It's possible that __ext4_new_inode will release the xattr block, so
it will trigger a warning since there is revoke credits will be 0 if
the handle == NULL. The below scripts can reproduce it easily.
------------[ cut here ]------------
WARNING: CPU: 0 PID: 3861 at fs/jbd2/revoke.c:374 jbd2_journal_revoke+0x30e/0x540 fs/jbd2/revoke.c:374
...
__ext4_forget+0x1d7/0x800 fs/ext4/ext4_jbd2.c:248
ext4_free_blocks+0x213/0x1d60 fs/ext4/mballoc.c:4743
ext4_xattr_release_block+0x55b/0x780 fs/ext4/xattr.c:1254
ext4_xattr_block_set+0x1c2c/0x2c40 fs/ext4/xattr.c:2112
ext4_xattr_set_handle+0xa7e/0x1090 fs/ext4/xattr.c:2384
__ext4_set_acl+0x54d/0x6c0 fs/ext4/acl.c:214
ext4_init_acl+0x218/0x2e0 fs/ext4/acl.c:293
__ext4_new_inode+0x352a/0x42b0 fs/ext4/ialloc.c:1151
ext4_mkdir+0x2e9/0xbd0 fs/ext4/namei.c:2774
vfs_mkdir+0x386/0x5f0 fs/namei.c:3811
do_mkdirat+0x11c/0x210 fs/namei.c:3834
do_syscall_64+0xa1/0x530 arch/x86/entry/common.c:294
...
-------------------------------------
scripts:
mkfs.ext4 /dev/vdb
mount /dev/vdb /mnt
cd /mnt && mkdir dir && for i in {1..8}; do setfacl -dm "u:user_"$i":rx" dir; done
mkdir dir/dir1 && mv dir/dir1 ./
sh repro.sh && add some user
[root@localhost ~]# cat repro.sh
while [ 1 -eq 1 ]; do
rm -rf dir
rm -rf dir1/dir1
mkdir dir
for i in {1..8}; do setfacl -dm "u:test"$i":rx" dir; done
setfacl -m "u:user_9:rx" dir &
mkdir dir1/dir1 &
done
Before exec repro.sh, dir1 has inherit the default acl from dir, and
xattr block of dir1 dir is not the same, so the h_refcount of these
two dir's xattr block will be 1. Then repro.sh can trigger the warning
with the situation show as below. The last h_refcount can be clear
with mkdir, and __ext4_new_inode has not reserved revoke credits, so
the warning will happened, fix it by reserve revoke credits in
__ext4_new_inode.
Thread 1 Thread 2
mkdir dir
set default acl(will create
a xattr block blk1 and the
refcount of ext4_xattr_header
will be 1)
...
mkdir dir1/dir1
->....->ext4_init_acl
->__ext4_set_acl(set default acl,
will reuse blk1, and h_refcount
will be 2)
setfacl->ext4_set_acl->...
->ext4_xattr_block_set(will create
new block blk2 to store xattr)
->__ext4_set_acl(set access acl, since
h_refcount of blk1 is 2, will create
blk3 to store xattr)
->ext4_xattr_release_block(dec
h_refcount of blk1 to 1)
->ext4_xattr_release_block(dec
h_refcount and since it is 0,
will release the block and trigger
the warning)
Link: https://lore.kernel.org/r/20191213014900.47228-1-yangerkun@huawei.com
Reported-by: Hulk Robot <hulkci@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: yangerkun <yangerkun@huawei.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Make __ext4_check_dir_entry() a bit easier to understand, and reduce
the object size of the function by over 11%.
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Link: https://lore.kernel.org/r/20191209004346.38526-1-tytso@mit.edu
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
ext4_check_dir_entry() currently does not catch a case when a directory
entry ends so close to the block end that the header of the next
directory entry would not fit in the remaining space. This can lead to
directory iteration code trying to access address beyond end of current
buffer head leading to oops.
CC: stable@vger.kernel.org
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191202170213.4761-3-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Function ext4_empty_dir() doesn't correctly handle directories with
holes and crashes on bh->b_data dereference when bh is NULL. Reorganize
the loop to use 'offset' variable all the times instead of comparing
pointers to current direntry with bh->b_data pointer. Also add more
strict checking of '.' and '..' directory entries to avoid entering loop
in possibly invalid state on corrupted filesystems.
