From 28b4c263961c47da84ed8b5be0b5116bad1133eb Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Sun, 7 Feb 2016 19:35:05 -0500 Subject: ext4 crypto: revalidate dentry after adding or removing the key Add a validation check for dentries for encrypted directory to make sure we're not caching stale data after a key has been added or removed. Also check to make sure that status of the encryption key is updated when readdir(2) is executed. Signed-off-by: Theodore Ts'o --- fs/ext4/crypto.c | 56 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ fs/ext4/dir.c | 6 ++++++ fs/ext4/ext4.h | 1 + fs/ext4/namei.c | 18 ++++++++++++++++++ 4 files changed, 81 insertions(+) (limited to 'fs') diff --git a/fs/ext4/crypto.c b/fs/ext4/crypto.c index c8021208a7eb..38f7562489bb 100644 --- a/fs/ext4/crypto.c +++ b/fs/ext4/crypto.c @@ -467,3 +467,59 @@ uint32_t ext4_validate_encryption_key_size(uint32_t mode, uint32_t size) return size; return 0; } + +/* + * Validate dentries for encrypted directories to make sure we aren't + * potentially caching stale data after a key has been added or + * removed. + */ +static int ext4_d_revalidate(struct dentry *dentry, unsigned int flags) +{ + struct inode *dir = d_inode(dentry->d_parent); + struct ext4_crypt_info *ci = EXT4_I(dir)->i_crypt_info; + int dir_has_key, cached_with_key; + + if (!ext4_encrypted_inode(dir)) + return 0; + + if (ci && ci->ci_keyring_key && + (ci->ci_keyring_key->flags & ((1 << KEY_FLAG_INVALIDATED) | + (1 << KEY_FLAG_REVOKED) | + (1 << KEY_FLAG_DEAD)))) + ci = NULL; + + /* this should eventually be an flag in d_flags */ + cached_with_key = dentry->d_fsdata != NULL; + dir_has_key = (ci != NULL); + + /* + * If the dentry was cached without the key, and it is a + * negative dentry, it might be a valid name. We can't check + * if the key has since been made available due to locking + * reasons, so we fail the validation so ext4_lookup() can do + * this check. + * + * We also fail the validation if the dentry was created with + * the key present, but we no longer have the key, or vice versa. + */ + if ((!cached_with_key && d_is_negative(dentry)) || + (!cached_with_key && dir_has_key) || + (cached_with_key && !dir_has_key)) { +#if 0 /* Revalidation debug */ + char buf[80]; + char *cp = simple_dname(dentry, buf, sizeof(buf)); + + if (IS_ERR(cp)) + cp = (char *) "???"; + pr_err("revalidate: %s %p %d %d %d\n", cp, dentry->d_fsdata, + cached_with_key, d_is_negative(dentry), + dir_has_key); +#endif + return 0; + } + return 1; +} + +const struct dentry_operations ext4_encrypted_d_ops = { + .d_revalidate = ext4_d_revalidate, +}; diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c index 1d1bca74f844..6d17f31a31d7 100644 --- a/fs/ext4/dir.c +++ b/fs/ext4/dir.c @@ -111,6 +111,12 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx) int dir_has_error = 0; struct ext4_str fname_crypto_str = {.name = NULL, .len = 0}; + if (ext4_encrypted_inode(inode)) { + err = ext4_get_encryption_info(inode); + if (err && err != -ENOKEY) + return err; + } + if (is_dx_dir(inode)) { err = ext4_dx_readdir(file, ctx); if (err != ERR_BAD_DX_DIR) { diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h index 0662b285dc8a..157b458a69d4 100644 --- a/fs/ext4/ext4.h +++ b/fs/ext4/ext4.h @@ -2302,6 +2302,7 @@ struct page *ext4_encrypt(struct inode *inode, int ext4_decrypt(struct page *page); int ext4_encrypted_zeroout(struct inode *inode, ext4_lblk_t lblk, ext4_fsblk_t pblk, ext4_lblk_t len); +extern const struct dentry_operations ext4_encrypted_d_ops; #ifdef CONFIG_EXT4_FS_ENCRYPTION