`pageofs_in` should be the compressed data offset of the page rather
than of the block.
Signed-off-by: Gao Xiang
---
fs/erofs/zdata.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index a2c3e87d2f81..8264936b8612 100644
--- a/fs/erofs
} else {
+#if FUSE_MINOR_VERSION < 2
Can we check FUSE_USE_VERSION directly?
if FUSE_USE_VERSION is neither 30 or 32, let's raise a build error with
#error unsupported FUSE_USE_VERSION
instead.
Thanks,
Gao Xiang
+
On 2023/12/12 15:35, Christoph Hellwig wrote:
On Tue, Dec 12, 2023 at 08:50:56AM +0800, Gao Xiang wrote:
For non-virtualization cases, I guess you could try to use `memmap`
kernel option [2] to specify a memory region by bootloaders which
contains an EROFS rootfs and a customized init
_metabuf() is not always called before erofs_bread() is used.
For example, buf->inode can be one of directory inodes other than
initialized by erofs_init_metabuf().
Thanks,
Gao Xiang
pgoff_t index = offset >> PAGE_SHIFT;
struct page *page = buf->page;
struct
ndividual filesystem anyway.
[1] https://www.qemu.org/docs/master/system/devices/virtio-pmem.html
[2]
https://docs.pmem.io/persistent-memory/getting-started-guide/creating-development-environments/linux-environments/linux-memmap
Thanks,
Gao Xiang
Echo'ing Lennart we must also "keep in mi
On 2023/12/6 17:10, Gao Xiang wrote:
Add a basic I/O submission path first to support sub-page blocks:
- Temporary short-lived pages will be used entirely;
- In-place I/O pages can be used partially, but compressed pages need
to be able to be mapped in contiguous virtual memory
On 2023/12/8 15:34, Gao Xiang wrote:
...
min_t(unsigned int, ,)?
../include/linux/minmax.h:21:28: error: comparison of distinct pointer types
lacks a cast [-Werror]
(!!(sizeof((typeof(x) *)1 == (typeof(y) *)1)))
What compiler version are you using? I didn't find any error
On 2023/12/8 13:20, Yue Hu wrote:
On Wed, 6 Dec 2023 17:10:56 +0800
Gao Xiang wrote:
Sub-page block support is still unusable even with previous commits if
interlaced PLAIN pclusters exist. Such pclusters can be found if the
fragment feature is enabled.
This commit tries to handle
Let's just disable cached decompression and inplace I/Os for partial
pages as a first step in order to enable sub-page block initial
support. In other words, currently it works primarily based on
temporary short-lived pages. Don't expect too much in terms of
performance.
Signed-off-by: Gao
ffc091e6f9 ("erofs:
support interlaced uncompressed data for compressed files").
It uses a unique way for both shifted and interlaced PLAIN pclusters.
As an added bonus, PLAIN pclusters larger than the block size is also
supported now for the upcoming large lclusters.
Signed-off-by: Gao
is explicitly disabled for
sub-page blocks, which will be supported in the future.
Signed-off-by: Gao Xiang
---
fs/erofs/zdata.c | 156 ++-
1 file changed, 74 insertions(+), 82 deletions(-)
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index a33cd6757f98
`Z_EROFS_LI_D0_CBLKCNT`.
To support smaller block sizes, `lobits` for each compacted lcluster is
now calculated as:
lobits = max(lclusterbits, ilog2(Z_EROFS_LI_D0_CBLKCNT) + 1)
Signed-off-by: Gao Xiang
---
fs/erofs/zmap.c | 32 ++--
1 file changed, 14 insertions
temporary
buffers all the time to keep compressed data if sub-page blocks are
identified. In other words, no inplace/cache decompression for
the preliminary support.
This patchset survives EROFS stress test on my own testfarms.
