Currently mkfs would ignore any unknown extended option, which keeps
silent if a mistyped option is given. Return failure in this case allows
users to catch their errors more easily.
---
mkfs/main.c | 27 ---
1 file changed, 16 insertions(+), 11 deletions(-)
diff --git
: realpath_stk (canonicalize.c:409)
by 0x48FFE2B: realpath@@GLIBC_2.3 (canonicalize.c:431)
by 0x10B7EB: mkfs_parse_options_cfg (main.c:587)
by 0x10B7EB: main (main.c:936)
Fix it by freeing the memory allocated by strdup() and realpath().
Signed-off-by: Yifan Zhao
---
lib
.
Signed-off-by: Yifan Zhao
---
Change since v1:
1. Polish commit message.
2. Iterate the extent list in a wrapper function to avoid massive change
of indentation.
3. Remove the redundant reset_clusterofs field in struct z_erofs_inmem_extent.
Instead, we check that if an uncompressed
,
this patch moves the writing of the lcluster indexes to take place after
the file's compression is completed.
Signed-off-by: Yifan Zhao
---
include/erofs/dedupe.h | 4 +-
lib/compress.c | 366 -
lib/dedupe.c | 1 +
3 files changed
This patch introduce multi-thread compression to accelerate image
packaging.
---
Hi all:
This is a very imperfect patch not ready for merging, and any suggestions would
be appreciated!
If it's on track, I'd like to follow up on that.
The inefficiency of EROFS compressed image creation is a much
If tmpfs is used for /tmp and blob device is not specified, we need to
copy data between two different file systems during mkfs, which is not
supported by the copy_file_range() syscall. In this case, let's give it
a second chance by fallback to __erofs_copy_file_range().
Signed-off-by: Yifan Zhao
last, allowing it to use the compression level and dictsize already set
so that the behavior of the functions matches their names.
Signed-off-by: Yifan Zhao
---
lib/compressor.c| 8
lib/compressor_deflate.c| 30 +++---
lib/compressor_libdeflate.c
On 1/24/24 17:16, Gao Xiang wrote:
Yifan reported a "segmentation fault (core dumped)" error days ago
with a large dataset (enwik9 x 5). Let's fix it.
Reported-by: Yifan Zhao
Fixes: 861037f4fc15 ("erofs-utils: add a built-in DEFLATE compressor")
Signed-off-by: Gao
On 2024/1/24 17:53, Gao Xiang wrote:
Hi Yifan,
On 2024/1/24 17:47, Yifan Zhao wrote:
On 1/24/24 17:16, Gao Xiang wrote:
Yifan reported a "segmentation fault (core dumped)" error days ago
with a large dataset (enwik9 x 5). Let's fix it.
Reported-by: Yifan Zhao
Fixes: 86
Change log since v1:
- Re-implement workqueue API instead of using xfsprogs' workqueue
- Use per-worker tmpfile for multi-threaded mkfs
Gao Xiang (1):
erofs-utils: add a helper to get available processors
Yifan Zhao (6):
erofs-utils: introduce multi-threading framework
erofs-utils: mkfs
This patch introduces a --worker=# parameter for the incoming
multi-threaded compression support. It also introduces a segment size
used in multi-threaded compression, which has the default value 16MB
and cannot be modified.
Signed-off-by: Yifan Zhao
---
include/erofs/config.h | 4
lib
From: Gao Xiang
In order to prepare for multi-threaded decompression.
Signed-off-by: Gao Xiang
---
configure.ac | 1 +
include/erofs/config.h | 1 +
lib/config.c | 12
3 files changed, 14 insertions(+)
diff --git a/configure.ac b/configure.ac
index
Add a workqueue implementation for multi-threading support inspired by
xfsprogs.
Signed-off-by: Yifan Zhao
Suggested-by: Gao Xiang
---
configure.ac | 16 +
include/erofs/internal.h | 3 +
include/erofs/workqueue.h | 38 +++
lib/Makefile.am | 4 ++
lib
of erofs_write_compressed_file() has been
modified to split the creation and completion logic of the compression
task.
