Merge tag 'fuse-update-5.8' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse

Pull fuse updates from Miklos Szeredi:

 - Fix a rare deadlock in virtiofs

 - Fix st_blocks in writeback cache mode

 - Fix wrong checks in splice move causing spurious warnings

 - Fix a race between a GETATTR request and a FUSE_NOTIFY_INVAL_INODE
   notification

 - Use rb-tree instead of linear search for pages currently under
   writeout by userspace

 - Fix copy_file_range() inconsistencies

* tag 'fuse-update-5.8' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse:
  fuse: copy_file_range should truncate cache
  fuse: fix copy_file_range cache issues
  fuse: optimize writepages search
  fuse: update attr_version counter on fuse_notify_inval_inode()
  fuse: don't check refcount after stealing page
  fuse: fix weird page warning
  fuse: use dump_page
  virtiofs: do not use fuse_fill_super_common() for device installation
  fuse: always allow query of st_dev
  fuse: always flush dirty data on close(2)
  fuse: invalidate inode attr in writeback cache mode
  fuse: Update stale comment in queue_interrupt()
  fuse: BUG_ON correction in fuse_dev_splice_write()
  virtiofs: Add mount option and atime behavior to the doc
  virtiofs: schedule blocking async replies in separate worker

Documentation/filesystems/virtiofs.rst

@@ -39,6 +39,20 @@ Mount file system with tag ``myfs`` on ``/mnt``:
 Please see https://virtio-fs.gitlab.io/ for details on how to configure QEMU
 and the virtiofsd daemon.
 
+Mount options
+-------------
+
+virtiofs supports general VFS mount options, for example, remount,
+ro, rw, context, etc. It also supports FUSE mount options.
+
+atime behavior
+^^^^^^^^^^^^^^
+
+The atime-related mount options, for example, noatime, strictatime,
+are ignored. The atime behavior for virtiofs is the same as the
+underlying filesystem of the directory that has been exported
+on the host.
+
 Internals
 =========
 Since the virtio-fs device uses the FUSE protocol for file system requests, the

fs/fuse/dev.c

@@ -342,7 +342,7 @@ static int queue_interrupt(struct fuse_iqueue *fiq, struct fuse_req *req)
 		list_add_tail(&req->intr_entry, &fiq->interrupts);
 		/*
 		 * Pairs with smp_mb() implied by test_and_set_bit()
-		 * from request_end().
+		 * from fuse_request_end().
 		 */
 		smp_mb();
 		if (test_bit(FR_FINISHED, &req->flags)) {
@@ -764,16 +764,15 @@ static int fuse_check_page(struct page *page)
 {
 	if (page_mapcount(page) ||
 	    page->mapping != NULL ||
-	    page_count(page) != 1 ||
 	    (page->flags & PAGE_FLAGS_CHECK_AT_PREP &
 	     ~(1 << PG_locked |
 	       1 << PG_referenced |
 	       1 << PG_uptodate |
 	       1 << PG_lru |
 	       1 << PG_active |
-	       1 << PG_reclaim))) {
-		pr_warn("trying to steal weird page\n");
-		pr_warn("  page=%p index=%li flags=%08lx, count=%i, mapcount=%i, mapping=%p\n", page, page->index, page->flags, page_count(page), page_mapcount(page), page->mapping);
+	       1 << PG_reclaim |
+	       1 << PG_waiters))) {
+		dump_page(page, "fuse: trying to steal weird page");
 		return 1;
 	}
 	return 0;
@@ -1977,8 +1976,9 @@ static ssize_t fuse_dev_splice_write(struct pipe_inode_info *pipe,
 		struct pipe_buffer *ibuf;
 		struct pipe_buffer *obuf;
 
-		BUG_ON(nbuf >= pipe->ring_size);
-		BUG_ON(tail == head);
+		if (WARN_ON(nbuf >= count || tail == head))
+			goto out_free;
+
 		ibuf = &pipe->bufs[tail & mask];
 		obuf = &bufs[nbuf];
 

fs/fuse/dir.c

@@ -1689,8 +1689,18 @@ static int fuse_getattr(const struct path *path, struct kstat *stat,
 	struct inode *inode = d_inode(path->dentry);
 	struct fuse_conn *fc = get_fuse_conn(inode);
 
