(dc, t, &c->cached_devs, list)
-			if (dc->bdev == bdev)
+			if (dc->bdev->bd_dev == dev)
 				return true;
 	list_for_each_entry_safe(dc, t, &uncached_devices, list)
-		if (dc->bdev == bdev)
+		if (dc->bdev->bd_dev == dev)
 			return true;
 	return false;
 }
 
-static bool bch_is_open_cache(struct block_device *bdev)
+static bool bch_is_open_cache(dev_t dev)
 {
 	struct cache_set *c, *tc;
 
 	list_for_each_entry_safe(c, tc, &bch_cache_sets, list) {
 		struct cache *ca = c->cache;
 
-		if (ca->bdev == bdev)
+		if (ca->bdev->bd_dev == dev)
 			return true;
 	}
 
 	return false;
 }
 
-static bool bch_is_open(struct block_device *bdev)
+static bool bch_is_open(dev_t dev)
 {
-	return bch_is_open_cache(bdev) || bch_is_open_backing(bdev);
+	return bch_is_open_cache(dev) || bch_is_open_backing(dev);
 }
 
 struct async_reg_args {
@@ -2535,9 +2535,11 @@ static ssize_t register_bcache(struct kobject *k, struct kobj_attribute *attr,
 				  sb);
 	if (IS_ERR(bdev)) {
 		if (bdev == ERR_PTR(-EBUSY)) {
-			bdev = lookup_bdev(strim(path));
+			dev_t dev;
+
 			mutex_lock(&bch_register_lock);
-			if (!IS_ERR(bdev) && bch_is_open(bdev))
+			if (lookup_bdev(strim(path), &dev) == 0 &&
+			    bch_is_open(dev))
 				err = "device already registered";
 			else
 				err = "device busy";
diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c
index ce543b761be7..dea677721710 100644
--- a/drivers/md/dm-table.c
+++ b/drivers/md/dm-table.c
@@ -348,16 +348,9 @@ static int upgrade_mode(struct dm_dev_internal *dd, fmode_t new_mode,
 dev_t dm_get_dev_t(const char *path)
 {
 	dev_t dev;
-	struct block_device *bdev;
 
-	bdev = lookup_bdev(path);
-	if (IS_ERR(bdev))
+	if (lookup_bdev(path, &dev))
 		dev = name_to_dev_t(path);
-	else {
-		dev = bdev->bd_dev;
-		bdput(bdev);
-	}
-
 	return dev;
 }
 EXPORT_SYMBOL_GPL(dm_get_dev_t);
diff --git a/drivers/mtd/mtdsuper.c b/drivers/mtd/mtdsuper.c
index c3e2098372f2..38b6aa849c63 100644
--- a/drivers/mtd/mtdsuper.c
+++ b/drivers/mtd/mtdsuper.c
@@ -120,8 +120,8 @@ int get_tree_mtd(struct fs_context *fc,
 				struct fs_context *fc))
 {
 #ifdef CONFIG_BLOCK
-	struct block_device *bdev;
-	int ret, major;
+	dev_t dev;
+	int ret;
 #endif
 	int mtdnr;
 
@@ -169,20 +169,15 @@ int get_tree_mtd(struct fs_context *fc,
 	/* try the old way - the hack where we allowed users to mount
 	 * /dev/mtdblock$(n) but didn't actually _use_ the blockdev
 	 */
-	bdev = lookup_bdev(fc->source);
-	if (IS_ERR(bdev)) {
-		ret = PTR_ERR(bdev);
+	ret = lookup_bdev(fc->source, &dev);
+	if (ret) {
 		errorf(fc, "MTD: Couldn't look up '%s': %d", fc->source, ret);
 		return ret;
 	}
 	pr_debug("MTDSB: lookup_bdev() returned 0\n");
 
-	major = MAJOR(bdev->bd_dev);
-	mtdnr = MINOR(bdev->bd_dev);
-	bdput(bdev);
-
-	if (major == MTD_BLOCK_MAJOR)
-		return mtd_get_sb_by_nr(fc, mtdnr, fill_super);
+	if (MAJOR(dev) == MTD_BLOCK_MAJOR)
+		return mtd_get_sb_by_nr(fc, MINOR(dev), fill_super);
 
 #endif /* CONFIG_BLOCK */
 
diff --git a/drivers/target/target_core_file.c b/drivers/target/target_core_file.c
index 7143d03f0e02..b0cb5b95e892 100644
--- a/drivers/target/target_core_file.c
+++ b/drivers/target/target_core_file.c
@@ -133,10 +133,10 @@ static int fd_configure_device(struct se_device *dev)
 	 */
 	inode = file->f_mapping->host;
 	if (S_ISBLK(inode->i_mode)) {
-		struct request_queue *q = bdev_get_queue(inode->i_bdev);
+		struct request_queue *q = bdev_get_queue(I_BDEV(inode));
 		unsigned long long dev_size;
 
-		fd_dev->fd_block_size = bdev_logical_block_size(inode->i_bdev);
+		fd_dev->fd_block_size = bdev_logical_block_size(I_BDEV(inode));
 		/*
 		 * Determine the number of bytes from i_size_read() minus
 		 * one (1) logical sector from underlying struct block_device
@@ -559,7 +559,7 @@ fd_execute_unmap(struct se_cmd *cmd, sector_t lba, sector_t nolb)
 
 	if (S_ISBLK(inode->i_mode)) {
 		/* The backend is block device, use discard */
-		struct block_device *bdev = inode->i_bdev;
+		struct block_device *bdev = I_BDEV(inode);
 		struct se_device *dev = cmd->se_dev;
 
 		ret = blkdev_issue_discard(bdev,
diff --git a/drivers/usb/gadget/function/storage_common.c b/drivers/usb/gadget/function/storage_common.c
index f7e6c42558eb..b859a158a414 100644
--- a/drivers/usb/gadget/function/storage_common.c
+++ b/drivers/usb/gadget/function/storage_common.c
@@ -204,7 +204,7 @@ int fsg_lun_open(struct fsg_lun *curlun, const char *filename)
 	if (!(filp->f_mode & FMODE_WRITE))
 		ro = 1;
 
