We just need to make sure ext4_filemap_fault() doesn't block in the
speculative case as it is called with an rcu read lock held.
Signed-off-by: Michel Lespinasse <michel@lespinasse.org>
Link: https://lore.kernel.org/all/20210407014502.24091-32-michel@lespinasse.org/
Conflicts:
fs/ext4/inode.c
1. The change in fs/ext4/inode.c is not needed since i_mmap_sem is not
used anymore.
Bug: 161210518
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Change-Id: Idafc81074cf7f4b31985bdb24e0cc1597c91b875
* aosp/upstream-f2fs-stable-linux-5.15.y:
f2fs: do not allow partial truncation on pinned file
f2fs: remove redunant invalidate compress pages
f2fs: Simplify bool conversion
f2fs: don't drop compressed page cache in .{invalidate,release}page
f2fs: fix to reserve space for IO align feature
f2fs: fix to check available space of CP area correctly in update_ckpt_flags()
f2fs: support fault injection to f2fs_trylock_op()
f2fs: clean up __find_inline_xattr() with __find_xattr()
f2fs: fix to do sanity check on last xattr entry in __f2fs_setxattr()
f2fs: do not bother checkpoint by f2fs_get_node_info
f2fs: avoid down_write on nat_tree_lock during checkpoint
f2fs: compress: fix potential deadlock of compress file
f2fs: avoid EINVAL by SBI_NEED_FSCK when pinning a file
f2fs: add gc_urgent_high_remaining sysfs node
f2fs: fix to do sanity check in is_alive()
f2fs: fix to avoid panic in is_alive() if metadata is inconsistent
f2fs: fix to do sanity check on inode type during garbage collection
f2fs: avoid duplicate call of mark_inode_dirty
f2fs: support POSIX_FADV_DONTNEED drop compressed page cache
f2fs: fix remove page failed in invalidate compress pages
f2fs: show more DIO information in tracepoint
f2fs: use iomap for direct I/O
f2fs: implement iomap operations
f2fs: fix the f2fs_file_write_iter tracepoint
f2fs: do not expose unwritten blocks to user by DIO
f2fs: reduce indentation in f2fs_file_write_iter()
f2fs: rework write preallocations
f2fs: compress: reduce one page array alloc and free when write compressed page
iomap: Add done_before argument to iomap_dio_rw
f2fs: show number of pending discard commands
Bug: 207156594
Signed-off-by: Jaegeuk Kim <jaegeuk@google.com>
Change-Id: I8d23c9b29c71f881cca894e72fabf094339ba202
Add a done_before argument to iomap_dio_rw that indicates how much of
the request has already been transferred. When the request succeeds, we
report that done_before additional bytes were tranferred. This is
useful for finishing a request asynchronously when part of the request
has already been completed synchronously.
We'll use that to allow iomap_dio_rw to be used with page faults
disabled: when a page fault occurs while submitting a request, we
synchronously complete the part of the request that has already been
submitted. The caller can then take care of the page fault and call
iomap_dio_rw again for the rest of the request, passing in the number of
bytes already tranferred.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Pull ext4 updates from Ted Ts'o:
"In addition to some ext4 bug fixes and cleanups, this cycle we add the
orphan_file feature, which eliminates bottlenecks when doing a large
number of parallel truncates and file deletions, and move the discard
operation out of the jbd2 commit thread when using the discard mount
option, to better support devices with slow discard operations"
* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (23 commits)
ext4: make the updating inode data procedure atomic
ext4: remove an unnecessary if statement in __ext4_get_inode_loc()
ext4: move inode eio simulation behind io completeion
ext4: Improve scalability of ext4 orphan file handling
ext4: Orphan file documentation
ext4: Speedup ext4 orphan inode handling
ext4: Move orphan inode handling into a separate file
ext4: Support for checksumming from journal triggers
ext4: fix race writing to an inline_data file while its xattrs are changing
jbd2: add sparse annotations for add_transaction_credits()
ext4: fix sparse warnings
ext4: Make sure quota files are not grabbed accidentally
ext4: fix e2fsprogs checksum failure for mounted filesystem
ext4: if zeroout fails fall back to splitting the extent node
ext4: reduce arguments of ext4_fc_add_dentry_tlv
ext4: flush background discard kwork when retry allocation
ext4: get discard out of jbd2 commit kthread contex
ext4: remove the repeated comment of ext4_trim_all_free
ext4: add new helper interface ext4_try_to_trim_range()
ext4: remove the 'group' parameter of ext4_trim_extent
...
