License cleanup: add SPDX GPL-2.0 license identifier to files with no license
Many source files in the tree are missing licensing information, which
makes it harder for compliance tools to determine the correct license.
By default all files without license information are under the default
license of the kernel, which is GPL version 2.
Update the files which contain no license information with the 'GPL-2.0'
SPDX license identifier. The SPDX identifier is a legally binding
shorthand, which can be used instead of the full boiler plate text.
This patch is based on work done by Thomas Gleixner and Kate Stewart and
Philippe Ombredanne.
How this work was done:
Patches were generated and checked against linux-4.14-rc6 for a subset of
the use cases:
- file had no licensing information it it.
- file was a */uapi/* one with no licensing information in it,
- file was a */uapi/* one with existing licensing information,
Further patches will be generated in subsequent months to fix up cases
where non-standard license headers were used, and references to license
had to be inferred by heuristics based on keywords.
The analysis to determine which SPDX License Identifier to be applied to
a file was done in a spreadsheet of side by side results from of the
output of two independent scanners (ScanCode & Windriver) producing SPDX
tag:value files created by Philippe Ombredanne. Philippe prepared the
base worksheet, and did an initial spot review of a few 1000 files.
The 4.13 kernel was the starting point of the analysis with 60,537 files
assessed. Kate Stewart did a file by file comparison of the scanner
results in the spreadsheet to determine which SPDX license identifier(s)
to be applied to the file. She confirmed any determination that was not
immediately clear with lawyers working with the Linux Foundation.
Criteria used to select files for SPDX license identifier tagging was:
- Files considered eligible had to be source code files.
- Make and config files were included as candidates if they contained >5
lines of source
- File already had some variant of a license header in it (even if <5
lines).
All documentation files were explicitly excluded.
The following heuristics were used to determine which SPDX license
identifiers to apply.
- when both scanners couldn't find any license traces, file was
considered to have no license information in it, and the top level
COPYING file license applied.
For non */uapi/* files that summary was:
SPDX license identifier # files
---------------------------------------------------|-------
GPL-2.0 11139
and resulted in the first patch in this series.
If that file was a */uapi/* path one, it was "GPL-2.0 WITH
Linux-syscall-note" otherwise it was "GPL-2.0". Results of that was:
SPDX license identifier # files
---------------------------------------------------|-------
GPL-2.0 WITH Linux-syscall-note 930
and resulted in the second patch in this series.
- if a file had some form of licensing information in it, and was one
of the */uapi/* ones, it was denoted with the Linux-syscall-note if
any GPL family license was found in the file or had no licensing in
it (per prior point). Results summary:
SPDX license identifier # files
---------------------------------------------------|------
GPL-2.0 WITH Linux-syscall-note 270
GPL-2.0+ WITH Linux-syscall-note 169
((GPL-2.0 WITH Linux-syscall-note) OR BSD-2-Clause) 21
((GPL-2.0 WITH Linux-syscall-note) OR BSD-3-Clause) 17
LGPL-2.1+ WITH Linux-syscall-note 15
GPL-1.0+ WITH Linux-syscall-note 14
((GPL-2.0+ WITH Linux-syscall-note) OR BSD-3-Clause) 5
LGPL-2.0+ WITH Linux-syscall-note 4
LGPL-2.1 WITH Linux-syscall-note 3
((GPL-2.0 WITH Linux-syscall-note) OR MIT) 3
((GPL-2.0 WITH Linux-syscall-note) AND MIT) 1
and that resulted in the third patch in this series.
- when the two scanners agreed on the detected license(s), that became
the concluded license(s).
- when there was disagreement between the two scanners (one detected a
license but the other didn't, or they both detected different
licenses) a manual inspection of the file occurred.
- In most cases a manual inspection of the information in the file
resulted in a clear resolution of the license that should apply (and
which scanner probably needed to revisit its heuristics).
- When it was not immediately clear, the license identifier was
confirmed with lawyers working with the Linux Foundation.
- If there was any question as to the appropriate license identifier,
the file was flagged for further research and to be revisited later
in time.
In total, over 70 hours of logged manual review was done on the
spreadsheet to determine the SPDX license identifiers to apply to the
source files by Kate, Philippe, Thomas and, in some cases, confirmation
by lawyers working with the Linux Foundation.
Kate also obtained a third independent scan of the 4.13 code base from
FOSSology, and compared selected files where the other two scanners
disagreed against that SPDX file, to see if there was new insights. The
Windriver scanner is based on an older version of FOSSology in part, so
they are related.
Thomas did random spot checks in about 500 files from the spreadsheets
for the uapi headers and agreed with SPDX license identifier in the
files he inspected. For the non-uapi files Thomas did random spot checks
in about 15000 files.
In initial set of patches against 4.14-rc6, 3 files were found to have
copy/paste license identifier errors, and have been fixed to reflect the
correct identifier.
Additionally Philippe spent 10 hours this week doing a detailed manual
inspection and review of the 12,461 patched files from the initial patch
version early this week with:
- a full scancode scan run, collecting the matched texts, detected
license ids and scores
- reviewing anything where there was a license detected (about 500+
files) to ensure that the applied SPDX license was correct
- reviewing anything where there was no detection but the patch license
was not GPL-2.0 WITH Linux-syscall-note to ensure that the applied
SPDX license was correct
This produced a worksheet with 20 files needing minor correction. This
worksheet was then exported into 3 different .csv files for the
different types of files to be modified.
These .csv files were then reviewed by Greg. Thomas wrote a script to
parse the csv files and add the proper SPDX tag to the file, in the
format that the file expected. This script was further refined by Greg
based on the output to detect more types of files automatically and to
distinguish between header and source .c files (which need different
comment types.) Finally Greg ran the script using the .csv files to
generate the patches.
Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org>
Reviewed-by: Philippe Ombredanne <pombredanne@nexb.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2017-11-01 15:07:57 +01:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
2006-10-11 01:20:53 -07:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2006-10-11 01:20:53 -07:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-10-01 17:58:54 -04:00
|
|
|
|
2009-06-13 10:09:48 -04:00
|
|
|
|
|
|
|
|
|
2015-09-08 14:58:40 -07:00
|
|
|
|
2026-01-08 12:13:01 -05:00
|
|
|
|
2010-03-03 09:05:07 -05:00
|
|
|
|
2015-02-22 08:58:50 -08:00
|
|
|
|
2017-11-01 16:36:45 +01:00
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
2008-04-29 18:13:32 -04:00
|
|
|
|
|
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
2022-08-26 23:58:47 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:37 +11:00
|
|
|
|
ext4: support direct I/O with fscrypt using blk-crypto
Encrypted files traditionally haven't supported DIO, due to the need to
encrypt/decrypt the data. However, when the encryption is implemented
using inline encryption (blk-crypto) instead of the traditional
filesystem-layer encryption, it is straightforward to support DIO.
Therefore, make ext4 support DIO on files that are using inline
encryption. Since ext4 uses iomap for DIO, and fscrypt support was
already added to iomap DIO, this just requires two small changes:
- Let DIO proceed when supported, by checking fscrypt_dio_supported()
instead of assuming that encrypted files never support DIO.
- In ext4_iomap_begin(), use fscrypt_limit_io_blocks() to limit the
length of the mapping in the rare case where a DUN discontiguity
occurs in the middle of an extent. The iomap DIO implementation
requires this, since it assumes that it can submit a bio covering (up
to) the whole mapping, without checking fscrypt constraints itself.
Co-developed-by: Satya Tangirala <satyat@google.com>
Signed-off-by: Satya Tangirala <satyat@google.com>
Acked-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Jaegeuk Kim <jaegeuk@kernel.org>
Link: https://lore.kernel.org/r/20220128233940.79464-4-ebiggers@kernel.org
Signed-off-by: Eric Biggers <ebiggers@google.com>
2022-01-28 15:39:38 -08:00
|
|
|
|
2022-08-26 23:58:47 -07:00
|
|
|
|
ext4: support direct I/O with fscrypt using blk-crypto
Encrypted files traditionally haven't supported DIO, due to the need to
encrypt/decrypt the data. However, when the encryption is implemented
using inline encryption (blk-crypto) instead of the traditional
filesystem-layer encryption, it is straightforward to support DIO.
Therefore, make ext4 support DIO on files that are using inline
encryption. Since ext4 uses iomap for DIO, and fscrypt support was
already added to iomap DIO, this just requires two small changes:
- Let DIO proceed when supported, by checking fscrypt_dio_supported()
instead of assuming that encrypted files never support DIO.
