From c755e251357a0cee0679081f08c3f4ba797a8009 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Wed, 11 Jan 2017 21:50:46 -0500 Subject: ext4: fix deadlock between inline_data and ext4_expand_extra_isize_ea() The xattr_sem deadlock problems fixed in commit 2e81a4eeedca: "ext4: avoid deadlock when expanding inode size" didn't include the use of xattr_sem in fs/ext4/inline.c. With the addition of project quota which added a new extra inode field, this exposed deadlocks in the inline_data code similar to the ones fixed by 2e81a4eeedca. The deadlock can be reproduced via: dmesg -n 7 mke2fs -t ext4 -O inline_data -Fq -I 256 /dev/vdc 32768 mount -t ext4 -o debug_want_extra_isize=24 /dev/vdc /vdc mkdir /vdc/a umount /vdc mount -t ext4 /dev/vdc /vdc echo foo > /vdc/a/foo and looks like this: [ 11.158815] [ 11.160276] ============================================= [ 11.161960] [ INFO: possible recursive locking detected ] [ 11.161960] 4.10.0-rc3-00015-g011b30a8a3cf #160 Tainted: G W [ 11.161960] --------------------------------------------- [ 11.161960] bash/2519 is trying to acquire lock: [ 11.161960] (&ei->xattr_sem){++++..}, at: [] ext4_expand_extra_isize_ea+0x3d/0x4cd [ 11.161960] [ 11.161960] but task is already holding lock: [ 11.161960] (&ei->xattr_sem){++++..}, at: [] ext4_try_add_inline_entry+0x3a/0x152 [ 11.161960] [ 11.161960] other info that might help us debug this: [ 11.161960] Possible unsafe locking scenario: [ 11.161960] [ 11.161960] CPU0 [ 11.161960] ---- [ 11.161960] lock(&ei->xattr_sem); [ 11.161960] lock(&ei->xattr_sem); [ 11.161960] [ 11.161960] *** DEADLOCK *** [ 11.161960] [ 11.161960] May be due to missing lock nesting notation [ 11.161960] [ 11.161960] 4 locks held by bash/2519: [ 11.161960] #0: (sb_writers#3){.+.+.+}, at: [] mnt_want_write+0x1e/0x3e [ 11.161960] #1: (&type->i_mutex_dir_key){++++++}, at: [] path_openat+0x338/0x67a [ 11.161960] #2: (jbd2_handle){++++..}, at: [] start_this_handle+0x582/0x622 [ 11.161960] #3: (&ei->xattr_sem){++++..}, at: [] ext4_try_add_inline_entry+0x3a/0x152 [ 11.161960] [ 11.161960] stack backtrace: [ 11.161960] CPU: 0 PID: 2519 Comm: bash Tainted: G W 4.10.0-rc3-00015-g011b30a8a3cf #160 [ 11.161960] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.1-1 04/01/2014 [ 11.161960] Call Trace: [ 11.161960] dump_stack+0x72/0xa3 [ 11.161960] __lock_acquire+0xb7c/0xcb9 [ 11.161960] ? kvm_clock_read+0x1f/0x29 [ 11.161960] ? __lock_is_held+0x36/0x66 [ 11.161960] ? __lock_is_held+0x36/0x66 [ 11.161960] lock_acquire+0x106/0x18a [ 11.161960] ? ext4_expand_extra_isize_ea+0x3d/0x4cd [ 11.161960] down_write+0x39/0x72 [ 11.161960] ? ext4_expand_extra_isize_ea+0x3d/0x4cd [ 11.161960] ext4_expand_extra_isize_ea+0x3d/0x4cd [ 11.161960] ? _raw_read_unlock+0x22/0x2c [ 11.161960] ? jbd2_journal_extend+0x1e2/0x262 [ 11.161960] ? __ext4_journal_get_write_access+0x3d/0x60 [ 11.161960] ext4_mark_inode_dirty+0x17d/0x26d [ 11.161960] ? ext4_add_dirent_to_inline.isra.12+0xa5/0xb2 [ 11.161960] ext4_add_dirent_to_inline.isra.12+0xa5/0xb2 [ 11.161960] ext4_try_add_inline_entry+0x69/0x152 [ 11.161960] ext4_add_entry+0xa3/0x848 [ 11.161960] ? __brelse+0x14/0x2f [ 11.161960] ? _raw_spin_unlock_irqrestore+0x44/0x4f [ 11.161960] ext4_add_nondir+0x17/0x5b [ 11.161960] ext4_create+0xcf/0x133 [ 11.161960] ? ext4_mknod+0x12f/0x12f [ 11.161960] lookup_open+0x39e/0x3fb [ 11.161960] ? __wake_up+0x1a/0x40 [ 11.161960] ? lock_acquire+0x11e/0x18a [ 11.161960] path_openat+0x35c/0x67a [ 11.161960] ? sched_clock_cpu+0xd7/0xf2 [ 11.161960] do_filp_open+0x36/0x7c [ 11.161960] ? _raw_spin_unlock+0x22/0x2c [ 11.161960] ? __alloc_fd+0x169/0x173 [ 11.161960] do_sys_open+0x59/0xcc [ 11.161960] SyS_open+0x1d/0x1f [ 11.161960] do_int80_syscall_32+0x4f/0x61 [ 11.161960] entry_INT80_32+0x2f/0x2f [ 11.161960] EIP: 0xb76ad469 [ 11.161960] EFLAGS: 00000286 CPU: 0 [ 11.161960] EAX: ffffffda EBX: 08168ac8 ECX: 00008241 EDX: 000001b6 [ 11.161960] ESI: b75e46bc EDI: b7755000 EBP: bfbdb108 ESP: bfbdafc0 [ 11.161960] DS: 007b ES: 007b FS: 0000 GS: 0033 SS: 007b Cc: stable@vger.kernel.org # 3.10 (requires 2e81a4eeedca as a prereq) Reported-by: George Spelvin Signed-off-by: Theodore Ts'o --- fs/ext4/inline.c | 66 ++++++++++++++++++++++++++------------------------------ 1 file changed, 30 insertions(+), 36 deletions(-) (limited to 'fs/ext4/inline.c') diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c index 437df6a1a841..99a5312ced52 100644 --- a/fs/ext4/inline.c +++ b/fs/ext4/inline.c @@ -381,7 +381,7 @@ out: static int ext4_prepare_inline_data(handle_t *handle, struct inode *inode, unsigned int len) { - int ret, size; + int ret, size, no_expand; struct ext4_inode_info *ei = EXT4_I(inode); if (!ext4_test_inode_state(inode, EXT4_STATE_MAY_INLINE_DATA)) @@ -391,15 +391,14 @@ static int ext4_prepare_inline_data(handle_t *handle, struct inode *inode, if (size < len) return -ENOSPC; - down_write(&EXT4_I(inode)->xattr_sem); + ext4_write_lock_xattr(inode, &no_expand); if (ei->i_inline_off) ret = ext4_update_inline_data(handle, inode, len); else ret = ext4_create_inline_data(handle, inode, len); - up_write(&EXT4_I(inode)->xattr_sem); - + ext4_write_unlock_xattr(inode, &no_expand); return ret; } @@ -533,7 +532,7 @@ static int ext4_convert_inline_data_to_extent(struct address_space *mapping, struct inode *inode, unsigned flags) { - int ret, needed_blocks; + int ret, needed_blocks, no_expand; handle_t *handle = NULL; int retries = 0, sem_held = 0; struct page *page = NULL; @@ -573,7 +572,7 @@ retry: goto out; } - down_write(&EXT4_I(inode)->xattr_sem); + ext4_write_lock_xattr(inode, &no_expand); sem_held = 1; /* If some one has already done this for us, just exit. */ if (!ext4_has_inline_data(inode)) { @@ -610,7 +609,7 @@ retry: put_page(page); page = NULL; ext4_orphan_add(handle, inode); - up_write(&EXT4_I(inode)->xattr_sem); + ext4_write_unlock_xattr(inode, &no_expand); sem_held = 0; ext4_journal_stop(handle); handle = NULL; @@ -636,7 +635,7 @@ out: put_page(page); } if (sem_held) - up_write(&EXT4_I(inode)->xattr_sem); + ext4_write_unlock_xattr(inode, &no_expand); if (handle) ext4_journal_stop(handle); brelse(iloc.bh); @@ -729,7 +728,7 @@ convert: int ext4_write_inline_data_end(struct inode *inode, loff_t pos, unsigned len, unsigned copied, struct page *page) { - int ret; + int ret, no_expand; void *kaddr; struct ext4_iloc iloc; @@ -747,7 +746,7 @@ int ext4_write_inline_data_end(struct inode *inode, loff_t pos, unsigned len, goto out; } - down_write(&EXT4_I(inode)->xattr_sem); + ext4_write_lock_xattr(inode, &no_expand); BUG_ON(!ext4_has_inline_data(inode)); kaddr = kmap_atomic(page); @@ -757,7 +756,7 @@ int ext4_write_inline_data_end(struct inode *inode, loff_t pos, unsigned len, /* clear page dirty so that writepages wouldn't work for us. */ ClearPageDirty(page); - up_write(&EXT4_I(inode)->xattr_sem); + ext4_write_unlock_xattr(inode, &no_expand); brelse(iloc.bh); out: return copied; @@ -768,7 +767,7 @@ ext4_journalled_write_inline_data(struct inode *inode, unsigned len, struct page *page) { - int ret; + int ret, no_expand; void *kaddr; struct ext4_iloc iloc; @@ -778,11 +777,11 @@ ext4_journalled_write_inline_data(struct inode *inode, return NULL; } - down_write(&EXT4_I(inode)->xattr_sem); + ext4_write_lock_xattr(inode, &no_expand); kaddr = kmap_atomic(page); ext4_write_inline_data(inode, &iloc, kaddr, 0, len); kunmap_atomic(kaddr); - up_write(&EXT4_I(inode)->xattr_sem); + ext4_write_unlock_xattr(inode, &no_expand); return iloc.bh; } @@ -1259,7 +1258,7 @@ out: int ext4_try_add_inline_entry(handle_t *handle, struct ext4_filename *fname, struct inode *dir, struct inode *inode) { - int ret, inline_size; + int ret, inline_size, no_expand; void *inline_start; struct ext4_iloc iloc; @@ -1267,7 +1266,7 @@ int ext4_try_add_inline_entry(handle_t *handle, struct ext4_filename *fname, if (ret) return ret; - down_write(&EXT4_I(dir)->xattr_sem); + ext4_write_lock_xattr(dir, &no_expand); if (!ext4_has_inline_data(dir)) goto out; @@ -1313,7 +1312,7 @@ int ext4_try_add_inline_entry(handle_t *handle, struct ext4_filename *fname, out: ext4_mark_inode_dirty(handle, dir); - up_write(&EXT4_I(dir)->xattr_sem); + ext4_write_unlock_xattr(dir, &no_expand); brelse(iloc.bh); return ret; } @@ -1673,7 +1672,7 @@ int ext4_delete_inline_entry(handle_t *handle, struct buffer_head *bh, int *has_inline_data) { - int err, inline_size; + int err, inline_size, no_expand; struct ext4_iloc iloc; void *inline_start; @@ -1681,7 +1680,7 @@ int ext4_delete_inline_entry(handle_t *handle, if (err) return err; - down_write(&EXT4_I(dir)->xattr_sem); + ext4_write_lock_xattr(dir, &no_expand); if (!ext4_has_inline_data(dir)) { *has_inline_data = 0; goto out; @@ -1715,7 +1714,7 @@ int ext4_delete_inline_entry(handle_t *handle, ext4_show_inline_dir(dir, iloc.bh, inline_start, inline_size); out: - up_write(&EXT4_I(dir)->xattr_sem); + ext4_write_unlock_xattr(dir, &no_expand); brelse(iloc.bh); if (err != -ENOENT) ext4_std_error(dir->i_sb, err); @@ -1814,11 +1813,11 @@ out: int ext4_destroy_inline_data(handle_t *handle, struct inode *inode) { - int ret; + int ret, no_expand; - down_write(&EXT4_I(inode)->xattr_sem); + ext4_write_lock_xattr(inode, &no_expand); ret = ext4_destroy_inline_data_nolock(handle, inode); - up_write(&EXT4_I(inode)->xattr_sem); + ext4_write_unlock_xattr(inode, &no_expand); return ret; } @@ -1903,7 +1902,7 @@ out: void ext4_inline_data_truncate(struct inode *inode, int *has_inline) { handle_t *handle; - int inline_size, value_len, needed_blocks; + int inline_size, value_len, needed_blocks, no_expand; size_t i_size; void *value = NULL; struct ext4_xattr_ibody_find is = { @@ -1920,7 +1919,7 @@ void ext4_inline_data_truncate(struct inode *inode, int *has_inline) if (IS_ERR(handle)) return; - down_write(&EXT4_I(inode)->xattr_sem); + ext4_write_lock_xattr(inode, &no_expand); if (!ext4_has_inline_data(inode)) { *has_inline = 0; ext4_journal_stop(handle); @@ -1978,7 +1977,7 @@ out_error: up_write(&EXT4_I(inode)->i_data_sem); out: brelse(is.iloc.bh); - up_write(&EXT4_I(inode)->xattr_sem); + ext4_write_unlock_xattr(inode, &no_expand); kfree(value); if (inode->i_nlink) ext4_orphan_del(handle, inode); @@ -1994,7 +1993,7 @@ out: int ext4_convert_inline_data(struct inode *inode) { - int error, needed_blocks; + int error, needed_blocks, no_expand; handle_t *handle; struct ext4_iloc iloc; @@ -2016,15 +2015,10 @@ int ext4_convert_inline_data(struct inode *inode) goto out_free; } - down_write(&EXT4_I(inode)->xattr_sem); - if (!ext4_has_inline_data(inode)) { - up_write(&EXT4_I(inode)->xattr_sem); - goto out; - } - - error = ext4_convert_inline_data_nolock(handle, inode, &iloc); - up_write(&EXT4_I(inode)->xattr_sem); -out: + ext4_write_lock_xattr(inode, &no_expand); + if (ext4_has_inline_data(inode)) + error = ext4_convert_inline_data_nolock(handle, inode, &iloc); + ext4_write_unlock_xattr(inode, &no_expand); ext4_journal_stop(handle); out_free: brelse(iloc.bh); -- cgit v1.2.3 From b907f2d5194c2636623415d89cfb91d692af0629 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Wed, 11 Jan 2017 22:14:49 -0500 Subject: ext4: avoid calling ext4_mark_inode_dirty() under unneeded semaphores There is no need to call ext4_mark_inode_dirty while holding xattr_sem or i_data_sem, so where it's easy to avoid it, move it out from the critical region. Signed-off-by: Theodore Ts'o --- fs/ext4/inline.c | 9 +++------ fs/ext4/inode.c | 2 +- 2 files changed, 4 insertions(+), 7 deletions(-) (limited to 'fs/ext4/inline.c') diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c index 99a5312ced52..31f98dd04e51 100644 --- a/fs/ext4/inline.c +++ b/fs/ext4/inline.c @@ -1042,7 +1042,6 @@ static int ext4_add_dirent_to_inline(handle_t *handle, dir->i_mtime = dir->i_ctime = current_time(dir); ext4_update_dx_flag(dir); dir->i_version++; - ext4_mark_inode_dirty(handle, dir); return 1; } @@ -1311,8 +1310,8 @@ int ext4_try_add_inline_entry(handle_t *handle, struct ext4_filename *fname, ret = ext4_convert_inline_data_nolock(handle, dir, &iloc); out: - ext4_mark_inode_dirty(handle, dir); ext4_write_unlock_xattr(dir, &no_expand); + ext4_mark_inode_dirty(handle, dir); brelse(iloc.bh); return ret; } @@ -1708,13 +1707,11 @@ int ext4_delete_inline_entry(handle_t *handle, if (err) goto out; - err = ext4_mark_inode_dirty(handle, dir); - if (unlikely(err)) - goto out; - ext4_show_inline_dir(dir, iloc.bh, inline_start, inline_size); out: ext4_write_unlock_xattr(dir, &no_expand); + if (likely(err == 0)) + err = ext4_mark_inode_dirty(handle, dir); brelse(iloc.bh); if (err != -ENOENT) ext4_std_error(dir->i_sb, err); diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 88d57af1b516..86dde0667ccc 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2464,8 +2464,8 @@ update_disksize: disksize = i_size; if (disksize > EXT4_I(inode)->i_disksize) EXT4_I(inode)->i_disksize = disksize; - err2 = ext4_mark_inode_dirty(handle, inode); up_write(&EXT4_I(inode)->i_data_sem); + err2 = ext4_mark_inode_dirty(handle, inode); if (err2) ext4_error(inode->i_sb, "Failed to mark inode %lu dirty", -- cgit v1.2.3 From 01daf9452569fe2e69e27fe3e617b43d2ebb1e93 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sun, 22 Jan 2017 19:35:49 -0500 Subject: ext4: propagate error values from ext4_inline_data_truncate() Signed-off-by: Theodore Ts'o --- fs/ext4/ext4.h | 2 +- fs/ext4/inline.c | 40 +++++++++++++++++++++++----------------- fs/ext4/inode.c | 4 +++- 3 files changed, 27 insertions(+), 19 deletions(-) (limited to 'fs/ext4/inline.c') diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 6bcb9622fdf9..1cd077e02517 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -3023,7 +3023,7 @@ extern int ext4_inline_data_fiemap(struct inode *inode, extern int ext4_try_to_evict_inline_data(handle_t *handle, struct inode *inode, int needed); -extern void ext4_inline_data_truncate(struct inode *inode, int *has_inline); +extern int ext4_inline_data_truncate(struct inode *inode, int *has_inline); extern int ext4_convert_inline_data(struct inode *inode); diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c index 31f98dd04e51..338cfd862adb 100644 --- a/fs/ext4/inline.c +++ b/fs/ext4/inline.c @@ -1896,10 +1896,10 @@ out: return error; } -void ext4_inline_data_truncate(struct inode *inode, int *has_inline) +int ext4_inline_data_truncate(struct inode *inode, int *has_inline) { handle_t *handle; - int inline_size, value_len, needed_blocks, no_expand; + int inline_size, value_len, needed_blocks, no_expand, err = 0; size_t i_size; void *value = NULL; struct ext4_xattr_ibody_find is = { @@ -1914,19 +1914,19 @@ void ext4_inline_data_truncate(struct inode *inode, int *has_inline) needed_blocks = ext4_writepage_trans_blocks(inode); handle = ext4_journal_start(inode, EXT4_HT_INODE, needed_blocks); if (IS_ERR(handle)) - return; + return PTR_ERR(handle); ext4_write_lock_xattr(inode, &no_expand); if (!ext4_has_inline_data(inode)) { *has_inline = 0; ext4_journal_stop(handle); - return; + return 0; } - if (ext4_orphan_add(handle, inode)) + if ((err = ext4_orphan_add(handle, inode)) != 0) goto out; - if (ext4_get_inode_loc(inode, &is.iloc)) + if ((err = ext4_get_inode_loc(inode, &is.iloc)) != 0) goto out; down_write(&EXT4_I(inode)->i_data_sem); @@ -1937,24 +1937,29 @@ void ext4_inline_data_truncate(struct inode *inode, int *has_inline) if (i_size < inline_size) { /* Clear the content in the xattr space. */ if (inline_size > EXT4_MIN_INLINE_DATA_SIZE) { - if (ext4_xattr_ibody_find(inode, &i, &is)) + if ((err = ext4_xattr_ibody_find(inode, &i, &is)) != 0) goto out_error; BUG_ON(is.s.not_found); value_len = le32_to_cpu(is.s.here->e_value_size); value = kmalloc(value_len, GFP_NOFS); - if (!value) + if (!value) { + err = -ENOMEM; goto out_error; + } - if (ext4_xattr_ibody_get(inode, i.name_index, i.name, - value, value_len)) + err = ext4_xattr_ibody_get(inode, i.name_index, + i.name, value, value_len); + if (err <= 0) goto out_error; i.value = value; i.value_len = i_size > EXT4_MIN_INLINE_DATA_SIZE ? i_size - EXT4_MIN_INLINE_DATA_SIZE : 0; - if (ext4_xattr_ibody_inline_set(handle, inode, &i, &is)) + err = ext4_xattr_ibody_inline_set(handle, inode, + &i, &is); + if (err) goto out_error; } @@ -1979,13 +1984,14 @@ out: if (inode->i_nlink) ext4_orphan_del(handle, inode); - inode->i_mtime = inode->i_ctime = current_time(inode); - ext4_mark_inode_dirty(handle, inode); - if (IS_SYNC(inode)) - ext4_handle_sync(handle); - + if (err == 0) { + inode->i_mtime = inode->i_ctime = current_time(inode); + err = ext4_mark_inode_dirty(handle, inode); + if (IS_SYNC(inode)) + ext4_handle_sync(handle); + } ext4_journal_stop(handle); - return; + return err; } int ext4_convert_inline_data(struct inode *inode) diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 86dde0667ccc..1e2c881f102d 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -4222,7 +4222,9 @@ int ext4_truncate(struct inode *inode) if (ext4_has_inline_data(inode)) { int has_inline = 1; - ext4_inline_data_truncate(inode, &has_inline); + err = ext4_inline_data_truncate(inode, &has_inline); + if (err) + return err; if (has_inline) return 0; } -- cgit v1.2.3 From eb5efbcb762aee4b454b04f7115f73ccbcf8f0ef Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sat, 4 Feb 2017 23:04:00 -0500 Subject: ext4: fix inline data error paths The write_end() function must always unlock the page and drop its ref count, even on an error. Signed-off-by: Theodore Ts'o Cc: stable@vger.kernel.org --- fs/ext4/inline.c | 9 ++++++++- fs/ext4/inode.c | 20 +++++++++++++++----- 2 files changed, 23 insertions(+), 6 deletions(-) (limited to 'fs/ext4/inline.c') diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c index 338cfd862adb..b777b8aa14ae 100644 --- a/fs/ext4/inline.c +++ b/fs/ext4/inline.c @@ -943,8 +943,15 @@ int ext4_da_write_inline_data_end(struct inode *inode, loff_t pos, struct page *page) { int i_size_changed = 0; + int ret; - copied = ext4_write_inline_data_end(inode, pos, len, copied, page); + ret = ext4_write_inline_data_end(inode, pos, len, copied, page); + if (ret < 0) { + unlock_page(page); + put_page(page); + return ret; + } + copied = ret; /* * No need to use i_size_read() here, the i_size diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 918d351d5b94..af97b9170358 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -1330,8 +1330,11 @@ static int ext4_write_end(struct file *file, if (ext4_has_inline_data(inode)) { ret = ext4_write_inline_data_end(inode, pos, len, copied, page); - if (ret < 0) + if (ret < 0) { + unlock_page(page); + put_page(page); goto errout; + } copied = ret; } else copied = block_write_end(file, mapping, pos, @@ -1433,10 +1436,16 @@ static int ext4_journalled_write_end(struct file *file, BUG_ON(!ext4_handle_valid(handle)); - if (ext4_has_inline_data(inode)) - copied = ext4_write_inline_data_end(inode, pos, len, - copied, page); - else if (unlikely(copied < len) && !PageUptodate(page)) { + if (ext4_has_inline_data(inode)) { + ret = ext4_write_inline_data_end(inode, pos, len, + copied, page); + if (ret < 0) { + unlock_page(page); + put_page(page); + goto errout; + } + copied = ret; + } else if (unlikely(copied < len) && !PageUptodate(page)) { copied = 0; ext4_journalled_zero_new_buffers(handle, page, from, to); } else { @@ -1471,6 +1480,7 @@ static int ext4_journalled_write_end(struct file *file, */ ext4_orphan_add(handle, inode); +errout: ret2 = ext4_journal_stop(handle); if (!ret) ret = ret2; -- cgit v1.2.3 From 0db1ff222d40f1601c961f0edb86d10426992595 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sun, 5 Feb 2017 01:28:48 -0500 Subject: ext4: add shutdown bit and check for it Add a shutdown bit that will cause ext4 processing to fail immediately with EIO. Signed-off-by: Theodore Ts'o --- fs/ext4/ext4.h | 6 ++++++ fs/ext4/ext4_jbd2.c | 11 +++++++++++ fs/ext4/file.c | 12 ++++++++++++ fs/ext4/fsync.c | 3 +++ fs/ext4/ialloc.c | 3 +++ fs/ext4/inline.c | 3 +++ fs/ext4/inode.c | 30 ++++++++++++++++++++++++++++-- fs/ext4/namei.c | 12 ++++++++++++ fs/ext4/page-io.c | 2 +- fs/ext4/super.c | 21 +++++++++++++++++++++ fs/ext4/xattr.c | 3 +++ 11 files changed, 103 insertions(+), 3 deletions(-) (limited to 'fs/ext4/inline.c') diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 2e7e02f2f771..35d93ab7f3fb 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -1836,6 +1836,12 @@ static inline bool ext4_has_incompat_features(struct super_block *sb) * Superblock flags */ #define EXT4_FLAGS_RESIZING 0 +#define EXT4_FLAGS_SHUTDOWN 1 + +static inline int ext4_forced_shutdown(struct ext4_sb_info *sbi) +{ + return test_bit(EXT4_FLAGS_SHUTDOWN, &sbi->s_ext4_flags); +} /* diff --git a/fs/ext4/ext4_jbd2.c b/fs/ext4/ext4_jbd2.c index e770c1ee4613..dd106b1d5d89 100644 --- a/fs/ext4/ext4_jbd2.c +++ b/fs/ext4/ext4_jbd2.c @@ -43,6 +43,10 @@ static int ext4_journal_check_start(struct super_block *sb) journal_t *journal; might_sleep(); + + if (unlikely(ext4_forced_shutdown(EXT4_SB(sb)))) + return -EIO; + if (sb->s_flags & MS_RDONLY) return -EROFS; WARN_ON(sb->s_writers.frozen == SB_FREEZE_COMPLETE); @@ -161,6 +165,13 @@ int __ext4_journal_get_write_access(const char *where, unsigned int line, might_sleep(); if (ext4_handle_valid(handle)) { + struct super_block *sb; + + sb = handle->h_transaction->t_journal->j_private; + if (unlikely(ext4_forced_shutdown(EXT4_SB(sb)))) { + jbd2_journal_abort_handle(handle); + return -EIO; + } err = jbd2_journal_get_write_access(handle, bh); if (err) ext4_journal_abort_handle(where, line, __func__, bh, diff --git a/fs/ext4/file.c b/fs/ext4/file.c index d663d3d7c81c..ff3f6107b0ba 100644 --- a/fs/ext4/file.c +++ b/fs/ext4/file.c @@ -57,6 +57,9 @@ static ssize_t ext4_dax_read_iter(struct kiocb *iocb, struct iov_iter *to) static ssize_t ext4_file_read_iter(struct kiocb *iocb, struct iov_iter *to) { + if (unlikely(ext4_forced_shutdown(EXT4_SB(file_inode(iocb->ki_filp)->i_sb)))) + return -EIO; + if (!iov_iter_count(to)) return 0; /* skip atime */ @@ -213,6 +216,9 @@ ext4_file_write_iter(struct kiocb *iocb, struct iov_iter *from) int overwrite = 0; ssize_t ret; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return -EIO; + #ifdef CONFIG_FS_DAX if (IS_DAX(inode)) return ext4_dax_write_iter(iocb, from); @@ -348,6 +354,9 @@ static int ext4_file_mmap(struct file *file, struct vm_area_struct *vma) { struct inode *inode = file->f_mapping->host; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return -EIO; + if (ext4_encrypted_inode(inode)) { int err = fscrypt_get_encryption_info(inode); if (err) @@ -375,6 +384,9 @@ static int ext4_file_open(struct inode * inode, struct file * filp) char buf[64], *cp; int ret; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return -EIO; + if (unlikely(!(sbi->s_mount_flags & EXT4_MF_MNTDIR_SAMPLED) && !(sb->s_flags & MS_RDONLY))) { sbi->s_mount_flags |= EXT4_MF_MNTDIR_SAMPLED; diff --git a/fs/ext4/fsync.c b/fs/ext4/fsync.c index 88effb1053c7..9d549608fd30 100644 --- a/fs/ext4/fsync.c +++ b/fs/ext4/fsync.c @@ -100,6 +100,9 @@ int ext4_sync_file(struct file *file, loff_t start, loff_t end, int datasync) tid_t commit_tid; bool needs_barrier = false; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return -EIO; + J_ASSERT(ext4_journal_current_handle() == NULL); trace_ext4_sync_file_enter(file, datasync); diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c index f372fc431b8e..b14bae2598bc 100644 --- a/fs/ext4/ialloc.c +++ b/fs/ext4/ialloc.c @@ -764,6 +764,9 @@ struct inode *__ext4_new_inode(handle_t *handle, struct inode *dir, if (!dir || !dir->i_nlink) return ERR_PTR(-EPERM); + if (unlikely(ext4_forced_shutdown(EXT4_SB(dir->i_sb)))) + return ERR_PTR(-EIO); + if ((ext4_encrypted_inode(dir) || DUMMY_ENCRYPTION_ENABLED(EXT4_SB(dir->i_sb))) && (S_ISREG(mode) || S_ISDIR(mode) || S_ISLNK(mode))) { diff --git a/fs/ext4/inline.c b/fs/ext4/inline.c index b777b8aa14ae..30a9f210d1e3 100644 --- a/fs/ext4/inline.c +++ b/fs/ext4/inline.c @@ -215,6 +215,9 @@ static void ext4_write_inline_data(struct inode *inode, struct ext4_iloc *iloc, struct ext4_inode *raw_inode; int cp_len = 0; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return; + BUG_ON(!EXT4_I(inode)->i_inline_off); BUG_ON(pos + len > EXT4_I(inode)->i_inline_size); diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index af97b9170358..bc282f9d0969 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -1189,6 +1189,9 @@ static int ext4_write_begin(struct file *file, struct address_space *mapping, pgoff_t index; unsigned from, to; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return -EIO; + trace_ext4_write_begin(inode, pos, len, flags); /* * Reserve one block more for addition to orphan list in case @@ -2047,6 +2050,12 @@ static int ext4_writepage(struct page *page, struct ext4_io_submit io_submit; bool keep_towrite = false; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) { + ext4_invalidatepage(page, 0, PAGE_SIZE); + unlock_page(page); + return -EIO; + } + trace_ext4_writepage(page); size = i_size_read(inode); if (page->index == size >> PAGE_SHIFT) @@ -2422,7 +2431,8 @@ static int mpage_map_and_submit_extent(handle_t *handle, if (err < 0) { struct super_block *sb = inode->i_sb; - if (EXT4_SB(sb)->s_mount_flags & EXT4_MF_FS_ABORTED) + if (ext4_forced_shutdown(EXT4_SB(sb)) || + EXT4_SB(sb)->s_mount_flags & EXT4_MF_FS_ABORTED) goto invalidate_dirty_pages; /* * Let the uper layers retry transient errors. @@ -2644,6 +2654,9 @@ static int ext4_writepages(struct address_space *mapping, struct blk_plug plug; bool give_up_on_write = false; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return -EIO; + percpu_down_read(&sbi->s_journal_flag_rwsem); trace_ext4_writepages(inode, wbc); @@ -2680,7 +2693,8 @@ static int ext4_writepages(struct address_space *mapping, * *never* be called, so if that ever happens, we would want * the stack trace. */ - if (unlikely(sbi->s_mount_flags & EXT4_MF_FS_ABORTED)) { + if (unlikely(ext4_forced_shutdown(EXT4_SB(mapping->host->i_sb)) || + sbi->s_mount_flags & EXT4_MF_FS_ABORTED)) { ret = -EROFS; goto out_writepages; } @@ -2905,6 +2919,9 @@ static int ext4_da_write_begin(struct file *file, struct address_space *mapping, struct inode *inode = mapping->host; handle_t *handle; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return -EIO; + index = pos >> PAGE_SHIFT; if (ext4_nonda_switch(inode->i_sb) || @@ -5212,6 +5229,9 @@ int ext4_setattr(struct dentry *dentry, struct iattr *attr) int orphan = 0; const unsigned int ia_valid = attr->ia_valid; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return -EIO; + error = setattr_prepare(dentry, attr); if (error) return error; @@ -5498,6 +5518,9 @@ int ext4_mark_iloc_dirty(handle_t *handle, { int err = 0; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return -EIO; + if (IS_I_VERSION(inode)) inode_inc_iversion(inode); @@ -5521,6 +5544,9 @@ ext4_reserve_inode_write(handle_t *handle, struct inode *inode, { int err; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return -EIO; + err = ext4_get_inode_loc(inode, iloc); if (!err) { BUFFER_TRACE(iloc->bh, "get_write_access"); diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c index 931da9d5d915..6ad612c576fc 100644 --- a/fs/ext4/namei.c +++ b/fs/ext4/namei.c @@ -2939,6 +2939,9 @@ static int ext4_rmdir(struct inode *dir, struct dentry *dentry) struct ext4_dir_entry_2 *de; handle_t *handle = NULL; + if (unlikely(ext4_forced_shutdown(EXT4_SB(dir->i_sb)))) + return -EIO; + /* Initialize quotas before so that eventual writes go in * separate transaction */ retval = dquot_initialize(dir); @@ -3012,6 +3015,9 @@ static int ext4_unlink(struct inode *dir, struct dentry *dentry) struct ext4_dir_entry_2 *de; handle_t *handle = NULL; + if (unlikely(ext4_forced_shutdown(EXT4_SB(dir->i_sb)))) + return -EIO; + trace_ext4_unlink_enter(dir, dentry); /* Initialize quotas before so that eventual writes go * in separate transaction */ @@ -3082,6 +3088,9 @@ static int ext4_symlink(struct inode *dir, struct fscrypt_str disk_link; struct fscrypt_symlink_data *sd = NULL; + if (unlikely(ext4_forced_shutdown(EXT4_SB(dir->i_sb)))) + return -EIO; + disk_link.len = len + 1; disk_link.name = (char *) symname; @@ -3874,6 +3883,9 @@ static int ext4_rename2(struct inode *old_dir, struct dentry *old_dentry, struct inode *new_dir, struct dentry *new_dentry, unsigned int flags) { + if (unlikely(ext4_forced_shutdown(EXT4_SB(old_dir->i_sb)))) + return -EIO; + if (flags & ~(RENAME_NOREPLACE | RENAME_EXCHANGE | RENAME_WHITEOUT)) return -EINVAL; diff --git a/fs/ext4/page-io.c b/fs/ext4/page-io.c index d83b0f3c5fe9..f8808835a28b 100644 --- a/fs/ext4/page-io.c +++ b/fs/ext4/page-io.c @@ -158,7 +158,7 @@ static int ext4_end_io(ext4_io_end_t *io) io->handle = NULL; /* Following call will use up the handle */ ret = ext4_convert_unwritten_extents(handle, inode, offset, size); - if (ret < 0) { + if (ret < 0 && !ext4_forced_shutdown(EXT4_SB(inode->i_sb))) { ext4_msg(inode->i_sb, KERN_EMERG, "failed to convert unwritten extents to written " "extents -- potential data loss! " diff --git a/fs/ext4/super.c b/fs/ext4/super.c index 514e5fc59893..cfa4ce5a1f80 100644 --- a/fs/ext4/super.c +++ b/fs/ext4/super.c @@ -438,6 +438,9 @@ void __ext4_error(struct super_block *sb, const char *function, struct va_format vaf; va_list args; + if (unlikely(ext4_forced_shutdown(EXT4_SB(sb)))) + return; + if (ext4_error_ratelimit(sb)) { va_start(args, fmt); vaf.fmt = fmt; @@ -459,6 +462,9 @@ void __ext4_error_inode(struct inode *inode, const char *function, struct va_format vaf; struct ext4_super_block *es = EXT4_SB(inode->i_sb)->s_es; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return; + es->s_last_error_ino = cpu_to_le32(inode->i_ino); es->s_last_error_block = cpu_to_le64(block); if (ext4_error_ratelimit(inode->i_sb)) { @@ -491,6 +497,9 @@ void __ext4_error_file(struct file *file, const char *function, struct inode *inode = file_inode(file); char pathname[80], *path; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return; + es = EXT4_SB(inode->i_sb)->s_es; es->s_last_error_ino = cpu_to_le32(inode->i_ino); if (ext4_error_ratelimit(inode->i_sb)) { @@ -567,6 +576,9 @@ void __ext4_std_error(struct super_block *sb, const char *function, char nbuf[16]; const char *errstr; + if (unlikely(ext4_forced_shutdown(EXT4_SB(sb)))) + return; + /* Special case: if the error is EROFS, and we're not already * inside a transaction, then there's really no point in logging * an error. */ @@ -600,6 +612,9 @@ void __ext4_abort(struct super_block *sb, const char *function, struct va_format vaf; va_list args; + if (unlikely(ext4_forced_shutdown(EXT4_SB(sb)))) + return; + save_error_info(sb, function, line); va_start(args, fmt); vaf.fmt = fmt; @@ -695,6 +710,9 @@ __acquires(bitlock) va_list args; struct ext4_super_block *es = EXT4_SB(sb)->s_es; + if (unlikely(ext4_forced_shutdown(EXT4_SB(sb)))) + return; + es->s_last_error_ino = cpu_to_le32(ino); es->s_last_error_block = cpu_to_le64(block); __save_error_info(sb, function, line); @@ -4717,6 +4735,9 @@ static int ext4_sync_fs(struct super_block *sb, int wait) bool needs_barrier = false; struct ext4_sb_info *sbi = EXT4_SB(sb); + if (unlikely(ext4_forced_shutdown(EXT4_SB(sb)))) + return 0; + trace_ext4_sync_fs(sb, wait); flush_workqueue(sbi->rsv_conversion_wq); /* diff --git a/fs/ext4/xattr.c b/fs/ext4/xattr.c index c40bd55b6400..67636acf7624 100644 --- a/fs/ext4/xattr.c +++ b/fs/ext4/xattr.c @@ -411,6 +411,9 @@ ext4_xattr_get(struct inode *inode, int name_index, const char *name, { int error; + if (unlikely(ext4_forced_shutdown(EXT4_SB(inode->i_sb)))) + return -EIO; + if (strlen(name) > 255) return -ERANGE; -- cgit v1.2.3