-	if (!fuse_allow_current_process(fc))
+	if (!fuse_allow_current_process(fc)) {
+		if (!request_mask) {
+			/*
+			 * If user explicitly requested *nothing* then don't
+			 * error out, but return st_dev only.
+			 */
+			stat->result_mask = 0;
+			stat->dev = inode->i_sb->s_dev;
+			return 0;
+		}
 		return -EACCES;
+	}
 
 	return fuse_update_get_attr(inode, NULL, stat, request_mask, flags);
 }

fs/fuse/file.c

@@ -357,7 +357,7 @@ u64 fuse_lock_owner_id(struct fuse_conn *fc, fl_owner_t id)
 
 struct fuse_writepage_args {
 	struct fuse_io_args ia;
-	struct list_head writepages_entry;
+	struct rb_node writepages_entry;
 	struct list_head queue_entry;
 	struct fuse_writepage_args *next;
 	struct inode *inode;
@@ -366,18 +366,24 @@ struct fuse_writepage_args {
 static struct fuse_writepage_args *fuse_find_writeback(struct fuse_inode *fi,
 					    pgoff_t idx_from, pgoff_t idx_to)
 {
-	struct fuse_writepage_args *wpa;
+	struct rb_node *n;
+
+	n = fi->writepages.rb_node;
 
-	list_for_each_entry(wpa, &fi->writepages, writepages_entry) {
+	while (n) {
+		struct fuse_writepage_args *wpa;
 		pgoff_t curr_index;
 
+		wpa = rb_entry(n, struct fuse_writepage_args, writepages_entry);
 		WARN_ON(get_fuse_inode(wpa->inode) != fi);
 		curr_index = wpa->ia.write.in.offset >> PAGE_SHIFT;
-		if (idx_from < curr_index + wpa->ia.ap.num_pages &&
-		    curr_index <= idx_to) {
+		if (idx_from >= curr_index + wpa->ia.ap.num_pages)
+			n = n->rb_right;
+		else if (idx_to < curr_index)
+			n = n->rb_left;
+		else
 			return wpa;
 	}
-	}
 	return NULL;
 }
 
@@ -445,9 +451,6 @@ static int fuse_flush(struct file *file, fl_owner_t id)
 	if (is_bad_inode(inode))
 		return -EIO;
 
-	if (fc->no_flush)
-		return 0;
-
 	err = write_inode_now(inode, 1);
 	if (err)
 		return err;
@@ -460,6 +463,10 @@ static int fuse_flush(struct file *file, fl_owner_t id)
 	if (err)
 		return err;
 
+	err = 0;
+	if (fc->no_flush)
+		goto inval_attr_out;
+
 	memset(&inarg, 0, sizeof(inarg));
 	inarg.fh = ff->fh;
 	inarg.lock_owner = fuse_lock_owner_id(fc, id);
@@ -475,6 +482,14 @@ static int fuse_flush(struct file *file, fl_owner_t id)
 		fc->no_flush = 1;
 		err = 0;
 	}
+
+inval_attr_out:
+	/*
+	 * In memory i_blocks is not maintained by fuse, if writeback cache is
+	 * enabled, i_blocks from cached attr may not be accurate.
+	 */
+	if (!err && fc->writeback_cache)
+		fuse_invalidate_attr(inode);
 	return err;
 }
 
@@ -712,6 +727,7 @@ static ssize_t fuse_async_req_send(struct fuse_conn *fc,
 	spin_unlock(&io->lock);
 
 	ia->ap.args.end = fuse_aio_complete_req;
+	ia->ap.args.may_block = io->should_dirty;
 	err = fuse_simple_background(fc, &ia->ap.args, GFP_KERNEL);
 	if (err)
 		fuse_aio_complete_req(fc, &ia->ap.args, err);
@@ -1570,7 +1586,7 @@ static void fuse_writepage_finish(struct fuse_conn *fc,
 	struct backing_dev_info *bdi = inode_to_bdi(inode);
 	int i;
 
