| Message ID | 20170922201551.15222-1-smayhew@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Fri, 2017-09-22 at 16:15 -0400, Scott Mayhew wrote: > RFC 1813 section 3.3.8 states > > Once the client has performed a successful exclusive > create, it must issue a SETATTR to set the correct file > attributes. Until it does so, it should not rely upon any > of the file attributes, since the server implementation > may need to overload file metadata to store the verifier. > > HP-UX expects the client to send the uid and gid in the post-create > setattr for exclusive creates. Customers are sometimes seeing bogus > values because the Linux NFS client is only updating the atime and > mtime. I'm sorry, but that is just insanely stupid server behaviour. How do you guarantee that a second client won't hijack the file by sending their own SETATTR call before your client has an opportunity to set the owner? -- Trond Myklebust Linux NFS client maintainer, PrimaryData trond.myklebust@primarydata.com
diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c index d1e87ec..4d77d2d 100644 --- a/fs/nfs/nfs3proc.c +++ b/fs/nfs/nfs3proc.c @@ -370,6 +370,13 @@ nfs3_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr, if (data->arg.create.createmode == NFS3_CREATE_EXCLUSIVE) { dprintk("NFS call setattr (post-create)\n"); + sattr->ia_uid = current_fsuid(); + sattr->ia_valid |= ATTR_UID; + if (dir->i_mode & S_ISGID) + sattr->ia_gid = dir->i_gid; + else + sattr->ia_gid = current_fsgid(); + sattr->ia_valid |= ATTR_GID; if (!(sattr->ia_valid & ATTR_ATIME_SET)) sattr->ia_valid |= ATTR_ATIME; if (!(sattr->ia_valid & ATTR_MTIME_SET))
RFC 1813 section 3.3.8 states Once the client has performed a successful exclusive create, it must issue a SETATTR to set the correct file attributes. Until it does so, it should not rely upon any of the file attributes, since the server implementation may need to overload file metadata to store the verifier. HP-UX expects the client to send the uid and gid in the post-create setattr for exclusive creates. Customers are sometimes seeing bogus values because the Linux NFS client is only updating the atime and mtime. Signed-off-by: Scott Mayhew <smayhew@redhat.com> --- fs/nfs/nfs3proc.c | 7 +++++++ 1 file changed, 7 insertions(+)