Received-SPF: pass (google.com: domain of linux-nfs-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Sun, 23 Jan 2022 08:29:19 +0200
From:   Dan Aloni <dan.aloni@vastdata.com>
To:     Chuck Lever III <chuck.lever@oracle.com>
Cc:     Anna Schumaker <Anna.Schumaker@netapp.com>,
        Linux NFS Mailing List <linux-nfs@vger.kernel.org>
Subject: Re: [PATCH v2] NFSD: trim reads past NFS_OFFSET_MAX and fix NFSv3
 check
Message-ID: <20220123062919.7yjz5ryousix6vxf@gmail.com>
References: <5DD3C5DF-52EF-4C84-894C-FCBB9A0F4259@oracle.com>
 <20220122124953.1606281-1-dan.aloni@vastdata.com>
 <ADEC85C2-8D72-4E25-A49B-2039C1FF82F2@oracle.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <ADEC85C2-8D72-4E25-A49B-2039C1FF82F2@oracle.com>
Precedence: bulk

On Sat, Jan 22, 2022 at 05:37:16PM +0000, Chuck Lever III wrote:
> Similar language in RFC 8881 section 18.22.4:
> 
> >>    If the server returns a "short read" (i.e., fewer data than requested
> >>    and eof is set to FALSE), the client should send another READ to get
> >>    the remaining data.  A server may return less data than requested
> >>    under several circumstances.  The file may have been truncated by
> >>    another client or perhaps on the server itself, changing the file
> >>    size from what the requesting client believes to be the case.  This
> >>    would reduce the actual amount of data available to the client.  It
> >>    is possible that the server reduce the transfer size and so return a
> >>    short read result.  Server resource exhaustion may also occur in a
> >>    short read.
> 
> So the server could be returning INVAL and leaving EOF set
> to false. That might be triggering the client to retry the
> READ. Does the server need to set the EOF flag if the READ
> arguments cross the limit of the range that the server can
> return? Does the client need to check for this case and
> stop retrying? The specs aren't particularly clear on this
> matter.

But eof only in `resok` and not in `resfail`. For reference, here's the
reply I got:

Network File System, READ Reply  Error: NFS3ERR_INVAL
    [Program Version: 3]
    [V3 Procedure: READ (6)]
    Status: NFS3ERR_INVAL (22)
    file_attributes  Regular File mode: 0600 uid: 0 gid: 0
        attributes_follow: value follows (1)
        attributes  Regular File mode: 0600 uid: 0 gid: 0
            Type: Regular File (1)
            Mode: 0600, S_IRUSR, S_IWUSR
            nlink: 1
            uid: 0
            gid: 0
            size: 9223372036854775807
            used: 4096
            rdev: 0,0
            fsid: 0x0000000000000002 (2)
            fileid: 69
            atime: Jan 18, 2022 20:20:33.611267439 IST
                seconds: 1642530033
                nano seconds: 611267439
            mtime: Jan 18, 2022 20:20:33.571266608 IST
                seconds: 1642530033
                nano seconds: 571266608
            ctime: Jan 18, 2022 20:20:33.571266608 IST
                seconds: 1642530033
                nano seconds: 571266608

-- 
Dan Aloni