int ext4_init_crypto(void); diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c index 06574dd77614..5de8483f0062 100644 --- a/fs/ext4/namei.c +++ b/fs/ext4/namei.c @@ -1558,6 +1558,24 @@ static struct dentry *ext4_lookup(struct inode *dir, struct dentry *dentry, unsi struct ext4_dir_entry_2 *de; struct buffer_head *bh; + if (ext4_encrypted_inode(dir)) { + int res = ext4_get_encryption_info(dir); + + /* + * This should be a properly defined flag for + * dentry->d_flags when we uplift this to the VFS. + * d_fsdata is set to (void *) 1 if if the dentry is + * created while the directory was encrypted and we + * don't have access to the key. + */ + dentry->d_fsdata = NULL; + if (ext4_encryption_info(dir)) + dentry->d_fsdata = (void *) 1; + d_set_d_op(dentry, &ext4_encrypted_d_ops); + if (res && res != -ENOKEY) + return ERR_PTR(res); + } + if (dentry->d_name.len > EXT4_NAME_LEN) return ERR_PTR(-ENAMETOOLONG); -- cgit v1.2.3-70-g09d2 From ff978b09f973db0d0597704eba350a994d7729e6 Mon Sep 17 00:00:00 2001 From: Theodore Ts'o Date: Mon, 8 Feb 2016 00:54:26 -0500 Subject: ext4 crypto: move context consistency check to ext4_file_open() In the case where the per-file key for the directory is cached, but root does not have access to the key needed to derive the per-file key for the files in the directory, we allow the lookup to succeed, so that lstat(2) and unlink(2) can suceed. However, if a program tries to open the file, it will get an ENOKEY error. Signed-off-by: Theodore Ts'o --- fs/ext4/file.c | 9 +++++++++ fs/ext4/namei.c | 8 ++++++-- 2 files changed, 15 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext4/file.c b/fs/ext4/file.c index 1126436dada1..474f1a4d2ca8 100644 --- a/fs/ext4/file.c +++ b/fs/ext4/file.c @@ -350,6 +350,7 @@ static int ext4_file_open(struct inode * inode, struct file * filp) struct super_block *sb = inode->i_sb; struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb); struct vfsmount *mnt = filp->f_path.mnt; + struct inode *dir = filp->f_path.dentry->d_parent->d_inode; struct path path; char buf[64], *cp; int ret; @@ -393,6 +394,14 @@ static int ext4_file_open(struct inode * inode, struct file * filp) if (ext4_encryption_info(inode) == NULL) return -ENOKEY; } + if (ext4_encrypted_inode(dir) && + !ext4_is_child_context_consistent_with_parent(dir, inode)) { + ext4_warning(inode->i_sb, + "Inconsistent encryption contexts: %lu/%lu\n", + (unsigned long) dir->i_ino, + (unsigned long) inode->i_ino); + return -EPERM; + } /* * Set up the jbd2_inode if we are opening the inode for * writing and the journal is present diff --git a/fs/ext4/namei.c b/fs/ext4/namei.c index 5de8483f0062..48e4b8907826 100644 --- a/fs/ext4/namei.c +++ b/fs/ext4/namei.c @@ -1603,11 +1603,15 @@ static struct dentry *ext4_lookup(struct inode *dir, struct dentry *dentry, unsi return ERR_PTR(-EFSCORRUPTED); } if (!IS_ERR(inode) && ext4_encrypted_inode(dir) && - (S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode) || - S_ISLNK(inode->i_mode)) && + (S_ISDIR(inode->i_mode) || S_ISLNK(inode->i_mode)) && !ext4_is_child_context_consistent_with_parent(dir, inode)) { + int nokey = ext4_encrypted_inode(inode) && + !ext4_encryption_info(inode); + iput(inode); + if (nokey) + return ERR_PTR(-ENOKEY); ext4_warning(inode->i_sb, "Inconsistent encryption contexts: %lu/%lu\n", (unsigned long) dir->i_ino, -- cgit v1.2.3-70-g09d2 From 05145bd799e498ce4e3b5145894174ee881f02b0 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Thu, 11 Feb 2016 23:15:12 -0500 Subject: ext4: fix scheduling in atomic