Signed-off-by: Yifan Zhao
Co-authored-by: Tong Xin
---
include/erofs/compress.h | 17 ++
include/erofs/internal.h | 3 +
include/erofs/list.h | 8 +
include/erofs/queue.h| 22 ++
lib
compresses a segment as if it were a separate
file. Finally, the main thread merges all the compressed segments.
Multi-threaded compression is not compatible with -Ededupe,
-E(all-)fragments and -Eztailpacking for now.
Signed-off-by: Yifan Zhao
Co-authored-by: Tong Xin
---
include/erofs/compress.h
() interface.
Signed-off-by: Yifan Zhao
---
lib/compress.c | 3 ++-
lib/compressor.c| 5 +++--
lib/compressor.h| 5 +++--
lib/compressor_deflate.c| 4 +++-
lib/compressor_libdeflate.c | 4 +++-
lib/compressor_liblzma.c| 5 -
lib/compressor_lz4.c| 2
the fallocate() syscall and
FALLOC_FL_PUNCH_HOLE flag.
Signed-off-by: Yifan Zhao
---
lib/compress.c | 68 +++---
1 file changed, 54 insertions(+), 14 deletions(-)
diff --git a/lib/compress.c b/lib/compress.c
index d5a5f16..3fae260 100644
--- a/lib/compress.c
by the same reason. Gao's patch successfully
fixes the problem that both Juhyung and I encountered in my machine.
Tested-by: Yifan Zhao
Thanks,
Yifan Zhao
On 2023/12/6 12:55, Gao Xiang wrote:
Currently EROFS can map another compressed buffer for inplace
decompression, that was used
be specified together with dictsize. The old -zX, way is still
supported for compatibility.
Suggested-by: Gao Xiang
Signed-off-by: Yifan Zhao
---
include/erofs/config.h | 10 ++--
include/erofs_fs.h | 4 ++
lib/compress.c | 45 +
lib/compress_hints.c | 2 +-
lib
last, allowing it to use the compression level and dictsize already set
so that the behavior of the functions matches their names.
Signed-off-by: Yifan Zhao
---
lib/compressor.c| 8
lib/compressor_deflate.c| 32
lib/compressor_libdeflate.c
it, let's report a warning rather than early aborting. Besides,
we do not need to assign the {default,best}_level for such an algorithm
in erofs_compressor struct.
Signed-off-by: Yifan Zhao
---
lib/compress.c | 6 +-
lib/compressor.c | 28 ++--
lib
Fix a misspelling in the version() function of mkfs.erofs.
Signed-off-by: Yifan Zhao
---
mkfs/main.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/mkfs/main.c b/mkfs/main.c
index 0517849..13fea41 100644
--- a/mkfs/main.c
+++ b/mkfs/main.c
@@ -175,7 +175,7 @@ static void
be specified together with dictsize. The old -zX, way is still
supported for compatibility.
Suggested-by: Gao Xiang
Signed-off-by: Yifan Zhao
---
change since v1:
- fix a bug in mkfs_parse_compress_algs() that "-zlzma,dictsize=" is not
correctly parsed as error.
include/erofs/config
of erofs_write_compressed_file() has been
modified to split the creation and completion logic of the compression
task.
Signed-off-by: Yifan Zhao
Co-authored-by: Tong Xin
---
include/erofs/compress.h | 17 ++
include/erofs/internal.h | 3 +
include/erofs/list.h | 8 +
include/erofs/queue.h| 22 ++
lib
compresses a segment as if it were a separate
file. Finally, the main thread merges all the compressed segments.
Multi-threaded compression is not compatible with -Ededupe,
-E(all-)fragments and -Eztailpacking for now.
Signed-off-by: Yifan Zhao
Co-authored-by: Tong Xin
---
include/erofs/compress.h
From: Gao Xiang
Add a workqueue implementation based on xfsprogs for multi-threading
support.