References: CVE-2019-19037
CC: stable@vger.kernel.org
Fixes: 4e19d6b65f ("ext4: allow directory holes")
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191202170213.4761-2-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
If the file system is corrupted such that a file's i_links_count is
too small, then it's possible that when unlinking that file, i_nlink
will already be zero. Previously we were working around this kind of
corruption by forcing i_nlink to one; but we were doing this before
trying to delete the directory entry --- and if the file system is
corrupted enough that ext4_delete_entry() fails, then we exit with
i_nlink elevated, and this causes the orphan inode list handling to be
FUBAR'ed, such that when we unmount the file system, the orphan inode
list can get corrupted.
A better way to fix this is to simply skip trying to call drop_nlink()
if i_nlink is already zero, thus moving the check to the place where
it makes the most sense.
https://bugzilla.kernel.org/show_bug.cgi?id=205433
Link: https://lore.kernel.org/r/20191112032903.8828-1-tytso@mit.edu
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@kernel.org
Reviewed-by: Andreas Dilger <adilger@dilger.ca>
The helper jbd2_handle_buffer_credits() doesn't correctly handle reserved
handles which can lead to crashes. Fix it getting of journal pointer to
work for reserved handles as well.
Fixes: a9a8344ee1 ("ext4, jbd2: Provide accessor function for handle credits")
Reported-by: Eric Biggers <ebiggers@kernel.org>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191115102210.29445-1-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
No need to wait for any commit once the page is fully truncated.
Besides, it may confuse e.g. concurrent ext4_writepage() with the page
still be dirty (will be cleared by truncate_pagecache() in
ext4_setattr()) but buffers has been freed; and then trigger a bug
show as below:
[ 26.057508] ------------[ cut here ]------------
[ 26.058531] kernel BUG at fs/ext4/inode.c:2134!
...
[ 26.088130] Call trace:
[ 26.088695] ext4_writepage+0x914/0xb28
[ 26.089541] writeout.isra.4+0x1b4/0x2b8
[ 26.090409] move_to_new_page+0x3b0/0x568
[ 26.091338] __unmap_and_move+0x648/0x988
[ 26.092241] unmap_and_move+0x48c/0xbb8
[ 26.093096] migrate_pages+0x220/0xb28
[ 26.093945] kernel_mbind+0x828/0xa18
[ 26.094791] __arm64_sys_mbind+0xc8/0x138
[ 26.095716] el0_svc_common+0x190/0x490
[ 26.096571] el0_svc_handler+0x60/0xd0
[ 26.097423] el0_svc+0x8/0xc
Run the procedure (generate by syzkaller) parallel with ext3.
void main()
{
int fd, fd1, ret;
void *addr;
size_t length = 4096;
int flags;
off_t offset = 0;
char *str = "12345";
fd = open("a", O_RDWR | O_CREAT);
assert(fd >= 0);
/* Truncate to 4k */
ret = ftruncate(fd, length);
assert(ret == 0);
/* Journal data mode */
flags = 0xc00f;
ret = ioctl(fd, _IOW('f', 2, long), &flags);
assert(ret == 0);
/* Truncate to 0 */
fd1 = open("a", O_TRUNC | O_NOATIME);
assert(fd1 >= 0);
addr = mmap(NULL, length, PROT_WRITE | PROT_READ,
MAP_SHARED, fd, offset);
assert(addr != (void *)-1);
memcpy(addr, str, 5);
mbind(addr, length, 0, 0, 0, MPOL_MF_MOVE);
}
And the bug will be triggered once we seen the below order.
reproduce1 reproduce2
... | ...
truncate to 4k |
change to journal data mode |
| memcpy(set page dirty)
truncate to 0: |
ext4_setattr: |
... |
ext4_wait_for_tail_page_commit |
| mbind(trigger bug)
truncate_pagecache(clean dirty)| ...
... |
mbind will call ext4_writepage() since the page still be dirty, and then
report the bug since the buffers has been free. Fix it by return
directly once offset equals to 0 which means the page has been fully
truncated.