JBD2 layer support triggers which are called when journaling layer moves
buffer to a certain state. We can use the frozen trigger, which gets
called when buffer data is frozen and about to be written out to the
journal, to compute block checksums for some buffer types (similarly as
does ocfs2). This avoids unnecessary repeated recomputation of the
checksum (at the cost of larger window where memory corruption won't be
caught by checksumming) and is even necessary when there are
unsynchronized updaters of the checksummed data.
So add superblock and journal trigger type arguments to
ext4_journal_get_write_access() and ext4_journal_get_create_access() so
that frozen triggers can be set accordingly. Also add inode argument to
ext4_walk_page_buffers() and all the callbacks used with that function
for the same purpose. This patch is mostly only a change of prototype of
the above mentioned functions and a few small helpers. Real checksumming
will come later.
Reviewed-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20210816095713.16537-1-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Convert ext4 to use mapping->invalidate_lock instead of its private
EXT4_I(inode)->i_mmap_sem. This is mostly search-and-replace. By this
conversion we fix a long standing race between hole punching and read(2)
/ readahead(2) paths that can lead to stale page cache contents.
CC: <linux-ext4@vger.kernel.org>
CC: Ted Tso <tytso@mit.edu>
Acked-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Jan Kara <jack@suse.cz>
Pull ext4 updates from Ted Ts'o:
"New features for ext4 this cycle include support for encrypted
casefold, ensure that deleted file names are cleared in directory
blocks by zeroing directory entries when they are unlinked or moved as
part of a hash tree node split. We also improve the block allocator's
performance on a freshly mounted file system by prefetching block
bitmaps.
There are also the usual cleanups and bug fixes, including fixing a
page cache invalidation race when there is mixed buffered and direct
I/O and the block size is less than page size, and allow the dax flag
to be set and cleared on inline directories"
* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (32 commits)
ext4: wipe ext4_dir_entry2 upon file deletion
ext4: Fix occasional generic/418 failure
fs: fix reporting supported extra file attributes for statx()
ext4: allow the dax flag to be set and cleared on inline directories
ext4: fix debug format string warning
ext4: fix trailing whitespace
ext4: fix various seppling typos
ext4: fix error return code in ext4_fc_perform_commit()
ext4: annotate data race in jbd2_journal_dirty_metadata()
ext4: annotate data race in start_this_handle()
ext4: fix ext4_error_err save negative errno into superblock
ext4: fix error code in ext4_commit_super
ext4: always panic when errors=panic is specified
ext4: delete redundant uptodate check for buffer
ext4: do not set SB_ACTIVE in ext4_orphan_cleanup()
ext4: make prefetch_block_bitmaps default
ext4: add proc files to monitor new structures
ext4: improve cr 0 / cr 1 group scanning
ext4: add MB_NUM_ORDERS macro
ext4: add mballoc stats proc file
...
Eric has noticed that after pagecache read rework, generic/418 is
occasionally failing for ext4 when blocksize < pagesize. In fact, the
pagecache rework just made hard to hit race in ext4 more likely. The
problem is that since ext4 conversion of direct IO writes to iomap
framework (commit 378f32bab3), we update inode size after direct IO
write only after invalidating page cache. Thus if buffered read sneaks
at unfortunate moment like:
CPU1 - write at offset 1k CPU2 - read from offset 0
iomap_dio_rw(..., IOMAP_DIO_FORCE_WAIT);
ext4_readpage();
ext4_handle_inode_extension()
the read will zero out tail of the page as it still sees smaller inode
size and thus page cache becomes inconsistent with on-disk contents with
all the consequences.
Fix the problem by moving inode size update into end_io handler which
gets called before the page cache is invalidated.