Thanks,
Gao Xiang
Gao Xiang (5):
erofs: support I/O submission
-by: Gao Xiang
---
fs/erofs/zdata.c | 64
1 file changed, 26 insertions(+), 38 deletions(-)
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index 421c0a88a0ca..d02989466711 100644
--- a/fs/erofs/zdata.c
+++ b/fs/erofs/zdata.c
@@ -56,6 +56,9 @@ struct
ompress big pcluster for lz4 backend")
Cc: stable # 5.4+
Signed-off-by: Gao Xiang
---
changes since v1:
- address some nits pointed out by Juhyung.
fs/erofs/decompressor.c | 31 ---
1 file changed, 16 insertions(+), 15 deletions(-)
diff --git a/fs/erofs/decompr
On 2023/12/6 12:39, Gao Xiang wrote:
On 2023/12/6 12:28, Juhyung Park wrote:
Hi Gao,
Thanks for the timely fix.
This also fixes the issue.
Reported-and-tested-by: Juhyung Park
Might wanna add
Cc: sta...@vger.kernel.org # 5.4+
but it seems like this doesn't apply cleanly for 5.4, 5.10
conflict resolution. :))
I will handle these LTSes manually, so don't worry about that :)
Wrote some nits below:
On Wed, Dec 6, 2023 at 12:08 PM Gao Xiang wrote:
Currently EROFS can map another compressed buffer for inplace
decompression, which was used to handle the cases that some pages
Hi Juhyung,
On 2023/12/5 22:43, Juhyung Park wrote:
On Tue, Dec 5, 2023 at 11:34 PM Gao Xiang wrote:
...
I'm still analyzing this behavior as well as the root cause and
I will also try to get a recent cloud server with FSRM myself
to find more clues.
Down the rabbit hole we go...
Let
On 2023/12/5 22:23, Juhyung Park wrote:
Hi Gao,
On Tue, Dec 5, 2023 at 4:32 PM Gao Xiang wrote:
Hi Juhyung,
On 2023/12/4 11:41, Juhyung Park wrote:
...
- Could you share the full message about the output of `lscpu`?
Sure:
Architecture:x86_64
CPU op-mode(s
startup latency
Thanks,
Gao Xiang
t of `lscpu`?
Thanks,
Gao Xiang
On 2023/12/4 01:01, Juhyung Park wrote:
Hi Gao,
On Mon, Dec 4, 2023 at 1:52 AM Gao Xiang wrote:
Hi Juhyung,
On 2023/12/4 00:22, Juhyung Park wrote:
(Cc'ing f2fs and crypto as I've noticed something similar with f2fs a
while ago, which may mean that this is not specific to EROFS:
https
nce like different compliers?
Thanks,
Gao Xiang
k like this, so it won't be impacted.
Fixes: 67139e36d970 ("erofs: introduce `z_erofs_parse_in_bvecs'")
Signed-off-by: Gao Xiang
---
v2:
- should be `be->compressed_pages`.
fs/erofs/zdata.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/fs/erofs/zdata.c b/fs/
k like this, so it won't be impacted.
Fixes: 67139e36d970 ("erofs: introduce `z_erofs_parse_in_bvecs'")
Signed-off-by: Gao Xiang
---
fs/erofs/zdata.c | 5 ++---
1 file changed, 2 insertions(+), 3 deletions(-)
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index a7e6847f6f8f..e5e
On 2023/11/24 13:30, Yue Hu wrote:
From: Yue Hu
No need to track these generated test files.
Signed-off-by: Yue Hu
---
.gitignore | 3 +++
1 file changed, 3 insertions(+)
diff --git a/.gitignore b/.gitignore
index 33e5d30..5b9ecd0 100644
--- a/.gitignore
+++ b/.gitignore
@@ -31,3
t; %d)\n",
~^
%ld
Signed-off-by: Yue Hu
Thanks, folded this with your SoB.
