linux-kernel - Re: [PATCH v6 00/16] nvmet: add target passthru commands support

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 for Android: free password hash cracker in your pocket

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

Message-ID: <322df1b1-dbba-2018-44da-c108336f8d55@grimberg.me>
Date:   Mon, 29 Jul 2019 09:15:06 -0700
From:   Sagi Grimberg <sagi@...mberg.me>
To:     Stephen Bates <sbates@...thlin.com>,
        Logan Gunthorpe <logang@...tatee.com>,
        Hannes Reinecke <hare@...e.de>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-nvme@...ts.infradead.org" <linux-nvme@...ts.infradead.org>,
        "linux-block@...r.kernel.org" <linux-block@...r.kernel.org>,
        "linux-fsdevel@...r.kernel.org" <linux-fsdevel@...r.kernel.org>
Cc:     Christoph Hellwig <hch@....de>, Keith Busch <kbusch@...nel.org>,
        Jens Axboe <axboe@...com>,
        Chaitanya Kulkarni <Chaitanya.Kulkarni@....com>,
        Max Gurtovoy <maxg@...lanox.com>
Subject: Re: [PATCH v6 00/16] nvmet: add target passthru commands support


>> This is different from multipath on say a multi-controller PCI device
>> and trying to expose both those controllers through passthru. this is
>> where the problems we are discussing come up.
> 
> I *think* there is some confusion. I *think* Sagi is talking about network multi-path (i.e. the ability for the host to connect to a controller on the target via two different network paths that fail-over as needed). I *think* Logan is talking about multi-port PCIe NVMe devices that allow namespaces to be accessed via more than one controller over PCIe (dual-port NVMe SSDs being the most obvious example of this today).

Yes, I was referring to fabrics multipathing which is somewhat
orthogonal to the backend pci multipathing (unless I'm missing
something).

>> But it's the multipath through different ports that
>>   seems important for fabrics. ie. If I have a host with a path through
>>   RDMA and a path through TCP they should both work and allow fail over.
> 
> Yes, or even two paths that are both RDMA or both TCP but which take a different path through the network from host to target.
> 
>> Our real-world use case is to support our PCI device which has a bunch
>> of vendor unique commands and isn't likely to support multiple
>> controllers in the foreseeable future.
> 
> I think solving passthru for single-port PCIe controllers would be a good start.

Me too.