-	list_del(&wpa->writepages_entry);
+	rb_erase(&wpa->writepages_entry, &fi->writepages);
 	for (i = 0; i < ap->num_pages; i++) {
 		dec_wb_stat(&bdi->wb, WB_WRITEBACK);
 		dec_node_page_state(ap->pages[i], NR_WRITEBACK_TEMP);
@@ -1658,6 +1674,36 @@ __acquires(fi->lock)
 	}
 }
 
+static void tree_insert(struct rb_root *root, struct fuse_writepage_args *wpa)
+{
+	pgoff_t idx_from = wpa->ia.write.in.offset >> PAGE_SHIFT;
+	pgoff_t idx_to = idx_from + wpa->ia.ap.num_pages - 1;
+	struct rb_node **p = &root->rb_node;
+	struct rb_node  *parent = NULL;
+
+	WARN_ON(!wpa->ia.ap.num_pages);
+	while (*p) {
+		struct fuse_writepage_args *curr;
+		pgoff_t curr_index;
+
+		parent = *p;
+		curr = rb_entry(parent, struct fuse_writepage_args,
+				writepages_entry);
+		WARN_ON(curr->inode != wpa->inode);
+		curr_index = curr->ia.write.in.offset >> PAGE_SHIFT;
+
+		if (idx_from >= curr_index + curr->ia.ap.num_pages)
+			p = &(*p)->rb_right;
+		else if (idx_to < curr_index)
+			p = &(*p)->rb_left;
+		else
+			return (void) WARN_ON(true);
+	}
+
+	rb_link_node(&wpa->writepages_entry, parent, p);
+	rb_insert_color(&wpa->writepages_entry, root);
+}
+
 static void fuse_writepage_end(struct fuse_conn *fc, struct fuse_args *args,
 			       int error)
 {
@@ -1676,7 +1722,7 @@ static void fuse_writepage_end(struct fuse_conn *fc, struct fuse_args *args,
 		wpa->next = next->next;
 		next->next = NULL;
 		next->ia.ff = fuse_file_get(wpa->ia.ff);
-		list_add(&next->writepages_entry, &fi->writepages);
+		tree_insert(&fi->writepages, next);
 
 		/*
 		 * Skip fuse_flush_writepages() to make it easy to crop requests
@@ -1811,7 +1857,7 @@ static int fuse_writepage_locked(struct page *page)
 	inc_node_page_state(tmp_page, NR_WRITEBACK_TEMP);
 
 	spin_lock(&fi->lock);
-	list_add(&wpa->writepages_entry, &fi->writepages);
+	tree_insert(&fi->writepages, wpa);
 	list_add_tail(&wpa->queue_entry, &fi->queued_writes);
 	fuse_flush_writepages(inode);
 	spin_unlock(&fi->lock);
@@ -1923,10 +1969,10 @@ static bool fuse_writepage_in_flight(struct fuse_writepage_args *new_wpa,
 	WARN_ON(new_ap->num_pages != 0);
 
 	spin_lock(&fi->lock);
-	list_del(&new_wpa->writepages_entry);
+	rb_erase(&new_wpa->writepages_entry, &fi->writepages);
 	old_wpa = fuse_find_writeback(fi, page->index, page->index);
 	if (!old_wpa) {
-		list_add(&new_wpa->writepages_entry, &fi->writepages);
+		tree_insert(&fi->writepages, new_wpa);
 		spin_unlock(&fi->lock);
 		return false;
 	}
@@ -2041,7 +2087,7 @@ static int fuse_writepages_fill(struct page *page,
 		wpa->inode = inode;
 
 		spin_lock(&fi->lock);
-		list_add(&wpa->writepages_entry, &fi->writepages);
+		tree_insert(&fi->writepages, wpa);
 		spin_unlock(&fi->lock);
 
 		data->wpa = wpa;
@@ -3235,13 +3281,11 @@ static ssize_t __fuse_copy_file_range(struct file *file_in, loff_t pos_in,
 	if (file_inode(file_in)->i_sb != file_inode(file_out)->i_sb)
 		return -EXDEV;
 