- In ext4_iomap_begin(), use fscrypt_limit_io_blocks() to limit the
length of the mapping in the rare case where a DUN discontiguity
occurs in the middle of an extent. The iomap DIO implementation
requires this, since it assumes that it can submit a bio covering (up
to) the whole mapping, without checking fscrypt constraints itself.
Co-developed-by: Satya Tangirala <satyat@google.com>
Signed-off-by: Satya Tangirala <satyat@google.com>
Acked-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Jaegeuk Kim <jaegeuk@kernel.org>
Link: https://lore.kernel.org/r/20220128233940.79464-4-ebiggers@kernel.org
Signed-off-by: Eric Biggers <ebiggers@google.com>
2022-01-28 15:39:38 -08:00
|
|
|
|
2022-08-26 23:58:47 -07:00
|
|
|
|
2019-11-05 23:01:37 +11:00
|
|
|
|
2022-08-26 23:58:47 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:37 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-08-26 23:58:47 -07:00
|
|
|
|
2019-11-05 23:01:37 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2026-07-29 12:27:18 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:37 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-20 17:36:06 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-12-12 11:25:55 +05:30
|
|
|
|
|
|
|
|
|
2017-06-20 07:05:47 -05:00
|
|
|
|
2019-12-12 11:25:55 +05:30
|
|
|
|
2017-06-20 07:05:47 -05:00
|
|
|
|
|
|
|
|
|
2016-11-20 17:36:06 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:37 +11:00
|
|
|
|
|
|
|
|
|
2023-06-16 18:50:49 +02:00
|
|
|
|
2017-02-05 01:28:48 -05:00
|
|
|
|
|
|
|
|
|
2016-11-20 17:36:06 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:37 +11:00
|
|
|
|
2016-11-20 17:36:06 -05:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:37 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-20 17:36:06 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-05-22 14:50:05 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-06-16 18:50:49 +02:00
|
|
|
|
2023-05-22 14:50:05 +01:00
|
|
|
|
2023-05-22 14:50:15 +01:00
|
|
|
|
2023-05-22 14:50:05 +01:00
|
|
|
|
|
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
|
|
|
|
|
2006-10-11 01:20:53 -07:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
|
|
|
|
|
2008-09-08 22:25:24 -04:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
2010-01-24 14:34:07 -05:00
|
|
|
|
2009-02-24 08:21:14 -05:00
|
|
|
|
2010-01-24 14:34:07 -05:00
|
|
|
|
2009-02-24 08:21:14 -05:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
|
|
|
|
|
2009-03-27 22:36:43 -04:00
|
|
|
|
2020-06-14 11:45:44 +07:00
|
|
|
|
2008-01-28 23:58:26 -05:00
|
|
|
|
2024-01-05 17:21:01 +08:00
|
|
|
|
2008-01-28 23:58:26 -05:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
|
|
|
|
|
2006-10-11 01:20:53 -07:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: serialize unaligned asynchronous DIO
ext4 has a data corruption case when doing non-block-aligned
asynchronous direct IO into a sparse file, as demonstrated
by xfstest 240.
The root cause is that while ext4 preallocates space in the
hole, mappings of that space still look "new" and
dio_zero_block() will zero out the unwritten portions. When
more than one AIO thread is going, they both find this "new"
block and race to zero out their portion; this is uncoordinated
and causes data corruption.
Dave Chinner fixed this for xfs by simply serializing all
unaligned asynchronous direct IO. I've done the same here.
The difference is that we only wait on conversions, not all IO.
This is a very big hammer, and I'm not very pleased with
stuffing this into ext4_file_write(). But since ext4 is
DIO_LOCKING, we need to serialize it at this high level.
I tried to move this into ext4_ext_direct_IO, but by then
we have the i_mutex already, and we will wait on the
work queue to do conversions - which must also take the
i_mutex. So that won't work.
This was originally exposed by qemu-kvm installing to
a raw disk image with a normal sector-63 alignment. I've
tested a backport of this patch with qemu, and it does
avoid the corruption. It is also quite a lot slower
(14 min for package installs, vs. 8 min for well-aligned)
but I'll take slow correctness over fast corruption any day.
Mingming suggested that we can track outstanding
conversions, and wait on those so that non-sparse
files won't be affected, and I've implemented that here;
unaligned AIO to nonsparse files won't take a perf hit.
[tytso@mit.edu: Keep the mutex as a hashed array instead
of bloating the ext4 inode]
[tytso@mit.edu: Fix up namespace issues so that global
variables are protected with an "ext4_" prefix.]
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-02-12 08:17:34 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
ext4: serialize unaligned asynchronous DIO
ext4 has a data corruption case when doing non-block-aligned
asynchronous direct IO into a sparse file, as demonstrated
by xfstest 240.
The root cause is that while ext4 preallocates space in the
hole, mappings of that space still look "new" and
dio_zero_block() will zero out the unwritten portions. When
more than one AIO thread is going, they both find this "new"
block and race to zero out their portion; this is uncoordinated
and causes data corruption.
Dave Chinner fixed this for xfs by simply serializing all
unaligned asynchronous direct IO. I've done the same here.
The difference is that we only wait on conversions, not all IO.
This is a very big hammer, and I'm not very pleased with
stuffing this into ext4_file_write(). But since ext4 is
DIO_LOCKING, we need to serialize it at this high level.
I tried to move this into ext4_ext_direct_IO, but by then
we have the i_mutex already, and we will wait on the
work queue to do conversions - which must also take the
i_mutex. So that won't work.
This was originally exposed by qemu-kvm installing to
a raw disk image with a normal sector-63 alignment. I've
tested a backport of this patch with qemu, and it does
avoid the corruption. It is also quite a lot slower
(14 min for package installs, vs. 8 min for well-aligned)
but I'll take slow correctness over fast corruption any day.
Mingming suggested that we can track outstanding
conversions, and wait on those so that non-sparse
files won't be affected, and I've implemented that here;
unaligned AIO to nonsparse files won't take a perf hit.
[tytso@mit.edu: Keep the mutex as a hashed array instead
of bloating the ext4 inode]
[tytso@mit.edu: Fix up namespace issues so that global
variables are protected with an "ext4_" prefix.]
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-02-12 08:17:34 -05:00
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
ext4: serialize unaligned asynchronous DIO
ext4 has a data corruption case when doing non-block-aligned
asynchronous direct IO into a sparse file, as demonstrated
by xfstest 240.
The root cause is that while ext4 preallocates space in the
hole, mappings of that space still look "new" and
dio_zero_block() will zero out the unwritten portions. When
more than one AIO thread is going, they both find this "new"
block and race to zero out their portion; this is uncoordinated
and causes data corruption.
Dave Chinner fixed this for xfs by simply serializing all
unaligned asynchronous direct IO. I've done the same here.
The difference is that we only wait on conversions, not all IO.
This is a very big hammer, and I'm not very pleased with
stuffing this into ext4_file_write(). But since ext4 is
DIO_LOCKING, we need to serialize it at this high level.
I tried to move this into ext4_ext_direct_IO, but by then
we have the i_mutex already, and we will wait on the
work queue to do conversions - which must also take the
i_mutex. So that won't work.
This was originally exposed by qemu-kvm installing to
a raw disk image with a normal sector-63 alignment. I've
tested a backport of this patch with qemu, and it does
avoid the corruption. It is also quite a lot slower
(14 min for package installs, vs. 8 min for well-aligned)
but I'll take slow correctness over fast corruption any day.
Mingming suggested that we can track outstanding
conversions, and wait on those so that non-sparse
files won't be affected, and I've implemented that here;
unaligned AIO to nonsparse files won't take a perf hit.
[tytso@mit.edu: Keep the mutex as a hashed array instead
of bloating the ext4 inode]
[tytso@mit.edu: Fix up namespace issues so that global
variables are protected with an "ext4_" prefix.]
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-02-12 08:17:34 -05:00
|
|
|
|
2014-04-17 16:09:22 -04:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
ext4: serialize unaligned asynchronous DIO
ext4 has a data corruption case when doing non-block-aligned
asynchronous direct IO into a sparse file, as demonstrated
by xfstest 240.
The root cause is that while ext4 preallocates space in the
hole, mappings of that space still look "new" and
dio_zero_block() will zero out the unwritten portions. When
more than one AIO thread is going, they both find this "new"
block and race to zero out their portion; this is uncoordinated
and causes data corruption.
Dave Chinner fixed this for xfs by simply serializing all
unaligned asynchronous direct IO. I've done the same here.