Thanks,
Gao Xiang
On 2023/11/24 11:50, Yue Hu wrote:
On Fri, 24 Nov 2023 11:46:53 +0800
Gao Xiang wrote:
Hi Yue,
On 2023/11/24 11:36, Yue Hu wrote:
From: Yue Hu
badlz4.c:72:58: warning: format ‘%d’ expects argument of type ‘int’, but
argument 4 has type ‘long unsigned int’ [-Wformat=]
printf("
cSize <= DstSize) {
printf("test LZ4_compress_HC_destSize(%d) error (%d < %d)\n",
- inlen, SrcSize, sizeof(dst));
Could we just use printf(, (int)sizeof(dst)); instead?
Thanks,
Gao Xiang
+ inlen, SrcSize, DstSize);
re
Hu
Thanks, folded this with your SoB.
Thanks,
Gao Xiang
indexes after the compression for the file is
completed.
Signed-off-by: Yifan Zhao
---
Sorry for the late reply. There are some issues on my side.
In order to make this work efficient. Let me to refine this commit as well.
Thanks,
Gao Xiang
, there is no enough room for
the special value `Z_EROFS_LI_D0_CBLKCNT`.
To support smaller block sizes, `lobits` for each compacted lcluster is
now calculated as:
lobits = max(lclusterbits, ilog2(Z_EROFS_LI_D0_CBLKCNT) + 1)
Signed-off-by: Gao Xiang
---
lib/zmap.c | 32 ++--
1 file
This commit also adds mkfs support of compact indexes for smaller
block sizes (less than 4096).
Signed-off-by: Gao Xiang
---
lib/compress.c | 15 ---
1 file changed, 8 insertions(+), 7 deletions(-)
diff --git a/lib/compress.c b/lib/compress.c
index 47f1c1d..61328ed 100644
--- a/lib
s have been in -next and no potential merge conflict is
observed.
Thanks,
Gao Xiang
The following changes since commit b85ea95d086471afb4ad062012a4d73cd328fa86:
Linux 6.7-rc1 (2023-11-12 16:19:07 -0800)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git
On 2023/11/20 11:31, Chao Yu wrote:
On 2023/11/20 11:23, Gao Xiang wrote:
On 2023/11/20 11:18, Chao Yu wrote:
On 2023/11/17 16:53, Gao Xiang wrote:
Add a new `W:` field of the EROFS entry points to the documentation
site at <https://erofs.docs.kernel.org>.
In addition,
On 2023/11/20 11:18, Chao Yu wrote:
On 2023/11/17 16:53, Gao Xiang wrote:
Add a new `W:` field of the EROFS entry points to the documentation
site at <https://erofs.docs.kernel.org>.
In addition, update the in-tree documentation and Kconfig too.
Signed-off-by: Gao Xiang
Nic
Add a new `W:` field of the EROFS entry points to the documentation
site at <https://erofs.docs.kernel.org>.
In addition, update the in-tree documentation and Kconfig too.
Signed-off-by: Gao Xiang
---
Documentation/filesystems/erofs.rst | 4
MAINTAINERS | 1
`-Eall-fragments` is broken if i_size is more than 32 bits.
Fixes: fcaa988a6ef6 ("erofs-utils: add `-Eall-fragments` option")
Signed-off-by: Gao Xiang
---
include/erofs/internal.h | 12
lib/compress.c | 9 +
2 files changed, 13 insertions(+), 8 deletion
2/0xf0
entry_SYSCALL_64_after_hwframe+0x6e/0x76
Reported-by: Yiqun Leng
Closes: https://bugzilla.openanolis.cn/show_bug.cgi?id=7245
Fixes: 49845720080d ("erofs: Convert to use bdev_open_by_path()")
Signed-off-by: Jingbo Xu
Reviewed-by: Gao Xiang
Thanks,
Gao Xiang
his up.
Assuming that is for the next cycle (no rush), I will also test
this patch and feedback later since I'm now working on other
stuffs.
