From b9ef22dedde08ab1b4ccd5f53344984c4dcb89f4 Mon Sep 17 00:00:00 2001 From: Feifei Xu Date: Wed, 1 Jun 2016 19:18:25 +0800 Subject: Btrfs: self-tests: Support non-4k page size self-tests code assumes 4k as the sectorsize and nodesize. This commit fix hardcoded 4K. Enables the self-tests code to be executed on non-4k page sized systems (e.g. ppc64). Reviewed-by: Josef Bacik Signed-off-by: Feifei Xu Signed-off-by: Chandan Rajendra Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 6628fca9f4ed..06ef433552b3 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1147,7 +1147,8 @@ struct extent_buffer *btrfs_find_create_tree_block(struct btrfs_root *root, u64 bytenr) { if (btrfs_test_is_dummy_root(root)) - return alloc_test_extent_buffer(root->fs_info, bytenr); + return alloc_test_extent_buffer(root->fs_info, bytenr, + root->nodesize); return alloc_extent_buffer(root->fs_info, bytenr); } @@ -1314,14 +1315,15 @@ static struct btrfs_root *btrfs_alloc_root(struct btrfs_fs_info *fs_info, #ifdef CONFIG_BTRFS_FS_RUN_SANITY_TESTS /* Should only be used by the testing infrastructure */ -struct btrfs_root *btrfs_alloc_dummy_root(void) +struct btrfs_root *btrfs_alloc_dummy_root(u32 sectorsize, u32 nodesize) { struct btrfs_root *root; root = btrfs_alloc_root(NULL, GFP_KERNEL); if (!root) return ERR_PTR(-ENOMEM); - __setup_root(4096, 4096, 4096, root, NULL, 1); + /* We don't use the stripesize in selftest, set it as sectorsize */ + __setup_root(nodesize, sectorsize, sectorsize, root, NULL, 1); set_bit(BTRFS_ROOT_DUMMY_ROOT, &root->state); root->alloc_bytenr = 0; -- cgit v1.2.2 From 99e3ecfcb9f4ca35192d20a5bea158b81f600062 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Fri, 3 Jun 2016 12:05:14 -0700 Subject: Btrfs: add more validation checks for superblock This adds validation checks for super_total_bytes, super_bytes_used and super_stripesize, super_num_devices. Reported-by: Vegard Nossum Reported-by: Quentin Casasnovas Reviewed-by: David Sterba Signed-off-by: Liu Bo Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 11 +++++++++++ 1 file changed, 11 insertions(+) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 6628fca9f4ed..2bd5f5e975cc 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -4130,6 +4130,17 @@ static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info, * Hint to catch really bogus numbers, bitflips or so, more exact checks are * done later */ + if (btrfs_super_bytes_used(sb) < 6 * btrfs_super_nodesize(sb)) { + btrfs_err(fs_info, "bytes_used is too small %llu", + btrfs_super_bytes_used(sb)); + ret = -EINVAL; + } + if (!is_power_of_2(btrfs_super_stripesize(sb)) || + btrfs_super_stripesize(sb) != sectorsize) { + btrfs_err(fs_info, "invalid stripesize %u", + btrfs_super_stripesize(sb)); + ret = -EINVAL; + } if (btrfs_super_num_devices(sb) > (1UL << 31)) printk(KERN_WARNING "BTRFS: suspicious number of devices: %llu\n", btrfs_super_num_devices(sb)); -- cgit v1.2.2 From ef9f2db365c31433e52b0c5863793273bb632666 Mon Sep 17 00:00:00 2001 From: Feifei Xu Date: Wed, 1 Jun 2016 19:18:28 +0800 Subject: Btrfs: self-tests: Use macros instead of constants and add missing newline This commit replaces numerical constants with appropriate preprocessor macros. Reviewed-by: Josef Bacik Signed-off-by: Chandan Rajendra Signed-off-by: Feifei Xu Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 06ef433552b3..c3764ddb2a46 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1323,7 +1323,8 @@ struct btrfs_root *btrfs_alloc_dummy_root(u32 sectorsize, u32 nodesize) if (!root) return ERR_PTR(-ENOMEM); /* We don't use the stripesize in selftest, set it as sectorsize */ - __setup_root(nodesize, sectorsize, sectorsize, root, NULL, 1); + __setup_root(nodesize, sectorsize, sectorsize, root, NULL, + BTRFS_ROOT_TREE_OBJECTID); set_bit(BTRFS_ROOT_DUMMY_ROOT, &root->state); root->alloc_bytenr = 0; -- cgit v1.2.2 From c871b0f2fd27e7f9097d507f47de5270f88003b9 Mon Sep 17 