The difference is that we only wait on conversions, not all IO.
This is a very big hammer, and I'm not very pleased with
stuffing this into ext4_file_write(). But since ext4 is
DIO_LOCKING, we need to serialize it at this high level.
I tried to move this into ext4_ext_direct_IO, but by then
we have the i_mutex already, and we will wait on the
work queue to do conversions - which must also take the
i_mutex. So that won't work.
This was originally exposed by qemu-kvm installing to
a raw disk image with a normal sector-63 alignment. I've
tested a backport of this patch with qemu, and it does
avoid the corruption. It is also quite a lot slower
(14 min for package installs, vs. 8 min for well-aligned)
but I'll take slow correctness over fast corruption any day.
Mingming suggested that we can track outstanding
conversions, and wait on those so that non-sparse
files won't be affected, and I've implemented that here;
unaligned AIO to nonsparse files won't take a perf hit.
[tytso@mit.edu: Keep the mutex as a hashed array instead
of bloating the ext4 inode]
[tytso@mit.edu: Fix up namespace issues so that global
variables are protected with an "ext4_" prefix.]
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-02-12 08:17:34 -05:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: serialize unaligned asynchronous DIO
ext4 has a data corruption case when doing non-block-aligned
asynchronous direct IO into a sparse file, as demonstrated
by xfstest 240.
The root cause is that while ext4 preallocates space in the
hole, mappings of that space still look "new" and
dio_zero_block() will zero out the unwritten portions. When
more than one AIO thread is going, they both find this "new"
block and race to zero out their portion; this is uncoordinated
and causes data corruption.
Dave Chinner fixed this for xfs by simply serializing all
unaligned asynchronous direct IO. I've done the same here.
The difference is that we only wait on conversions, not all IO.
This is a very big hammer, and I'm not very pleased with
stuffing this into ext4_file_write(). But since ext4 is
DIO_LOCKING, we need to serialize it at this high level.
I tried to move this into ext4_ext_direct_IO, but by then
we have the i_mutex already, and we will wait on the
work queue to do conversions - which must also take the
i_mutex. So that won't work.
This was originally exposed by qemu-kvm installing to
a raw disk image with a normal sector-63 alignment. I've
tested a backport of this patch with qemu, and it does
avoid the corruption. It is also quite a lot slower
(14 min for package installs, vs. 8 min for well-aligned)
but I'll take slow correctness over fast corruption any day.
Mingming suggested that we can track outstanding
conversions, and wait on those so that non-sparse
files won't be affected, and I've implemented that here;
unaligned AIO to nonsparse files won't take a perf hit.
[tytso@mit.edu: Keep the mutex as a hashed array instead
of bloating the ext4 inode]
[tytso@mit.edu: Fix up namespace issues so that global
variables are protected with an "ext4_" prefix.]
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
2011-02-12 08:17:34 -05:00
|
|
|
|
|
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-20 17:29:51 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-20 17:29:51 -05:00
|
|
|
|
|
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
2016-11-20 17:29:51 -05:00
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-20 17:29:51 -05:00
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-20 17:29:51 -05:00
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
2016-11-20 17:29:51 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-20 17:29:51 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-06-09 22:04:33 -04:00
|
|
|
|
2016-11-20 17:29:51 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2026-03-27 18:29:39 +08:00
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2026-06-29 19:38:26 +08:00
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
2026-03-27 18:29:39 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
2016-11-20 17:29:51 -05:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: drain in-flight DIO before buffered write fallback
generic/746 started failing intermittently on ext3 (no-extent inodes).
The test triggers 'Page cache invalidation failure on direct I/O'
warnings and subsequent fsync returns -EIO. Adding a 50ms delay
between ext4_buffered_write_iter() and filemap_write_and_wait_range()
in ext4_dio_write_iter() makes the race almost always reproducible.
On no-extent inodes, DIO writes to holes cannot use unwritten extents,
so ext4_iomap_alloc() leaves m_flags=0 and ext4_map_blocks() returns 0.
The iomap layer then returns -ENOTBLK, causing fallback to buffered I/O.
The fallback path in ext4_dio_write_iter() calls
ext4_buffered_write_iter() which dirties pages, then does flush and
invalidate. However, there's an unprotected window between
ext4_buffered_write_iter() returning (with inode lock released) and
the subsequent flush+invalidate.
Concurrent async DIO completions from other threads can run
kiocb_invalidate_post_direct_write() during this window. If pages have
been re-dirtied, post-invalidation finds dirty pages and triggers the
warning, setting -EIO in the error sequence.
Consider a file with two 4k extents: [hole][written]. Thread A does
DIO to the written extent, while thread B does DIO spanning both:
kworker A (4k DIO, allocated block) kworker B (8k DIO, fallback)
----------------------------------- ----------------------------
inode_lock_shared() inode_lock_shared()
iomap_dio_rw(): iomap_dio_rw():
kiocb_invalidate_pages -> clean iomap_begin -> -ENOTBLK
submit_bio (async) dio->size = 0
inode_unlock_shared() inode_unlock_shared()
[bio pending in block layer] /* fallback: lock released */
ext4_buffered_write_iter()
inode_lock(exclusive)
generic_perform_write()
-> dirty pages [0, 8k]
inode_unlock(exclusive)
/* pages dirty, no lock */
[bio completes] filemap_write_and_wait_range()
iomap_dio_complete() -> flush dirty pages
kiocb_invalidate_post_direct_write() invalidate_mapping_pages()
invalidate_inode_pages2_range()
-> finds dirty page!
-> dio_warn_stale_pagecache()
-> errseq_set(-EIO)
This issue can be triggered through normal I/O paths, not just
intentionally overlapping DIO writes from userspace. For example,
generic/746 uses a loop device where multiple kworkers issue concurrent
I/O to the backing file. Additionally, when block_size < folio_size,
non-overlapping DIO writes that share a large folio can also trigger
the race.
Add inode_dio_wait() in ext4_buffered_write_iter() before
ext4_write_checks() to drain all in-flight DIO. This ensures that
all DIO clears existing pages before submitting IO (via
kiocb_invalidate_pages()), all BIO waits for all DIO to complete
(via inode_dio_wait()), and ext4_write_checks() observes the inode
size after all completed DIO so that ext4_block_zero_eof() does not
race with in-flight DIO, thus eliminating the race.
Fixes: 378f32bab371 ("ext4: introduce direct I/O write using iomap infrastructure")
Suggested-by: Zhang Yi <yi.zhang@huawei.com>
Link: https://patch.msgid.link/d1adcf7c-c276-458d-9cac-68a4410f7626@gmail.com
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-3-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:23 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2022-02-19 23:19:49 -05:00
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-06-01 16:58:55 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
2024-08-29 19:02:22 +08:00
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
2024-08-29 19:02:22 +08:00
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
2020-04-26 18:34:37 -07:00
|
|
|
|
|
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
2020-04-26 18:34:37 -07:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
2024-08-29 19:02:22 +08:00
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-08-29 19:02:22 +08:00
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-08-09 20:15:32 +08:00
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
|
|
|
|
|
2024-08-09 20:15:32 +08:00
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
2023-11-30 10:56:53 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
2025-09-25 14:30:39 +02:00
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-04-15 17:54:17 +02:00
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
ext4: Add multi-fsblock atomic write support with bigalloc
EXT4 supports bigalloc feature which allows the FS to work in size of
clusters (group of blocks) rather than individual blocks. This patch
adds atomic write support for bigalloc so that systems with bs = ps can
also create FS using -
mkfs.ext4 -F -O bigalloc -b 4096 -C 16384 <dev>
With bigalloc ext4 can support multi-fsblock atomic writes. We will have to
adjust ext4's atomic write unit max value to cluster size. This can then support
atomic write of size anywhere between [blocksize, clustersize]. This
patch adds the required changes to enable multi-fsblock atomic write
support using bigalloc in the next patch.
In this patch for block allocation:
we first query the underlying region of the requested range by calling
ext4_map_blocks() call. Here are the various cases which we then handle
depending upon the underlying mapping type:
1. If the underlying region for the entire requested range is a mapped extent,
then we don't call ext4_map_blocks() to allocate anything. We don't need to
even start the jbd2 txn in this case.
2. For an append write case, we create a mapped extent.
3. If the underlying region is entirely a hole, then we create an unwritten
extent for the requested range.
4. If the underlying region is a large unwritten extent, then we split the
extent into 2 unwritten extent of required size.