Thanks,
Gao Xiang
y: Ferry Meng
Reviewed-by: Gao Xiang
Thanks,
Gao Xiang
On 2023/11/9 21:45, Zizhi Wo wrote:
在 2023/11/9 21:14, Gao Xiang 写道:
Hi,
On 2023/11/10 03:48, WoZ1zh1 wrote:
Because variables "die" and "copied" only appear in case
EROFS_INODE_LAYOUT_EXTENDED, move them from the outer space into this
case. Also, call "kfree(
the code
2) We'd like to avoid defining variables like this (in the
switch block), and I even don't think this patch can compile.
3) The logic itself is also broken...
Thanks,
Gao Xiang
+
vi->inode_isize = sizeof(struct erofs_inode_extended);
/* check if
UPP;
goto err_out;
}
- if (vi->datalayout == EROFS_INODE_CHUNK_BASED) {
+ switch (inode->i_mode & S_IFMT) {
+ case S_IFREG:
+ case S_IFDIR:
+ case S_IFLNK:
+ vi->raw_blkaddr = le32_to_cpu(iu->raw_blkaddr);
may cause use-after-free for acrossing block boundary case due to kfree(copied)
before?
Yeah, good catch, I think we could just make a copy instead.
Thanks,
Gao Xiang
Add `--offset` option to allows users to specify an offset in the file
where the filesystem will begin.
Suggested-by: Pavel Otchertsov
Closes:
https://lore.kernel.org/r/caaxntogtd2nkknbphz+ver7nvnwvt0u02e+c8pn8zvfcxp5...@mail.gmail.com
Signed-off-by: Gao Xiang
---
I plan to apply this version
Add `--offset` option to allows users to specify an offset in the file
where the filesystem will begin.
Suggested-by: Pavel Otchertsov
Closes:
https://lore.kernel.org/r/caaxntogtd2nkknbphz+ver7nvnwvt0u02e+c8pn8zvfcxp5...@mail.gmail.com
Signed-off-by: Gao Xiang
---
Hi Pavel,
Could you check
mation or
assist in any way.
Thanks for the reference! Let's implement this feature.
Thanks,
Gao Xiang
Best regards,
Pavel Otchertsov mailto:pavel.otchert...@gmail.com>>
.
Thanks, applied for testing.
Thanks,
Gao Xiang
Luke T. Shumaker (3):
erofs-utils: have each non-fuse command take -h, --help, -V, and
--version
erofs-utils: improve the usage and version text of non-fuse commands
erofs-utils: fsck: Add -a, -A, and -y flags
dump/main.c
() and realpath().
Signed-off-by: Yifan Zhao
Reviewed-by: Gao Xiang
BTW, would you mind adding a configuration to enable Valgrind?
Thanks,
Gao Xiang
, and there are
minor cleanups around documentation and codebase. In addition, it
also fixes incorrect lockref usage in erofs_insert_workgroup().
All commits have been in -next and no potential merge conflict is
observed.
Thanks,
Gao Xiang
The following changes since commit
On 2023/10/31 14:20, Linus Torvalds wrote:
On Mon, 30 Oct 2023 at 20:08, Gao Xiang wrote:
As Linus pointed out [1], lockref_put_return() is fundamentally
designed to be something that can fail. It behaves as a fastpath-only
thing, and the failure case needs to be handled anyway.
Actually
place handcrafted
approach")
Signed-off-by: Gao Xiang
---
fs/erofs/utils.c | 8 +---
fs/erofs/zdata.c | 1 +
2 files changed, 2 insertions(+), 7 deletions(-)
diff --git a/fs/erofs/utils.c b/fs/erofs/utils.c
index cc6fb9e98899..4256a85719a1 100644
--- a/fs/erofs/utils.c
+++ b/fs/ero
On 2023/10/27 15:06, Gao Xiang wrote:
Drop the old workaround logic to prepare for the following development.
(I've checked the Linux 6.1.53 source code and an AOSP rootfs without
any image size change and strange behavior.)
