Many userspace programs (including erofs-utils itself) uses SEEK_DATA /
SEEK_HOLE to parse hole extents in addition to FIEMAP.
Signed-off-by: Gao Xiang
---
fs/erofs/data.c | 26 +-
1 file changed, 25 insertions(+), 1 deletion(-)
diff --git a/fs/erofs/data.c b/fs/erofs/da
On 2024/10/11 7:58, Gao Xiang wrote:
`kaddr` becomes useless after switching to metabuf.
Signed-off-by: Gao Xiang
Reviewed-by: Chao Yu
Thanks,
On 2024/10/10 17:04, Gao Xiang wrote:
Just fold it into the caller for simplicity.
Signed-off-by: Gao Xiang
Reviewed-by: Chao Yu
Thanks,
On 2024/9/17 21:08, Gao Xiang wrote:
Only regular inodes are allowed for file-backed mounts, not directories
(as seen in the original syzbot case) or special inodes.
Also ensure that .read_folio() is implemented on the underlying fs
for the primary device.
Fixes: fb176750266a ("erofs: add file-
在 2024/10/10 22:52, David Howells 写道:
Zizhi Wo wrote:
在 2024/10/10 19:26, David Howells 写道:
Zizhi Wo wrote:
+ spin_lock(&object->lock);
if (object->file) {
fput(object->file);
object->file = NULL;
}
+ spin_unlock(&object->lock)
`kaddr` becomes useless after switching to metabuf.
Signed-off-by: Gao Xiang
---
change since v1:
- shouldn't move `m->lcn` downwards since `lcn` has been changed.
fs/erofs/zmap.c | 32
1 file changed, 12 insertions(+), 20 deletions(-)
diff --git a/fs/erofs/zm
tree/branch: https://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs.git
dev-test
branch HEAD: 56bd565ea59192bbc7d5bbcea155e861a20393f4 erofs: get rid of kaddr
in `struct z_erofs_maprecorder`
elapsed time: 737m
configs tested: 54
configs skipped: 1
The following configs have been built su
Zizhi Wo wrote:
> 在 2024/10/10 19:26, David Howells 写道:
> > Zizhi Wo wrote:
> >
> >> + spin_lock(&object->lock);
> >>if (object->file) {
> >>fput(object->file);
> >>object->file = NULL;
> >>}
> >> + spin_unlock(&object->lock);
> > I would suggest stashing the f
在 2024/10/10 19:36, David Howells 写道:
Zizhi Wo wrote:
For scenarios such as nfs/cifs, the corresponding bsize is PAGE_SIZE aligned
cache->bsize is a property of the cache device, not the network filesystems
that might be making use of it (and it might be shared between multiple
volumes fr
在 2024/10/10 19:36, David Howells 写道:
Zizhi Wo wrote:
For scenarios such as nfs/cifs, the corresponding bsize is PAGE_SIZE aligned
cache->bsize is a property of the cache device, not the network filesystems
that might be making use of it (and it might be shared between multiple
volumes fr
在 2024/10/10 19:26, David Howells 写道:
Zizhi Wo wrote:
+ spin_lock(&object->lock);
if (object->file) {
fput(object->file);
object->file = NULL;
}
+ spin_unlock(&object->lock);
I would suggest stashing the file pointer in a local v
On 2024/10/10 17:51, Hongzhen Luo wrote:
When checking with _require_erofs_compression(), the error "xxx is
not a block device" occurs. This patch adds "-o loop" to address this
issue.
Fixes: ace04f5bbc5c ("erofs-utils: tests: add compression algorithms
check for tests")
Signed-off-by: Hongzh
在 2024/10/10 19:31, David Howells 写道:
Zizhi Wo wrote:
In cachefiles_daemon_secctx(), if it is detected that secctx has been
written to the cache, the error code returned is -EINVAL, which is
inappropriate and does not distinguish the situation well.