-	inode = file_inode(filp);
+	inode = filp->f_mapping->host;
 	if ((!S_ISREG(inode->i_mode) && !S_ISBLK(inode->i_mode))) {
 		LINFO(curlun, "invalid file type: %s\n", filename);
 		goto out;
@@ -221,7 +221,7 @@ int fsg_lun_open(struct fsg_lun *curlun, const char *filename)
 	if (!(filp->f_mode & FMODE_CAN_WRITE))
 		ro = 1;
 
-	size = i_size_read(inode->i_mapping->host);
+	size = i_size_read(inode);
 	if (size < 0) {
 		LINFO(curlun, "unable to find file size: %s\n", filename);
 		rc = (int) size;
@@ -231,8 +231,8 @@ int fsg_lun_open(struct fsg_lun *curlun, const char *filename)
 	if (curlun->cdrom) {
 		blksize = 2048;
 		blkbits = 11;
-	} else if (inode->i_bdev) {
-		blksize = bdev_logical_block_size(inode->i_bdev);
+	} else if (S_ISBLK(inode->i_mode)) {
+		blksize = bdev_logical_block_size(I_BDEV(inode));
 		blkbits = blksize_bits(blksize);
 	} else {
 		blksize = 512;
diff --git a/fs/block_dev.c b/fs/block_dev.c
index 2b8c0586314f..6d6e4d50834c 100644
--- a/fs/block_dev.c
+++ b/fs/block_dev.c
@@ -883,7 +883,6 @@ static struct block_device *bdget(dev_t dev)
 		bdev->bd_dev = dev;
 		inode->i_mode = S_IFBLK;
 		inode->i_rdev = dev;
-		inode->i_bdev = bdev;
 		inode->i_data.a_ops = &def_blk_aops;
 		mapping_set_gfp_mask(&inode->i_data, GFP_USER);
 		unlock_new_inode(inode);
@@ -928,67 +927,8 @@ void bdput(struct block_device *bdev)
 {
 	iput(bdev->bd_inode);
 }
-
 EXPORT_SYMBOL(bdput);
  
-static struct block_device *bd_acquire(struct inode *inode)
-{
-	struct block_device *bdev;
-
-	spin_lock(&bdev_lock);
-	bdev = inode->i_bdev;
-	if (bdev && !inode_unhashed(bdev->bd_inode)) {
-		bdgrab(bdev);
-		spin_unlock(&bdev_lock);
-		return bdev;
-	}
-	spin_unlock(&bdev_lock);
-
-	/*
-	 * i_bdev references block device inode that was already shut down
-	 * (corresponding device got removed).  Remove the reference and look
-	 * up block device inode again just in case new device got
-	 * reestablished under the same device number.
-	 */
-	if (bdev)
-		bd_forget(inode);
-
-	bdev = bdget(inode->i_rdev);
-	if (bdev) {
-		spin_lock(&bdev_lock);
-		if (!inode->i_bdev) {
-			/*
-			 * We take an additional reference to bd_inode,
-			 * and it's released in clear_inode() of inode.
-			 * So, we can access it via ->i_mapping always
-			 * without igrab().
-			 */
-			bdgrab(bdev);
-			inode->i_bdev = bdev;
-			inode->i_mapping = bdev->bd_inode->i_mapping;
-		}
-		spin_unlock(&bdev_lock);
-	}
-	return bdev;
-}
-
-/* Call when you free inode */
-
-void bd_forget(struct inode *inode)
-{
-	struct block_device *bdev = NULL;
-
-	spin_lock(&bdev_lock);
-	if (!sb_is_blkdev_sb(inode->i_sb))
-		bdev = inode->i_bdev;
-	inode->i_bdev = NULL;
-	inode->i_mapping = &inode->i_data;
-	spin_unlock(&bdev_lock);
-
-	if (bdev)
-		bdput(bdev);
-}
-
 /**
  * bd_may_claim - test whether a block device can be claimed
  * @bdev: block device of interest
@@ -1497,38 +1437,45 @@ static int __blkdev_get(struct block_device *bdev, struct gendisk *disk,
 }
 
 /**
- * blkdev_get - open a block device
- * @bdev: block_device to open
+ * blkdev_get_by_dev - open a block device by device number
+ * @dev: device number of block device to open
  * @mode: FMODE_* mask
  * @holder: exclusive holder identifier
  *
- * Open @bdev with @mode.  If @mode includes %FMODE_EXCL, @bdev is
- * open with exclusive access.  Specifying %FMODE_EXCL with %NULL
- * @holder is invalid.  Exclusive opens may nest for the same @holder.
+ * Open the block device described by device number @dev. If @mode includes
+ * %FMODE_EXCL, the block device is opened with exclusive access.  Specifying
+ * %FMODE_EXCL with a %NULL @holder is invalid.  Exclusive opens may nest for
+ * the same @holder.
  *
- * On success, the reference count of @bdev is unchanged.  On failure,
- * @bdev is put.
+ * Use this interface ONLY if you really do not have anything better - i.e. when
+ * you are behind a truly sucky interface and all you are given is a device
+ * number.  Everything else should use blkdev_get_by_path().
  *
  * CONTEXT:
  * Might sleep.
  *
  * RETURNS:
- * 0 on success, -errno on failure.
+ * Reference to the block_device on success, ERR_PTR(-errno) on failure.
  */
-static int blkdev_get(struct block_device *bdev, fmode_t mode, void *holder)
+struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode, void *holder)
 {
 	struct block_device *claiming;
 	bool unblock_events = true;
+	struct block_device *bdev;
 	struct gendisk *disk;
 	int partno;
 	int ret;
 
