[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <jfjchxptpfcpolpy2etiigflqvfcp2nzk53iwb6rx2zsdfd6by@h6ezfayjfqai>
Date: Fri, 16 Feb 2024 17:44:02 +0100
From: Daniel Wagner <dwagner@...e.de>
To: Christoph Hellwig <hch@....de>
Cc: James Smart <james.smart@...adcom.com>,
Keith Busch <kbusch@...nel.org>, Sagi Grimberg <sagi@...mberg.me>, Hannes Reinecke <hare@...e.de>,
linux-nvme@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v0 1/6] nvme-fabrics: introduce connect_sync option
On Fri, Feb 16, 2024 at 10:49:09AM +0100, Christoph Hellwig wrote:
> On Fri, Feb 16, 2024 at 09:45:21AM +0100, Daniel Wagner wrote:
> > The TCP and RDMA transport are doing a synchronous connect, meaning the
> > syscal returns with the final result, that is. it either failed or
> > succeeded.
> >
> > This isn't the case for FC. This transport just setups and triggers
> > the connect and returns without waiting on the result.
>
> That's really weird and unexpected. James, can you explain the reason
> behind this?
James answered this point on my attempt to make this synchronous:
https://lore.kernel.org/linux-nvme/0605ac36-16d5-2026-d3c6-62d346db6dfb@gmail.com/
> > Introduce a flag
> > to allow user space to control the behavior, wait or don't wait.
>
> I'd expect this to be the default, but I'll wait to hear more about
> the rationale. If we keep the async default the option looks sensible.
Ideally, we could agree on behavior which is the same for all
transports.
Powered by blists - more mailing lists