00:00:00 2001 From: Liu Bo Date: Mon, 6 Jun 2016 12:01:23 -0700 Subject: Btrfs: check if extent buffer is aligned to sectorsize Thanks to fuzz testing, we can pass an invalid bytenr to extent buffer via alloc_extent_buffer(). An unaligned eb can have more pages than it should have, which ends up extent buffer's leak or some corrupted content in extent buffer. This adds a warning to let us quickly know what was happening. Now that alloc_extent_buffer() no more returns NULL, this changes its caller and callers of its caller to match with the new error handling. Signed-off-by: Liu Bo Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 1142127f6e5e..7b5d5e8efde6 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1098,7 +1098,7 @@ void readahead_tree_block(struct btrfs_root *root, u64 bytenr) struct inode *btree_inode = root->fs_info->btree_inode; buf = btrfs_find_create_tree_block(root, bytenr); - if (!buf) + if (IS_ERR(buf)) return; read_extent_buffer_pages(&BTRFS_I(btree_inode)->io_tree, buf, 0, WAIT_NONE, btree_get_extent, 0); @@ -1114,7 +1114,7 @@ int reada_tree_block_flagged(struct btrfs_root *root, u64 bytenr, int ret; buf = btrfs_find_create_tree_block(root, bytenr); - if (!buf) + if (IS_ERR(buf)) return 0; set_bit(EXTENT_BUFFER_READAHEAD, &buf->bflags); @@ -1172,8 +1172,8 @@ struct extent_buffer *read_tree_block(struct btrfs_root *root, u64 bytenr, int ret; buf = btrfs_find_create_tree_block(root, bytenr); - if (!buf) - return ERR_PTR(-ENOMEM); + if (IS_ERR(buf)) + return buf; ret = btree_read_extent_buffer_pages(root, buf, 0, parent_transid); if (ret) { -- cgit v1.2.2 From 90c711ab380d633bf85249bd5d819edb601feda7 Mon Sep 17 00:00:00 2001 From: Zygo Blaxell Date: Sun, 12 Jun 2016 23:39:58 -0400 Subject: btrfs: avoid blocking open_ctree from cleaner_kthread This fixes a problem introduced in commit 2f3165ecf103599f82bf0ea254039db335fb5005 "btrfs: don't force mounts to wait for cleaner_kthread to delete one or more subvolumes". open_ctree eventually calls btrfs_replay_log which in turn calls btrfs_commit_super which tries to lock the cleaner_mutex, causing a recursive mutex deadlock during mount. Instead of playing whack-a-mole trying to keep up with all the functions that may want to lock cleaner_mutex, put all the cleaner_mutex lockers back where they were, and attack the problem more directly: keep cleaner_kthread asleep until the filesystem is mounted. When filesystems are mounted read-only and later remounted read-write, open_ctree did not set fs_info->open and neither does anything else. Set this flag in btrfs_remount so that neither btrfs_delete_unused_bgs nor cleaner_kthread get confused by the common case of "/" filesystem read-only mount followed by read-write remount. Signed-off-by: Zygo Blaxell Reviewed-by: David Sterba Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 25 ++++++++++--------------- 1 file changed, 10 insertions(+), 15 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 7b5d5e8efde6..789f5f233940 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -1806,6 +1806,13 @@ static int cleaner_kthread(void *arg) if (btrfs_need_cleaner_sleep(root)) goto sleep; + /* + * Do not do anything if we might cause open_ctree() to block + * before we have finished mounting the filesystem. + */ + if (!root->fs_info->open) + goto sleep; + if (!mutex_trylock(&root->fs_info->cleaner_mutex)) goto sleep; @@ -2520,7 +2527,6 @@ int open_ctree(struct super_block *sb, int num_backups_tried = 0; int backup_index = 0; int max_active; - bool cleaner_mutex_locked = false; tree_root = fs_info->tree_root = btrfs_alloc_root(fs_info, GFP_KERNEL); chunk_root = fs_info->chunk_root = btrfs_alloc_root(fs_info, GFP_KERNEL); @@ -2999,13 +3005,6 @@ retry_root_backup: goto fail_sysfs; } - /* - * Hold the cleaner_mutex thread here so that we don't block - * for a long time on btrfs_recover_relocation. cleaner_kthread - * will wait for us to finish mounting