Re: [RFC] fhandle implementation.

From: Alexei I. Adamovich (
Date: Sun Jun 18 2000 - 15:34:27 EST

"Roman V. Shaposhnick" <> (RVS) wrote on Fri,
16 Jun 2000 00:40:34 +0400:

RVS> On Thu, Jun 15, 2000 at 06:00:17PM +0200, Trond Myklebust wrote:
RVS> > My first and immediate question though was whether you think the
RVS> > struct fhandle is generic enough?
RVS> Well, I hope it is.
Certainly not. Roman, your "fhandle" contains only 2 ulongs: id (AKA
inode number) and version. Passing this info to fhandle_to_dentry
method could in general case be not sufficient for to access inode--if
correspondent on-disk data structures are placed there in any order
other then id-order. For instance, in ReiserFS now they are ordered
according to the value of 64-bit field that can be called "inode key"
(ikey); to access them in an effective manner ReiserFS needs this
64-bit value (to make things more interesting, ikey contains id as the
least significant part).

RVS> I can say that it is like 4-d coordinates. Where 'id' is a 3-d
RVS> space component
RVS> ...
In mathematical sence, it's 1-dimentional. Ikey can add at least one
more dimension. And it's hard to predict what number of dimensions
would be sufficient in general case. From this aspect correspondent
Neil's proposal seems more accurate.


   Alexei I.Adamovich
Res. Centre for Multiprocessor Systems, | e-mail:
PSI RAS, Pereslavl-Zalessky 152140 Russia |

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at

This archive was generated by hypermail 2b29 : Fri Jun 23 2000 - 21:00:16 EST