Reported-and-tested-by: Eric Whitney <enwlinux@gmail.com>
Fixes: 378f32bab3 ("ext4: introduce direct I/O write using iomap infrastructure")
CC: stable@vger.kernel.org
Signed-off-by: Jan Kara <jack@suse.cz>
Acked-by: Dave Chinner <dchinner@redhat.com>
Link: https://lore.kernel.org/r/20210415155417.4734-1-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Use the fileattr API to let the VFS handle locking, permission checking and
conversion.
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Wire up ext4 with fscrypt direct I/O support. Direct I/O with fscrypt is
only supported through blk-crypto (i.e. CONFIG_BLK_INLINE_ENCRYPTION must
have been enabled, the 'inlinecrypt' mount option must have been specified,
and either hardware inline encryption support must be present or
CONFIG_BLK_INLINE_ENCYRPTION_FALLBACK must have been enabled). Further,
direct I/O on encrypted files is only supported when I/O is aligned
to the filesystem block size (which is *not* necessarily the same as the
block device's block size).
fscrypt_limit_io_blocks() is called before setting up the iomap to ensure
that the blocks of each bio that iomap will submit will have contiguous
DUNs. Note that fscrypt_limit_io_blocks() is normally a no-op, as normally
the DUNs simply increment along with the logical blocks. But it's needed
to handle an edge case in one of the fscrypt IV generation methods.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Co-developed-by: Satya Tangirala <satyat@google.com>
Signed-off-by: Satya Tangirala <satyat@google.com>
Reviewed-by: Jaegeuk Kim <jaegeuk@kernel.org>
Bug: 162255927
Link: https://lore.kernel.org/r/20200724184501.1651378-5-satyat@google.com
Change-Id: Ia3d869cefabdff070f4e77c46190351f6cb5d74c
Signed-off-by: Eric Biggers <ebiggers@google.com>
Revert the direct I/O support for encrypted files so that we can bring
in the latest version of the patches from the mailing list. This is
needed because in v5.5 and later, the ext4 support (via fs/iomap/) is
broken as-is -- not only is the second call to fscrypt_limit_dio_pages()
in the wrong place, but bios can exceed the intended nr_pages limit due
to multipage bvecs. In order to fix this we need the v6 patches which
make fs/ext4/ handle the limiting instead of fs/iomap/.
On android-mainline, this fixes a failure in vts_kernel_encryption_test
(specifically, FBEPolicyTest#TestAesEmmcOptimizedPolicy) when run on a
device that uses the inlinecrypt mount option on ext4 (e.g. db845c).
Bug: 162255927
Bug: 171462575
Change-Id: I0da753dc9e0e7bc8d84bbcadfdfcdb9328cdb8d8
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Satya Tangirala <satyat@google.com>
Pass a set of flags to iomap_dio_rw instead of the boolean
wait_for_completion argument. The IOMAP_DIO_FORCE_WAIT flag
replaces the wait_for_completion, but only needs to be passed
when the iocb isn't synchronous to start with to simplify the
callers.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
[djwong: rework xfs_file.c so that we can push iomap changes separately]
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
When the first file is opened, ext4 samples the mountpoint of the
filesystem in 64 bytes of the super block. It does so using
strlcpy(), this means that the remaining bytes in the super block
string buffer are untouched. If the mount point before had a longer
path than the current one, it can be reconstructed.
Consider the case where the fs was mounted to "/media/johnjdeveloper"
and later to "/". The super block buffer then contains
"/\x00edia/johnjdeveloper".
This case was seen in the wild and caused confusion how the name
of a developer ands up on the super block of a filesystem used
in production...
Fix this by using strncpy() instead of strlcpy(). The superblock
field is defined to be a fixed-size char array, and it is already
marked using __nonstring in fs/ext4/ext4.h. The consumer of the field
in e2fsprogs already assumes that in the case of a 64+ byte mount
path, that s_last_mounted will not be NUL terminated.
Link: https://lore.kernel.org/r/X9ujIOJG/HqMr88R@mit.edu
Reported-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: stable@kernel.org
Protect all superblock modifications (including checksum computation)
with a superblock buffer lock. That way we are sure computed checksum
matches current superblock contents (a mismatch could cause checksum
failures in nojournal mode or if an unjournalled superblock update races
with a journalled one). Also we avoid modifying superblock contents
while it is being written out (which can cause DIF/DIX failures if we
are running in nojournal mode).