 	ret = devcgroup_check_permission(DEVCG_DEV_BLOCK,
-			imajor(bdev->bd_inode), iminor(bdev->bd_inode),
+			MAJOR(dev), MINOR(dev),
 			((mode & FMODE_READ) ? DEVCG_ACC_READ : 0) |
 			((mode & FMODE_WRITE) ? DEVCG_ACC_WRITE : 0));
 	if (ret)
-		goto bdput;
+		return ERR_PTR(ret);
+
+	bdev = bdget(dev);
+	if (!bdev)
+		return ERR_PTR(-ENOMEM);
 
 	/*
 	 * If we lost a race with 'disk' being deleted, try again.  See md.c.
@@ -1589,10 +1536,13 @@ put_disk:
 	if (ret == -ERESTARTSYS)
 		goto retry;
 bdput:
-	if (ret)
+	if (ret) {
 		bdput(bdev);
-	return ret;
+		return ERR_PTR(ret);
+	}
+	return bdev;
 }
+EXPORT_SYMBOL(blkdev_get_by_dev);
 
 /**
  * blkdev_get_by_path - open a block device by name
@@ -1600,32 +1550,30 @@ bdput:
  * @mode: FMODE_* mask
  * @holder: exclusive holder identifier
  *
- * Open the blockdevice described by the device file at @path.  @mode
- * and @holder are identical to blkdev_get().
- *
- * On success, the returned block_device has reference count of one.
+ * Open the block device described by the device file at @path.  If @mode
+ * includes %FMODE_EXCL, the block device is opened with exclusive access.
+ * Specifying %FMODE_EXCL with a %NULL @holder is invalid.  Exclusive opens may
+ * nest for the same @holder.
  *
  * CONTEXT:
  * Might sleep.
  *
  * RETURNS:
- * Pointer to block_device on success, ERR_PTR(-errno) on failure.
+ * Reference to the block_device on success, ERR_PTR(-errno) on failure.
  */
 struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
 					void *holder)
 {
 	struct block_device *bdev;
-	int err;
-
-	bdev = lookup_bdev(path);
-	if (IS_ERR(bdev))
-		return bdev;
+	dev_t dev;
+	int error;
 
-	err = blkdev_get(bdev, mode, holder);
-	if (err)
-		return ERR_PTR(err);
+	error = lookup_bdev(path, &dev);
+	if (error)
+		return ERR_PTR(error);
 
-	if ((mode & FMODE_WRITE) && bdev_read_only(bdev)) {
+	bdev = blkdev_get_by_dev(dev, mode, holder);
+	if (!IS_ERR(bdev) && (mode & FMODE_WRITE) && bdev_read_only(bdev)) {
 		blkdev_put(bdev, mode);
 		return ERR_PTR(-EACCES);
 	}
@@ -1634,45 +1582,6 @@ struct block_device *blkdev_get_by_path(const char *path, fmode_t mode,
 }
 EXPORT_SYMBOL(blkdev_get_by_path);
 
-/**
- * blkdev_get_by_dev - open a block device by device number
- * @dev: device number of block device to open
- * @mode: FMODE_* mask
- * @holder: exclusive holder identifier
- *
- * Open the blockdevice described by device number @dev.  @mode and
- * @holder are identical to blkdev_get().
- *
- * Use it ONLY if you really do not have anything better - i.e. when
- * you are behind a truly sucky interface and all you are given is a
- * device number.  _Never_ to be used for internal purposes.  If you
- * ever need it - reconsider your API.
- *
- * On success, the returned block_device has reference count of one.
- *
- * CONTEXT:
- * Might sleep.
- *
- * RETURNS:
- * Pointer to block_device on success, ERR_PTR(-errno) on failure.
- */
-struct block_device *blkdev_get_by_dev(dev_t dev, fmode_t mode, void *holder)
-{
-	struct block_device *bdev;
-	int err;
-
-	bdev = bdget(dev);
-	if (!bdev)
-		return ERR_PTR(-ENOMEM);
-
-	err = blkdev_get(bdev, mode, holder);
-	if (err)
-		return ERR_PTR(err);
-
-	return bdev;
-}
-EXPORT_SYMBOL(blkdev_get_by_dev);
-
 static int blkdev_open(struct inode * inode, struct file * filp)
 {
 	struct block_device *bdev;
@@ -1694,14 +1603,12 @@ static int blkdev_open(struct inode * inode, struct file * filp)
 	if ((filp->f_flags & O_ACCMODE) == 3)
 		filp->f_mode |= FMODE_WRITE_IOCTL;
 