5. If the underlying region has any type of mixed mapping, then we call
ext4_map_blocks() in a loop to zero out the unwritten and the hole regions
within the requested range. This then provide a single mapped extent type
mapping for the requested range.
Note: We invoke ext4_map_blocks() in a loop with the EXT4_GET_BLOCKS_ZERO
flag only when the underlying extent mapping of the requested range is
not entirely a hole, an unwritten extent, or a fully mapped extent. That
is, if the underlying region contains a mix of hole(s), unwritten
extent(s), and mapped extent(s), we use this loop to ensure that all the
short mappings are zeroed out. This guarantees that the entire requested
range becomes a single, uniformly mapped extent. It is ok to do so
because we know this is being done on a bigalloc enabled filesystem
where the block bitmap represents the entire cluster unit.
Note having a single contiguous underlying region of type mapped,
unwrittn or hole is not a problem. But the reason to avoid writing on
top of mixed mapping region is because, atomic writes requires all or
nothing should get written for the userspace pwritev2 request. So if at
any point in time during the write if a crash or a sudden poweroff
occurs, the region undergoing atomic write should read either complete
old data or complete new data. But it should never have a mix of both
old and new data.
So, we first convert any mixed mapping region to a single contiguous
mapped extent before any data gets written to it. This is because
normally FS will only convert unwritten extents to written at the end of
the write in ->end_io() call. And if we allow the writes over a mixed
mapping and if a sudden power off happens in between, we will end up
reading mix of new data (over mapped extents) and old data (over
unwritten extents), because unwritten to written conversion never went
through.
So to avoid this and to avoid writes getting torned due to mixed
mapping, we first allocate a single contiguous block mapping and then
do the write.
Acked-by: Darrick J. Wong <djwong@kernel.org>
Co-developed-by: Ojaswin Mujoo <ojaswin@linux.ibm.com>
Signed-off-by: Ojaswin Mujoo <ojaswin@linux.ibm.com>
Signed-off-by: Ritesh Harjani (IBM) <ritesh.list@gmail.com>
Link: https://patch.msgid.link/c4965ac3407cbc773f0bc954d0966d9696f5038a.1747337952.git.ritesh.list@gmail.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2025-05-16 01:20:53 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2026-07-29 16:59:17 +08:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
2021-04-15 17:54:17 +02:00
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-11-30 10:56:53 +01:00
|
|
|
|
|
|
|
|
|
2021-04-15 17:54:17 +02:00
|
|
|
|
2023-11-30 10:56:53 +01:00
|
|
|
|
|
|
|
|
|
2024-09-19 16:25:39 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: skip overwrite check for aligned non-extending DIO writes
Currently, ext4_dio_write_checks() calls ext4_overwrite_io() to
determine if a write is a pure overwrite, and upgrades to exclusive
i_rwsem if not. However, ext4_overwrite_io() uses a single
ext4_map_blocks() call which only returns the first contiguous extent of
the same type. A write spanning multiple pre-allocated extents (e.g.
written + unwritten, or two physically discontiguous written extents)
produces a false negative, forcing an unnecessary exclusive lock upgrade.
After commit 5d87c7fca2c1 ("ext4: avoid starting handle when dio
writing an unwritten extent") and commit 012924f0eeef ("ext4: remove
useless ext4_iomap_overwrite_ops"), ext4_iomap_begin()'s fast path
accepts both EXT4_MAP_MAPPED and EXT4_MAP_UNWRITTEN without starting a
journal transaction. The iomap iteration naturally handles multi-extent
ranges: each call returns the mapping for the current segment, and
unwritten-to-written conversion is deferred to ext4_dio_write_end_io().
This means the common case of mixed written/unwritten extents never
reaches ext4_iomap_alloc() at all.
Even for the less common case where the range contains a hole and
ext4_iomap_alloc() is needed, exclusive i_rwsem is still unnecessary for
aligned non-extending writes:
- truncate/punch_hole are kept out: they require exclusive i_rwsem
(blocked by our shared lock during allocation), and inode_dio_begin()
keeps their inode_dio_wait() blocked until in-flight bios complete.
- i_data_sem write-lock inside ext4_map_blocks() serializes concurrent
extent tree modifications (parallel writers to the same hole).
- The journal handle is per-thread and does not require i_rwsem
exclusion.
- i_disksize and orphan list are not involved in non-extending writes.
Skip the ext4_overwrite_io() check entirely for aligned writes by
initializing overwrite to true and only calling ext4_overwrite_io() for
unaligned writes. Unaligned writes still need the extent state check
because concurrent partial block zeroing in the DIO layer requires
exclusive serialization unless the range is a pure written-extent
overwrite.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Aligned 8K DIO writes spanning written+unwritten extent boundaries.
Each thread writes its own 1G region sequentially; the file is rebuilt
between runs so every block is written exactly once. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 42,322 43,329 1.02x
2 68,516 70,677 1.03x
4 62,489 97,072 1.55x
8 58,701 110,819 1.89x
16 58,569 116,392 1.99x
32 60,860 117,244 1.93x
Wall time at JOBS=32: 69.2s (Before) -> 35.4s (After), 1.96x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-4-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:24 +08:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
2026-06-29 19:38:26 +08:00
|
|
|
|
ext4: skip overwrite check for aligned non-extending DIO writes
Currently, ext4_dio_write_checks() calls ext4_overwrite_io() to
determine if a write is a pure overwrite, and upgrades to exclusive
i_rwsem if not. However, ext4_overwrite_io() uses a single
ext4_map_blocks() call which only returns the first contiguous extent of
the same type. A write spanning multiple pre-allocated extents (e.g.
written + unwritten, or two physically discontiguous written extents)
produces a false negative, forcing an unnecessary exclusive lock upgrade.
After commit 5d87c7fca2c1 ("ext4: avoid starting handle when dio
writing an unwritten extent") and commit 012924f0eeef ("ext4: remove
useless ext4_iomap_overwrite_ops"), ext4_iomap_begin()'s fast path
accepts both EXT4_MAP_MAPPED and EXT4_MAP_UNWRITTEN without starting a
journal transaction. The iomap iteration naturally handles multi-extent
ranges: each call returns the mapping for the current segment, and
unwritten-to-written conversion is deferred to ext4_dio_write_end_io().
This means the common case of mixed written/unwritten extents never
reaches ext4_iomap_alloc() at all.
Even for the less common case where the range contains a hole and
ext4_iomap_alloc() is needed, exclusive i_rwsem is still unnecessary for
aligned non-extending writes:
- truncate/punch_hole are kept out: they require exclusive i_rwsem
(blocked by our shared lock during allocation), and inode_dio_begin()
keeps their inode_dio_wait() blocked until in-flight bios complete.
- i_data_sem write-lock inside ext4_map_blocks() serializes concurrent
extent tree modifications (parallel writers to the same hole).
- The journal handle is per-thread and does not require i_rwsem
exclusion.
- i_disksize and orphan list are not involved in non-extending writes.
Skip the ext4_overwrite_io() check entirely for aligned writes by
initializing overwrite to true and only calling ext4_overwrite_io() for
unaligned writes. Unaligned writes still need the extent state check
because concurrent partial block zeroing in the DIO layer requires
exclusive serialization unless the range is a pure written-extent
overwrite.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Aligned 8K DIO writes spanning written+unwritten extent boundaries.
Each thread writes its own 1G region sequentially; the file is rebuilt
between runs so every block is written exactly once. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 42,322 43,329 1.02x
2 68,516 70,677 1.03x
4 62,489 97,072 1.55x
8 58,701 110,819 1.89x
16 58,569 116,392 1.99x
32 60,860 117,244 1.93x
Wall time at JOBS=32: 69.2s (Before) -> 35.4s (After), 1.96x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-4-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:24 +08:00
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
ext4: skip overwrite check for aligned non-extending DIO writes
Currently, ext4_dio_write_checks() calls ext4_overwrite_io() to
determine if a write is a pure overwrite, and upgrades to exclusive
i_rwsem if not. However, ext4_overwrite_io() uses a single
ext4_map_blocks() call which only returns the first contiguous extent of
the same type. A write spanning multiple pre-allocated extents (e.g.
written + unwritten, or two physically discontiguous written extents)
produces a false negative, forcing an unnecessary exclusive lock upgrade.