on group checksum failure When block group checksum is wrong, we call ext4_error() while holding group spinlock from ext4_init_block_bitmap() or ext4_init_inode_bitmap() which results in scheduling while in atomic. Fix the issue by calling ext4_error() later after dropping the spinlock. CC: stable@vger.kernel.org Reported-by: Dmitry Vyukov Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o Reviewed-by: Darrick J. Wong --- fs/ext4/balloc.c | 7 ++++--- fs/ext4/ialloc.c | 6 ++++-- 2 files changed, 8 insertions(+), 5 deletions(-) (limited to 'fs') diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c index ec0668a60678..fe1f50fe764f 100644 --- a/fs/ext4/balloc.c +++ b/fs/ext4/balloc.c @@ -191,7 +191,6 @@ static int ext4_init_block_bitmap(struct super_block *sb, /* If checksum is bad mark all blocks used to prevent allocation * essentially implementing a per-group read-only flag. */ if (!ext4_group_desc_csum_verify(sb, block_group, gdp)) { - ext4_error(sb, "Checksum bad for group %u", block_group); grp = ext4_get_group_info(sb, block_group); if (!EXT4_MB_GRP_BBITMAP_CORRUPT(grp)) percpu_counter_sub(&sbi->s_freeclusters_counter, @@ -442,14 +441,16 @@ ext4_read_block_bitmap_nowait(struct super_block *sb, ext4_group_t block_group) } ext4_lock_group(sb, block_group); if (desc->bg_flags & cpu_to_le16(EXT4_BG_BLOCK_UNINIT)) { - err = ext4_init_block_bitmap(sb, bh, block_group, desc); set_bitmap_uptodate(bh); set_buffer_uptodate(bh); ext4_unlock_group(sb, block_group); unlock_buffer(bh); - if (err) + if (err) { + ext4_error(sb, "Failed to init block bitmap for group " + "%u: %d", block_group, err); goto out; + } goto verify; } ext4_unlock_group(sb, block_group); diff --git a/fs/ext4/ialloc.c b/fs/ext4/ialloc.c index 3fcfd50a2e8a..acc0ad56bf2f 100644 --- a/fs/ext4/ialloc.c +++ b/fs/ext4/ialloc.c @@ -76,7 +76,6 @@ static int ext4_init_inode_bitmap(struct super_block *sb, /* If checksum is bad mark all blocks and inodes use to prevent * allocation, essentially implementing a per-group read-only flag. */ if (!ext4_group_desc_csum_verify(sb, block_group, gdp)) { - ext4_error(sb, "Checksum bad for group %u", block_group); grp = ext4_get_group_info(sb, block_group); if (!EXT4_MB_GRP_BBITMAP_CORRUPT(grp)) percpu_counter_sub(&sbi->s_freeclusters_counter, @@ -191,8 +190,11 @@ ext4_read_inode_bitmap(struct super_block *sb, ext4_group_t block_group) set_buffer_verified(bh); ext4_unlock_group(sb, block_group); unlock_buffer(bh); - if (err) + if (err) { + ext4_error(sb, "Failed to init inode bitmap for group " + "%u: %d", block_group, err); goto out; + } return bh; } ext4_unlock_group(sb, block_group); -- cgit v1.2.3-70-g09d2 From fdde368e7c2963e319f919fa57908537060847fc Mon Sep 17 00:00:00 2001 From: Anton Protopopov Date: Thu, 11 Feb 2016 23:57:21 -0500 Subject: ext4: ioctl: fix erroneous return value The ext4_ioctl_setflags() function which is used in the ioctls EXT4_IOC_SETFLAGS and EXT4_IOC_FSSETXATTR may return the positive value EPERM instead of -EPERM in case of error. This bug was introduced by a recent commit 9b7365fc. The following program can be used to illustrate the wrong behavior: #include #include #include #include #include #define FS_IOC_GETFLAGS _IOR('f', 1, long) #define FS_IOC_SETFLAGS _IOW('f', 2, long) #define FS_IMMUTABLE_FL 0x00000010 int main(void) { int fd; long flags; fd = open("file", O_RDWR|O_CREAT, 0600); if (fd < 0) err(1, "open"); if (ioctl(fd, FS_IOC_GETFLAGS, &flags) < 0) err(1, "ioctl: FS_IOC_GETFLAGS"); flags |= FS_IMMUTABLE_FL; if (ioctl(fd, FS_IOC_SETFLAGS, &flags) < 