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: yangerkun <yangerkun@huawei.com>
Link: https://lore.kernel.org/r/20190919063508.1045-1-yangerkun@huawei.com
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Now the checks in ext4_get_next_id() and dquot_get_next_id()
are almost the same, so just call dquot_get_next_id() instead
of ext4_get_next_id().
Signed-off-by: Chengguang Xu <cgxu519@mykernel.net>
Link: https://lore.kernel.org/r/20191006103028.31299-1-cgxu519@mykernel.net
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Commit 8fcc3a5806 ("ext4: rework reserved cluster accounting when
invalidating pages") moved freeing of delayed allocation reservations
from dirty page invalidation time to time when we evict corresponding
status extent from extent status tree. For inodes which don't have any
blocks allocated this may actually happen only in ext4_clear_blocks()
which is after we've dropped references to quota structures from the
inode. Thus reservation of quota leaked. Fix the problem by clearing
quota information from the inode only after evicting extent status tree
in ext4_clear_inode().
Link: https://lore.kernel.org/r/20191108115420.GI20863@quack2.suse.cz
Reported-by: Konstantin Khlebnikov <khlebnikov@yandex-team.ru>
Fixes: 8fcc3a5806 ("ext4: rework reserved cluster accounting when invalidating pages")
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Commit c33fbe8f67 ("ext4: Enable blocksize < pagesize for
dioread_nolock") removed the only user of 'sbi' outside of the ifdef,
so it caused a new warning:
fs/ext4/super.c:2068:23: warning: unused variable 'sbi' [-Wunused-variable]
Fixes: c33fbe8f67 ("ext4: Enable blocksize < pagesize for dioread_nolock")
Signed-off-by: Olof Johansson <olof@lixom.net>
Link: https://lore.kernel.org/r/20191111022523.34256-1-olof@lixom.net
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Ritesh Harjani <riteshh@linux.ibm.com>
Now that we have the code to support encryption for subpage-sized
blocks, this commit removes the conditional check in filesystem mount
code.
The commit also changes the support statement in
Documentation/filesystems/fscrypt.rst to reflect the fact that
encryption on filesystems with blocksize less than page size now works.
[EB: Tested with 'gce-xfstests -c ext4/encrypt_1k -g auto', using the
new "encrypt_1k" config I created. All tests pass except for those that
already fail or are excluded with the encrypt or 1k configs, and 2 tests
that try to create 1023-byte symlinks which fails since encrypted
symlinks are limited to blocksize-3 bytes. Also ran the dedicated
encryption tests using 'kvm-xfstests -c ext4/1k -g encrypt'; all pass,
including the on-disk ciphertext verification tests.]
Signed-off-by: Chandan Rajendra <chandan@linux.ibm.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20191023033312.361355-3-ebiggers@kernel.org
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
After each filesystem block (as represented by a buffer_head) has been
read from disk by block_read_full_page(), decrypt it if needed. The
decryption is done on the fscrypt_read_workqueue.
This is the final change needed to support ext4 encryption with
blocksize != PAGE_SIZE, and it's a fairly small change now that
CONFIG_FS_ENCRYPTION is a bool and fs/crypto/ exposes functions to
decrypt individual blocks and to enqueue work on the fscrypt workqueue.
Don't try to add fs-verity support yet, as the fs/verity/ support layer
isn't ready for sub-page blocks yet. Just add fscrypt support for now.
Almost all the new code is compiled away when CONFIG_FS_ENCRYPTION=n.
Cc: Chandan Rajendra <chandan@linux.ibm.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Link: https://lore.kernel.org/r/20191023033312.361355-2-ebiggers@kernel.org
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
This patch adds the error handling in case of any memory allocation
failure for io_end_vec. This was missing in original
patch series which enables dioread_nolock for blocksize < pagesize.
