Message-ID: <CA+1E3rJqtpfDbKVGBjSYhX=WfzZ2bE8b0U3drUNz_=bp0u9Vuw@mail.gmail.com>
Date:   Mon, 22 Feb 2021 19:12:18 +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>,
        Selvakumar S <selvakuma.s1@...sung.com>
Subject: Re: [RFC PATCH 0/4] Asynchronous passthrough ioctl

On Thu, Jan 28, 2021 at 10:54 PM Jens Axboe <axboe@...nel.dk> wrote:
>
> On 1/28/21 10:13 AM, Kanchan Joshi wrote:
> > On Thu, Jan 28, 2021 at 8:08 PM Jens Axboe <axboe@...nel.dk> wrote:
> >>
> >> On 1/28/21 5:04 AM, Kanchan Joshi wrote:
> >>> On Wed, Jan 27, 2021 at 9:32 PM Pavel Begunkov <asml.silence@...il.com> wrote:
> >>>>
> >>>> On 27/01/2021 15:42, Pavel Begunkov wrote:
> >>>>> On 27/01/2021 15:00, Kanchan Joshi wrote:
> >>>>>> This RFC patchset adds asynchronous ioctl capability for NVMe devices.
> >>>>>> Purpose of RFC is to get the feedback and optimize the path.
> >>>>>>
> >>>>>> At the uppermost io-uring layer, a new opcode IORING_OP_IOCTL_PT is
> >>>>>> presented to user-space applications. Like regular-ioctl, it takes
> >>>>>> ioctl opcode and an optional argument (ioctl-specific input/output
> >>>>>> parameter). Unlike regular-ioctl, it is made to skip the block-layer
> >>>>>> and reach directly to the underlying driver (nvme in the case of this
> >>>>>> patchset). This path between io-uring and nvme is via a newly
> >>>>>> introduced block-device operation "async_ioctl". This operation
> >>>>>> expects io-uring to supply a callback function which can be used to
> >>>>>> report completion at later stage.
> >>>>>>
> >>>>>> For a regular ioctl, NVMe driver submits the command to the device and
> >>>>>> the submitter (task) is made to wait until completion arrives. For
> >>>>>> async-ioctl, completion is decoupled from submission. Submitter goes
> >>>>>> back to its business without waiting for nvme-completion. When
> >>>>>> nvme-completion arrives, it informs io-uring via the registered
> >>>>>> completion-handler. But some ioctls may require updating certain
> >>>>>> ioctl-specific fields which can be accessed only in context of the
> >>>>>> submitter task. For that reason, NVMe driver uses task-work infra for
> >>>>>> that ioctl-specific update. Since task-work is not exported, it cannot
> >>>>>> be referenced when nvme is compiled as a module. Therefore, one of the
> >>>>>> patch exports task-work API.
> >>>>>>
> >>>>>> Here goes example of usage (pseudo-code).
> >>>>>> Actual nvme-cli source, modified to issue all ioctls via this opcode
> >>>>>> is present at-
> >>>>>> https://github.com/joshkan/nvme-cli/commit/a008a733f24ab5593e7874cfbc69ee04e88068c5
> >>>>>
> >>>>> see https://git.kernel.dk/cgit/linux-block/log/?h=io_uring-fops
> >>>>>
> >>>>> Looks like good time to bring that branch/discussion back
> >>>>
> >>>> a bit more context:
> >>>> https://github.com/axboe/liburing/issues/270
> >>>
> >>> Thanks, it looked good. It seems key differences (compared to
> >>> uring-patch that I posted) are -
> >>> 1. using file-operation instead of block-dev operation.
> >>
> >> Right, it's meant to span wider than just block devices.
> >>
> >>> 2. repurpose the sqe memory for ioctl-cmd. If an application does
> >>> ioctl with <=40 bytes of cmd, it does not have to allocate ioctl-cmd.
> >>> That's nifty. We still need to support passing larger-cmd (e.g.
> >>> nvme-passthru ioctl takes 72 bytes) but that shouldn't get too
> >>> difficult I suppose.
> >>
> >> It's actually 48 bytes in the as-posted version, and I've bumped it to
> >> 56 bytes in the latest branch. So not quite enough for everything,
> >> nothing ever will be, but should work for a lot of cases without
> >> requiring per-command allocations just for the actual command.
> >
> > Agreed. But if I got it right, you are open to support both in-the-sqe
> > command (<= 56 bytes) and out-of-sqe command (> 56 bytes) with this
> > interface.
> > Driver processing the ioctl can fetch the cmd from user-space in one
> > case (as it does now), and skips in another.
>
> Your out-of-seq command would be none of io_urings business, outside of
> the fact that we'd need to ensure it's stable if we need to postpone
> it. So yes, that would be fine, it just means your actual command is
> passed in as a pointer, and you would be responsible for copying it
> in for execution
>
> We're going to need something to handle postponing, and something
> for ensuring that eg cancelations free the allocated memory.

I have few doubts about allocation/postponing. Are you referring to
uring allocating memory for this case, similar to the way
"req->async_data" is managed for few other opcodes?
Or can it (i.e. larger cmd) remain a user-space pointer, and the
underlying driver fetches the command in.
If submission context changes (for sqo/io-wq case), uring seemed to
apply context-grabbing techniques to make that work.

> >>> 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.
> >>>
> >>> @Jens - will it be fine if I start looking at plumbing nvme-part of
> >>> this series on top of your work?
> >>
> >> Sure, go ahead. Just beware that things are still changing, so you might
> >> have to adapt it a few times. It's still early days, but I do think
> >> that's the way forward in providing controlled access to what is
> >> basically async ioctls.
> >
> > Sounds good, I will start with the latest branch that you posted. Thanks.
>
> It's io_uring-fops.v2 for now, use that one.

Moved to v3 now.
nvme_user_io is 48 bytes, while nvme passthrough requires 72 or 80
bytes (passthru with 64 bit result).
The block_uring_cmd has 32 bytes of available space. If NVMe defines
its own "nvme_uring_cmd" (which can be used for nvme char interface)
that will buy some more space, but still won't be enough for passthru
command.

So I am looking at adding support for large-cmd in uring. And felt the
need to clear those doubts I mentioned above.


Thanks

Powered by blists - more mailing lists