the filesystem. - */ - mutex_lock(&fs_info->cleaner_mutex); - cleaner_mutex_locked = true; fs_info->cleaner_kthread = kthread_run(cleaner_kthread, tree_root, "btrfs-cleaner"); if (IS_ERR(fs_info->cleaner_kthread)) @@ -3065,8 +3064,10 @@ retry_root_backup: ret = btrfs_cleanup_fs_roots(fs_info); if (ret) goto fail_qgroup; - /* We locked cleaner_mutex before creating cleaner_kthread. */ + + mutex_lock(&fs_info->cleaner_mutex); ret = btrfs_recover_relocation(tree_root); + mutex_unlock(&fs_info->cleaner_mutex); if (ret < 0) { btrfs_warn(fs_info, "failed to recover relocation: %d", ret); @@ -3074,8 +3075,6 @@ retry_root_backup: goto fail_qgroup; } } - mutex_unlock(&fs_info->cleaner_mutex); - cleaner_mutex_locked = false; location.objectid = BTRFS_FS_TREE_OBJECTID; location.type = BTRFS_ROOT_ITEM_KEY; @@ -3189,10 +3188,6 @@ fail_cleaner: filemap_write_and_wait(fs_info->btree_inode->i_mapping); fail_sysfs: - if (cleaner_mutex_locked) { - mutex_unlock(&fs_info->cleaner_mutex); - cleaner_mutex_locked = false; - } btrfs_sysfs_remove_mounted(fs_info); fail_fsdev_sysfs: -- cgit v1.2.2 From dd5c93111dc9d26e038ac437f7a403d617e82c62 Mon Sep 17 00:00:00 2001 From: Chandan Rajendra Date: Thu, 16 Jun 2016 22:07:58 +0530 Subject: Btrfs: btrfs_check_super_valid: Allow 4096 as stripesize Older btrfs-progs/mkfs.btrfs sets 4096 as the stripesize. Hence restricting stripesize to be equal to sectorsize would cause super block validation to return an error on architectures where PAGE_SIZE is not equal to 4096. Hence as a workaround, this commit allows stripesize to be set to 4096 bytes. Signed-off-by: Chandan Rajendra Signed-off-by: David Sterba --- fs/btrfs/disk-io.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 789f5f233940..54cca7a1572b 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -4134,7 +4134,8 @@ static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info, ret = -EINVAL; } if (!is_power_of_2(btrfs_super_stripesize(sb)) || - btrfs_super_stripesize(sb) != sectorsize) { + ((btrfs_super_stripesize(sb) != sectorsize) && + (btrfs_super_stripesize(sb) != 4096))) { btrfs_err(fs_info, "invalid stripesize %u", btrfs_super_stripesize(sb)); ret = -EINVAL; -- cgit v1.2.2 From b7f67055d2df9b8f68f02e49d256ee3973999bd2 Mon Sep 17 00:00:00 2001 From: Chandan Rajendra Date: Thu, 23 Jun 2016 15:16:44 +0530 Subject: Btrfs: Force stripesize to the value of sectorsize Btrfs code currently assumes stripesize to be same as sectorsize. However Btrfs-progs (until commit df05c7ed455f519e6e15e46196392e4757257305) has been setting btrfs_super_block->stripesize to a value of 4096. This commit makes sure that the value of btrfs_super_block->stripesize is a power of 2. Later, it unconditionally sets btrfs_root->stripesize to sectorsize. Signed-off-by: Chandan Rajendra Reviewed-by: David Sterba Signed-off-by: David Sterba Signed-off-by: Chris Mason --- fs/btrfs/disk-io.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) (limited to 'fs/btrfs/disk-io.c') diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c index 54cca7a1572b..60ce1190307b 100644 --- a/fs/btrfs/disk-io.c +++ b/fs/btrfs/disk-io.c @@ -2806,7 +2806,7 @@ int open_ctree(struct super_block *sb, nodesize = btrfs_super_nodesize(disk_super); sectorsize = btrfs_super_sectorsize(disk_super); - stripesize = btrfs_super_stripesize(disk_super); + stripesize = sectorsize; fs_info->dirty_metadata_batch = nodesize * (1 + ilog2(nr_cpu_ids)); fs_info->delalloc_batch = sectorsize * 512 * (1 + ilog2(nr_cpu_ids)); @@ -4133,9 +4133,7 @@ static int btrfs_check_super_valid(struct btrfs_fs_info *fs_info, btrfs_super_bytes_used(sb)); ret = -EINVAL; } - if (!is_power_of_2(btrfs_super_stripesize(sb)) || - ((btrfs_super_stripesize(sb) != sectorsize) && - (btrfs_super_stripesize(sb) != 4096))) { + if (!is_power_of_2(btrfs_super_stripesize(sb))) { btrfs_err(fs_info, "invalid stripesize %u", btrfs_super_stripesize(sb)); ret = -EINVAL; -- cgit v1.2.2