-	if (fc->writeback_cache) {
 	inode_lock(inode_in);
-		err = fuse_writeback_range(inode_in, pos_in, pos_in + len);
+	err = fuse_writeback_range(inode_in, pos_in, pos_in + len - 1);
 	inode_unlock(inode_in);
 	if (err)
 		return err;
-	}
 
 	inode_lock(inode_out);
 
@@ -3249,11 +3293,27 @@ static ssize_t __fuse_copy_file_range(struct file *file_in, loff_t pos_in,
 	if (err)
 		goto out;
 
-	if (fc->writeback_cache) {
-		err = fuse_writeback_range(inode_out, pos_out, pos_out + len);
+	/*
+	 * Write out dirty pages in the destination file before sending the COPY
+	 * request to userspace.  After the request is completed, truncate off
+	 * pages (including partial ones) from the cache that have been copied,
+	 * since these contain stale data at that point.
+	 *
+	 * This should be mostly correct, but if the COPY writes to partial
+	 * pages (at the start or end) and the parts not covered by the COPY are
+	 * written through a memory map after calling fuse_writeback_range(),
+	 * then these partial page modifications will be lost on truncation.
+	 *
+	 * It is unlikely that someone would rely on such mixed style
+	 * modifications.  Yet this does give less guarantees than if the
+	 * copying was performed with write(2).
+	 *
+	 * To fix this a i_mmap_sem style lock could be used to prevent new
+	 * faults while the copy is ongoing.
+	 */
+	err = fuse_writeback_range(inode_out, pos_out, pos_out + len - 1);
 	if (err)
 		goto out;
-	}
 
 	if (is_unstable)
 		set_bit(FUSE_I_SIZE_UNSTABLE, &fi_out->state);
@@ -3274,6 +3334,10 @@ static ssize_t __fuse_copy_file_range(struct file *file_in, loff_t pos_in,
 	if (err)
 		goto out;
 
+	truncate_inode_pages_range(inode_out->i_mapping,
+				   ALIGN_DOWN(pos_out, PAGE_SIZE),
+				   ALIGN(pos_out + outarg.size, PAGE_SIZE) - 1);
+
 	if (fc->writeback_cache) {
 		fuse_write_update_size(inode_out, pos_out + outarg.size);
 		file_update_time(file_out);
@@ -3351,5 +3415,5 @@ void fuse_init_file_inode(struct inode *inode)
 	INIT_LIST_HEAD(&fi->queued_writes);
 	fi->writectr = 0;
 	init_waitqueue_head(&fi->page_waitq);
-	INIT_LIST_HEAD(&fi->writepages);
+	fi->writepages = RB_ROOT;
 }

fs/fuse/fuse_i.h

@@ -111,7 +111,7 @@ struct fuse_inode {
 			wait_queue_head_t page_waitq;
 
 			/* List of writepage requestst (pending or sent) */
-			struct list_head writepages;
+			struct rb_root writepages;
 		};
 
 		/* readdir cache (directory only) */
@@ -249,6 +249,7 @@ struct fuse_args {
 	bool out_argvar:1;
 	bool page_zeroing:1;
 	bool page_replace:1;
+	bool may_block:1;
 	struct fuse_in_arg in_args[3];
 	struct fuse_arg out_args[2];
 	void (*end)(struct fuse_conn *fc, struct fuse_args *args, int error);

fs/fuse/inode.c

@@ -321,6 +321,8 @@ struct inode *fuse_iget(struct super_block *sb, u64 nodeid,
 int fuse_reverse_inval_inode(struct super_block *sb, u64 nodeid,
 			     loff_t offset, loff_t len)
 {
+	struct fuse_conn *fc = get_fuse_conn_super(sb);
+	struct fuse_inode *fi;
 	struct inode *inode;
 	pgoff_t pg_start;
 	pgoff_t pg_end;
@@ -329,6 +331,11 @@ int fuse_reverse_inval_inode(struct super_block *sb, u64 nodeid,
 	if (!inode)
 		return -ENOENT;
 