Fixes: c8cc88163f ("ext4: Add support for blocksize < pagesize in dioread_nolock")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Link: https://lore.kernel.org/r/20191106093809.10673-1-riteshh@linux.ibm.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Currently we reserve j_max_transaction_buffers / 32 for transaction
descriptor blocks. Now that revoke descriptors are accounted for
separately this estimate is unnecessarily high and we can actually
compute much tighter estimate. In the common case of 32k journal blocks
and 4k blocksize this actually reduces the amount of reserved descriptor
blocks from 256 to ~25 which allows us to fit more real data into a
transaction.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-25-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Provide trace event for handle restarts to ease debugging.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-24-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
So far we have reserved only relatively high fixed amount of revoke
credits for each transaction. We over-reserved by large amount for most
cases but when freeing large directories or files with data journalling,
the fixed amount is not enough. In fact the worst case estimate is
inconveniently large (maximum extent size) for freeing of one extent.
We fix this by doing proper estimate of the amount of blocks that need
to be revoked when removing blocks from the inode due to truncate or
hole punching and otherwise reserve just a small amount of revoke
credits for each transaction to accommodate freeing of xattrs block or
so.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-23-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Make checking of available credits in jbd2_journal_dirty_metadata() more
strict. There should be always enough credits in the handle to write all
potential revoke descriptors. Also we warn in case there are not enough
credits since this is a bug in the filesystem.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-22-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The credit counter now contains both buffer and revoke descriptor block
credits. Rename to counter to h_total_credits to reflect that. No
functional change.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-21-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Extend functions for starting, extending, and restarting transaction
handles to take number of revoke records handle must be able to
accommodate. These functions then make sure transaction has enough
credits to be able to store resulting revoke descriptor blocks. Also
revoke code tracks number of revoke records created by a handle to catch
situation where some place didn't reserve enough space for revoke
records. Similarly to standard transaction credits, space for unused
reserved revoke records is released when the handle is stopped.
On the ext4 side we currently take a simplistic approach of reserving
space for 1024 revoke records for any transaction. This grows amount of
credits reserved for each handle only by a few and is enough for any
normal workload so that we don't hit warnings in jbd2. We will refine
the logic in following commits.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-20-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
The function is now just a trivial wrapper returning
journal->j_max_transaction_buffers. Drop it.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-19-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Currently, journal descriptor blocks were not accounted in
transaction->t_outstanding_credits and we were just leaving some slack
space in the journal for them (in jbd2_log_space_left() and
jbd2_space_needed()). This is making proper accounting (and reservation
we want to add) of descriptor blocks difficult so switch to accounting
descriptor blocks in transaction->t_outstanding_credits and just reserve
the same amount of credits in t_outstanding credits for journal
descriptor blocks when creating transaction.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-18-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
jbd2__journal_restart() has quite some code that is common with
jbd2_journal_stop(). Factor this functionality into stop_this_handle()
helper and use it from both functions. Note that this also drops
t_handle_lock protection from jbd2__journal_restart() as
jbd2_journal_stop() does the same thing without it.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-17-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
When we drop last handle from a transaction and journal->j_barrier_count
> 0, jbd2_journal_stop() wakes up journal->j_wait_transaction_locked
wait queue. This looks pointless - wait for outstanding handles always
happens on journal->j_wait_updates waitqueue.
journal->j_wait_transaction_locked is used to wait for transaction state
changes and by start_this_handle() for waiting until
journal->j_barrier_count drops to 0. The first case is clearly
irrelevant here since only jbd2 thread changes transaction state. The
second case looks related but jbd2_journal_unlock_updates() is
responsible for the wakeup in this case. So just drop the wakeup.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-16-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
If a transaction is larger than journal->j_max_transaction_buffers, that
is a bug and not a trigger for transaction commit. Also the very next
attempt to start new handle will start transaction commit anyway. So
just remove the pointless check. Arguably, we could start transaction
commit whenever the transaction size is *close* to
journal->j_max_transaction_buffers. This has a potential to reduce
latency of the next jbd2_journal_start() at the cost of somewhat smaller
transactions. However for this to have any effect, it would mean that
there isn't someone already waiting in jbd2_journal_start() which means
metadata load for the fs is pretty light anyway so probably this
optimization is not worth it.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-15-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Move code in jbd2_journal_stop() around a bit. It removes some
unnecessary code duplication and will make factoring out parts common
with jbd2__journal_restart() easier.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-14-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
jbd2 statistics counting number of blocks logged in a transaction was
wrong. It didn't count the commit block and more importantly it didn't
count revoke descriptor blocks. Make sure these get properly counted.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-13-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Use the jbd2 accessor function for h_buffer_credits.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-12-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Provide accessor function to get number of credits available in a handle
and use it from ext4. Later, computation of available credits won't be
so straightforward.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-11-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Provide ext4_journal_ensure_credits_fn() function to ensure transaction
has given amount of credits and call helper function to prepare for
restarting a transaction. This allows to remove some boilerplate code
from various places, add proper error handling for the case where
transaction extension or restart fails, and reduces following changes
needed for proper revoke record reservation tracking.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-10-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Error cleanup path in ext4_alloc_branch() calls ext4_forget() on freshly
allocated indirect blocks with 'metadata' set to 1. This results in
generating revoke records for these blocks. However this is unnecessary
as the freed blocks are only allocated in the current transaction and
thus they will never be journalled. Make this cleanup path similar to
e.g. cleanup in ext4_splice_branch() and use ext4_free_blocks() to
handle block forgetting by passing EXT4_FREE_BLOCKS_FORGET and not
EXT4_FREE_BLOCKS_METADATA to ext4_free_blocks(). This also allows
allocating transaction not to reserve any credits for revoke records.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-9-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Use ext4 helper ext4_journal_extend() instead of opencoding it in
ext4_try_to_expand_extra_isize().
