[1/1] nfs: Move ACL xattr definitions to linux/xattr.h

Commit Message

Ondrej Valousek Oct. 27, 2022, 7:43 a.m. UTC

Hi kernel maintainers,

Please help to submit the following patch into kernel
---
Signed-off-by: Ondrej Valousek <[e-mail]> 
Short description:

The XATTR_NAME_NFSV4_ACL definition is also useful for userspace (i.e. nfs4_acl_tools/libacl/coreutils) 
so makes a sense to move the definition to the linux/xattr.h

Comments

Christoph Hellwig Oct. 31, 2022, 8:02 a.m. UTC | #1

On Thu, Oct 27, 2022 at 07:43:58AM +0000, Ondrej Valousek wrote:
> Hi kernel maintainers,
> 
> Please help to submit the following patch into kernel

Well, for that a proper formatted mail really helps.

And these identifiers do not belong into xattr.h.  They are a
non-sanctioned NFS side API, not a VFS API.

Ondrej Valousek Nov. 2, 2022, 1:37 p.m. UTC | #2

Hi Christoph,
> Well, for that a proper formatted mail really helps.

Yeah for the first main I used Outlook to send, the second email with patch was sent directly using GIT so should be formatted well.
Sorry I am beginner here.

> And these identifiers do not belong into xattr.h.  They are a non-sanctioned NFS side API, not a VFS API.

Ah yes, I forgot about VFS, true. What is the best place to keep them then?
Or is it a big issue to extend VFS with them?
I mean NFSv4 - style ACLs are IMHO even better than Posix ACLs - if not for anything else then because Posix ACLs have never been ratified - maybe we'll see them on local FS at some stage, too.

Thanks,
Ondrej

-----Original Message-----
From: Christoph Hellwig <[e-mail]> 
Sent: pondělí 31. října 2022 9:03
To: Ondrej Valousek <[e-mail]>
Cc: [e-mail]; [e-mail]; [e-mail]; [e-mail]; Chuck Lever III <[e-mail]>
Subject: Re: [PATCH 1/1] nfs: Move ACL xattr definitions to linux/xattr.h

On Thu, Oct 27, 2022 at 07:43:58AM +0000, Ondrej Valousek wrote:
> Hi kernel maintainers,
> 
> Please help to submit the following patch into kernel

Well, for that a proper formatted mail really helps.

And these identifiers do not belong into xattr.h.  They are a non-sanctioned NFS side API, not a VFS API.

Patch

diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index e2efcd26336c..07c3d8572912 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -7680,8 +7680,6 @@  nfs4_release_lockowner(struct nfs_server *server, struct nfs4_lock_state *lsp)
        rpc_call_async(server->client, &msg, 0, &nfs4_release_lockowner_ops, data);
 }
 
-#define XATTR_NAME_NFSV4_ACL "system.nfs4_acl"
-
 static int nfs4_xattr_set_nfs4_acl(const struct xattr_handler *handler,
                                   struct user_namespace *mnt_userns,
                                   struct dentry *unused, struct inode *inode,
diff --git a/include/uapi/linux/xattr.h b/include/uapi/linux/xattr.h
index 9463db2dfa9d..77eb8c885861 100644
--- a/include/uapi/linux/xattr.h
+++ b/include/uapi/linux/xattr.h
@@ -81,5 +81,7 @@ 
 #define XATTR_POSIX_ACL_DEFAULT  "posix_acl_default"
 #define XATTR_NAME_POSIX_ACL_DEFAULT XATTR_SYSTEM_PREFIX XATTR_POSIX_ACL_DEFAULT
 
+#define XATTR_NFSV4_ACL "nfs4_acl"
+#define XATTR_NAME_NFSV4_ACL XATTR_SYSTEM_PREFIX XATTR_NFSV4_ACL
 
 #endif /* _UAPI_LINUX_XATTR_H */