+	fi = get_fuse_inode(inode);
+	spin_lock(&fi->lock);
+	fi->attr_version = atomic64_inc_return(&fc->attr_version);
+	spin_unlock(&fi->lock);
+
 	fuse_invalidate_attr(inode);
 	forget_all_cached_acls(inode);
 	if (offset >= 0) {
@@ -1113,7 +1120,7 @@ EXPORT_SYMBOL_GPL(fuse_dev_free);
 
 int fuse_fill_super_common(struct super_block *sb, struct fuse_fs_context *ctx)
 {
-	struct fuse_dev *fud;
+	struct fuse_dev *fud = NULL;
 	struct fuse_conn *fc = get_fuse_conn_super(sb);
 	struct inode *root;
 	struct dentry *root_dentry;
@@ -1155,9 +1162,12 @@ int fuse_fill_super_common(struct super_block *sb, struct fuse_fs_context *ctx)
 	if (sb->s_user_ns != &init_user_ns)
 		sb->s_xattr = fuse_no_acl_xattr_handlers;
 
+	if (ctx->fudptr) {
+		err = -ENOMEM;
 		fud = fuse_dev_alloc_install(fc);
 		if (!fud)
 			goto err;
+	}
 
 	fc->dev = sb->s_dev;
 	fc->sb = sb;
@@ -1191,7 +1201,7 @@ int fuse_fill_super_common(struct super_block *sb, struct fuse_fs_context *ctx)
 
 	mutex_lock(&fuse_mutex);
 	err = -EINVAL;
-	if (*ctx->fudptr)
+	if (ctx->fudptr && *ctx->fudptr)
 		goto err_unlock;
 
 	err = fuse_ctl_add_conn(fc);
@@ -1200,6 +1210,7 @@ int fuse_fill_super_common(struct super_block *sb, struct fuse_fs_context *ctx)
 
 	list_add_tail(&fc->entry, &fuse_conn_list);
 	sb->s_root = root_dentry;
+	if (ctx->fudptr)
 		*ctx->fudptr = fud;
 	mutex_unlock(&fuse_mutex);
 	return 0;
@@ -1208,6 +1219,7 @@ int fuse_fill_super_common(struct super_block *sb, struct fuse_fs_context *ctx)
 	mutex_unlock(&fuse_mutex);
 	dput(root_dentry);
  err_dev_free:
+	if (fud)
 		fuse_dev_free(fud);
  err:
 	return err;

fs/fuse/virtio_fs.c

@@ -60,6 +60,12 @@ struct virtio_fs_forget {
 	struct virtio_fs_forget_req req;
 };
 
+struct virtio_fs_req_work {
+	struct fuse_req *req;
+	struct virtio_fs_vq *fsvq;
+	struct work_struct done_work;
+};
+
 static int virtio_fs_enqueue_req(struct virtio_fs_vq *fsvq,
 				 struct fuse_req *req, bool in_flight);
 
@@ -485,36 +491,16 @@ static void copy_args_from_argbuf(struct fuse_args *args, struct fuse_req *req)
 }
 
 /* Work function for request completion */
-static void virtio_fs_requests_done_work(struct work_struct *work)
+static void virtio_fs_request_complete(struct fuse_req *req,
+				       struct virtio_fs_vq *fsvq)
 {
-	struct virtio_fs_vq *fsvq = container_of(work, struct virtio_fs_vq,
-						 done_work);
 	struct fuse_pqueue *fpq = &fsvq->fud->pq;
 	struct fuse_conn *fc = fsvq->fud->fc;
-	struct virtqueue *vq = fsvq->vq;
-	struct fuse_req *req;
-	struct fuse_args_pages *ap;
-	struct fuse_req *next;
 	struct fuse_args *args;
+	struct fuse_args_pages *ap;
 	unsigned int len, i, thislen;
 	struct page *page;
-	LIST_HEAD(reqs);
 
