From 4000339dc01d99d959b5fcbeaff9715c9c53100d Mon Sep 17 00:00:00 2001 From: Miklos Szeredi Date: Fri, 22 Apr 2022 15:48:53 +0200 Subject: [PATCH] UPSTREAM: fuse: fix deadlock between atomic O_TRUNC and page invalidation commit 2fdbb8dd01556e1501132b5ad3826e8f71e24a8b upstream. fuse_finish_open() will be called with FUSE_NOWRITE set in case of atomic O_TRUNC open(), so commit 76224355db75 ("fuse: truncate pagecache on atomic_o_trunc") replaced invalidate_inode_pages2() by truncate_pagecache() in such a case to avoid the A-A deadlock. However, we found another A-B-B-A deadlock related to the case above, which will cause the xfstests generic/464 testcase hung in our virtio-fs test environment. For example, consider two processes concurrently open one same file, one with O_TRUNC and another without O_TRUNC. The deadlock case is described below, if open(O_TRUNC) is already set_nowrite(acquired A), and is trying to lock a page (acquiring B), open() could have held the page lock (acquired B), and waiting on the page writeback (acquiring A). This would lead to deadlocks. open(O_TRUNC) ---------------------------------------------------------------- fuse_open_common inode_lock [C acquire] fuse_set_nowrite [A acquire] fuse_finish_open truncate_pagecache lock_page [B acquire] truncate_inode_page unlock_page [B release] fuse_release_nowrite [A release] inode_unlock [C release] ---------------------------------------------------------------- open() ---------------------------------------------------------------- fuse_open_common fuse_finish_open invalidate_inode_pages2 lock_page [B acquire] fuse_launder_page fuse_wait_on_page_writeback [A acquire & release] unlock_page [B release] ---------------------------------------------------------------- Besides this case, all calls of invalidate_inode_pages2() and invalidate_inode_pages2_range() in fuse code also can deadlock with open(O_TRUNC). Fix by moving the truncate_pagecache() call outside the nowrite protected region. The nowrite protection is only for delayed writeback (writeback_cache) case, where inode lock does not protect against truncation racing with writes on the server. Write syscalls racing with page cache truncation still get the inode lock protection. This patch also changes the order of filemap_invalidate_lock() vs. fuse_set_nowrite() in fuse_open_common(). This new order matches the order found in fuse_file_fallocate() and fuse_do_setattr(). Reported-by: Jiachen Zhang Tested-by: Jiachen Zhang Fixes: e4648309b85a ("fuse: truncate pending writes on O_TRUNC") Cc: Change-Id: I7e01c787503114b4cd671ef7283fb0a1143887a1 Signed-off-by: Miklos Szeredi Signed-off-by: Yang Bo Signed-off-by: Greg Kroah-Hartman --- fs/fuse/dir.c | 5 +++++ fs/fuse/file.c | 29 +++++++++++++++++------------ 2 files changed, 22 insertions(+), 12 deletions(-) diff --git a/fs/fuse/dir.c b/fs/fuse/dir.c index 4a336096b153..b1ee3a6b5b83 100644 --- a/fs/fuse/dir.c +++ b/fs/fuse/dir.c @@ -713,6 +713,7 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, struct fuse_entry_out outentry; struct fuse_inode *fi; struct fuse_file *ff; + bool trunc = flags & O_TRUNC; /* Userspace expects S_IFREG in create mode */ BUG_ON((mode & S_IFMT) != S_IFREG); @@ -795,6 +796,10 @@ static int fuse_create_open(struct inode *dir, struct dentry *entry, } else { file->private_data = ff; fuse_finish_open(inode, file); + if (fm->fc->atomic_o_trunc && trunc) + truncate_pagecache(inode, 0); + else if (!(ff->open_flags & FOPEN_KEEP_CACHE)) + invalidate_inode_pages2(inode->i_mapping); } return err; diff --git a/fs/fuse/file.c b/fs/fuse/file.c index fea39e0b037c..bb76ba48ef32 100644 --- a/fs/fuse/file.c +++ b/fs/fuse/file.c @@ -211,14 +211,10 @@ void fuse_finish_open(struct inode *inode, struct file *file) fi->attr_version = atomic64_inc_return(&fc->attr_version); i_size_write(inode, 0); spin_unlock(&fi->lock); - truncate_pagecache(inode, 0); fuse_invalidate_attr(inode); if (fc->writeback_cache) file_update_time(file); - } else if (!(ff->open_flags & FOPEN_KEEP_CACHE)) { - invalidate_inode_pages2(inode->i_mapping); } - if ((file->f_mode & FMODE_WRITE) && fc->writeback_cache) fuse_link_write_file(file); } @@ -255,30 +251,39 @@ int fuse_open_common(struct inode *inode, struct file *file, bool isdir) } #endif - if (is_wb_truncate || dax_truncate) { + if (is_wb_truncate || dax_truncate) inode_lock(inode); - fuse_set_nowrite(inode); - } if (dax_truncate) { down_write(&get_fuse_inode(inode)->i_mmap_sem); err = fuse_dax_break_layouts(inode, 0, 0); if (err) - goto out; + goto out_inode_unlock; } + if (is_wb_truncate || dax_truncate) + fuse_set_nowrite(inode); + err = fuse_do_open(fm, get_node_id(inode), file, isdir); if (!err) fuse_finish_open(inode, file); -out: + if (is_wb_truncate || dax_truncate) + fuse_release_nowrite(inode); + if (!err) { + struct fuse_file *ff = file->private_data; + + if (fc->atomic_o_trunc && (file->f_flags & O_TRUNC)) + truncate_pagecache(inode, 0); + else if (!(ff->open_flags & FOPEN_KEEP_CACHE)) + invalidate_inode_pages2(inode->i_mapping); + } if (dax_truncate) up_write(&get_fuse_inode(inode)->i_mmap_sem); - if (is_wb_truncate | dax_truncate) { - fuse_release_nowrite(inode); +out_inode_unlock: + if (is_wb_truncate || dax_truncate) inode_unlock(inode); - } return err; }