Re: [PATCH v2 1/8] fs: don't use IOP_XATTR for posix acls

[Christoph Hellwig <hch@lst.de>]

Sorry for not keeping up with your flow of ideas, so chiming in now:

> diff --git a/fs/reiserfs/inode.c b/fs/reiserfs/inode.c
> index c7d1fa526dea..e293eaaed185 100644
> --- a/fs/reiserfs/inode.c
> +++ b/fs/reiserfs/inode.c
> @@ -2090,6 +2090,7 @@ int reiserfs_new_inode(struct reiserfs_transaction_handle *th,
>  	if (IS_PRIVATE(dir) || dentry == REISERFS_SB(sb)->priv_root) {
>  		inode->i_flags |= S_PRIVATE;
>  		inode->i_opflags &= ~IOP_XATTR;
> +		inode->i_op = &reiserfs_privdir_inode_operations;

I wonder if there is any way to set this where reiserfs assigns
the other ops. 

> +const struct inode_operations reiserfs_privdir_inode_operations = {
> +	.create = reiserfs_create,
> +	.lookup = reiserfs_lookup,
> +	.link = reiserfs_link,
> +	.unlink = reiserfs_unlink,
> +	.symlink = reiserfs_symlink,
> +	.mkdir = reiserfs_mkdir,
> +	.rmdir = reiserfs_rmdir,
> +	.mknod = reiserfs_mknod,
> +	.rename = reiserfs_rename,
> +	.setattr = reiserfs_setattr,
> +	.permission = reiserfs_permission,
> +	.fileattr_get = reiserfs_fileattr_get,
> +	.fileattr_set = reiserfs_fileattr_set,
> +};

I suspect many other ops aren't need either, but I really need input
from people that known reiserfs better.

> +	if (likely(!is_bad_inode(inode)))
>  		error = set_posix_acl(mnt_userns, dentry, acl_type, kacl);
>  	else
> +		error = -EIO;

I wonder if the is_bad_inode check should simplify move into
get/set_posix_acl.  But otherwise this patch looks good.