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20201216101844.22917-4-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Steps on the way to 5.10-rc1
Resolves conflicts in:
fs/ext4/dir.c
fs/ext4/ext4.h
fs/ext4/namei.c
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: I397787046920175eb183fa5342d2923f819bb2f4
This patch adds main fast commit commit path handlers. The overall
patch can be divided into two inter-related parts:
(A) Metadata updates tracking
This part consists of helper functions to track changes that need
to be committed during a commit operation. These updates are
maintained by Ext4 in different in-memory queues. Following are
the APIs and their short description that are implemented in this
patch:
- ext4_fc_track_link/unlink/creat() - Track unlink. link and creat
operations
- ext4_fc_track_range() - Track changed logical block offsets
inodes
- ext4_fc_track_inode() - Track inodes
- ext4_fc_mark_ineligible() - Mark file system fast commit
ineligible()
- ext4_fc_start_update() / ext4_fc_stop_update() /
ext4_fc_start_ineligible() / ext4_fc_stop_ineligible() These
functions are useful for co-ordinating inode updates with
commits.
(B) Main commit Path
This part consists of functions to convert updates tracked in
in-memory data structures into on-disk commits. Function
ext4_fc_commit() is the main entry point to commit path.
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Harshad Shirwadkar <harshadshirwadkar@gmail.com>
Link: https://lore.kernel.org/r/20201015203802.3597742-6-harshadshirwadkar@gmail.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Pull ext4 updates from Ted Ts'o:
"Improvements to ext4's block allocator performance for very large file
systems, especially when the file system or files which are highly
fragmented. There is a new mount option, prefetch_block_bitmaps which
will pull in the block bitmaps and set up the in-memory buddy bitmaps
when the file system is initially mounted.
Beyond that, a lot of bug fixes and cleanups. In particular, a number
of changes to make ext4 more robust in the face of write errors or
file system corruptions"
* tag 'ext4_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4: (46 commits)
ext4: limit the length of per-inode prealloc list
ext4: reorganize if statement of ext4_mb_release_context()
ext4: add mb_debug logging when there are lost chunks
ext4: Fix comment typo "the the".
jbd2: clean up checksum verification in do_one_pass()
ext4: change to use fallthrough macro
ext4: remove unused parameter of ext4_generic_delete_entry function
mballoc: replace seq_printf with seq_puts
ext4: optimize the implementation of ext4_mb_good_group()
ext4: delete invalid comments near ext4_mb_check_limits()
ext4: fix typos in ext4_mb_regular_allocator() comment
ext4: fix checking of directory entry validity for inline directories
fs: prevent BUG_ON in submit_bh_wbc()
ext4: correctly restore system zone info when remount fails
ext4: handle add_system_zone() failure in ext4_setup_system_zone()
ext4: fold ext4_data_block_valid_rcu() into the caller
ext4: check journal inode extents more carefully
ext4: don't allow overlapping system zones
ext4: handle error of ext4_setup_system_zone() on remount
ext4: delete the invalid BUGON in ext4_mb_load_buddy_gfp()
...
In the scenario of writing sparse files, the per-inode prealloc list may
be very long, resulting in high overhead for ext4_mb_use_preallocated().
To circumvent this problem, we limit the maximum length of per-inode
prealloc list to 512 and allow users to modify it.
