Restating this, the code is unified for tcp v4 and v3. It consults
xid and the checksum, which I think is a correct intent. The
slot-reply cache in 4.1+ is, we hope, the improved mechanism.
Matt
On Tue, Jan 30, 2018 at 9:45 AM, Matt Benjamin wrote:
> I don't think that's the case. the behavio
innovations? ok. there's other innovation needed in the gss impl,
that would be welcome as well
Matt
On Tue, Jan 30, 2018 at 9:50 AM, William Allen Simpson
wrote:
> On 1/30/18 9:36 AM, William Allen Simpson wrote:
>>
>> But the code is obscure, so I could be missing something.
>
>
> Also, it b
On 1/30/18 9:36 AM, William Allen Simpson wrote:
But the code is obscure, so I could be missing something.
Also, it bears repeating that the dupreq cache wasn't working for
secure connections. Pre-V2.6 checksummed the ciphertext, which is by
definition different on every request. We'd never s
I don't think that's the case. the behavior is distinct for v4 and v3
(which does not have clientids). DRC is bypassed for v4.1+
Matt
On Tue, Jan 30, 2018 at 9:36 AM, William Allen Simpson
wrote:
> On 1/30/18 9:22 AM, William Allen Simpson wrote:
>>
>> On 1/29/18 3:32 PM, Frank Filz wrote:
>>>
On 1/30/18 9:22 AM, William Allen Simpson wrote:
On 1/29/18 3:32 PM, Frank Filz wrote:
I haven't looked at how the SVCXPRT structure has changed, but if there's a
field in there we can attach a Ganesha structure to that would be cool, or
if not, if we could add one.
There are two: xp_u1, and x
That sounds right, I figured we could reclaim xp_u1.
Matt
On Tue, Jan 30, 2018 at 9:22 AM, William Allen Simpson
wrote:
> On 1/29/18 3:32 PM, Frank Filz wrote:
>>
>> I haven't looked at how the SVCXPRT structure has changed, but if there's
>> a
>> field in there we can attach a Ganesha structure
On 1/29/18 3:32 PM, Frank Filz wrote:
I haven't looked at how the SVCXPRT structure has changed, but if there's a
field in there we can attach a Ganesha structure to that would be cool, or
if not, if we could add one.
There are two: xp_u1, and xp_u2.
Right now, Ganesha is using xp_u2 for dup r
Bill,
I have been talking with Matt about one thing that might be easy to do to
improve latency for NFS v4 clients. We do a lookup into a table of clientids
for most operations (anything using a stated for NFS v4.0 and anything in a
session for V4.1+). Most clients will only have one clientid, so