-	bdev = bd_acquire(inode);
-	if (bdev == NULL)
-		return -ENOMEM;
-
+	bdev = blkdev_get_by_dev(inode->i_rdev, filp->f_mode, filp);
+	if (IS_ERR(bdev))
+		return PTR_ERR(bdev);
 	filp->f_mapping = bdev->bd_inode->i_mapping;
 	filp->f_wb_err = filemap_sample_wb_err(filp->f_mapping);
-
-	return blkdev_get(bdev, filp->f_mode, filp);
+	return 0;
 }
 
 static void __blkdev_put(struct block_device *bdev, fmode_t mode, int for_part)
@@ -2010,37 +1917,32 @@ const struct file_operations def_blk_fops = {
  * namespace if possible and return it.  Return ERR_PTR(error)
  * otherwise.
  */
-struct block_device *lookup_bdev(const char *pathname)
+int lookup_bdev(const char *pathname, dev_t *dev)
 {
-	struct block_device *bdev;
 	struct inode *inode;
 	struct path path;
 	int error;
 
 	if (!pathname || !*pathname)
-		return ERR_PTR(-EINVAL);
+		return -EINVAL;
 
 	error = kern_path(pathname, LOOKUP_FOLLOW, &path);
 	if (error)
-		return ERR_PTR(error);
+		return error;
 
 	inode = d_backing_inode(path.dentry);
 	error = -ENOTBLK;
 	if (!S_ISBLK(inode->i_mode))
-		goto fail;
+		goto out_path_put;
 	error = -EACCES;
 	if (!may_open_dev(&path))
-		goto fail;
-	error = -ENOMEM;
-	bdev = bd_acquire(inode);
-	if (!bdev)
-		goto fail;
-out:
+		goto out_path_put;
+
+	*dev = inode->i_rdev;
+	error = 0;
+out_path_put:
 	path_put(&path);
-	return bdev;
-fail:
-	bdev = ERR_PTR(error);
-	goto out;
+	return error;
 }
 EXPORT_SYMBOL(lookup_bdev);
 
diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c
index a6406b3b8c2b..fbc4b58228f7 100644
--- a/fs/btrfs/volumes.c
+++ b/fs/btrfs/volumes.c
@@ -929,16 +929,16 @@ static noinline struct btrfs_device *device_list_add(const char *path,
 		 * make sure it's the same device if the device is mounted
 		 */
 		if (device->bdev) {
-			struct block_device *path_bdev;
+			int error;
+			dev_t path_dev;
 
-			path_bdev = lookup_bdev(path);
-			if (IS_ERR(path_bdev)) {
+			error = lookup_bdev(path, &path_dev);
+			if (error) {
 				mutex_unlock(&fs_devices->device_list_mutex);
-				return ERR_CAST(path_bdev);
+				return ERR_PTR(error);
 			}
 
-			if (device->bdev != path_bdev) {
-				bdput(path_bdev);
+			if (device->bdev->bd_dev != path_dev) {
 				mutex_unlock(&fs_devices->device_list_mutex);
 				btrfs_warn_in_rcu(device->fs_info,
 	"duplicate device %s devid %llu generation %llu scanned by %s (%d)",
@@ -947,7 +947,6 @@ static noinline struct btrfs_device *device_list_add(const char *path,
 						  task_pid_nr(current));
 				return ERR_PTR(-EEXIST);
 			}
-			bdput(path_bdev);
 			btrfs_info_in_rcu(device->fs_info,
 	"devid %llu device path %s changed to %s scanned by %s (%d)",
 					  devid, rcu_str_deref(device->name),
diff --git a/fs/inode.c b/fs/inode.c
index 9d78c37b00b8..cb008acf0efd 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -155,7 +155,6 @@ int inode_init_always(struct super_block *sb, struct inode *inode)
 	inode->i_bytes = 0;
 	inode->i_generation = 0;
 	inode->i_pipe = NULL;
-	inode->i_bdev = NULL;
 	inode->i_cdev = NULL;
 	inode->i_link = NULL;
 	inode->i_dir_seq = 0;
@@ -580,8 +579,6 @@ static void evict(struct inode *inode)
 		truncate_inode_pages_final(&inode->i_data);
 		clear_inode(inode);
 	}
-	if (S_ISBLK(inode->i_mode) && inode->i_bdev)
-		bd_forget(inode);
 	if (S_ISCHR(inode->i_mode) && inode->i_cdev)
 		cd_forget(inode);
 
diff --git a/fs/internal.h b/fs/internal.h
index 47be21dfeebe..53f890446e75 100644
--- a/fs/internal.h
+++ b/fs/internal.h
@@ -25,7 +25,6 @@ extern void __init bdev_cache_init(void);
 extern int __sync_blockdev(struct block_device *bdev, int wait);
 void iterate_bdevs(void (*)(struct block_device *, void *), void *);
 void emergency_thaw_bdev(struct super_block *sb);
-void bd_forget(struct inode *inode);
 #else
 static inline void bdev_cache_init(void)
 {
@@ -43,9 +42,6 @@ static inline int emergency_thaw_bdev(struct super_block *sb)
 {
 	return 0;
 }
-static inline void bd_forget(struct inode *inode)
-{
-}
 #endif /* CONFIG_BLOCK */
 
 /*
@@ -114,8 +110,7 @@ extern struct file *alloc_empty_file_noaccount(int, const struct cred *);
  */
 extern int reconfigure_super(struct fs_context *);
 extern bool trylock_super(struct super_block *sb);
-struct super_block *__get_super(struct block_device *bdev, bool excl);
-extern struct super_block *user_get_super(dev_t);
+struct super_block *user_get_super(dev_t, bool excl);
 void put_super(struct super_block *sb);
 extern bool mount_capable(struct fs_context *);
 
diff --git a/fs/io_uring.c b/fs/io_uring.c
index 4ead291b2976..8f13c0417f94 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -2716,11 +2716,7 @@ static struct file *__io_file_get(struct io_submit_state *state, int fd)
 
 static bool io_bdev_nowait(struct block_device *bdev)
 {
-#ifdef CONFIG_BLOCK
 	return !bdev || blk_queue_nowait(bdev_get_queue(bdev));
-#else
-	return true;
-#endif
 }
 