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-8-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Similarly to directories, EA inodes do only journalled modifications to
their data. Change ext4_should_journal_data() to return true for them so
that we don't have to special-case them during truncate.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-7-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Estimate for the number of credits needed for final freeing of inode in
ext4_evict_inode() was to small. We may modify 4 blocks (inode & sb for
orphan deletion, bitmap & group descriptor for inode freeing) and not
just 3.
[ Fixed minor whitespace nit. -- TYT ]
Fixes: e50e5129f3 ("ext4: xattr-in-inode support")
CC: stable@vger.kernel.org
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-6-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
This patch introduces a new direct I/O write path which makes use of
the iomap infrastructure.
All direct I/O writes are now passed from the ->write_iter() callback
through to the new direct I/O handler ext4_dio_write_iter(). This
function is responsible for calling into the iomap infrastructure via
iomap_dio_rw().
Code snippets from the existing direct I/O write code within
ext4_file_write_iter() such as, checking whether the I/O request is
unaligned asynchronous I/O, or whether the write will result in an
overwrite have effectively been moved out and into the new direct I/O
->write_iter() handler.
The block mapping flags that are eventually passed down to
ext4_map_blocks() from the *_get_block_*() suite of routines have been
taken out and introduced within ext4_iomap_alloc().
For inode extension cases, ext4_handle_inode_extension() is
effectively the function responsible for performing such metadata
updates. This is called after iomap_dio_rw() has returned so that we
can safely determine whether we need to potentially truncate any
allocated blocks that may have been prepared for this direct I/O
write. We don't perform the inode extension, or truncate operations
from the ->end_io() handler as we don't have the original I/O 'length'
available there. The ->end_io() however is responsible fo converting
allocated unwritten extents to written extents.
In the instance of a short write, we fallback and complete the
remainder of the I/O using buffered I/O via
ext4_buffered_write_iter().
The existing buffer_head direct I/O implementation has been removed as
it's now redundant.
[ Fix up ext4_dio_write_iter() per Jan's comments at
https://lore.kernel.org/r/20191105135932.GN22379@quack2.suse.cz -- TYT ]
Signed-off-by: Matthew Bobrowski <mbobrowski@mbobrowski.org>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Ritesh Harjani <riteshh@linux.ibm.com>
Link: https://lore.kernel.org/r/e55db6f12ae6ff017f36774135e79f3e7b0333da.1572949325.git.mbobrowski@mbobrowski.org
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
When ext4_mkdir(), ext4_symlink(), ext4_create(), or ext4_mknod() fail
to add entry into directory, it ends up dropping freshly created inode
under the running transaction and thus inode truncation happens under
that transaction. That breaks assumptions that evict() does not get
called from a transaction context and at least in ext4_symlink() case it
can result in inode eviction deadlocking in inode_wait_for_writeback()
when flush worker finds symlink inode, starts to write it back and
blocks on starting a transaction. So change the code in ext4_mkdir() and
ext4_add_nondir() to drop inode reference only after the transaction is
stopped. We also have to add inode to the orphan list in that case as
otherwise the inode would get leaked in case we crash before inode
deletion is committed.
