netdev - Re: [PATCH rdma-next 00/10] Enable relaxed ordering for ULPs

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

Message-ID: <1ac705a6-0504-fa6e-4d4d-5256b40c363d@nvidia.com>
Date:   Sun, 11 Apr 2021 13:09:03 +0300
From:   Max Gurtovoy <mgurtovoy@...dia.com>
To:     Jason Gunthorpe <jgg@...dia.com>, Leon Romanovsky <leon@...nel.org>
CC:     Honggang LI <honli@...hat.com>, Doug Ledford <dledford@...hat.com>,
        "Adit Ranadive" <aditr@...are.com>,
        Anna Schumaker <anna.schumaker@...app.com>,
        Ariel Elior <aelior@...vell.com>,
        Avihai Horon <avihaih@...dia.com>,
        "Bart Van Assche" <bvanassche@....org>,
        Bernard Metzler <bmt@...ich.ibm.com>,
        "Christoph Hellwig" <hch@....de>,
        Chuck Lever <chuck.lever@...cle.com>,
        "David S. Miller" <davem@...emloft.net>,
        Dennis Dalessandro <dennis.dalessandro@...nelisnetworks.com>,
        Devesh Sharma <devesh.sharma@...adcom.com>,
        Faisal Latif <faisal.latif@...el.com>,
        "Jack Wang" <jinpu.wang@...os.com>,
        Jakub Kicinski <kuba@...nel.org>,
        "J. Bruce Fields" <bfields@...ldses.org>,
        Jens Axboe <axboe@...com>,
        Karsten Graul <kgraul@...ux.ibm.com>,
        Keith Busch <kbusch@...nel.org>, Lijun Ou <oulijun@...wei.com>,
        <linux-cifs@...r.kernel.org>, <linux-kernel@...r.kernel.org>,
        <linux-nfs@...r.kernel.org>, <linux-nvme@...ts.infradead.org>,
        <linux-rdma@...r.kernel.org>, <linux-s390@...r.kernel.org>,
        Max Gurtovoy <maxg@...lanox.com>,
        "Md. Haris Iqbal" <haris.iqbal@...os.com>,
        Michael Guralnik <michaelgur@...dia.com>,
        Michal Kalderon <mkalderon@...vell.com>,
        Mike Marciniszyn <mike.marciniszyn@...nelisnetworks.com>,
        Naresh Kumar PBS <nareshkumar.pbs@...adcom.com>,
        <netdev@...r.kernel.org>, Potnuri Bharat Teja <bharat@...lsio.com>,
        <rds-devel@....oracle.com>, Sagi Grimberg <sagi@...mberg.me>,
        <samba-technical@...ts.samba.org>,
        Santosh Shilimkar <santosh.shilimkar@...cle.com>,
        Selvin Xavier <selvin.xavier@...adcom.com>,
        Shiraz Saleem <shiraz.saleem@...el.com>,
        Somnath Kotur <somnath.kotur@...adcom.com>,
        Sriharsha Basavapatna <sriharsha.basavapatna@...adcom.com>,
        Steve French <sfrench@...ba.org>,
        "Trond Myklebust" <trond.myklebust@...merspace.com>,
        VMware PV-Drivers <pv-drivers@...are.com>,
        Weihang Li <liweihang@...wei.com>,
        Yishai Hadas <yishaih@...dia.com>,
        Zhu Yanjun <zyjzyj2000@...il.com>
Subject: Re: [PATCH rdma-next 00/10] Enable relaxed ordering for ULPs


On 4/6/2021 2:53 PM, Jason Gunthorpe wrote:
> On Tue, Apr 06, 2021 at 08:09:43AM +0300, Leon Romanovsky wrote:
>> On Tue, Apr 06, 2021 at 10:37:38AM +0800, Honggang LI wrote:
>>> On Mon, Apr 05, 2021 at 08:23:54AM +0300, Leon Romanovsky wrote:
>>>> From: Leon Romanovsky <leonro@...dia.com>
>>>>
>>>>  From Avihai,
>>>>
>>>> Relaxed Ordering is a PCIe mechanism that relaxes the strict ordering
>>>> imposed on PCI transactions, and thus, can improve performance.
>>>>
>>>> Until now, relaxed ordering could be set only by user space applications
>>>> for user MRs. The following patch series enables relaxed ordering for the
>>>> kernel ULPs as well. Relaxed ordering is an optional capability, and as
>>>> such, it is ignored by vendors that don't support it.
>>>>
>>>> The following test results show the performance improvement achieved
>>> Did you test this patchset with CPU does not support relaxed ordering?
>> I don't think so, the CPUs that don't support RO are Intel's fourth/fifth-generation
>> and they are not interesting from performance point of view.
>>
>>> We observed significantly performance degradation when run perftest with
>>> relaxed ordering enabled over old CPU.
>>>
>>> https://github.com/linux-rdma/perftest/issues/116
>> The perftest is slightly different, but you pointed to the valid point.
>> We forgot to call pcie_relaxed_ordering_enabled() before setting RO bit
>> and arguably this was needed to be done in perftest too.
> No, the PCI device should not have the RO bit set in this situation.
> It is something mlx5_core needs to do. We can't push this into
> applications.

pcie_relaxed_ordering_enabled is called in 
drivers/net/ethernet/mellanox/mlx5/core/en_common.c so probably need to 
move it to

mlx5_core in this series.



>
> There should be no performance difference from asking for
> IBV_ACCESS_RELAXED_ORDERING when RO is disabled at the PCI config and
> not asking for it at all.
>
> Either the platform has working relaxed ordering that gives a
> performance gain and the RO config spec bit should be set, or it
> doesn't and the bit should be clear.

is this the case today ?

>
> This is not something to decide in userspace, or in RDMA. At worst it
> becomes another platform specific PCI tunable people have to set.
>
> I thought the old haswell systems were quirked to disable RO globally
> anyhow?
>
> Jason