0) err(1, "ioctl: FS_IOC_SETFLAGS"); warnx("ioctl returned no error"); return 0; } Running it gives the following result: $ strace -e ioctl ./test ioctl(3, FS_IOC_GETFLAGS, 0x7ffdbd8bfd38) = 0 ioctl(3, FS_IOC_SETFLAGS, 0x7ffdbd8bfd38) = 1 test: ioctl returned no error +++ exited with 0 +++ Running the program on a kernel with the bug fixed gives the proper result: $ strace -e ioctl ./test ioctl(3, FS_IOC_GETFLAGS, 0x7ffdd2768258) = 0 ioctl(3, FS_IOC_SETFLAGS, 0x7ffdd2768258) = -1 EPERM (Operation not permitted) test: ioctl: FS_IOC_SETFLAGS: Operation not permitted +++ exited with 1 +++ Signed-off-by: Anton Protopopov Signed-off-by: Theodore Ts'o --- fs/ext4/ioctl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c index 0f6c36922c24..a99b010e2194 100644 --- a/fs/ext4/ioctl.c +++ b/fs/ext4/ioctl.c @@ -208,7 +208,7 @@ static int ext4_ioctl_setflags(struct inode *inode, { struct ext4_inode_info *ei = EXT4_I(inode); handle_t *handle = NULL; - int err = EPERM, migrate = 0; + int err = -EPERM, migrate = 0; struct ext4_iloc iloc; unsigned int oldflags, mask, i; unsigned int jflag; -- cgit v1.2.3-70-g09d2 From 802cf1f9f504e09f7118c9acc69ab3e825449e52 Mon Sep 17 00:00:00 2001 From: Huaitong Han Date: Fri, 12 Feb 2016 00:17:16 -0500 Subject: ext4: add a line break for proc mb_groups display This patch adds a line break for proc mb_groups display. Signed-off-by: Huaitong Han Signed-off-by: Theodore Ts'o Reviewed-by: Andreas Dilger --- fs/ext4/mballoc.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c index 61eaf74dca37..4424b7bf8ac6 100644 --- a/fs/ext4/mballoc.c +++ b/fs/ext4/mballoc.c @@ -2285,7 +2285,7 @@ static int ext4_mb_seq_groups_show(struct seq_file *seq, void *v) if (group == 0) seq_puts(seq, "#group: free frags first [" " 2^0 2^1 2^2 2^3 2^4 2^5 2^6 " - " 2^7 2^8 2^9 2^10 2^11 2^12 2^13 ]"); + " 2^7 2^8 2^9 2^10 2^11 2^12 2^13 ]\n"); i = (sb->s_blocksize_bits + 2) * sizeof(sg.info.bb_counters[0]) + sizeof(struct ext4_group_info); -- cgit v1.2.3-70-g09d2 From 46901760b46064964b41015d00c140c83aa05bcf Mon Sep 17 00:00:00 2001 From: Insu Yun Date: Fri, 12 Feb 2016 01:15:59 -0500 Subject: ext4: fix potential integer overflow Since sizeof(ext_new_group_data) > sizeof(ext_new_flex_group_data), integer overflow could be happened. Therefore, need to fix integer overflow sanitization. Cc: stable@vger.kernel.org Signed-off-by: Insu Yun Signed-off-by: Theodore Ts'o --- fs/ext4/resize.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'fs') diff --git a/fs/ext4/resize.c b/fs/ext4/resize.c index ad62d7acc315..34038e3598d5 100644 --- a/fs/ext4/resize.c +++ b/fs/ext4/resize.c @@ -198,7 +198,7 @@ static struct ext4_new_flex_group_data *alloc_flex_gd(unsigned long flexbg_size) if (flex_gd == NULL) goto out3; - if (flexbg_size >= UINT_MAX / sizeof(struct ext4_new_flex_group_data)) + if (flexbg_size >= UINT_MAX / sizeof(struct ext4_new_group_data)) goto out2; flex_gd->count = flexbg_size; -- cgit v1.2.3-70-g09d2 From bcff24887d00bce102e0857d7b0a8c44a40f53d1 Mon Sep 17 00:00:00 2001 From: Eryu Guan Date: Fri, 12 Feb 2016 01:20:43 -0500 Subject: ext4: don't read blocks from disk after extents being swapped I notice ext4/307 fails occasionally on ppc64 host, reporting md5 checksum mismatch after moving data from original file to donor file. The reason is that move_extent_per_page() calls __block_write_begin() and block_commit_write() to write saved data from original inode blocks to donor inode blocks, but __block_write_begin() not only maps buffer heads but also reads block content from disk if the