[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CA+1E3rLD2e4yFsA-gxPGtp2Fg2Z0sdfpS3Rz+WsUC4i5aF43LA@mail.gmail.com>
Date: Thu, 28 Jan 2021 22:55:41 +0530
From: Kanchan Joshi <joshiiitr@...il.com>
To: Jens Axboe <axboe@...nel.dk>
Cc: Pavel Begunkov <asml.silence@...il.com>,
Kanchan Joshi <joshi.k@...sung.com>,
Keith Busch <kbusch@...nel.org>,
Christoph Hellwig <hch@....de>, sagi@...mberg.me,
linux-nvme@...ts.infradead.org, io-uring@...r.kernel.org,
linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org,
Javier Gonzalez <javier.gonz@...sung.com>,
Nitesh Shetty <nj.shetty@...sung.com>, anuj20.g@...sung.com
Subject: Re: [RFC PATCH 0/4] Asynchronous passthrough ioctl
On Thu, Jan 28, 2021 at 8:20 PM Jens Axboe <axboe@...nel.dk> wrote:
>
> On 1/28/21 5:04 AM, Kanchan Joshi wrote:
> > And for some ioctls, driver may still need to use task-work to update
> > the user-space pointers (embedded in uring/ioctl cmd) during
> > completion.
>
> For this use case, we should ensure that just io_uring handles this
> part. It's already got everything setup for it, and I'd rather avoid
> having drivers touch any of those parts. Could be done by having an
> io_uring helper ala:
>
> io_uring_cmd_complete_in_task(cmd, handler);
>
> which takes care of the nitty gritty details.
Ah right. With that, I can do away with exporting task-work.
NVMe completion can invoke (depending on ioctl) this uring-helper with
a handler that does the ioctl-specific update in task context.
--
Kanchan
Powered by blists - more mailing lists