Re: change of persistent_fh attribute (attempt 2)

New Message Reply About this list Date view Thread view Subject view Author view Attachment view

From: Mike Eisler (mre@eng.sun.com)
Date: 11/04/99-07:22:40 PM Z


Date: Thu, 4 Nov 1999 17:22:40 -0800 (PST)
From: Mike Eisler <mre@eng.sun.com>
Subject: Re: change of persistent_fh attribute (attempt 2)
Message-ID: <Roam.SIMC.2.0.6.941764960.10004.mre@eng.sun.com>

> > 
> > FH4_VOL_MIGRATION - fh may expire during file system migration.  May
> > only be set if FH4_VOLATILE_ANY is not set.
> > 
> > FH4_VOL_RENAME - fh may expire due to a rename.  This includes a
> > rename by the requesting client or a rename by another client.  May
> > only be set if FH4_VOLATILE_ANY is not set.

> Could there be some explanation about how the client could make use of
> the last two.  I cannot seem them being, from the clients
> perspective, functionally different from FH4_VOLATILE_ANY (as rename
> and migration could happen at any time).

We should call these the Linux bits. :-)

The notion behind FH4_VOL_RENAME is that I can cache file handles as long as I
don't rename my files. if I do rename my files then I should tack a
LOOKUP/GETFH onto the end of RENAME compound. Of course, another client could
rename files from under me, if I'm concered about that, then I'll treat
FH4_VOL_RENAME as FH4_VOLATILE_ANY, i.e. cache path names in my fh caches.

If FH4_VOL_MIGRATION on, the client will cache the path name of every file is
caches an fh for, but if FH4_VOL_RENAME is off it need not append a
LOOKUP/GETFH to the RENAME. If the client doesn't support migration, then it
need not do anything special at all.

FH4_VOLATILE_ANY simply makes it very clear how tenuous things are. I'd use
FH4_VOLATILE_ANY when doing file transfer to a DOS floppy, but unless
FH4_NOEXPIRE_WITH_OPEN, is also set I'd think twice about writable file access.

> On that note, Mike, your minor versioning proposal says:
> 
> 	         Note that this rule doesn't mean we can't add bits to
> 	flag fields, such as new attributes to GETATTR's bitmap4 data
> 
> This seems to imply that any bit field, such as persistent_fh and the
> various bit fields to do with ACLs, may be extended by a minor
> version.  Is this true? Should it be explicit?

I'll make it explicit. Any other feedback?

	-mre


New Message Reply About this list Date view Thread view Subject view Author view Attachment view

This archive was generated by hypermail 2.1.2 : 03/04/05-01:47:50 AM Z CST