linux-kernel - Re: [PATCH v1] NFSD: Decode NFSv4 birth time attribute

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:   Tue, 12 Jul 2022 13:46:33 +0000
From:   Chuck Lever III <chuck.lever@...cle.com>
To:     Igor Mammedov <imammedo@...hat.com>
CC:     Linux NFS Mailing List <linux-nfs@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v1] NFSD: Decode NFSv4 birth time attribute


> On Jul 12, 2022, at 4:09 AM, Igor Mammedov <imammedo@...hat.com> wrote:
> 
> On Mon, 11 Jul 2022 17:18:38 +0000
> Chuck Lever III <chuck.lever@...cle.com> wrote:
> 
>>> On Jul 11, 2022, at 1:14 PM, Igor Mammedov <imammedo@...hat.com> wrote:
>>> 
>>> on tangent:
>>> when copying file from Mac (used 'cp') there is a delay ~4sec/file
>>> 'cp' does first triggers create then extra open and then setattr
>>> which returns
>>> SETATTR Status: NFS4ERR_DELAY
>>> after which the client stalls for a few seconds before repeating setattr.
>>> So question is what makes server unhappy to trigger this error
>>> and if it could be fixed on server side.
>>> 
>>> it seems to affect other methods of copying. So if one extracted
>>> an archive with multiple files or copied multiple files, that
>>> would be a pain.
>>> 
>>> With vers=3 copying is 'instant'
>>> with linux client and vers=4.0 copying is 'instant' as well but it
>>> doesn't use the same call sequence.
>>> 
>>> PS:
>>> it is not regression (I think slowness was there for a long time) 
>> 
>> A network capture would help diagnose this further, but it
>> sounds like it's delegation-related.
> yep, there was delegation request/response right after SETATTR failure
> possibly prompted by NFS4ERR_DELAY
> 
> shall I provide a network capture (I guess pcap file) from test env
> I have?

Yes, you can send it directly to me.


--
Chuck Lever