 /*
@@ -2733,14 +2729,16 @@ static bool io_file_supports_async(struct file *file, int rw)
 	umode_t mode = file_inode(file)->i_mode;
 
 	if (S_ISBLK(mode)) {
-		if (io_bdev_nowait(file->f_inode->i_bdev))
+		if (IS_ENABLED(CONFIG_BLOCK) &&
+		    io_bdev_nowait(I_BDEV(file->f_mapping->host)))
 			return true;
 		return false;
 	}
 	if (S_ISCHR(mode) || S_ISSOCK(mode))
 		return true;
 	if (S_ISREG(mode)) {
-		if (io_bdev_nowait(file->f_inode->i_sb->s_bdev) &&
+		if (IS_ENABLED(CONFIG_BLOCK) &&
+		    io_bdev_nowait(file->f_inode->i_sb->s_bdev) &&
 		    file->f_op != &io_uring_fops)
 			return true;
 		return false;
diff --git a/fs/pipe.c b/fs/pipe.c
index 0ac197658a2d..c5989cfd564d 100644
--- a/fs/pipe.c
+++ b/fs/pipe.c
@@ -1342,9 +1342,8 @@ out_revert_acct:
 }
 
 /*
- * After the inode slimming patch, i_pipe/i_bdev/i_cdev share the same
- * location, so checking ->i_pipe is not enough to verify that this is a
- * pipe.
+ * Note that i_pipe and i_cdev share the same location, so checking ->i_pipe is
+ * not enough to verify that this is a pipe.
  */
 struct pipe_inode_info *get_pipe_info(struct file *file, bool for_splice)
 {
diff --git a/fs/quota/quota.c b/fs/quota/quota.c
index f3d32b0d9008..6d16b2be5ac4 100644
--- a/fs/quota/quota.c
+++ b/fs/quota/quota.c
@@ -866,17 +866,18 @@ static bool quotactl_cmd_onoff(int cmd)
 static struct super_block *quotactl_block(const char __user *special, int cmd)
 {
 #ifdef CONFIG_BLOCK
-	struct block_device *bdev;
 	struct super_block *sb;
 	struct filename *tmp = getname(special);
 	bool excl = false, thawed = false;
+	int error;
+	dev_t dev;
 
 	if (IS_ERR(tmp))
 		return ERR_CAST(tmp);
-	bdev = lookup_bdev(tmp->name);
+	error = lookup_bdev(tmp->name, &dev);
 	putname(tmp);
-	if (IS_ERR(bdev))
-		return ERR_CAST(bdev);
+	if (error)
+		return ERR_PTR(error);
 
 	if (quotactl_cmd_onoff(cmd)) {
 		excl = true;
@@ -886,8 +887,10 @@ static struct super_block *quotactl_block(const char __user *special, int cmd)
 	}
 
 retry:
-	sb = __get_super(bdev, excl);
-	if (thawed && sb && sb->s_writers.frozen != SB_UNFROZEN) {
+	sb = user_get_super(dev, excl);
+	if (!sb)
+		return ERR_PTR(-ENODEV);
+	if (thawed && sb->s_writers.frozen != SB_UNFROZEN) {
 		if (excl)
 			up_write(&sb->s_umount);
 		else
@@ -897,10 +900,6 @@ retry:
 		put_super(sb);
 		goto retry;
 	}
-
-	bdput(bdev);
-	if (!sb)
-		return ERR_PTR(-ENODEV);
 	return sb;
 
 #else
diff --git a/fs/statfs.c b/fs/statfs.c
index 59f33752c131..68cb07788750 100644
--- a/fs/statfs.c
+++ b/fs/statfs.c
@@ -235,7 +235,7 @@ SYSCALL_DEFINE3(fstatfs64, unsigned int, fd, size_t, sz, struct statfs64 __user
 
 static int vfs_ustat(dev_t dev, struct kstatfs *sbuf)
 {
-	struct super_block *s = user_get_super(dev);
+	struct super_block *s = user_get_super(dev, false);
 	int err;
 	if (!s)
 		return -EINVAL;
diff --git a/fs/super.c b/fs/super.c
index 343e5c1e538d..2c6cdea2ab2d 100644
--- a/fs/super.c
+++ b/fs/super.c
@@ -740,7 +740,14 @@ void iterate_supers_type(struct file_system_type *type,
 
 EXPORT_SYMBOL(iterate_supers_type);
 
-struct super_block *__get_super(struct block_device *bdev, bool excl)
+/**
+ * get_super - get the superblock of a device
+ * @bdev: device to get the superblock for
+ *
+ * Scans the superblock list and finds the superblock of the file system
+ * mounted on the device given. %NULL is returned if no match is found.
+ */
+struct super_block *get_super(struct block_device *bdev)
 {
 	struct super_block *sb;
 
@@ -755,17 +762,11 @@ rescan:
 		if (sb->s_bdev == bdev) {
 			sb->s_count++;
 			spin_unlock(&sb_lock);
-			if (!excl)
-				down_read(&sb->s_umount);
-			else
-				down_write(&sb->s_umount);
+			down_read(&sb->s_umount);
 			/* still alive? */
 			if (sb->s_root && (sb->s_flags & SB_BORN))
 				return sb;
-			if (!excl)
-				up_read(&sb->s_umount);
-			else
-				up_write(&sb->s_umount);
+			up_read(&sb->s_umount);
 			/* nope, got unmounted */
 			spin_lock(&sb_lock);
 			__put_super(sb);
@@ -776,19 +777,6 @@ rescan:
 	return NULL;
 }
 
