lwn.git/fs/9p/xattr.c, branch docs-next-merge

switch xattr_handler->set() to passing dentry and inode separately

2016-05-27T19:39:43+00:00

preparation for similar switch in ->setxattr() (see the next commit for rationale). Signed-off-by: Al Viro

xattr_handler: pass dentry and inode as separate arguments of ->get()

2016-04-11T00:48:24+00:00

... and do not assume they are already attached to each other Signed-off-by: Al Viro

vfs: Distinguish between full xattr names and proper prefixes

2015-12-07T02:33:52+00:00

Add an additional "name" field to struct xattr_handler. When the name is set, the handler matches attributes with exactly that name. When the prefix is set instead, the handler matches attributes with the given prefix and with a non-empty suffix. This patch should avoid bugs like the one fixed in commit c361016a in the future. Signed-off-by: Andreas Gruenbacher Reviewed-by: James Morris Signed-off-by: Al Viro

9p: xattr simplifications

2015-11-14T01:34:33+00:00

Now that the xattr handler is passed to the xattr handler operations, we can use the same get and set operations for the user, trusted, and security xattr namespaces. In those namespaces, we can access the full attribute name by "reattaching" the name prefix the vfs has skipped for us. Add a xattr_full_name helper to make this obvious in the code. For the "system.posix_acl_access" and "system.posix_acl_default" attributes, handler->prefix is the full attribute name; the suffix is the empty string. Signed-off-by: Andreas Gruenbacher Cc: Eric Van Hensbergen Cc: Ron Minnich Cc: Latchesar Ionkov Cc: v9fs-developer@lists.sourceforge.net Signed-off-by: Al Viro

9p: switch p9_client_read() to passing struct iov_iter *

2015-04-12T02:28:27+00:00

... and make it loop Signed-off-by: Al Viro

9p: switch p9_client_write() to passing it struct iov_iter *

2015-04-12T02:28:25+00:00

... and make it loop until it's done Signed-off-by: Al Viro

9P: fix return value in v9fs_fid_xattr_set

2014-06-01T20:31:34+00:00

v9fs_fid_xattr_set is supposed to return 0 on success. This corrects the behaviour introduced in commit bdd5c28dcb8330b9074404cc92a0b83aae5606a "9p: fix return value in case in v9fs_fid_xattr_set()" (The function returns a negative error on error, as expected) Signed-off-by: Dominique Martinet Signed-off-by: Eric Van Hensbergen

9p: fix return value in case in v9fs_fid_xattr_set()

2013-11-23T22:10:31+00:00

In case of error in the p9_client_write, the function v9fs_fid_xattr_set should return its negative value, what was never being done. In case of success it only retuned 0. Now it returns the 'offset' variable (write_count total). Signed-off-by: Geyslan G. Bem Signed-off-by: Eric Van Hensbergen

fs/9p: xattr: add trusted and security namespaces

2013-07-08T03:02:18+00:00

Allow requests for security.* and trusted.* xattr name spaces to pass through to server. The new files are 99% cut and paste from fs/9p/xattr_user.c with the namespaces changed. It has the intended effect in superficial testing. I do not know much detail about how these namespaces are used, but passing them through to the server, which can decide whether to handle them or not, seems reasonable. I want to support a use case where an ext4 file system is mounted via 9P, then re-exported via samba to windows clients in a cluster. Windows wants to store xattrs such as security.NTACL. This works when ext4 directly backs samba, but not when 9P is inserted. This use case is documented here: http://code.google.com/p/diod/issues/detail?id=95 Signed-off-by: Jim Garlick Signed-off-by: Eric Van Hensbergen

9p: add fid-based variant of v9fs_xattr_set()

2013-02-26T07:46:05+00:00

... making v9fs_xattr_set() a wrapper for it. Signed-off-by: Al Viro