After commit 5d87c7fca2c1 ("ext4: avoid starting handle when dio
writing an unwritten extent") and commit 012924f0eeef ("ext4: remove
useless ext4_iomap_overwrite_ops"), ext4_iomap_begin()'s fast path
accepts both EXT4_MAP_MAPPED and EXT4_MAP_UNWRITTEN without starting a
journal transaction. The iomap iteration naturally handles multi-extent
ranges: each call returns the mapping for the current segment, and
unwritten-to-written conversion is deferred to ext4_dio_write_end_io().
This means the common case of mixed written/unwritten extents never
reaches ext4_iomap_alloc() at all.
Even for the less common case where the range contains a hole and
ext4_iomap_alloc() is needed, exclusive i_rwsem is still unnecessary for
aligned non-extending writes:
- truncate/punch_hole are kept out: they require exclusive i_rwsem
(blocked by our shared lock during allocation), and inode_dio_begin()
keeps their inode_dio_wait() blocked until in-flight bios complete.
- i_data_sem write-lock inside ext4_map_blocks() serializes concurrent
extent tree modifications (parallel writers to the same hole).
- The journal handle is per-thread and does not require i_rwsem
exclusion.
- i_disksize and orphan list are not involved in non-extending writes.
Skip the ext4_overwrite_io() check entirely for aligned writes by
initializing overwrite to true and only calling ext4_overwrite_io() for
unaligned writes. Unaligned writes still need the extent state check
because concurrent partial block zeroing in the DIO layer requires
exclusive serialization unless the range is a pure written-extent
overwrite.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Aligned 8K DIO writes spanning written+unwritten extent boundaries.
Each thread writes its own 1G region sequentially; the file is rebuilt
between runs so every block is written exactly once. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 42,322 43,329 1.02x
2 68,516 70,677 1.03x
4 62,489 97,072 1.55x
8 58,701 110,819 1.89x
16 58,569 116,392 1.99x
32 60,860 117,244 1.93x
Wall time at JOBS=32: 69.2s (Before) -> 35.4s (After), 1.96x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-4-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:24 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: dio take shared inode lock when overwriting preallocated blocks
In the dio write path, we only take shared inode lock for the case of
aligned overwriting initialized blocks inside EOF. But for overwriting
preallocated blocks, it may only need to split unwritten extents, this
procedure has been protected under i_data_sem lock, it's safe to
release the exclusive inode lock and take shared inode lock.
This could give a significant speed up for multi-threaded writes. Test
on Intel Xeon Gold 6140 and nvme SSD with below fio parameters.
direct=1
ioengine=libaio
iodepth=10
numjobs=10
runtime=60
rw=randwrite
size=100G
And the test result are:
Before:
bs=4k IOPS=11.1k, BW=43.2MiB/s
bs=16k IOPS=11.1k, BW=173MiB/s
bs=64k IOPS=11.2k, BW=697MiB/s
After:
bs=4k IOPS=41.4k, BW=162MiB/s
bs=16k IOPS=41.3k, BW=646MiB/s
bs=64k IOPS=13.5k, BW=843MiB/s
Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20221226062015.3479416-1-yi.zhang@huaweicloud.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2022-12-26 14:20:15 +08:00
|
|
|
|
ext4: skip overwrite check for aligned non-extending DIO writes
Currently, ext4_dio_write_checks() calls ext4_overwrite_io() to
determine if a write is a pure overwrite, and upgrades to exclusive
i_rwsem if not. However, ext4_overwrite_io() uses a single
ext4_map_blocks() call which only returns the first contiguous extent of
the same type. A write spanning multiple pre-allocated extents (e.g.
written + unwritten, or two physically discontiguous written extents)
produces a false negative, forcing an unnecessary exclusive lock upgrade.
After commit 5d87c7fca2c1 ("ext4: avoid starting handle when dio
writing an unwritten extent") and commit 012924f0eeef ("ext4: remove
useless ext4_iomap_overwrite_ops"), ext4_iomap_begin()'s fast path
accepts both EXT4_MAP_MAPPED and EXT4_MAP_UNWRITTEN without starting a
journal transaction. The iomap iteration naturally handles multi-extent
ranges: each call returns the mapping for the current segment, and
unwritten-to-written conversion is deferred to ext4_dio_write_end_io().
This means the common case of mixed written/unwritten extents never
reaches ext4_iomap_alloc() at all.
Even for the less common case where the range contains a hole and
ext4_iomap_alloc() is needed, exclusive i_rwsem is still unnecessary for
aligned non-extending writes:
- truncate/punch_hole are kept out: they require exclusive i_rwsem
(blocked by our shared lock during allocation), and inode_dio_begin()
keeps their inode_dio_wait() blocked until in-flight bios complete.
- i_data_sem write-lock inside ext4_map_blocks() serializes concurrent
extent tree modifications (parallel writers to the same hole).
- The journal handle is per-thread and does not require i_rwsem
exclusion.
- i_disksize and orphan list are not involved in non-extending writes.
Skip the ext4_overwrite_io() check entirely for aligned writes by
initializing overwrite to true and only calling ext4_overwrite_io() for
unaligned writes. Unaligned writes still need the extent state check
because concurrent partial block zeroing in the DIO layer requires
exclusive serialization unless the range is a pure written-extent
overwrite.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Aligned 8K DIO writes spanning written+unwritten extent boundaries.
Each thread writes its own 1G region sequentially; the file is rebuilt
between runs so every block is written exactly once. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 42,322 43,329 1.02x
2 68,516 70,677 1.03x
4 62,489 97,072 1.55x
8 58,701 110,819 1.89x
16 58,569 116,392 1.99x
32 60,860 117,244 1.93x
Wall time at JOBS=32: 69.2s (Before) -> 35.4s (After), 1.96x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-4-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:24 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
ext4: dio take shared inode lock when overwriting preallocated blocks
In the dio write path, we only take shared inode lock for the case of
aligned overwriting initialized blocks inside EOF. But for overwriting
preallocated blocks, it may only need to split unwritten extents, this
procedure has been protected under i_data_sem lock, it's safe to
release the exclusive inode lock and take shared inode lock.
This could give a significant speed up for multi-threaded writes. Test
on Intel Xeon Gold 6140 and nvme SSD with below fio parameters.
direct=1
ioengine=libaio
iodepth=10
numjobs=10
runtime=60
rw=randwrite
size=100G
And the test result are:
Before:
bs=4k IOPS=11.1k, BW=43.2MiB/s
bs=16k IOPS=11.1k, BW=173MiB/s
bs=64k IOPS=11.2k, BW=697MiB/s
After:
bs=4k IOPS=41.4k, BW=162MiB/s
bs=16k IOPS=41.3k, BW=646MiB/s
bs=64k IOPS=13.5k, BW=843MiB/s
Signed-off-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20221226062015.3479416-1-yi.zhang@huaweicloud.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2022-12-26 14:20:15 +08:00
|
|
|
|
2026-01-05 09:45:20 +08:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-03-14 09:07:59 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: drop dio overwrite only flag and associated warning
The commit referenced below opened up concurrent unaligned dio under
shared locking for pure overwrites. In doing so, it enabled use of
the IOMAP_DIO_OVERWRITE_ONLY flag and added a warning on unexpected
-EAGAIN returns as an extra precaution, since ext4 does not retry
writes in such cases. The flag itself is advisory in this case since
ext4 checks for unaligned I/Os and uses appropriate locking up
front, rather than on a retry in response to -EAGAIN.
As it turns out, the warning check is susceptible to false positives
because there are scenarios where -EAGAIN can be expected from lower
layers without necessarily having IOCB_NOWAIT set on the iocb. For
example, one instance of the warning has been seen where io_uring
sets IOCB_HIPRI, which in turn results in REQ_POLLED|REQ_NOWAIT on
the bio. This results in -EAGAIN if the block layer is unable to
allocate a request, etc. [Note that there is an outstanding patch to
untangle REQ_POLLED and REQ_NOWAIT such that the latter relies on
IOCB_NOWAIT, which would also address this instance of the warning.]
Another instance of the warning has been reproduced by syzbot. A dio
write is interrupted down in __get_user_pages_locked() waiting on
the mm lock and returns -EAGAIN up the stack. If the iomap dio
iteration layer has made no progress on the write to this point,
-EAGAIN returns up to the filesystem and triggers the warning.
This use of the overwrite flag in ext4 is precautionary and
half-baked. I.e., ext4 doesn't actually implement overwrite checking
in the iomap callbacks when the flag is set, so the only extra
verification it provides are i_size checks in the generic iomap dio
layer. Combined with the tendency for false positives, the added
verification is not worth the extra trouble. Remove the flag,
associated warning, and update the comments to document when
concurrent unaligned dio writes are allowed and why said flag is not
used.