-/**
- *	get_super - get the superblock of a device
- *	@bdev: device to get the superblock for
- *
- *	Scans the superblock list and finds the superblock of the file system
- *	mounted on the device given. %NULL is returned if no match is found.
- */
-struct super_block *get_super(struct block_device *bdev)
-{
-	return __get_super(bdev, false);
-}
-EXPORT_SYMBOL(get_super);
-
 /**
  * get_active_super - get an active reference to the superblock of a device
  * @bdev: device to get the superblock for
@@ -820,7 +808,7 @@ restart:
 	return NULL;
 }
 
-struct super_block *user_get_super(dev_t dev)
+struct super_block *user_get_super(dev_t dev, bool excl)
 {
 	struct super_block *sb;
 
@@ -832,11 +820,17 @@ rescan:
 		if (sb->s_dev ==  dev) {
 			sb->s_count++;
 			spin_unlock(&sb_lock);
-			down_read(&sb->s_umount);
+			if (excl)
+				down_write(&sb->s_umount);
+			else
+				down_read(&sb->s_umount);
 			/* still alive? */
 			if (sb->s_root && (sb->s_flags & SB_BORN))
 				return sb;
-			up_read(&sb->s_umount);
+			if (excl)
+				up_write(&sb->s_umount);
+			else
+				up_read(&sb->s_umount);
 			/* nope, got unmounted */
 			spin_lock(&sb_lock);
 			__put_super(sb);
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 12810a19edeb..bdd7339bcda4 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -1973,7 +1973,7 @@ int bdev_read_only(struct block_device *bdev);
 int set_blocksize(struct block_device *bdev, int size);
 
 const char *bdevname(struct block_device *bdev, char *buffer);
-struct block_device *lookup_bdev(const char *);
+int lookup_bdev(const char *pathname, dev_t *dev);
 
 void blkdev_show(struct seq_file *seqf, off_t offset);
 
diff --git a/include/linux/fs.h b/include/linux/fs.h
index a61df0dd4f19..b0b358309657 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -696,7 +696,6 @@ struct inode {
 	struct list_head	i_devices;
 	union {
 		struct pipe_inode_info	*i_pipe;
-		struct block_device	*i_bdev;
 		struct cdev		*i_cdev;
 		char			*i_link;
 		unsigned		i_dir_seq;
-- 
cgit v1.3-14-g43fede


From 608af703519a58f5a7da4273809211cac27edfb2 Mon Sep 17 00:00:00 2001
From: Daniel Rosenberg <drosen@google.com>
Date: Thu, 19 Nov 2020 06:09:02 +0000
Subject: libfs: Add generic function for setting dentry_ops

This adds a function to set dentry operations at lookup time that will
work for both encrypted filenames and casefolded filenames.

A filesystem that supports both features simultaneously can use this
function during lookup preparations to set up its dentry operations once
fscrypt no longer does that itself.

Currently the casefolding dentry operation are always set if the
filesystem defines an encoding because the features is toggleable on
empty directories. Unlike in the encryption case, the dentry operations
used come from the parent. Since we don't know what set of functions
we'll eventually need, and cannot change them later, we enable the
casefolding operations if the filesystem supports them at all.

By splitting out the various cases, we support as few dentry operations
as we can get away with, maximizing compatibility with overlayfs, which
will not function if a filesystem supports certain dentry_operations.

Signed-off-by: Daniel Rosenberg <drosen@google.com>
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Gabriel Krisman Bertazi <krisman@collabora.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
---
 fs/libfs.c         | 70 ++++++++++++++++++++++++++++++++++++++++++++++++++++++
 include/linux/fs.h |  1 +
 2 files changed, 71 insertions(+)

(limited to 'include/linux/fs.h')

diff --git a/fs/libfs.c b/fs/libfs.c
index fc34361c1489..bac918699022 100644
--- a/fs/libfs.c
+++ b/fs/libfs.c
@@ -1449,4 +1449,74 @@ int generic_ci_d_hash(const struct dentry *dentry, struct qstr *str)
 	return 0;
 }
 EXPORT_SYMBOL(generic_ci_d_hash);
+
+static const struct dentry_operations generic_ci_dentry_ops = {
+	.d_hash = generic_ci_d_hash,
+	.d_compare = generic_ci_d_compare,
+};
+#endif
+
+#ifdef CONFIG_FS_ENCRYPTION
+static const struct dentry_operations generic_encrypted_dentry_ops = {
+	.d_revalidate = fscrypt_d_revalidate,
+};
+#endif
+
+#if defined(CONFIG_FS_ENCRYPTION) && defined(CONFIG_UNICODE)
+static const struct dentry_operations generic_encrypted_ci_dentry_ops = {
+	.d_hash = generic_ci_d_hash,
+	.d_compare = generic_ci_d_compare,
+	.d_revalidate = fscrypt_d_revalidate,
+};
+#endif
+
+/**
+ * generic_set_encrypted_ci_d_ops - helper for setting d_ops for given dentry
+ * @dentry:	dentry to set ops on
+ *
+ * Casefolded directories need d_hash and d_compare set, so that the dentries
+ * contained in them are handled case-insensitively.  Note that these operations
+ * are needed on the parent directory rather than on the dentries in it, and
+ * while the casefolding flag can be toggled on and off on an empty directory,
+ * dentry_operations can't be changed later.  As a result, if the filesystem has
+ * casefolding support enabled at all, we have to give all dentries the
+ * casefolding operations even if their inode doesn't have the casefolding flag
+ * currently (and thus the casefolding ops would be no-ops for now).
+ *
+ * Encryption works differently in that the only dentry operation it needs is
+ * d_revalidate, which it only needs on dentries that have the no-key name flag.
+ * The no-key flag can't be set "later", so we don't have to worry about that.
+ *
+ * Finally, to maximize compatibility with overlayfs (which isn't compatible
+ * with certain dentry operations) and to avoid taking an unnecessary
+ * performance hit, we use custom dentry_operations for each possible
+ * combination rather than always installing all operations.
+ */
+void generic_set_encrypted_ci_d_ops(struct dentry *dentry)
+{
+#ifdef CONFIG_FS_ENCRYPTION
+	bool needs_encrypt_ops = dentry->d_flags & DCACHE_NOKEY_NAME;
+#endif
+#ifdef CONFIG_UNICODE
+	bool needs_ci_ops = dentry->d_sb->s_encoding;
+#endif
+#if defined(CONFIG_FS_ENCRYPTION) && defined(CONFIG_UNICODE)
+	if (needs_encrypt_ops && needs_ci_ops) {
+		d_set_d_op(dentry, &generic_encrypted_ci_dentry_ops);
+		return;
+	}
 #endif
+#ifdef CONFIG_FS_ENCRYPTION
+	if (needs_encrypt_ops) {
+		d_set_d_op(dentry, &generic_encrypted_dentry_ops);
+		return;
+	}
+#endif
+#ifdef CONFIG_UNICODE
+	if (needs_ci_ops) {
+		d_set_d_op(dentry, &generic_ci_dentry_ops);
+		return;
+	}
+#endif
+}
+EXPORT_SYMBOL(generic_set_encrypted_ci_d_ops);
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 21cc971fd960..4a25ab4dbd3e 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -3186,6 +3186,7 @@ extern int generic_ci_d_hash(const struct dentry *dentry, struct qstr *str);
 extern int generic_ci_d_compare(const struct dentry *dentry, unsigned int len,
 				const char *str, const struct qstr *name);
 #endif
+extern void generic_set_encrypted_ci_d_ops(struct dentry *dentry);
 