Signed-off-by: Gao Xiang
I found that It could cause
Drop the old workaround logic to prepare for the following development.
(I've checked the Linux 6.1.53 source code and an AOSP rootfs without
any image size change and strange behavior.)
Signed-off-by: Gao Xiang
---
lib/compress.c | 71 +++-
lib
On 2023/10/26 10:16, Ferry Meng wrote:
- Remove unused includes like and ;
- Move common includes into "internal.h".
Signed-off-by: Ferry Meng
Reviewed-by: Gao Xiang
Thanks,
Gao Xiang
On 2023/10/26 10:16, Ferry Meng wrote:
Let's open code this helper for simplicity.
Signed-off-by: Ferry Meng
Reviewed-by: Gao Xiang
Thanks,
Gao Xiang
Instead of just using a boolean.
Signed-off-by: Gao Xiang
---
include/erofs/cache.h | 8
lib/cache.c | 21 ++---
lib/compress.c| 1 +
lib/inode.c | 18 +-
mkfs/main.c | 14 +++---
5 files changed, 35
It's actually never used.
Signed-off-by: Gao Xiang
---
include/erofs/cache.h | 1 -
lib/cache.c | 5 -
2 files changed, 6 deletions(-)
diff --git a/include/erofs/cache.h b/include/erofs/cache.h
index de5584e..31f54a4 100644
--- a/include/erofs/cache.h
+++ b/include/erofs/cache.h
Instead of just using a boolean.
Signed-off-by: Gao Xiang
---
include/erofs/cache.h | 8
lib/cache.c | 21 ++---
lib/inode.c | 18 +-
mkfs/main.c | 14 +++---
4 files changed, 34 insertions(+), 27 deletions(-)
diff
Also avoid division in BLK_ROUND_UP().
Signed-off-by: Gao Xiang
---
include/erofs/internal.h | 3 ++-
lib/cache.c | 15 ---
2 files changed, 10 insertions(+), 8 deletions(-)
diff --git a/include/erofs/internal.h b/include/erofs/internal.h
index c1ff582..4d794ae 100644
From: Gao Xiang
Move erofs_load_compr_cfgs() into decompressor.c as well as introduce
a callback instead of a hard-coded switch for each algorithm for
simplicity.
Signed-off-by: Gao Xiang
---
v2:
- fix build warning of `-Wmissing-declarations`
https://lore.kernel.org/r
The LZMA algorithm support has been landed for more than one year since
Linux 5.16. Besides, the new XZ Utils 5.4 has been available in most
Linux distributions.
Signed-off-by: Gao Xiang
---
fs/erofs/Kconfig | 7 ++-
fs/erofs/decompressor_lzma.c | 2 --
2 files changed, 2
Move erofs_load_compr_cfgs() into decompressor.c as well as introduce
a callback instead of a hard-coded switch for each algorithm for
simplicity.
Signed-off-by: Gao Xiang
---
fs/erofs/compress.h | 6 +++
fs/erofs/decompressor.c | 60 ++-
fs/erofs
commit 8cbc205185a1 ("erofs-utils: mkfs: fix corrupted directories
with hardlinks") introduced a reference leak although it has no real
impact to users. Fix it now.
Signed-off-by: Gao Xiang
---
lib/inode.c | 13 +
1 file changed, 9 insertions(+), 4 deletions(-)
diff -
adata")
Signed-off-by: Gao Xiang
---
include/erofs/internal.h | 5 +++--
lib/inode.c | 29 -
2 files changed, 15 insertions(+), 19 deletions(-)
diff --git a/include/erofs/internal.h b/include/erofs/internal.h
index d859905..c1ff582 100644
--- a/inc
Hi Chao,
On 2023/10/12 09:10, Chao Yu wrote:
On 2023/10/10 21:06, Gao Xiang wrote:
Signed-off-by: Tiwei Bie
Reviewed-by: Gao Xiang
Looks fine to me for the version in dev-test branch.