Signed-off-by: Gao Xiang
---
configure.ac | 16 +++
include/erofs/internal.h | 3 +
include/erofs/workqueue.h | 44
lib/Makefile.am | 4 +
lib/workqueue.c
() interface.
Signed-off-by: Yifan Zhao
---
lib/compress.c | 3 ++-
lib/compressor.c| 5 +++--
lib/compressor.h| 5 +++--
lib/compressor_deflate.c| 10 ++
lib/compressor_libdeflate.c | 6 --
lib/compressor_liblzma.c| 9 ++---
lib
Currently each worker thread is stateless. This patch allows us to bind
some per-worker fields to each worker thread, which could be reused
accross different tasks.
One of the examples is the compressor and buffer used in multi-threaded
compression.
Signed-off-by: Yifan Zhao
---
include/erofs
This patch introduces a --worker=# parameter for the incoming
multi-threaded compression support. It also introduces a segment size
used in multi-threaded compression, which has the default value 16MB
and cannot be modified.
Signed-off-by: Yifan Zhao
---
include/erofs/config.h | 4
lib
From: Gao Xiang
In order to prepare for multi-threaded decompression.
Signed-off-by: Gao Xiang
---
configure.ac | 1 +
include/erofs/config.h | 1 +
lib/config.c | 12
3 files changed, 14 insertions(+)
diff --git a/configure.ac b/configure.ac
index
| 183.26 | 4.55x |
Gao Xiang (2):
erofs-utils: introduce multi-threading framework
erofs-utils: add a helper to get available processors
Yifan Zhao (5):
erofs-utils: mkfs: add --worker=# parameter
erofs-utils: mkfs: optionally print warning in erofs_compressor_init
erofs
, reorganize and code style fix
Yifan Zhao (3):
erofs-utils: mkfs: merge erofs_compressor_setlevel() into
erofs_compressor_init()
erofs-utils: mkfs: allow to specify dictionary size for compression
algorithms
erofs-utils: mkfs: reorganize logic in erofs_compressor_init()
include/erofs/config.h
does not support the compression level in its
erofs_compressor struct.
Signed-off-by: Yifan Zhao
---
lib/compress.c | 6 +-
lib/compressor.c | 31 +--
lib/compressor.h | 5 ++---
lib/compressor_lz4.c | 2 --
4 files changed, 20 insertions(+), 24
together with dictsize. The old -zX, form is still
supported for compatibility.
Suggested-by: Gao Xiang
Signed-off-by: Yifan Zhao
---
include/erofs/config.h | 10 ++--
lib/compress.c | 33 -
lib/compress_hints.c | 2 +-
lib/compressor.c | 16 +-
lib
last, allowing it to use the compression level and dictsize already set
so that the behavior of the functions matches their names.
Signed-off-by: Yifan Zhao
---
lib/compressor.c| 8
lib/compressor_deflate.c| 30 +++---
lib/compressor_libdeflate.c
and compares the
sha256sum of the extracted files with the original ones.
Signed-off-by: Yifan Zhao
---
Hi forks,
I believe it's time to introduce a CI for erofs-utils, which could help
us to ensure the correctness of the code and avoid regressions.
Currently this simple CI covers only a limited
together with dictsize. The old -zX, form is still
supported for compatibility.
Suggested-by: Gao Xiang
Signed-off-by: Yifan Zhao
---
include/erofs/config.h | 10 ++--
lib/compress.c | 33 -
lib/compress_hints.c | 2 +-
lib/compressor.c | 16 +-
lib
This patchset allows to specify dictionary size for compression algorithms.
change since v3:
- remove the third patch as is unrelated to this patchset, will send as a
separate patch later
- Now patch 1 and 2 are able to be `git am` without any conflict
Yifan Zhao (2):
erofs-utils: mkfs
does not support the compression level in its
erofs_compressor struct.