 #ifdef CONFIG_MIGRATION
 extern int buffer_migrate_page(struct address_space *,
-- 
cgit v1.3-14-g43fede


From 88149082bb8ef31b289673669e080ec6a00c2e59 Mon Sep 17 00:00:00 2001
From: Hao Li <lihao2018.fnst@cn.fujitsu.com>
Date: Tue, 8 Dec 2020 10:08:43 +0800
Subject: fs: Handle I_DONTCACHE in iput_final() instead of
 generic_drop_inode()

If generic_drop_inode() returns true, it means iput_final() can evict
this inode regardless of whether it is dirty or not. If we check
I_DONTCACHE in generic_drop_inode(), any inode with this bit set will be
evicted unconditionally. This is not the desired behavior because
I_DONTCACHE only means the inode shouldn't be cached on the LRU list.
As for whether we need to evict this inode, this is what
generic_drop_inode() should do. This patch corrects the usage of
I_DONTCACHE.

This patch was proposed in [1].

[1]: https://lore.kernel.org/linux-fsdevel/20200831003407.GE12096@dread.disaster.area/

Fixes: dae2f8ed7992 ("fs: Lift XFS_IDONTCACHE to the VFS layer")
Signed-off-by: Hao Li <lihao2018.fnst@cn.fujitsu.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
---
 fs/inode.c         | 4 +++-
 include/linux/fs.h | 3 +--
 2 files changed, 4 insertions(+), 3 deletions(-)

(limited to 'include/linux/fs.h')

diff --git a/fs/inode.c b/fs/inode.c
index 9d78c37b00b8..5eea9912a0b9 100644
--- a/fs/inode.c
+++ b/fs/inode.c
@@ -1627,7 +1627,9 @@ static void iput_final(struct inode *inode)
 	else
 		drop = generic_drop_inode(inode);
 
-	if (!drop && (sb->s_flags & SB_ACTIVE)) {
+	if (!drop &&
+	    !(inode->i_state & I_DONTCACHE) &&
+	    (sb->s_flags & SB_ACTIVE)) {
 		inode_add_lru(inode);
 		spin_unlock(&inode->i_lock);
 		return;
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 8667d0cdc71e..8bde32cf9711 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -2878,8 +2878,7 @@ extern int inode_needs_sync(struct inode *inode);
 extern int generic_delete_inode(struct inode *inode);
 static inline int generic_drop_inode(struct inode *inode)
 {
-	return !inode->i_nlink || inode_unhashed(inode) ||
-		(inode->i_state & I_DONTCACHE);
+	return !inode->i_nlink || inode_unhashed(inode);
 }
 extern void d_mark_dontcache(struct inode *inode);
 
-- 
cgit v1.3-14-g43fede


From 52650c8b466bac399aec213c61d74bfe6f7af1a4 Mon Sep 17 00:00:00 2001
From: Jason Gunthorpe <jgg@nvidia.com>
Date: Mon, 14 Dec 2020 19:05:48 -0800
Subject: mm/gup: remove the vma allocation from gup_longterm_locked()

Long ago there wasn't a FOLL_LONGTERM flag so this DAX check was done by
post-processing the VMA list.

These days it is trivial to just check each VMA to see if it is DAX before
processing it inside __get_user_pages() and return failure if a DAX VMA is
encountered with FOLL_LONGTERM.

Removing the allocation of the VMA list is a significant speed up for many
call sites.

Add an IS_ENABLED to vma_is_fsdax so that code generation is unchanged
when DAX is compiled out.

Remove the dummy version of __gup_longterm_locked() as !CONFIG_CMA already
makes memalloc_nocma_save(), check_and_migrate_cma_pages(), and
memalloc_nocma_restore() into a NOP.