Cc: stable@kernel.org
Reported-by: syzbot+5050ad0fb47527b1808a@syzkaller.appspotmail.com
Reported-by: Pengfei Xu <pengfei.xu@intel.com>
Fixes: 310ee0902b8d ("ext4: allow concurrent unaligned dio overwrites")
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20230810165559.946222-1-bfoster@redhat.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2023-08-10 12:55:59 -04:00
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
|
|
|
|
|
ext4: skip overwrite check for aligned non-extending DIO writes
Currently, ext4_dio_write_checks() calls ext4_overwrite_io() to
determine if a write is a pure overwrite, and upgrades to exclusive
i_rwsem if not. However, ext4_overwrite_io() uses a single
ext4_map_blocks() call which only returns the first contiguous extent of
the same type. A write spanning multiple pre-allocated extents (e.g.
written + unwritten, or two physically discontiguous written extents)
produces a false negative, forcing an unnecessary exclusive lock upgrade.
After commit 5d87c7fca2c1 ("ext4: avoid starting handle when dio
writing an unwritten extent") and commit 012924f0eeef ("ext4: remove
useless ext4_iomap_overwrite_ops"), ext4_iomap_begin()'s fast path
accepts both EXT4_MAP_MAPPED and EXT4_MAP_UNWRITTEN without starting a
journal transaction. The iomap iteration naturally handles multi-extent
ranges: each call returns the mapping for the current segment, and
unwritten-to-written conversion is deferred to ext4_dio_write_end_io().
This means the common case of mixed written/unwritten extents never
reaches ext4_iomap_alloc() at all.
Even for the less common case where the range contains a hole and
ext4_iomap_alloc() is needed, exclusive i_rwsem is still unnecessary for
aligned non-extending writes:
- truncate/punch_hole are kept out: they require exclusive i_rwsem
(blocked by our shared lock during allocation), and inode_dio_begin()
keeps their inode_dio_wait() blocked until in-flight bios complete.
- i_data_sem write-lock inside ext4_map_blocks() serializes concurrent
extent tree modifications (parallel writers to the same hole).
- The journal handle is per-thread and does not require i_rwsem
exclusion.
- i_disksize and orphan list are not involved in non-extending writes.
Skip the ext4_overwrite_io() check entirely for aligned writes by
initializing overwrite to true and only calling ext4_overwrite_io() for
unaligned writes. Unaligned writes still need the extent state check
because concurrent partial block zeroing in the DIO layer requires
exclusive serialization unless the range is a pure written-extent
overwrite.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Aligned 8K DIO writes spanning written+unwritten extent boundaries.
Each thread writes its own 1G region sequentially; the file is rebuilt
between runs so every block is written exactly once. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 42,322 43,329 1.02x
2 68,516 70,677 1.03x
4 62,489 97,072 1.55x
8 58,701 110,819 1.89x
16 58,569 116,392 1.99x
32 60,860 117,244 1.93x
Wall time at JOBS=32: 69.2s (Before) -> 35.4s (After), 1.96x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-4-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:24 +08:00
|
|
|
|
|
|
|
|
|
2023-03-14 09:07:59 -04:00
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
2020-07-08 17:35:16 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-03-14 09:07:59 -04:00
|
|
|
|
|
|
|
|
|
ext4: drop dio overwrite only flag and associated warning
The commit referenced below opened up concurrent unaligned dio under
shared locking for pure overwrites. In doing so, it enabled use of
the IOMAP_DIO_OVERWRITE_ONLY flag and added a warning on unexpected
-EAGAIN returns as an extra precaution, since ext4 does not retry
writes in such cases. The flag itself is advisory in this case since
ext4 checks for unaligned I/Os and uses appropriate locking up
front, rather than on a retry in response to -EAGAIN.
As it turns out, the warning check is susceptible to false positives
because there are scenarios where -EAGAIN can be expected from lower
layers without necessarily having IOCB_NOWAIT set on the iocb. For
example, one instance of the warning has been seen where io_uring
sets IOCB_HIPRI, which in turn results in REQ_POLLED|REQ_NOWAIT on
the bio. This results in -EAGAIN if the block layer is unable to
allocate a request, etc. [Note that there is an outstanding patch to
untangle REQ_POLLED and REQ_NOWAIT such that the latter relies on
IOCB_NOWAIT, which would also address this instance of the warning.]
Another instance of the warning has been reproduced by syzbot. A dio
write is interrupted down in __get_user_pages_locked() waiting on
the mm lock and returns -EAGAIN up the stack. If the iomap dio
iteration layer has made no progress on the write to this point,
-EAGAIN returns up to the filesystem and triggers the warning.
This use of the overwrite flag in ext4 is precautionary and
half-baked. I.e., ext4 doesn't actually implement overwrite checking
in the iomap callbacks when the flag is set, so the only extra
verification it provides are i_size checks in the generic iomap dio
layer. Combined with the tendency for false positives, the added
verification is not worth the extra trouble. Remove the flag,
associated warning, and update the comments to document when
concurrent unaligned dio writes are allowed and why said flag is not
used.
Cc: stable@kernel.org
Reported-by: syzbot+5050ad0fb47527b1808a@syzkaller.appspotmail.com
Reported-by: Pengfei Xu <pengfei.xu@intel.com>
Fixes: 310ee0902b8d ("ext4: allow concurrent unaligned dio overwrites")
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20230810165559.946222-1-bfoster@redhat.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2023-08-10 12:55:59 -04:00
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-03-14 09:07:59 -04:00
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
2023-03-14 09:07:59 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: base unaligned DIO lock decision on partial block zeroing
For unaligned DIO writes, the previous ext4_overwrite_io() required the
entire range to fall within a single written extent. This was overly
conservative: the DIO layer only performs partial block zeroing for the
head and tail blocks when they are partially covered by the write.
Middle blocks that are fully covered are written as whole blocks
without any zeroing, so they are safe regardless of extent state.
Therefore exclusive lock is only required when partial block zeroing
will actually happen:
- The head partial block (if any) lands on a hole or unwritten extent.
- The tail partial block (if any) lands on a hole or unwritten extent.
Middle full-cover blocks can be in any state (hole, unwritten, or
written) - block allocation under shared lock is safe per the previous
patch's analysis (inode_dio_begin + i_data_sem protection).
Replace ext4_overwrite_io() with ext4_dio_needs_zeroing(), which
directly answers the question driving the lock decision. It uses at
most two ext4_map_blocks() calls: one for the head partial block (also
catching the case where it spans through the tail), and one for the
tail partial block if not already covered.
This enables shared lock for previously-rejected scenarios such as:
- Unaligned write spanning written extent + mid-range hole + written
extent at the tail.
- Unaligned write where the partial blocks land on written extents but
the middle has unwritten extents.
Performance:
Hardware: /dev/sda (rotational disk, ~1 GB/s sustained write)
Filesystem: ext4 default mkfs
Unaligned DIO writes (14336 bytes at +512 within each 16K stripe).
Each stripe is laid out as [written][unwritten][unwritten][written],
so the head and tail partial blocks land on written extents but the
middle is unwritten. Metric: IOPS.
JOBS Before After speedup
---- -------- --------- -------
1 15,547 17,381 1.12x
2 15,910 34,172 2.15x
4 15,014 57,567 3.83x
8 15,022 81,947 5.46x
16 14,586 99,126 6.80x
32 14,047 92,519 6.59x
Wall time at JOBS=32: 149.3s (Before) -> 22.7s (After), 6.58x faster.