Signed-off-by: Yifan Zhao
---
lib/compress.c | 6 +-
lib/compressor.c | 31 +--
lib/compressor.h | 5 ++---
lib/compressor_lz4.c | 2 --
4 files changed, 20 insertions(+), 24
the patch related to pring warning from this patchset, which may be
supported later with atomic variables
Gao Xiang (1):
erofs-utils: add a helper to get available processors
Yifan Zhao (3):
erofs-utils: introduce multi-threading framework
erofs-utils: mkfs: add --worker=# parameter
erofs
Add a workqueue implementation for multi-threading support inspired by
xfsprogs.
Signed-off-by: Yifan Zhao
Suggested-by: Gao Xiang
---
configure.ac | 16 +
include/erofs/internal.h | 3 +
include/erofs/workqueue.h | 37 +++
lib/Makefile.am | 4 ++
lib
From: Gao Xiang
In order to prepare for multi-threaded decompression.
Signed-off-by: Gao Xiang
---
configure.ac | 1 +
include/erofs/config.h | 1 +
lib/config.c | 12
3 files changed, 14 insertions(+)
diff --git a/configure.ac b/configure.ac
index
for multi-threading, which has the default value 16MB for now.
Signed-off-by: Yifan Zhao
---
include/erofs/config.h | 4
lib/config.c | 4
mkfs/main.c| 38 ++
3 files changed, 46 insertions(+)
diff --git a/include/erofs
compresses a segment as if it were a separate
file. Finally, the main thread merges all the compressed segments.
Multi-threaded compression is not compatible with -Ededupe,
-E(all-)fragments and -Eztailpacking for now.
Signed-off-by: Yifan Zhao
Co-authored-by: Tong Xin
---
include/erofs/compress.h
will be printed only once by
adding `__warnonce` atomic booleans to each erofs_compressor_init().
Cc: Yifan Zhao
Signed-off-by: Gao Xiang
---
include/erofs/atomic.h | 27 +++
lib/compressor_deflate.c| 11 ---
lib/compressor_libdeflate.c | 6 +-
lib
On 2/28/24 17:57, Gao Xiang wrote:
On 2024/2/25 22:27, Yifan Zhao wrote:
Currently, the creation of EROFS compressed image creation is
single-threaded, which suffers from performance issues. This patch
attempts to address it by compressing the large file in parallel.
Specifically, each
compresses a segment as if it were a separate
file. Finally, the main thread merges all the compressed segments.
Multi-threaded compression is not compatible with -Ededupe,
-E(all-)fragments and -Eztailpacking for now.
Signed-off-by: Yifan Zhao
Co-authored-by: Tong Xin
---
include/erofs/compress.h
Add a workqueue implementation for multi-threading support inspired by
xfsprogs.
Signed-off-by: Yifan Zhao
Suggested-by: Gao Xiang
---
configure.ac | 16 +
include/erofs/internal.h | 3 +
include/erofs/workqueue.h | 37 +++
lib/Makefile.am | 4 ++
lib
once by
adding `__warnonce` atomic booleans to each erofs_compressor_init().
Cc: Yifan Zhao
Signed-off-by: Gao Xiang
Signed-off-by: Yifan Zhao
---
include/erofs/atomic.h | 28
lib/compressor_deflate.c| 11 ---
lib/compressor_libdeflate.c | 6
inode_ctx
- leave tryrecompress_trailing() unchanged as it's only used in
ztailpacking scenario which is not supported yet
Gao Xiang (2):
erofs-utils: add a helper to get available processors
erofs-utils: lib: introduce atomic operations
Yifan Zhao (3):
erofs-utils: introduce multi-
From: Gao Xiang
In order to prepare for multi-threaded decompression.
Signed-off-by: Gao Xiang
---
configure.ac | 1 +
include/erofs/config.h | 1 +
lib/config.c | 12
3 files changed, 14 insertions(+)
diff --git a/configure.ac b/configure.ac
index
for multi-threading, which has the default value 16MB for now.