Link: https://lkml.kernel.org/r/0-v1-5551df3ed12e+b8-gup_dax_speedup_jgg@nvidia.com
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Pavel Tatashin <pasha.tatashin@soleen.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
---
 include/linux/fs.h |  2 +-
 mm/gup.c           | 83 ++++++++++--------------------------------------------
 2 files changed, 16 insertions(+), 69 deletions(-)

(limited to 'include/linux/fs.h')

diff --git a/include/linux/fs.h b/include/linux/fs.h
index 8667d0cdc71e..1fcc2b00582b 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -3230,7 +3230,7 @@ static inline bool vma_is_fsdax(struct vm_area_struct *vma)
 {
 	struct inode *inode;
 
-	if (!vma->vm_file)
+	if (!IS_ENABLED(CONFIG_FS_DAX) || !vma->vm_file)
 		return false;
 	if (!vma_is_dax(vma))
 		return false;
diff --git a/mm/gup.c b/mm/gup.c
index 9c6a2f5001c5..311a44ff41ff 100644
--- a/mm/gup.c
+++ b/mm/gup.c
@@ -923,6 +923,9 @@ static int check_vma_flags(struct vm_area_struct *vma, unsigned long gup_flags)
 	if (gup_flags & FOLL_ANON && !vma_is_anonymous(vma))
 		return -EFAULT;
 
+	if ((gup_flags & FOLL_LONGTERM) && vma_is_fsdax(vma))
+		return -EOPNOTSUPP;
+
 	if (write) {
 		if (!(vm_flags & VM_WRITE)) {
 			if (!(gup_flags & FOLL_FORCE))
@@ -1060,10 +1063,14 @@ static long __get_user_pages(struct mm_struct *mm,
 				goto next_page;
 			}
 
-			if (!vma || check_vma_flags(vma, gup_flags)) {
+			if (!vma) {
 				ret = -EFAULT;
 				goto out;
 			}
+			ret = check_vma_flags(vma, gup_flags);
+			if (ret)
+				goto out;
+
 			if (is_vm_hugetlb_page(vma)) {
 				i = follow_hugetlb_page(mm, vma, pages, vmas,
 						&start, &nr_pages, i,
@@ -1567,26 +1574,6 @@ struct page *get_dump_page(unsigned long addr)
 }
 #endif /* CONFIG_ELF_CORE */
 
-#if defined(CONFIG_FS_DAX) || defined (CONFIG_CMA)
-static bool check_dax_vmas(struct vm_area_struct **vmas, long nr_pages)
-{
-	long i;
-	struct vm_area_struct *vma_prev = NULL;
-
-	for (i = 0; i < nr_pages; i++) {
-		struct vm_area_struct *vma = vmas[i];
-
-		if (vma == vma_prev)
-			continue;
-
-		vma_prev = vma;
-
-		if (vma_is_fsdax(vma))
-			return true;
-	}
-	return false;
-}
-
 #ifdef CONFIG_CMA
 static long check_and_migrate_cma_pages(struct mm_struct *mm,
 					unsigned long start,
@@ -1705,63 +1692,23 @@ static long __gup_longterm_locked(struct mm_struct *mm,
 				  struct vm_area_struct **vmas,
 				  unsigned int gup_flags)
 {
-	struct vm_area_struct **vmas_tmp = vmas;
 	unsigned long flags = 0;
-	long rc, i;
+	long rc;
 
-	if (gup_flags & FOLL_LONGTERM) {
-		if (!pages)
-			return -EINVAL;
-
-		if (!vmas_tmp) {
-			vmas_tmp = kcalloc(nr_pages,
-					   sizeof(struct vm_area_struct *),
-					   GFP_KERNEL);
-			if (!vmas_tmp)
-				return -ENOMEM;
-		}
+	if (gup_flags & FOLL_LONGTERM)
 		flags = memalloc_nocma_save();
-	}
 
-	rc = __get_user_pages_locked(mm, start, nr_pages, pages,
-				     vmas_tmp, NULL, gup_flags);
+	rc = __get_user_pages_locked(mm, start, nr_pages, pages, vmas, NULL,
+				     gup_flags);
 
 	if (gup_flags & FOLL_LONGTERM) {
-		if (rc < 0)
-			goto out;
-
-		if (check_dax_vmas(vmas_tmp, rc)) {
-			if (gup_flags & FOLL_PIN)
-				unpin_user_pages(pages, rc);
-			else
-				for (i = 0; i < rc; i++)
-					put_page(pages[i]);
-			rc = -EOPNOTSUPP;
-			goto out;
-		}
-
-		rc = check_and_migrate_cma_pages(mm, start, rc, pages,
-						 vmas_tmp, gup_flags);
-out:
+		if (rc > 0)
+			rc = check_and_migrate_cma_pages(mm, start, rc, pages,
+							 vmas, gup_flags);
 		memalloc_nocma_restore(flags);
 	}
-
-	if (vmas_tmp != vmas)
-		kfree(vmas_tmp);
 	return rc;
 }
-#else /* !CONFIG_FS_DAX && !CONFIG_CMA */
-static __always_inline long __gup_longterm_locked(struct mm_struct *mm,
-						  unsigned long start,
-						  unsigned long nr_pages,
-						  struct page **pages,
-						  struct vm_area_struct **vmas,
-						  unsigned int flags)
-{
-	return __get_user_pages_locked(mm, start, nr_pages, pages, vmas,
-				       NULL, flags);
-}
-#endif /* CONFIG_FS_DAX || CONFIG_CMA */
 
 static bool is_valid_gup_flags(unsigned int gup_flags)
 {
-- 
cgit v1.3-14-g43fede