Reviewed-by: Zhang Yi <yi.zhang@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Baokun Li <libaokun@linux.alibaba.com>
Link: https://patch.msgid.link/20260629113827.4074335-5-libaokun@linux.alibaba.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
2026-06-29 19:38:25 +08:00
|
|
|
|
2023-03-14 09:07:59 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2026-06-29 19:38:26 +08:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
2026-01-05 09:45:20 +08:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
2023-03-14 09:07:59 -04:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
2022-08-26 23:58:47 -07:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-10-02 14:50:20 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-03-14 09:07:59 -04:00
|
|
|
|
2026-01-05 09:45:20 +08:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-09-25 21:47:39 +02:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
2026-01-05 09:45:19 +08:00
|
|
|
|
2023-03-14 09:07:59 -04:00
|
|
|
|
2020-07-23 22:45:59 -07:00
|
|
|
|
|
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-08-09 20:15:32 +08:00
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
ext4: Start with shared i_rwsem in case of DIO instead of exclusive
Earlier there was no shared lock in DIO read path. But this patch
(16c54688592ce: 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>
2019-12-12 11:25:56 +05:30
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-11-04 16:22:59 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-20 18:09:11 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:08 +11:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
2016-11-20 18:09:11 -05:00
|
|
|
|
2019-12-12 11:25:55 +05:30
|
|
|
|
|
|
|
|
|
2017-06-20 07:05:47 -05:00
|
|
|
|
2019-12-12 11:25:55 +05:30
|
|
|
|
2017-06-20 07:05:47 -05:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
2016-11-20 18:09:11 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:01:51 +11:00
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:08 +11:00
|
|
|
|
|
|
|
|
|
2026-06-29 19:38:27 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:08 +11:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-11-20 18:09:11 -05:00
|
|
|
|
2019-11-05 23:02:08 +11:00
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
2024-08-29 19:02:22 +08:00
|
|
|
|
2024-08-09 20:15:32 +08:00
|
|
|
|
2023-10-13 14:13:50 +02:00
|
|
|
|
2016-11-20 18:09:11 -05:00
|
|
|
|
2017-02-08 14:39:27 -05:00
|
|
|
|
2016-11-20 18:09:11 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
2014-04-17 16:09:22 -04:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
2025-01-22 19:41:26 +08:00
|
|
|
|
2013-01-23 17:07:38 -05:00
|
|
|
|
2014-04-21 14:26:28 -04:00
|
|
|
|
2025-01-22 19:41:26 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-02-05 01:28:48 -05:00
|
|
|
|
2016-11-20 18:09:11 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-11-04 16:22:58 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:02:39 +11:00
|
|
|
|
|
|
|
|
|
2020-10-15 13:37:57 -07:00
|
|
|
|
|
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
|
|
|
|
|
2015-02-16 15:59:38 -08:00
|
|
|
|
2023-08-18 21:23:35 +01:00
|
|
|
|
2015-02-16 15:59:38 -08:00
|
|
|
|
2018-05-13 16:01:49 -04:00
|
|
|
|
|
|
|
|
|
2018-01-07 16:41:01 -05:00
|
|
|
|
2017-05-12 15:46:54 -07:00
|
|
|
|
2017-02-24 14:56:41 -08:00
|
|
|
|
2015-12-07 14:28:03 -05:00
|
|
|
|
2017-08-24 15:26:01 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2023-08-18 21:23:35 +01:00
|
|
|
|
2017-08-24 15:26:01 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2021-02-04 18:05:42 +01:00
|
|
|
|
2025-06-19 18:58:05 +10:00
|
|
|
|
2015-09-08 14:59:22 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-02-24 14:56:41 -08:00
|
|
|
|
2021-02-04 18:05:42 +01:00
|
|
|
|
2018-01-07 16:41:01 -05:00
|
|
|
|
2017-05-12 15:46:54 -07:00
|
|
|
|
|
|
|
|
|
2017-11-01 16:36:44 +01:00
|
|
|
|
2021-02-04 18:05:42 +01:00
|
|
|
|
2017-11-01 16:36:44 +01:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-05-12 15:46:54 -07:00
|
|
|
|
2021-02-04 18:05:42 +01:00
|
|
|
|
2016-10-21 11:33:49 +02:00
|
|
|
|
2023-08-18 21:23:35 +01:00
|
|
|
|
2017-05-12 15:46:54 -07:00
|
|
|
|
2017-11-01 16:36:44 +01:00
|
|
|
|
2018-01-07 16:41:01 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-11-01 16:36:45 +01:00
|
|
|
|
|
|
|
|
|
2023-08-18 21:23:35 +01:00
|
|
|
|
2021-02-04 18:05:42 +01:00
|
|
|
|
2015-09-08 14:59:22 -07:00
|
|
|
|
2017-05-12 15:46:54 -07:00
|
|
|
|
2021-02-04 18:05:42 +01:00
|
|
|
|
2017-05-12 15:46:54 -07:00
|
|
|
|
2015-09-08 14:59:22 -07:00
|
|
|
|
|
|
|
|
|
2015-02-16 15:59:38 -08:00
|
|
|
|
|
|
|
|
|
2018-05-13 16:01:49 -04:00
|
|
|
|
2017-02-24 14:57:08 -08:00
|
|
|
|
2023-08-18 21:23:35 +01:00
|
|
|
|
2017-02-24 14:57:08 -08:00
|
|
|
|
|
|
|
|
|
2015-02-16 15:59:38 -08:00
|
|
|
|
|
|
|
|
|
2017-02-24 14:57:08 -08:00
|
|
|
|
2016-02-27 14:01:13 -05:00
|
|
|
|
dax: use common 4k zero page for dax mmap reads
When servicing mmap() reads from file holes the current DAX code
allocates a page cache page of all zeroes and places the struct page
pointer in the mapping->page_tree radix tree.
This has three major drawbacks:
1) It consumes memory unnecessarily. For every 4k page that is read via
a DAX mmap() over a hole, we allocate a new page cache page. This
means that if you read 1GiB worth of pages, you end up using 1GiB of
zeroed memory. This is easily visible by looking at the overall
memory consumption of the system or by looking at /proc/[pid]/smaps:
7f62e72b3000-7f63272b3000 rw-s 00000000 103:00 12 /root/dax/data
Size: 1048576 kB
Rss: 1048576 kB
Pss: 1048576 kB
Shared_Clean: 0 kB
Shared_Dirty: 0 kB
Private_Clean: 1048576 kB
Private_Dirty: 0 kB
Referenced: 1048576 kB
Anonymous: 0 kB
LazyFree: 0 kB
AnonHugePages: 0 kB
ShmemPmdMapped: 0 kB
Shared_Hugetlb: 0 kB
Private_Hugetlb: 0 kB
Swap: 0 kB
SwapPss: 0 kB
KernelPageSize: 4 kB
MMUPageSize: 4 kB
Locked: 0 kB
2) It is slower than using a common zero page because each page fault
has more work to do. Instead of just inserting a common zero page we
have to allocate a page cache page, zero it, and then insert it. Here
are the average latencies of dax_load_hole() as measured by ftrace on
a random test box:
Old method, using zeroed page cache pages: 3.4 us
New method, using the common 4k zero page: 0.8 us
This was the average latency over 1 GiB of sequential reads done by
this simple fio script:
[global]
size=1G
filename=/root/dax/data
fallocate=none
[io]
rw=read
ioengine=mmap
3) The fact that we had to check for both DAX exceptional entries and
for page cache pages in the radix tree made the DAX code more
complex.
Solve these issues by following the lead of the DAX PMD code and using a
common 4k zero page instead. As with the PMD code we will now insert a
DAX exceptional entry into the radix tree instead of a struct page
pointer which allows us to remove all the special casing in the DAX
code.
Note that we do still pretty aggressively check for regular pages in the
DAX radix tree, especially where we take action based on the bits set in
the page. If we ever find a regular page in our radix tree now that
most likely means that someone besides DAX is inserting pages (which has
happened lots of times in the past), and we want to find that out early
and fail loudly.
This solution also removes the extra memory consumption. Here is that
same /proc/[pid]/smaps after 1GiB of reading from a hole with the new
code:
7f2054a74000-7f2094a74000 rw-s 00000000 103:00 12 /root/dax/data
Size: 1048576 kB
Rss: 0 kB
Pss: 0 kB
Shared_Clean: 0 kB
Shared_Dirty: 0 kB
Private_Clean: 0 kB
Private_Dirty: 0 kB
Referenced: 0 kB
Anonymous: 0 kB
LazyFree: 0 kB
AnonHugePages: 0 kB
ShmemPmdMapped: 0 kB
Shared_Hugetlb: 0 kB
Private_Hugetlb: 0 kB
Swap: 0 kB
SwapPss: 0 kB
KernelPageSize: 4 kB
MMUPageSize: 4 kB
Locked: 0 kB
Overall system memory consumption is similarly improved.
Another major change is that we remove dax_pfn_mkwrite() from our fault
flow, and instead rely on the page fault itself to make the PTE dirty
and writeable. The following description from the patch adding the
vm_insert_mixed_mkwrite() call explains this a little more:
"To be able to use the common 4k zero page in DAX we need to have our
PTE fault path look more like our PMD fault path where a PTE entry
can be marked as dirty and writeable as it is first inserted rather
than waiting for a follow-up dax_pfn_mkwrite() =>
finish_mkwrite_fault() call.