Signed-off-by: Yifan Zhao
---
include/erofs/config.h | 4
lib/config.c | 4
mkfs/main.c| 38 ++
3 files changed, 46 insertions(+)
diff --git a/include/erofs
On 2/29/24 17:43, Gao Xiang wrote:
Hi Yifan,
On 2024/2/29 00:16, Yifan Zhao wrote:
Add a workqueue implementation for multi-threading support inspired by
xfsprogs.
Signed-off-by: Yifan Zhao
Suggested-by: Gao Xiang
---
configure.ac | 16 +
include/erofs/internal.h
images for now.
(I couldn't apply v2 patch set without rejects.)
2024年2月25日(日) 23:28 Yifan Zhao :
We split the inter-file compression functionality as a separate patch
set for ease of review. It will be re-sent shortly (maybe a few days)
after we finish polishing the inner-file patch set.
I
, the logic of erofs_write_compressed_file() has been
modified to split the creation and completion logic of the compression
task.
Signed-off-by: Yifan Zhao
Co-authored-by: Tong Xin
---
include/erofs/compress.h | 16 ++
include/erofs/internal.h | 3 +
include/erofs/list.h | 8 +
include
This patch splits part of the logic in function erofs_mkfs_build_tree()
and erofs_mkfs_build_tree_from_path() into several new functions. This
is in preparation for the upcoming inter-file multi-threaded compression
feature.
Signed-off-by: Yifan Zhao
---
lib/inode.c | 161
This patchset introduces inter-file multi-threaded compression.
Yifan Zhao (2):
erofs-utils: lib: split function logic in inode.c
erofs-utils: mkfs: introduce inter-file multi-threaded compression
include/erofs/compress.h | 16 ++
include/erofs/internal.h | 3 +
include/erofs/list.h
compresses a segment as if it were a separate
file. Finally, the main thread merges all the compressed segments.
Multi-threaded compression is not compatible with -Ededupe,
-E(all-)fragments and -Eztailpacking for now.
Signed-off-by: Yifan Zhao
Co-authored-by: Tong Xin
---
include/erofs/compress.h
for multi-threading, which has the default value 16MB for now.
Signed-off-by: Yifan Zhao
---
include/erofs/config.h | 4
lib/config.c | 4
mkfs/main.c| 28
3 files changed, 36 insertions(+)
diff --git a/include/erofs/config.h b
Add a workqueue implementation for multi-threading support inspired by
xfsprogs.
Signed-off-by: Yifan Zhao
Suggested-by: Gao Xiang
---
configure.ac | 16 +
include/erofs/internal.h | 3 +
include/erofs/workqueue.h | 35 +++
lib/Makefile.am | 4 ++
lib
From: Gao Xiang
In order to prepare for multi-threaded decompression.
Signed-off-by: Gao Xiang
---
configure.ac | 1 +
include/erofs/config.h | 1 +
lib/config.c | 12
3 files changed, 14 insertions(+)
diff --git a/configure.ac b/configure.ac
index
once by
adding `__warnonce` atomic booleans to each erofs_compressor_init().
Cc: Yifan Zhao
Signed-off-by: Gao Xiang
Signed-off-by: Yifan Zhao
---
include/erofs/atomic.h | 28
lib/compressor_deflate.c| 11 ---
lib/compressor_libdeflate.c | 6
changelog since v5:
- use memory buffer instead of tmpfile to store intermediate compressed data
- re-strucutre the 5th patch
Gao Xiang (2):
erofs-utils: add a helper to get available processors
erofs-utils: lib: introduce atomic operations
Yifan Zhao (3):
erofs-utils: introduce multi
changelog since v1:
- apply fixes from Jianan's review
- remove misincluded changes in list.h
Yifan Zhao (2):
erofs-utils: lib: split function logic in inode.c
erofs-utils: mkfs: introduce inter-file multi-threaded compression
include/erofs/compress.h | 16 ++
include/erofs/internal.h
This patch splits part of the logic in function erofs_mkfs_build_tree()
and erofs_mkfs_build_tree_from_path() into several new functions. This
is in preparation for the upcoming inter-file multi-threaded compression
feature.