I disagree: it is an invalid parameter,
Zizhi Wo wrote:
> For scenarios such as nfs/cifs, the corresponding bsize is PAGE_SIZE aligned
cache->bsize is a property of the cache device, not the network filesystems
that might be making use of it (and it might be shared between multiple
volumes from multiple network filesystems, all of whi
Zizhi Wo wrote:
> In cachefiles_daemon_secctx(), if it is detected that secctx has been
> written to the cache, the error code returned is -EINVAL, which is
> inappropriate and does not distinguish the situation well.
I disagree: it is an invalid parameter, not an already extant file, and a
mess
Zizhi Wo wrote:
> In the erofs on-demand loading scenario, read and write operations are
> usually delivered through "off" and "len" contained in read req in user
> mode. Naturally, pwrite is used to specify a specific offset to complete
> write operations.
>
> However, if the write(not pwrite)
Zizhi Wo wrote:
> + spin_lock(&object->lock);
> if (object->file) {
> fput(object->file);
> object->file = NULL;
> }
> + spin_unlock(&object->lock);
I would suggest stashing the file pointer in a local var and then doing the
fput() outside of the l
Zizhi Wo wrote:
> In fscache_create_volume(), there is a missing memory barrier between the
> bit-clearing operation and the wake-up operation. This may cause a
> situation where, after a wake-up, the bit-clearing operation hasn't been
> detected yet, leading to an indefinite wait. The triggering
Zizhi Wo wrote:
> Currently, cachefiles_commit_tmpfile() will only be called if object->flags
> is set to CACHEFILES_OBJECT_USING_TMPFILE. Only cachefiles_create_file()
> and cachefiles_invalidate_cookie() set this flag. Both of these functions
> replace object->file with the new tmpfile, and bot
Zizhi Wo wrote:
> In the current on-demand loading scenario, when umount is called, the
> cachefiles_commit_tmpfile() is invoked. When checking the inode
> corresponding to object->file is inconsistent with the dentry,
> cachefiles_unlink() is called to perform cleanup to prevent invalid data
> f
在 2024/10/10 18:34, David Howells 写道:
Zizhi Wo wrote:
In the __cachefiles_prepare_write function, DIO aligns blocks using
PAGE_SIZE as the unit. And currently cachefiles_add_cache() binds
cache->bsize with the requirement that it must not exceed PAGE_SIZE.
However, if cache->bsize is smalle
Zizhi Wo wrote:
> In the __cachefiles_prepare_write function, DIO aligns blocks using
> PAGE_SIZE as the unit. And currently cachefiles_add_cache() binds
> cache->bsize with the requirement that it must not exceed PAGE_SIZE.
> However, if cache->bsize is smaller than PAGE_SIZE, the calculated blo
When checking with _require_erofs_compression(), the error "xxx is
not a block device" occurs. This patch adds "-o loop" to address this
issue.
Fixes: ace04f5bbc5c ("erofs-utils: tests: add compression algorithms
check for tests")
Signed-off-by: Hongzhen Luo
---
tests/common/rc | 2 +-
1 file c
On Wed, 09 Oct 2024 11:31:50 +0800, Gao Xiang wrote:
> As Allison reported [1], currently get_tree_bdev() will store
> "Can't lookup blockdev" error message. Although it makes sense for
> pure bdev-based fses, this message may mislead users who try to use
> EROFS file-backed mounts since get_tree_
Just fold it into the caller for simplicity.
Signed-off-by: Gao Xiang
---
fs/erofs/zdata.c | 29 +
1 file changed, 9 insertions(+), 20 deletions(-)
diff --git a/fs/erofs/zdata.c b/fs/erofs/zdata.c
index 8936790618c6..a569ff9dfd04 100644
--- a/fs/erofs/zdata.c
+++ b/f
`kaddr` becomes useless after switching to metabuf.
Signed-off-by: Gao Xiang
---
fs/erofs/zmap.c | 34 +-
1 file changed, 13 insertions(+), 21 deletions(-)
diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
index e980e29873a5..d329a38de1b9 100644
--- a/fs/erofs/zmap.
26 matches
Mail list logo