size is not block size aligned. At this time the physical block number in mapped buffer head is pointing to the donor file not the original file, and that results in reading wrong data to page, which get written to disk in following block_commit_write call. This also can be reproduced by the following script on 1k block size ext4 on x86_64 host: mnt=/mnt/ext4 donorfile=$mnt/donor testfile=$mnt/testfile e4compact=~/xfstests/src/e4compact rm -f $donorfile $testfile # reserve space for donor file, written by 0xaa and sync to disk to # avoid EBUSY on EXT4_IOC_MOVE_EXT xfs_io -fc "pwrite -S 0xaa 0 1m" -c "fsync" $donorfile # create test file written by 0xbb xfs_io -fc "pwrite -S 0xbb 0 1023" -c "fsync" $testfile # compute initial md5sum md5sum $testfile | tee md5sum.txt # drop cache, force e4compact to read data from disk echo 3 > /proc/sys/vm/drop_caches # test defrag echo "$testfile" | $e4compact -i -v -f $donorfile # check md5sum md5sum -c md5sum.txt Fix it by creating & mapping buffer heads only but not reading blocks from disk, because all the data in page is guaranteed to be up-to-date in mext_page_mkuptodate(). Cc: stable@vger.kernel.org Signed-off-by: Eryu Guan Signed-off-by: Theodore Ts'o --- fs/ext4/move_extent.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-) (limited to 'fs') diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c index fb6f11709ae6..e032a0423e35 100644 --- a/fs/ext4/move_extent.c +++ b/fs/ext4/move_extent.c @@ -265,11 +265,12 @@ move_extent_per_page(struct file *o_filp, struct inode *donor_inode, ext4_lblk_t orig_blk_offset, donor_blk_offset; unsigned long blocksize = orig_inode->i_sb->s_blocksize; unsigned int tmp_data_size, data_size, replaced_size; - int err2, jblocks, retries = 0; + int i, err2, jblocks, retries = 0; int replaced_count = 0; int from = data_offset_in_page << orig_inode->i_blkbits; int blocks_per_page = PAGE_CACHE_SIZE >> orig_inode->i_blkbits; struct super_block *sb = orig_inode->i_sb; + struct buffer_head *bh = NULL; /* * It needs twice the amount of ordinary journal buffers because @@ -380,8 +381,16 @@ data_copy: } /* Perform all necessary steps similar write_begin()/write_end() * but keeping in mind that i_size will not change */ - *err = __block_write_begin(pagep[0], from, replaced_size, - ext4_get_block); + if (!page_has_buffers(pagep[0])) + create_empty_buffers(pagep[0], 1 << orig_inode->i_blkbits, 0); + bh = page_buffers(pagep[0]); + for (i = 0; i < data_offset_in_page; i++) + bh = bh->b_this_page; + for (i = 0; i < block_len_in_page; i++) { + *err = ext4_get_block(orig_inode, orig_blk_offset + i, bh, 0); + if (*err < 0) + break; + } if (!*err) *err = block_commit_write(pagep[0], from, from + replaced_size); -- cgit v1.2.3-70-g09d2 From 56263b4ceba90b1db77a6b31e22726f101bca807 Mon Sep 17 00:00:00 2001 From: Eryu Guan Date: Fri, 12 Feb 2016 01:23:00 -0500 Subject: ext4: remove unused parameter "newblock" in convert_initialized_extent() The "newblock" parameter is not used in convert_initialized_extent(), remove it. Signed-off-by: Eryu Guan Signed-off-by: Theodore Ts'o --- fs/ext4/extents.