Signed-off-by: Yifan Zhao
---
lib/inode.c | 161
, the logic of erofs_write_compressed_file() has been
modified to split the creation and completion logic of the compression
task.
Signed-off-by: Yifan Zhao
Co-authored-by: Tong Xin
---
include/erofs/compress.h | 16 ++
include/erofs/internal.h | 3 +
include/erofs/queue.h| 22 +++
lib
Thanks for the review!
On 2024/3/21 2:15, Huang Jianan wrote:
On 2024/3/17 22:41, Yifan Zhao wrote:
This patch allows parallelizing the compression process of different
files in mkfs. Specifically, a traverser thread traverses the files and
issues the compression task, which is handled
On 2024/3/21 10:07, Gao Xiang wrote:
On 2024/3/21 02:15, Huang Jianan wrote:
On 2024/3/17 22:41, Yifan Zhao wrote:
This patch allows parallelizing the compression process of different
files in mkfs. Specifically, a traverser thread traverses the files and
issues the compression task
This patch splits part of the logic in function erofs_mkfs_build_tree()
and erofs_mkfs_build_tree_from_path() into several new functions. This
is in preparation for the upcoming inter-file multi-threaded compression
feature.
Signed-off-by: Yifan Zhao
---
lib/inode.c | 161
, the logic of erofs_write_compressed_file() has been
modified to split the creation and completion logic of the compression
task.
Signed-off-by: Yifan Zhao
Co-authored-by: Tong Xin
---
include/erofs/compress.h | 16 ++
include/erofs/inode.h| 17 ++
include/erofs/internal.h | 3 +
lib
change log since v2:
- erofs_queue => erofs_inode_fifo, defined in inode.c
Yifan Zhao (2):
erofs-utils: lib: split function logic in inode.c
erofs-utils: mkfs: introduce inter-file multi-threaded compression
include/erofs/compress.h | 16 ++
include/erofs/inode.h| 17 ++
include/er
If a segment is smaller than the block size, sizeof(sctx->membuf) should
be at least as large as the block size, as memory write into the buffer
is done in block size.
Signed-off-by: Yifan Zhao
---
lib/compress.c | 9 +
1 file changed, 5 insertions(+), 4 deletions(-)
diff --git a/
changelog since v4:
- Fix more format issues in inode.c reported by checkpatch.pl
- Add missing cfg.c_chunkbits in erofs_mkfs_issue_compress
Yifan Zhao (2):
erofs-utils: lib: split function logic in inode.c
erofs-utils: mkfs: introduce inter-file multi-threaded compression
include/erofs
, the logic of erofs_write_compressed_file() has been
modified to split the creation and completion logic of the compression
task.
Signed-off-by: Yifan Zhao
Co-authored-by: Tong Xin
---
include/erofs/compress.h | 16 ++
include/erofs/inode.h| 30
include/erofs/internal.h | 3 +
lib
This patch splits part of the logic in function erofs_mkfs_build_tree()
and erofs_mkfs_build_tree_from_path() into several new functions. This
is in preparation for the upcoming inter-file multi-threaded compression
feature.
Signed-off-by: Yifan Zhao
---
lib/inode.c | 161
On 2024/3/23 10:51, Huang Jianan wrote:
Hi, Yifan,
I got some warnings and errors via checkpatch.pl on this patchset. You can
check and fix them since erofs-utils follows kernel coding style.
Roger that. Will do it in next patches.
Thanks,
Yifan Zhao
Thanks,
Jianan
Yifan Zhao 于2024年
, the logic of erofs_write_compressed_file() has been
modified to split the creation and completion logic of the compression
task.