CC: stable@vger.kernel.org
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-5-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Every caller of ext4_add_nondir() marks handle as sync if directory has
DIRSYNC set. Move this marking to ext4_add_nondir() so reduce some
duplication.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-4-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
With 32-bit block numbers, we don't allocate the array for journal
buffer heads large enough for corresponding descriptor tags to fill the
descriptor block. Thus we end up writing out half-full descriptor blocks
to the journal unnecessarily growing the transaction. Fix the logic to
allocate the array large enough.
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-3-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
jbd2_journal_next_log_block() does not look at
transaction->t_outstanding_credits. Remove the misleading comment.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-2-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
When number of free space in the journal is very low, the arithmetic in
jbd2_log_space_left() could underflow resulting in very high number of
free blocks and thus triggering assertion failure in transaction commit
code complaining there's not enough space in the journal:
J_ASSERT(journal->j_free > 1);
Properly check for the low number of free blocks.
CC: stable@vger.kernel.org
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191105164437.32602-1-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
When the filesystem is created without a journal, we eventually call
into __generic_file_fsync() in order to write out all the modified
in-core data to the permanent storage device. This function happens to
try and obtain an inode_lock() while synchronizing the files buffer
and it's associated metadata.
Generally, this is fine, however it becomes a problem when there is
higher level code that has already obtained an inode_lock() as this
leads to a recursive lock situation. This case is especially true when
porting across direct I/O to iomap infrastructure as we obtain an
inode_lock() early on in the I/O within ext4_dio_write_iter() and hold
it until the I/O has been completed. Consequently, to not run into
this specific issue, we move away from calling into
__generic_file_fsync() and perform the necessary synchronization tasks
within ext4_sync_file().
Signed-off-by: Matthew Bobrowski <mbobrowski@mbobrowski.org>
Reviewed-by: Ritesh Harjani <riteshh@linux.ibm.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/3495f35ef67f2021b567e28e6f59222e583689b8.1572949325.git.mbobrowski@mbobrowski.org
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Lift the inode extension/orphan list handling code out from
ext4_iomap_alloc() and apply it within the ext4_dax_write_iter().
Signed-off-by: Matthew Bobrowski <mbobrowski@mbobrowski.org>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Ritesh Harjani <riteshh@linux.ibm.com>
Link: https://lore.kernel.org/r/fd5c84db25d5d0da87d97ed4c36fd844f57da759.1572949325.git.mbobrowski@mbobrowski.org
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
In preparation for implementing the iomap direct I/O modifications,
the inode extension/truncate code needs to be moved out from the
ext4_iomap_end() callback. For direct I/O, if the current code
remained, it would behave incorrrectly. Updating the inode size prior
to converting unwritten extents would potentially allow a racing
direct I/O read to find unwritten extents before being converted
correctly.
The inode extension/truncate code now resides within a new helper
ext4_handle_inode_extension(). This function has been designed so that
it can accommodate for both DAX and direct I/O extension/truncate
operations.
Signed-off-by: Matthew Bobrowski <mbobrowski@mbobrowski.org>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Ritesh Harjani <riteshh@linux.ibm.com>
Link: https://lore.kernel.org/r/d41ffa26e20b15b12895812c3cad7c91a6a59bc6.1572949325.git.mbobrowski@mbobrowski.org
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
This patch introduces a new direct I/O read path which makes use of
the iomap infrastructure.
The new function ext4_do_read_iter() is responsible for calling into
the iomap infrastructure via iomap_dio_rw(). If the read operation
performed on the inode is not supported, which is checked via
ext4_dio_supported(), then we simply fallback and complete the I/O
using buffered I/O.
Existing direct I/O read code path has been removed, as it is now
redundant.
Signed-off-by: Matthew Bobrowski <mbobrowski@mbobrowski.org>
Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: Ritesh Harjani <riteshh@linux.ibm.com>
Link: https://lore.kernel.org/r/f98a6f73fadddbfbad0fc5ed04f712ca0b799f37.1572949325.git.mbobrowski@mbobrowski.org
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