After patching, we observed that the sys ratio of cpu has dropped, and
the system throughput has increased significantly. We created a process
to write the sparse file, and the running time of the process on the
fixed kernel was significantly reduced, as follows:
Running time on unfixed kernel：
[root@TENCENT64 ~]# time taskset 0x01 ./sparse /data1/sparce.dat
real 0m2.051s
user 0m0.008s
sys 0m2.026s
Running time on fixed kernel：
[root@TENCENT64 ~]# time taskset 0x01 ./sparse /data1/sparce.dat
real 0m0.471s
user 0m0.004s
sys 0m0.395s
Signed-off-by: Chunguang Xu <brookxu@tencent.com>
Link: https://lore.kernel.org/r/d7a98178-056b-6db5-6bce-4ead23f4a257@gmail.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Running with some debug patches to detect illegal blocking triggered the
extend/unaligned condition in ext4. If ext4 needs to extend the file (and
hence go to buffered IO), or if the app is doing unaligned IO, then ext4
asks the iomap code to wait for IO completion. If the caller asked for
no-wait semantics by setting IOCB_NOWAIT, then ext4 should return -EAGAIN
instead.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Link: https://lore.kernel.org/r/76152096-2bbb-7682-8fce-4cb498bcd909@kernel.dk
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
ext4_mark_inode_dirty() can fail for real reasons. Ignoring its return
value may lead ext4 to ignore real failures that would result in
corruption / crashes. Harden ext4_mark_inode_dirty error paths to fail
as soon as possible and return errors to the caller whenever
appropriate.
One of the possible scnearios when this bug could affected is that
while creating a new inode, its directory entry gets added
successfully but while writing the inode itself mark_inode_dirty
returns error which is ignored. This would result in inconsistency
that the directory entry points to a non-existent inode.
Ran gce-xfstests smoke tests and verified that there were no
regressions.
Signed-off-by: Harshad Shirwadkar <harshadshirwadkar@gmail.com>
Link: https://lore.kernel.org/r/20200427013438.219117-1-harshadshirwadkar@gmail.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
With the existing fscrypt IV generation methods, each file's data blocks
have contiguous DUNs. Therefore the direct I/O code "just worked"
because it only submits logically contiguous bios. But with
IV_INO_LBLK_32, the direct I/O code breaks because the DUN can wrap from
0xffffffff to 0. We can't submit bios across such boundaries.
This is especially difficult to handle when block_size != PAGE_SIZE,
since in that case the DUN can wrap in the middle of a page. Punt on
this case for now and just handle block_size == PAGE_SIZE.
Add and use a new function fscrypt_dio_supported() to check whether a
direct I/O request is unsupported due to encryption constraints.
Then, update fs/direct-io.c (used by f2fs, and by ext4 in kernel v5.4
and earlier) and fs/iomap/direct-io.c (used by ext4 in kernel v5.5 and
later) to avoid submitting I/O across a DUN discontinuity.
(This is needed in ACK now because ACK already supports direct I/O with
inline crypto. I'll be sending this upstream along with the encrypted
direct I/O support itself once its prerequisites are closer to landing.)
Test: For now, just manually tested direct I/O on ext4 and f2fs in the
DUN discontinuity case.
Bug: 144046242
Change-Id: I0c0b0b20a73ade35c3660cc6f9c09d49d3853ba5
Signed-off-by: Eric Biggers <ebiggers@google.com>
git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4") into
android-mainline
Handle the ext4 merge issues in one small merge to make it more obvious.
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Change-Id: I988ab727a579da9475cdf907032e233a403b6fa1
ext4 and f2fs have traditionally not supported direct I/O on encrypted
files, since it's difficult to implement with the traditional
filesystem-layer encryption. But when inline encryption is used
instead, it's straightforward to support direct I/O, as long as the I/O
is fully filesystem-block-aligned. Add support for it by:
- Making the two generic direct I/O implementations in the kernel,
__blockdev_direct_IO() and iomap_dio_rw(), set the encryption context
on bios for inline-encrypted files. __blockdev_direct_IO() is used by
f2fs, and was used by ext4 in kernel v5.4 and earlier. iomap_dio_rw()
is used by ext4 in kernel v5.5 and later.
- Making ext4 and f2fs allow direct I/O to encrypted files (rather the
current behavior of falling back to buffered I/O) when the file is
using inline encryption and the I/O is fully filesystem-block-aligned.
Bug: 137270441
Change-Id: I4c8f7497eb8f829d03611d24281113d68c21d4d1
Signed-off-by: Eric Biggers <ebiggers@google.com>
Currently we start transaction for mapping every extent for writing
using direct IO. This is unnecessary when we know we are overwriting
already allocated blocks and the overhead of starting a transaction can
be significant especially for multithreaded workloads doing small writes.
Use iomap operations that avoid starting a transaction for direct IO
overwrites.
