linux.git/fs/cachefiles, branch v6.3

fs: port ->permission() to pass mnt_idmap

2023-01-19T08:24:28+00:00

Convert to struct mnt_idmap.

Last cycle we merged the necessary infrastructure in
256c8aed2b42 ("fs: introduce dedicated idmap type for mounts").
This is just the conversion to struct mnt_idmap.

Currently we still pass around the plain namespace that was attached to a
mount. This is in general pretty convenient but it makes it easy to
conflate namespaces that are relevant on the filesystem with namespaces
that are relevent on the mount level. Especially for non-vfs developers
without detailed knowledge in this area this can be a potential source for
bugs.

Once the conversion to struct mnt_idmap is done all helpers down to the
really low-level helpers will take a struct mnt_idmap argument instead of
two namespace arguments. This way it becomes impossible to conflate the two
eliminating the possibility of any bugs. All of the vfs and all filesystems
only operate on struct mnt_idmap.

Acked-by: Dave Chinner 
Reviewed-by: Christoph Hellwig 
Signed-off-by: Christian Brauner (Microsoft)

fs: port vfs_*() helpers to struct mnt_idmap

2023-01-18T16:51:45+00:00

Convert to struct mnt_idmap.

Last cycle we merged the necessary infrastructure in
256c8aed2b42 ("fs: introduce dedicated idmap type for mounts").
This is just the conversion to struct mnt_idmap.

Currently we still pass around the plain namespace that was attached to a
mount. This is in general pretty convenient but it makes it easy to
conflate namespaces that are relevant on the filesystem with namespaces
that are relevent on the mount level. Especially for non-vfs developers
without detailed knowledge in this area this can be a potential source for
bugs.

Once the conversion to struct mnt_idmap is done all helpers down to the
really low-level helpers will take a struct mnt_idmap argument instead of
two namespace arguments. This way it becomes impossible to conflate the two
eliminating the possibility of any bugs. All of the vfs and all filesystems
only operate on struct mnt_idmap.

Acked-by: Dave Chinner 
Reviewed-by: Christoph Hellwig 
Signed-off-by: Christian Brauner (Microsoft)

fscache,cachefiles: add prepare_ondemand_read() callback

2022-12-07T02:56:29+00:00

Add prepare_ondemand_read() callback dedicated for the on-demand read
scenario, so that callers from this scenario can be decoupled from
netfs_io_subrequest.

The original cachefiles_prepare_read() is now refactored to a generic
routine accepting a parameter list instead of netfs_io_subrequest.
There's no logic change, except that the debug id of subrequest and
request is removed from trace_cachefiles_prep_read().

Reviewed-by: Jeff Layton 
Signed-off-by: Jingbo Xu 
Acked-by: David Howells 
Link: https://lore.kernel.org/r/20221124034212.81892-2-jefflexu@linux.alibaba.com
Signed-off-by: Gao Xiang

cachefiles: use vfs_tmpfile_open() helper

2022-09-24T05:00:00+00:00

Use the vfs_tmpfile_open() helper instead of doing tmpfile creation and
opening separately.

The only minor difference is that previously no permission checking was
done, while vfs_tmpfile_open() will call may_open() with zero access mask
(i.e. no access is checked).  Even if this would make a difference with
callers caps (don't see how it could, even in the LSM codepaths) cachfiles
raises caps before performing the tmpfile creation, so this extra
permission check will not result in any regression.

Reviewed-by: Christian Brauner (Microsoft) 
Signed-off-by: Miklos Szeredi

cachefiles: only pass inode to *mark_inode_inuse() helpers

2022-09-24T04:59:59+00:00

The only reason to pass dentry was because of a pr_notice() text.  Move
that to the two callers where it makes sense and add a WARN_ON() to the
third.

file_inode(file) is never NULL on an opened file. Remove check in
cachefiles_unmark_inode_in_use().

Do not open code cachefiles_do_unmark_inode_in_use() in
cachefiles_put_directory().

Signed-off-by: Miklos Szeredi

cachefiles: tmpfile error handling cleanup

2022-09-24T04:59:59+00:00

Separate the error labels from the success path and use 'ret' to store the
error value before jumping to the error label.

Signed-off-by: Miklos Szeredi

cachefiles: make on-demand request distribution fairer

2022-08-31T15:41:10+00:00

For now, enqueuing and dequeuing on-demand requests all start from
idx 0, this makes request distribution unfair. In the weighty
concurrent I/O scenario, the request stored in higher idx will starve.

Searching requests cyclically in cachefiles_ondemand_daemon_read,
makes distribution fairer.

Fixes: c8383054506c ("cachefiles: notify the user daemon when looking up cookie")
Reported-by: Yongqing Li 
Signed-off-by: Xin Yin 
Signed-off-by: David Howells 
Reviewed-by: Jeffle Xu 
Reviewed-by: Gao Xiang 
Link: https://lore.kernel.org/r/20220817065200.11543-1-yinxin.x@bytedance.com/ # v1
Link: https://lore.kernel.org/r/20220825020945.2293-1-yinxin.x@bytedance.com/ # v2

cachefiles: fix error return code in cachefiles_ondemand_copen()

2022-08-31T15:41:10+00:00

The cache_size field of copen is specified by the user daemon.
If cache_size < 0, then the OPEN request is expected to fail,
while copen itself shall succeed. However, returning 0 is indeed
unexpected when cache_size is an invalid error code.

Fix this by returning error when cache_size is an invalid error code.

Changes
=======
v4: update the code suggested by Dan
v3: update the commit log suggested by Jingbo.

Fixes: c8383054506c ("cachefiles: notify the user daemon when looking up cookie")
Signed-off-by: Sun Ke 
Suggested-by: Jeffle Xu 
Suggested-by: Dan Carpenter 
Signed-off-by: David Howells 
Reviewed-by: Gao Xiang 
Reviewed-by: Jingbo Xu 
Reviewed-by: Dan Carpenter 
Link: https://lore.kernel.org/r/20220818111935.1683062-1-sunke32@huawei.com/ # v2
Link: https://lore.kernel.org/r/20220818125038.2247720-1-sunke32@huawei.com/ # v3
Link: https://lore.kernel.org/r/20220826023515.3437469-1-sunke32@huawei.com/ # v4

cachefiles: narrow the scope of flushed requests when releasing fd

2022-07-05T15:12:21+00:00

When an anonymous fd is released, only flush the requests
associated with it, rather than all of requests in xarray.

Fixes: 9032b6e8589f ("cachefiles: implement on-demand read")
Signed-off-by: Jia Zhu 
Signed-off-by: David Howells 
Reviewed-by: Jeffle Xu 
Reviewed-by: Gao Xiang 
Link: https://listman.redhat.com/archives/linux-cachefs/2022-June/006937.html

cachefiles: add tracepoints for on-demand read mode

2022-05-17T16:11:18+00:00

Add tracepoints for on-demand read mode. Currently following tracepoints
are added:

	OPEN request / COPEN reply
	CLOSE request
	READ request / CREAD reply
	write through anonymous fd
	release of anonymous fd

Signed-off-by: Jeffle Xu 
Acked-by: David Howells 
Link: https://lore.kernel.org/r/20220425122143.56815-8-jefflexu@linux.alibaba.com
Signed-off-by: Gao Xiang