Right now we can rely on having a dax_pfn_mkwrite() call because we
can distinguish between these two cases in do_wp_page():
case 1: 4k zero page => writable DAX storage
case 2: read-only DAX storage => writeable DAX storage
This distinction is made by via vm_normal_page(). vm_normal_page()
returns false for the common 4k zero page, though, just as it does
for DAX ptes. Instead of special casing the DAX + 4k zero page case
we will simplify our DAX PTE page fault sequence so that it matches
our DAX PMD sequence, and get rid of the dax_pfn_mkwrite() helper.
We will instead use dax_iomap_fault() to handle write-protection
faults.
This means that insert_pfn() needs to follow the lead of
insert_pfn_pmd() and allow us to pass in a 'mkwrite' flag. If
'mkwrite' is set insert_pfn() will do the work that was previously
done by wp_page_reuse() as part of the dax_pfn_mkwrite() call path"
Link: http://lkml.kernel.org/r/20170724170616.25810-4-ross.zwisler@linux.intel.com
Signed-off-by: Ross Zwisler <ross.zwisler@linux.intel.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Cc: "Darrick J. Wong" <darrick.wong@oracle.com>
Cc: "Theodore Ts'o" <tytso@mit.edu>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Andreas Dilger <adilger.kernel@dilger.ca>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Chinner <david@fromorbit.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Matthew Wilcox <mawilcox@microsoft.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-09-06 16:18:43 -07:00
|
|
|
|
2015-02-16 15:59:38 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2009-09-27 22:29:37 +04:00
|
|
|
|
2021-02-04 18:05:42 +01:00
|
|
|
|
2014-04-07 15:37:19 -07:00
|
|
|
|
2008-07-11 19:27:31 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2025-06-16 20:33:24 +01:00
|
|
|
|
2008-07-11 19:27:31 -04:00
|
|
|
|
2025-01-22 19:41:26 +08:00
|
|
|
|
2025-06-16 20:33:24 +01:00
|
|
|
|
2015-04-12 00:56:10 -04:00
|
|
|
|
2023-06-16 18:50:49 +02:00
|
|
|
|
2015-04-12 00:56:10 -04:00
|
|
|
|
2025-01-22 19:41:26 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-02-05 01:28:48 -05:00
|
|
|
|
2017-11-01 16:36:45 +01:00
|
|
|
|
2019-07-05 19:33:27 +05:30
|
|
|
|
|
|
|
|
|
2017-11-01 16:36:45 +01:00
|
|
|
|
2026-01-22 16:06:18 +00:00
|
|
|
|
2017-11-01 16:36:45 +01:00
|
|
|
|
|
|
|
|
|
2008-07-11 19:27:31 -04:00
|
|
|
|
2015-02-16 15:59:38 -08:00
|
|
|
|
2025-06-16 20:33:24 +01:00
|
|
|
|
2026-01-22 16:06:18 +00:00
|
|
|
|
2015-02-16 15:59:38 -08:00
|
|
|
|
2025-06-16 20:33:24 +01:00
|
|
|
|
2015-02-16 15:59:38 -08:00
|
|
|
|
2008-07-11 19:27:31 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
|
|
|
|
|
2009-06-13 10:09:48 -04:00
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
2009-06-13 10:09:48 -04:00
|
|
|
|
|
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-11-05 19:59:09 -08:00
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
|
|
|
|
|
2025-01-22 19:41:27 +08:00
|
|
|
|
|
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
|
|
|
|
|
2020-11-05 19:59:09 -08:00
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2018-05-13 22:54:44 -04:00
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
2018-05-13 22:54:44 -04:00
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
|
|
|
|
|
2018-05-13 22:54:44 -04:00
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
2018-05-13 22:54:44 -04:00
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
2021-08-16 11:57:04 +02:00
|
|
|
|
|
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
2018-05-13 22:54:44 -04:00
|
|
|
|
2020-12-16 11:18:39 +01:00
|
|
|
|
2024-03-21 01:03:10 +00:00
|
|
|
|
2020-12-16 11:18:39 +01:00
|
|
|
|
|
|
|
|
|
2020-12-16 11:18:44 +01:00
|
|
|
|
2018-05-13 22:54:44 -04:00
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
2018-05-13 22:54:44 -04:00
|
|
|
|
|
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2020-06-14 11:45:44 +07:00
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
2015-04-12 00:56:10 -04:00
|
|
|
|
2009-06-13 10:09:48 -04:00
|
|
|
|
2025-01-22 19:41:26 +08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-02-05 01:28:48 -05:00
|
|
|
|
2018-05-13 22:44:23 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2016-03-26 16:14:41 -04:00
|
|
|
|
2017-10-18 20:21:57 -04:00
|
|
|
|
|
|
|
|
|
2019-07-22 09:26:24 -07:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2017-10-18 20:21:57 -04:00
|
|
|
|
|
|
|
|
|
2011-01-10 12:29:43 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2013-08-16 21:19:41 -04:00
|
|
|
|
2015-04-12 00:56:10 -04:00
|
|
|
|
2013-08-16 21:19:41 -04:00
|
|
|
|
|
|
|
|
|
2011-01-10 12:29:43 -05:00
|
|
|
|
2017-06-20 07:05:47 -05:00
|
|
|
|
2024-11-04 16:22:59 -08:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2024-05-18 14:11:54 -07:00
|
|
|
|
2015-05-31 13:35:39 -04:00
|
|
|
|
2009-06-13 10:09:48 -04:00
|
|
|
|
|
|
|
|
|
2010-10-27 21:30:06 -04:00
|
|
|
|
2012-04-30 13:14:03 -05:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-10-27 21:30:06 -04:00
|
|
|
|
2012-12-17 15:59:39 -08:00
|
|
|
|
2010-10-27 21:30:06 -04:00
|
|
|
|
|
|
|
|
|
2025-05-06 09:20:08 +08:00
|
|
|
|
2010-10-27 21:30:06 -04:00
|
|
|
|
2012-12-17 15:59:39 -08:00
|
|
|
|
2017-10-01 17:58:54 -04:00
|
|
|
|
2012-12-17 15:59:39 -08:00
|
|
|
|
2012-11-08 21:57:40 -05:00
|
|
|
|
|
|
|
|
|
2017-10-01 17:58:54 -04:00
|
|
|
|
2019-11-05 23:03:31 +11:00
|
|
|
|
|
|
|
|
|
2017-10-01 17:58:54 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2019-11-05 23:03:31 +11:00
|
|
|
|
|
|
|
|
|
2017-10-01 17:58:54 -04:00
|
|
|
|
|
|
|
|
|
2012-11-08 21:57:40 -05:00
|
|
|
|
|
|
|
|
|
2017-10-01 17:58:54 -04:00
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2010-10-27 21:30:06 -04:00
|
|
|
|
|
|
|
|
|
2006-10-11 01:20:53 -07:00
|
|
|
|
2010-10-27 21:30:06 -04:00
|
|
|
|
2016-11-20 17:36:06 -05:00
|
|
|
|
2014-04-17 16:09:22 -04:00
|
|
|
|
2021-10-12 13:12:24 +02:00
|
|
|
|
2008-04-29 22:03:54 -04:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
2006-10-11 01:20:53 -07:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
2025-06-16 20:33:24 +01:00
|
|
|
|
2009-06-13 10:09:48 -04:00
|
|
|
|
2006-10-11 01:20:53 -07:00
|
|
|
|
|
|
|
|
|
2016-10-07 16:59:59 -07:00
|
|
|
|
2023-05-22 14:50:05 +01:00
|
|
|
|
2014-04-05 04:27:08 -04:00
|
|
|
|
2011-01-14 13:07:43 +01:00
|
|
|
|
2024-03-28 13:27:24 +01:00
|
|
|
|
2025-07-16 09:36:09 +00:00
|
|
|
|
|
|
|
|
|
2026-01-08 12:13:01 -05:00
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
|
|
|
|
|
2007-02-12 00:55:38 -08:00
|
|
|
|
2006-10-11 01:20:53 -07:00
|
|
|
|
2017-03-31 18:31:56 +01:00
|
|
|
|
2006-10-11 01:20:53 -07:00
|
|
|
|
2022-09-22 17:17:00 +02:00
|
|
|
|
2013-12-20 05:16:44 -08:00
|
|
|
|
2008-10-07 00:46:36 -04:00
|
|
|
|
2021-04-07 14:36:43 +02:00
|
|
|
|
|
|
|
|
|
2006-10-11 01:20:50 -07:00
|
|
|
|
|
|
|
|
|
