From: Noveck, Dave (Dave.Noveck@netapp.com)
Date: 10/17/01-09:51:54 AM Z
Message-ID: <8C610D86AF6CD4119C9800B0D0499E333356B0@red.nane.netapp.com> From: "Noveck, Dave" <Dave.Noveck@netapp.com> Subject: RE: Out-of-range data Date: Wed, 17 Oct 2001 07:51:54 -0700 Calling it incorrect seems a bit strong, but I think it is better to return an error in this case. That's what I'm trying to do in such cases in my server implementation. Having said that, I just looked at my current code, and, in the specific case you mentioned, I don't check either :-( I'll fix it, but it's not at the top of my list right now. There is some question about what the proper error to return in some of these cases is. In the case of access, we have an int rather than a enum argument. I think NFS4ERR_INVAL is the right thing to return for something that has undefined flags set. In the case of something that is defined as an enum within the .x file, I'm currently normally returning NFS4ERR_INVAL as well. However, since the .x file specifies that is an enum, I'm wondering if perhaps, NFS4ERR_BADXDR would be more correct in the case that something that is outside what the .x file specifies is sent. In the case of a discriminated union (without default) however, I think NFS4ERR_BADXDR should be returned if the value is one that has no associated case. In that case you've reached a point where the rest of the XDR stream can't be decoded. -----Original Message----- From: Peter Åstrand [mailto:peter@cendio.se] Sent: Wednesday, October 17, 2001 10:04 AM To: nfsv4-wg@sunroof.eng.sun.com Subject: Out-of-range data How should a NFSv4 server respond upon a request with enum arguments, with an invalid integer? For example, calling ACCESS with accessreq=0x00000040. The CITI implementation accepts this and returns OK. Is this correct behavior? -- Peter Åstrand Telephone: +46-13-29 08 61 Cendio Systems E-mail: peter@cendio.se Teknikringen 3 583 30 Linköping
This archive was generated by hypermail 2.1.2 : 03/04/05-01:49:17 AM Z CST