Reviewed-by: Chao Yu
Yeah, thanks! I will add the tag when applying to -next.
Thanks,
Gao Xiang
ument")
I'm not sure if it's necessary to tag document fixes anyway
since docs.kernel.org already uses the latest version and
`.rst` format is adapted much later..
I will drop this tag for the next merge window if not urgent.
Signed-off-by: Tiwei Bie
Reviewed-by: Gao Xiang
Thanks,
_VERSION >= 3
+ fputs("\nFUSE options:\n", stderr);
fuse_cmdline_help();
#else
+ struct fuse_args args = FUSE_ARGS_INIT(0, NULL);
It's not a good idea to mix definitions and the rest code.
I will fix manually.
Thanks,
Gao Xiang
On 2023/10/6 02:43, Gao Xiang wrote:
Hi Erik,
On Mon, Oct 02, 2023 at 07:36:08PM +0200, Erik Sjölund wrote:
Hi,
Does this patch make sense?
(I thought errno should be set to a non-negative value)
Best regards,
Erik Sjölund
Thanks for the patch.
I'm on vacation, sorry for late reply
On 2023/10/7 15:16, Sandeep Dhavale wrote:
On Fri, Oct 6, 2023 at 8:27 PM Gao Xiang wrote:
Hi Sandeep!
On 2023/10/6 06:40, Sandeep Dhavale wrote:
AC_RUN_IFELSE expects the action if cross compiling. If not provided
cross compilation fails with error "configure: error: cannot run
On 2023/9/30 13:00, Wedson Almeida Filho wrote:
From: Wedson Almeida Filho
This makes it harder for accidental or malicious changes to
erofs_xattr_handlers or xattr_handler_map at runtime.
Cc: Gao Xiang
Cc: Chao Yu
Cc: Yue Hu
Cc: Jeffle Xu
Cc: linux-erofs@lists.ozlabs.org
Signed-off
_block_size=$MAX_BLOCK_SIZE])
Actually the following check will reset erofs_cv_max_block_size to 4096
if needed. But it seems that it has syntax errors.
I wonder if the following diff could fix the issue too?
Thanks,
Gao Xiang
diff --git a/configure.ac b/configure.ac
index 13ee616..94eec01 1006
for a while and no potential merge
conflict is observed.
Thanks,
Gao Xiang
The following changes since commit 0bb80ecc33a8fb5a682236443c1e740d5c917d1d:
Linux 6.6-rc1 (2023-09-10 16:28:41 -0700)
are available in the Git repository at:
git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs.git
I will address it when I'm back.
Thanks,
Gao Xiang
>
> diff --git a/lib/tar.c b/lib/tar.c
> index 0744972..8204939 100644
> --- a/lib/tar.c
> +++ b/lib/tar.c
> @@ -241,7 +241,7 @@ static long long tarerofs_otoi(const char *ptr, int len)
> val = strtol(ptr, , 8)
On 2023/9/28 21:48, Jingbo Xu wrote:
- update new features like bloom filter and DEFLATE.
- add documentation for the long xattr name prefixes, which was
landed upstream since v6.4.
Signed-off-by: Jingbo Xu
Reviewed-by: Gao Xiang
Thanks,
Gao Xiang
efix is stored in the format: ALIGN({__le16 len, data}, 4),
> +where len represents the total size of the data part. The data part is
> actually
> +represented by 'struct erofs_xattr_long_prefix', where base_index represents
> the
> +index of the predefined xattr name prefix, e.g. EROFS_XATTR_INDEX_TRUSTED for
> +"trusted.overlay." long name prefix, while the infix string kepts the string
^ keeps
Thanks,
Gao Xiang
People care more about speed when building. Later, we could also
try to use file-backed memory to avoid memory pressure.
As a result, it could decrease time by 50% (0m58.255s -> 0m29.120s).
Signed-off-by: Gao Xiang
---
lib/dedupe.c | 15 +--
1 file changed, 5 insertions(+),
Let's check if xxh64 equals when rolling back on global compressed
deduplication.