This improves throughput of 4k random writes - fio jobfile:
[global]
rw=randrw
norandommap=1
invalidate=0
bs=4k
numjobs=16
time_based=1
ramp_time=30
runtime=120
group_reporting=1
ioengine=psync
direct=1
size=16G
filename=file1.0.0:file1.0.1:file1.0.2:file1.0.3:file1.0.4:file1.0.5:file1.0.6:file1.0.7:file1.0.8:file1.0.9:file1.0.10:file1.0.11:file1.0.12:file1.0.13:file1.0.14:file1.0.15:file1.0.16:file1.0.17:file1.0.18:file1.0.19:file1.0.20:file1.0.21:file1.0.22:file1.0.23:file1.0.24:file1.0.25:file1.0.26:file1.0.27:file1.0.28:file1.0.29:file1.0.30:file1.0.31
file_service_type=random
nrfiles=32
from 3018MB/s to 4059MB/s in my test VM running test against simulated
pmem device (note that before iomap conversion, this workload was able
to achieve 3708MB/s because old direct IO path avoided transaction start
for overwrites as well). For dax, the win is even larger improving
throughput from 3042MB/s to 4311MB/s.
Reported-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20191218174433.19380-1-jack@suse.cz
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
We were using shared locking only in case of dioread_nolock mount option in case
of DIO overwrites. This mount condition is not needed anymore with current code,
since:-
1. No race between buffered writes & DIO overwrites. Since buffIO writes takes
exclusive lock & DIO overwrites will take shared locking. Also DIO path will
make sure to flush and wait for any dirty page cache data.
2. No race between buffered reads & DIO overwrites, since there is no block
allocation that is possible with DIO overwrites. So no stale data exposure
should happen. Same is the case between DIO reads & DIO overwrites.
3. Also other paths like truncate is protected, since we wait there for any DIO
in flight to be over.
Reviewed-by: Jan Kara <jack@suse.cz>
Tested-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Link: https://lore.kernel.org/r/20191212055557.11151-4-riteshh@linux.ibm.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Earlier there was no shared lock in DIO read path. But this patch
(16c5468859: ext4: Allow parallel DIO reads)
simplified some of the locking mechanism while still allowing for parallel DIO
reads by adding shared lock in inode DIO read path.
But this created problem with mixed read/write workload. It is due to the fact
that in DIO path, we first start with exclusive lock and only when we determine
that it is a ovewrite IO, we downgrade the lock. This causes the problem, since
we still have shared locking in DIO reads.
So, this patch tries to fix this issue by starting with shared lock and then
switching to exclusive lock only when required based on ext4_dio_write_checks().
Other than that, it also simplifies below cases:-
1. Simplified ext4_unaligned_aio API to ext4_unaligned_io. Previous API was
abused in the sense that it was not really checking for AIO anywhere also it
used to check for extending writes. So this API was renamed and simplified to
ext4_unaligned_io() which actully only checks if the IO is really unaligned.
Now, in case of unaligned direct IO, iomap_dio_rw needs to do zeroing of partial
block and that will require serialization against other direct IOs in the same
block. So we take a exclusive inode lock for any unaligned DIO. In case of AIO
we also need to wait for any outstanding IOs to complete so that conversion from
unwritten to written is completed before anyone try to map the overlapping block.
Hence we take exclusive inode lock and also wait for inode_dio_wait() for
unaligned DIO case. Please note since we are anyway taking an exclusive lock in
unaligned IO, inode_dio_wait() becomes a no-op in case of non-AIO DIO.
2. Added ext4_extending_io(). This checks if the IO is extending the file.
3. Added ext4_dio_write_checks(). In this we start with shared inode lock and
only switch to exclusive lock if required. So in most cases with aligned,
non-extending, dioread_nolock & overwrites, it tries to write with a shared
lock. If not, then we restart the operation in ext4_dio_write_checks(), after
acquiring exclusive lock.
Reviewed-by: Jan Kara <jack@suse.cz>
Tested-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Signed-off-by: Ritesh Harjani <riteshh@linux.ibm.com>
Link: https://lore.kernel.org/r/20191212055557.11151-3-riteshh@linux.ibm.com
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>
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>