-	/* Collect completed requests off the virtqueue */
-	spin_lock(&fsvq->lock);
-	do {
-		virtqueue_disable_cb(vq);
-
-		while ((req = virtqueue_get_buf(vq, &len)) != NULL) {
-			spin_lock(&fpq->lock);
-			list_move_tail(&req->list, &reqs);
-			spin_unlock(&fpq->lock);
-		}
-	} while (!virtqueue_enable_cb(vq) && likely(!virtqueue_is_broken(vq)));
-	spin_unlock(&fsvq->lock);
-
-	/* End requests */
-	list_for_each_entry_safe(req, next, &reqs, list) {
 	/*
 	 * TODO verify that server properly follows FUSE protocol
 	 * (oh.uniq, oh.len)
@@ -540,13 +526,63 @@ static void virtio_fs_requests_done_work(struct work_struct *work)
 
 	spin_lock(&fpq->lock);
 	clear_bit(FR_SENT, &req->flags);
-		list_del_init(&req->list);
 	spin_unlock(&fpq->lock);
 
 	fuse_request_end(fc, req);
 	spin_lock(&fsvq->lock);
 	dec_in_flight_req(fsvq);
 	spin_unlock(&fsvq->lock);
+}
+
+static void virtio_fs_complete_req_work(struct work_struct *work)
+{
+	struct virtio_fs_req_work *w =
+		container_of(work, typeof(*w), done_work);
+
+	virtio_fs_request_complete(w->req, w->fsvq);
+	kfree(w);
+}
+
+static void virtio_fs_requests_done_work(struct work_struct *work)
+{
+	struct virtio_fs_vq *fsvq = container_of(work, struct virtio_fs_vq,
+						 done_work);
+	struct fuse_pqueue *fpq = &fsvq->fud->pq;
+	struct virtqueue *vq = fsvq->vq;
+	struct fuse_req *req;
+	struct fuse_req *next;
+	unsigned int len;
+	LIST_HEAD(reqs);
+
+	/* Collect completed requests off the virtqueue */
+	spin_lock(&fsvq->lock);
+	do {
+		virtqueue_disable_cb(vq);
+
+		while ((req = virtqueue_get_buf(vq, &len)) != NULL) {
+			spin_lock(&fpq->lock);
+			list_move_tail(&req->list, &reqs);
+			spin_unlock(&fpq->lock);
+		}
+	} while (!virtqueue_enable_cb(vq) && likely(!virtqueue_is_broken(vq)));
+	spin_unlock(&fsvq->lock);
+
+	/* End requests */
+	list_for_each_entry_safe(req, next, &reqs, list) {
+		list_del_init(&req->list);
+
+		/* blocking async request completes in a worker context */
+		if (req->args->may_block) {
+			struct virtio_fs_req_work *w;
+
+			w = kzalloc(sizeof(*w), GFP_NOFS | __GFP_NOFAIL);
+			INIT_WORK(&w->done_work, virtio_fs_complete_req_work);
+			w->fsvq = fsvq;
+			w->req = req;
+			schedule_work(&w->done_work);
+		} else {
+			virtio_fs_request_complete(req, fsvq);
+		}
 	}
 }
 
@@ -1067,7 +1103,7 @@ static int virtio_fs_fill_super(struct super_block *sb)
 
 	err = -ENOMEM;
 	/* Allocate fuse_dev for hiprio and notification queues */
-	for (i = 0; i < VQ_REQUEST; i++) {
+	for (i = 0; i < fs->nvqs; i++) {
 		struct virtio_fs_vq *fsvq = &fs->vqs[i];
 
 		fsvq->fud = fuse_dev_alloc();
@@ -1075,18 +1111,15 @@ static int virtio_fs_fill_super(struct super_block *sb)
 			goto err_free_fuse_devs;
 	}
 
-	ctx.fudptr = (void **)&fs->vqs[VQ_REQUEST].fud;
+	/* virtiofs allocates and installs its own fuse devices */
+	ctx.fudptr = NULL;
 	err = fuse_fill_super_common(sb, &ctx);
 	if (err < 0)
 		goto err_free_fuse_devs;
 
-	fc = fs->vqs[VQ_REQUEST].fud->fc;
-
 	for (i = 0; i < fs->nvqs; i++) {
 		struct virtio_fs_vq *fsvq = &fs->vqs[i];
 
-		if (i == VQ_REQUEST)
-			continue; /* already initialized */
 		fuse_dev_install(fsvq->fud, fc);
 	}