As a result, it could decrease time by 26.4% (6m57.990s -> 5m7.755s)
on a dataset with "-Ededupe -C8192".
Signed-off-by: Gao Xiang
---
include/erofs/defs.h | 5 +++
include/erofs/
This rb-tree implementation is too slow and there is no benefits of it.
As a result, it could decrease time by 81.1% (5m7.755s -> 0m58.255s)
with the same dataset, sigh.
Signed-off-by: Gao Xiang
---
lib/Makefile.am | 2 +-
lib/dedupe.c| 116 +--
lib/rb_tree.c |
rnel jobs). It's always worth trying, and any contribution is much
appreciated.
Thanks,
Gao Xiang
[1] https://lore.kernel.org/r/20230918090306.2524624-1-lyy0...@sjtu.edu.cn
ed.
Fixes: f3f9a2ce3137 ("erofs-utils: mkfs: introduce global compressed data
deduplication")
Signed-off-by: Gao Xiang
---
lib/compress.c | 25 ++---
1 file changed, 18 insertions(+), 7 deletions(-)
diff --git a/lib/compress.c b/lib/compress.c
index 81f277a..f6dc12a 100644
From: Gao Xiang
Actually, it's false positive and only used for build assertion.
Reported-by: Kelvin Zhang
Signed-off-by: Gao Xiang
---
include/erofs_fs.h | 14 --
1 file changed, 8 insertions(+), 6 deletions(-)
diff --git a/include/erofs_fs.h b/include/erofs_fs.h
index
Don't output more than 255 shared xattrs for a single inode due to the
EROFS on-disk format limitation.
Fixes: 116ac0a254fc ("erofs-utils: introduce shared xattr support")
Signed-off-by: Gao Xiang
---
lib/xattr.c | 8 ++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff -
Don't output more than 255 shared xattrs for a single inode due to the
EROFS on-disk format limitation.
Fixes: 116ac0a254fc ("erofs-utils: introduce shared xattr support")
Signed-off-by: Gao Xiang
---
lib/xattr.c | 8 ++--
1 file changed, 6 insertions(+), 2 deletions(-)
diff -
-b block-size
-E ^xattr-name-filter
--gzip
--tar=[fi]
--xattr-prefix=X
Signed-off-by: Gao Xiang
---
man/mkfs.erofs.1 | 23 +++
1 file changed, 23 insertions(+)
diff --git a/man/mkfs.erofs.1 b/man/mkfs.erofs.1
index 1cfde28..5dd718d 100644
--- a/man/mkfs.erofs.1
+++ b
On Wed, Sep 20, 2023 at 02:59:47AM +0800, Gao Xiang wrote:
> Introduce iostream to wrap up the input tarball stream for tarerofs.
>
> Besides, add bultin tgz support if zlib is linked to mkfs.
>
> Signed-off-by: Gao Xiang
Apply the following diff to fix pax header regression:
d
Hi Daan,
On 2023/8/31 17:48, Gao Xiang wrote:
Hi Daan,
On 2023/8/31 17:33, Daan De Meyer wrote:
Hi,
For hacking on systemd, we build disk images using mkosi, and use an erofs
filesystem for the /usr directory. When hacking on systemd, we would like to be
able to rebuild the disk image
Temporarily move the common part into __erofs_fill_inode() for tarerofs.
Signed-off-by: Gao Xiang
---
include/erofs/inode.h | 2 ++
lib/inode.c | 19 ---
lib/tar.c | 16 ++--
3 files changed, 28 insertions(+), 9 deletions(-)
diff --git
GNU tar volume labels (by using `-V`) will be applied to EROFS.