Signed-off-by: Yifan Zhao
Co-authored-by: Tong Xin
---
include/erofs/compress.h | 16 ++
include/erofs/inode.h| 30
include/erofs/internal.h | 3 +
lib
This patch splits part of the logic in function erofs_mkfs_build_tree()
and erofs_mkfs_build_tree_from_path() into several new functions. This
is in preparation for the upcoming inter-file multi-threaded compression
feature.
Signed-off-by: Yifan Zhao
---
lib/inode.c | 161
On 2024/3/23 11:46, Huang Jianan wrote:
Yifan Zhao 于2024年3月22日周五 18:25写道：
This patch allows parallelizing the compression process of different
files in mkfs. Specifically, a traverser thread traverses the files and
issues the compression task, which is handled by the workers. Then, the
main
change log since v2:
- Give erofs_*_inode_fifo functions empty implementation if !EROFS_MT_ENABLED
- Fix format issue reported by checkpatch.pl
- rename: z_erofs_mt_queue => z_erofs_mt_inode_fifo
Yifan Zhao (2):
erofs-utils: lib: split function logic in inode.c
erofs-utils: mkfs: introd
` modifier is removed in the first version of the
patchset, but the change
is reversed during the review. I think Xiang may share his opinion about
this.
Thanks,
Yifan Zhao
On 4/8/24 5:16 PM, Noboru Asai wrote:
In multi-threaded mode, each thread must use a different buffer
z_erofs_merge_segment() doesn't consider the ztailpacking block in the
extent list and unnecessarily writes it back to the disk. This patch
fixes this issue by introducing a new `inlined` field in the struct
`z_erofs_inmem_extent`.
Signed-off-by: Yifan Zhao
---
include/erofs/dedupe.h | 2
int err;
+
+ q = malloc(sizeof(*q));
+ if (!q)
+ return ERR_PTR(-ENOMEM);
+
+ q->queue = malloc(q->size * sizeof(*q->queue));
q->size is not initialized here. Should move `q->size =
EROFS_MT_QUEUE_SIZE;` before it.
Th
Currently fsck skips file extraction if it finds that EROFS_MAP_MAPPED
is unset, which is not the case for chunk-based files with hole. This
patch handles the corner case correctly.
Signed-off-by: Yifan Zhao
---
fsck/main.c | 13 +
1 file changed, 9 insertions(+), 4 deletions
Currently fsck skips file extraction if it finds that EROFS_MAP_MAPPED
is unset, which is not the case for chunk-based files with hole. This
patch handles the corner case correctly.
Signed-off-by: Yifan Zhao
---
changelog since v1:
- use lseek instead of write zero to the hole
fsck/main.c | 11
_init(struct erofs_sb_info *sbi,
struct erofs_buffer_head *s
#ifdef EROFS_MT_ENABLED
if (cfg.c_mt_workers > 1) {
pthread_mutex_init(_erofs_mt_ctrl.mutex, NULL);
Remove this line too.
Thanks,
Yifan Zhao
- pthread_cond_init(_erofs_mt_ctrl.cond, NULL);
ret = erofs_alloc_workqueue(_erofs_mt_ctrl.wq,
cfg.c_mt_workers,
cfg.c_mt_workers << 2,
line above does not exist in the logic replaced by
`erofs_mkfs_jobfn`, should it appear in this patch, or need further
explanation in the commit msg?
Thanks,
Yifan Zhao
+ return 0;
+ }
+
+ if (item->type == EROFS_MKFS_JOB_DIR_BH) {
+ erofs_wr
is
irrelevant and we don't need to care about it.
For the mkfs side I think things could be polished according to Gao's
review opinion.
Thanks,
Yifan Zhao
2024年4月17日(水) 23:43 Yifan Zhao :
z_erofs_merge_segment() doesn't consider the ztailpacking block in the
extent list and unnecessarily
e based on current
+ * compression ratio.
+*/
+ m = (dstsize * m) / csize;
+ } else {
+doesnt_fit:
+ /* Doesn't fit */
+ r = m;
+
89 matches
Mail list logo