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c index 0ffabaf90aa5..3753ceb0b0dd 100644 --- a/fs/ext4/extents.c +++ b/fs/ext4/extents.c @@ -3928,7 +3928,7 @@ static int convert_initialized_extent(handle_t *handle, struct inode *inode, struct ext4_map_blocks *map, struct ext4_ext_path **ppath, int flags, - unsigned int allocated, ext4_fsblk_t newblock) + unsigned int allocated) { struct ext4_ext_path *path = *ppath; struct ext4_extent *ex; @@ -4347,7 +4347,7 @@ int ext4_ext_map_blocks(handle_t *handle, struct inode *inode, (flags & EXT4_GET_BLOCKS_CONVERT_UNWRITTEN)) { allocated = convert_initialized_extent( handle, inode, map, &path, - flags, allocated, newblock); + flags, allocated); goto out2; } else if (!ext4_ext_is_unwritten(ex)) goto out; -- cgit v1.2.3-70-g09d2 From c906f38e8853cfd407b30d2f4756a93c1d8f698f Mon Sep 17 00:00:00 2001 From: Kirill Tkhai Date: Tue, 16 Feb 2016 00:20:19 -0500 Subject: ext4: fix memleak in ext4_readdir() When ext4_bread() fails, fname_crypto_str remains allocated after return. Fix that. Signed-off-by: Kirill Tkhai Signed-off-by: Theodore Ts'o CC: Dmitry Monakhov --- fs/ext4/dir.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext4/dir.c b/fs/ext4/dir.c index 6d17f31a31d7..33f5e2a50cf8 100644 --- a/fs/ext4/dir.c +++ b/fs/ext4/dir.c @@ -163,8 +163,11 @@ static int ext4_readdir(struct file *file, struct dir_context *ctx) index, 1); file->f_ra.prev_pos = (loff_t)index << PAGE_CACHE_SHIFT; bh = ext4_bread(NULL, inode, map.m_lblk, 0); - if (IS_ERR(bh)) - return PTR_ERR(bh); + if (IS_ERR(bh)) { + err = PTR_ERR(bh); + bh = NULL; + goto errout; + } } if (!bh) { -- cgit v1.2.3-70-g09d2 From ed8ad83808f009ade97ebbf6519bc3a97fefbc0c Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 19 Feb 2016 00:18:25 -0500 Subject: ext4: fix bh->b_state corruption ext4 can update bh->b_state non-atomically in _ext4_get_block() and ext4_da_get_block_prep(). Usually this is fine since bh is just a temporary storage for mapping information on stack but in some cases it can be fully living bh attached to a page. In such case non-atomic update of bh->b_state can race with an atomic update which then gets lost. Usually when we are mapping bh and thus updating bh->b_state non-atomically, nobody else touches the bh and so things work out fine but there is one case to especially worry about: ext4_finish_bio() uses BH_Uptodate_Lock on the first bh in the page to synchronize handling of PageWriteback state. So when blocksize < pagesize, we can be atomically modifying bh->b_state of a buffer that actually isn't under IO and thus can race e.g. with delalloc trying to map that buffer. The result is that we can mistakenly set / clear BH_Uptodate_Lock bit resulting in the corruption of PageWriteback state or missed unlock of BH_Uptodate_Lock. Fix the problem by always updating bh->b_state bits atomically. CC: stable@vger.kernel.org Reported-by: Nikolay Borisov Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/ext4/inode.c | 32 ++++++++++++++++++++++++++++++-- 1 file changed, 30 insertions(+), 2 deletions(-) (limited to 'fs') diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index 83bc8bfb3bea..d6674479269d 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -686,6 +686,34 @@ out_sem: return retval; } +/* + * Update EXT4_MAP_FLAGS in bh->b_state. For buffer heads attached to pages + * we have to be careful as someone else may be manipulating b_state as well. + */ +static void ext4_update_bh_state(struct buffer_head *bh, unsigned long flags) +{ + unsigned long old_state; + unsigned long new_state; + + flags &= EXT4_MAP_FLAGS; + + /* Dummy buffer_head? Set non-atomically. */ + if (!bh->b_page) { + bh->b_state = (bh->b_state & ~EXT4_MAP_FLAGS) | flags; + return; + } + /* + * Someone else may be modifying b_state. Be careful! This is ugly but + * once we get rid of using bh as a container for mapping information + * to pass to / from get_block functions, this can go away. + */ + do { + old_state = READ_ONCE(bh->b_state); + new_state = (old_state & ~EXT4_MAP_FLAGS) | flags; + } while (unlikely( + cmpxchg(&bh->b_state, old_state, new_state) != old_state)); +} + /* Maximum number of blocks we map for direct IO at once. */ #define DIO_MAX_BLOCKS 4096 @@ -722,7 +750,7 @@ static int _ext4_get_block(struct inode *inode, sector_t iblock, ext4_io_end_t *io_end = ext4_inode_aio(inode); map_bh(bh, inode->i_sb, map.m_pblk); - bh->b_state = (bh->b_state & ~EXT4_MAP_FLAGS) | map.m_flags; + ext4_update_bh_state(bh, map.m_flags); if (io_end && io_end->flag & EXT4_IO_END_UNWRITTEN) set_buffer_defer_completion(bh); bh->b_size = inode->i_sb->s_blocksize * map.m_len; @@ -1685,7 +1713,7 @@ int ext4_da_get_block_prep(struct inode *inode, sector_t iblock, return ret; map_bh(bh, inode->i_sb, map.m_pblk); - bh->b_state = (bh->b_state & ~EXT4_MAP_FLAGS) | map.m_flags; + ext4_update_bh_state(bh, map.m_flags); if (buffer_unwritten(bh)) { /* A delayed write to unwritten bh should be marked -- cgit v1.2.3-70-g09d2 From 74dae4278546b897eb81784fdfcce872ddd8b2b8 Mon Sep 17 00:00:00 2001 From: Jan Kara Date: Fri, 19 Feb 2016 00:33:21 -0500 Subject: ext4: fix crashes in dioread_nolock mode Competing overwrite DIO in dioread_nolock mode will just overwrite pointer to io_end in the inode. This may result in data corruption or extent conversion happening from IO completion interrupt because we don't properly set buffer_defer_completion() when unlocked DIO races with locked DIO to unwritten extent. Since unlocked DIO doesn't need io_end for anything, just avoid allocating it and corrupting pointer from inode for locked DIO. A cleaner fix would be to avoid these games with io_end pointer from the inode but that requires more intrusive changes so we leave that for later. Cc: stable@vger.kernel.org Signed-off-by: Jan Kara Signed-off-by: Theodore Ts'o --- fs/ext4/inode.c | 40 ++++++++++++++++++++-------------------- 1 file changed, 20 insertions(+), 20 deletions(-) (limited to 'fs') diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index d6674479269d..9cc57c3b4661 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -3281,29 +3281,29 @@ static ssize_t ext4_ext_direct_IO(struct kiocb *iocb, struct iov_iter *iter, * case, we allocate an io_end structure to hook to the iocb. */ iocb->private = NULL; - ext4_inode_aio_set(inode, NULL); - if (!is_sync_kiocb(iocb)) { - io_end = ext4_init_io_end(inode, GFP_NOFS); - if (!io_end) { - ret = -ENOMEM; - goto retake_lock; - } - /* - * Grab reference for DIO. Will be dropped in ext4_end_io_dio() - */ - iocb->private = ext4_get_io_end(io_end); - /* - * we save the io structure for current async direct - * IO, so that later ext4_map_blocks() could flag the - * io structure whether there is a unwritten extents - * needs to be converted when IO is completed. - */ - ext4_inode_aio_set(inode, io_end); - } - if (overwrite) { get_block_func = ext4_get_block_overwrite; } else { + ext4_inode_aio_set(inode, NULL); + if (!is_sync_kiocb(iocb)) { + io_end = ext4_init_io_end(inode, GFP_NOFS); + if (!io_end) { + ret = -ENOMEM; + goto retake_lock; + } + /* + * Grab reference for DIO. Will be dropped in + * ext4_end_io_dio() + */ + iocb->private = ext4_get_io_end(io_end); + /* + * we save the io structure for current async direct + * IO, so that later ext4_map_blocks() could flag the + * io structure whether there is a unwritten extents + * needs to be converted when IO is completed. + */ + ext4_inode_aio_set(inode, io_end); + } get_block_func = ext4_get_block_write; dio_flags = DIO_LOCKING; } -- cgit v1.2.3-70-g09d2