Signed-off-by: Gao Xiang
---
lib/tar.c | 24 ++--
1 file changed, 18 insertions(+), 6 deletions(-)
diff --git a/lib/tar.c b/lib/tar.c
index 8e71f11..1cd820f 100644
--- a/lib/tar.c
+++ b/lib/tar.c
@@ -17,6 +17,9
From: Gao Xiang
Previously, each tar data file will be kept as a temporary file before
landing to the target image since the input stream may be non-seekable.
It's somewhat ineffective. Let's introduce a new diskbuf approach to
manage those buffers. Laterly, each stream can be redirected
On 2023/9/20 11:35, Gao Xiang wrote:
On 2023/9/20 11:32, Jingbo Xu wrote:
On 9/15/23 11:29 PM, Gao Xiang wrote:
Previously, each tar data file will be kept as a temporary file before
landing to the target image since the input stream may be non-seekable.
It's somewhat ineffective
On 2023/9/20 11:32, Jingbo Xu wrote:
On 9/15/23 11:29 PM, Gao Xiang wrote:
Previously, each tar data file will be kept as a temporary file before
landing to the target image since the input stream may be non-seekable.
It's somewhat ineffective. Let's introduce a new diskbuf approach
ply like this directly.
Thanks,
Gao Xiang
inode->chunkindexes = NULL;
return ret;
Introduce iostream to wrap up the input tarball stream for tarerofs.
Besides, add bultin tgz support if zlib is linked to mkfs.
Signed-off-by: Gao Xiang
---
include/erofs/tar.h | 21 +++
lib/tar.c | 309
mkfs/main.c | 34 -
3
to
apply it after erofs-utils v1.7 is released.
Thanks,
Gao Xiang
: Structure needs cleaning
[xize.gx@e69b19392.et15sqa /home/xize.gx/erofs-utils]
$fssum -MACUG mnt
stat failed for mnt//Documentation/.gitignore: Structure needs cleaning
[xize.gx@e69b19392.et15sqa /home/xize.gx/erofs-utils]
$fssum -MACUG mnt
stat failed for mnt//LICENSES/deprecated: Structure needs cleaning
[xize.gx@e69b19392.et15sqa /home/xize.gx/erofs-utils]
$fssum -MACUG mnt
stat failed for mnt//drivers/staging/most: Structure needs cleaning
[xize.gx@e69b19392.et15sqa /home/xize.gx/erofs-utils]
$fssum -MACUG mnt
stat failed for mnt//drivers/staging/most: Structure needs cleaning
[xize.gx@e69b19392.et15sqa /home/xize.gx/erofs-utils]
$fssum -MACUG mnt
stat failed for mnt//tools/testing/selftests/drivers/net/bonding/settings:
Structure needs cleaning
I don't think it's a correct sign. I try to debug more but
it seems more issues here, I gave up eveuatually.
Thanks,
Gao Xiang
for external reference.
Signed-off-by: Gao Xiang
---
include/erofs/diskbuf.h | 30
include/erofs/internal.h | 8 ++-
lib/Makefile.am | 3 +-
lib/diskbuf.c| 147 +++
lib/inode.c | 33 +
lib/io.c
vs[src->dev - 1].blocks = nblocks;
+ idx = src->dev - 1;
+ sbi.devs[idx].blocks = nblocks;
+ memcpy(sbi.devs[idx].tag, src->uuid, sizeof(src->uuid));
Should we use string conversion instead? Should we use uuid only
if sbi.devs[idx].tag is empty?
Thanks,
Gao Xiang
actually required in flatdev mode, thus
fix mount failure due to empty device tags in flatdev mode.
Fixes: 8b465fecc35a ("erofs: support flattened block device for multi-blob
images")
Reviewed-by: Gao Xiang
Thanks,
Gao Xiang
Each tar header keeps the full file path. It's useful to move most
recently used intermediate dirs to list heads.
User time of tarerofs index mode can be reduced by 19%.
Signed-off-by: Gao Xiang
---
include/erofs/rebuild.h | 2 +-
lib/rebuild.c | 9 +++--
lib/tar.c
301 - 400 of 3544 matches
Mail list logo
