public inbox for [email protected]
 help / color / mirror / Atom feed
From: Pavel Begunkov <[email protected]>
To: Jens Axboe <[email protected]>, Kanchan Joshi <[email protected]>
Cc: Kanchan Joshi <[email protected]>,
	[email protected], [email protected],
	Matthew Wilcox <[email protected]>,
	Christoph Hellwig <[email protected]>,
	Damien Le Moal <[email protected]>,
	[email protected], [email protected],
	[email protected], [email protected],
	[email protected], [email protected],
	SelvaKumar S <[email protected]>,
	Nitesh Shetty <[email protected]>,
	Javier Gonzalez <[email protected]>
Subject: Re: [PATCH v4 6/6] io_uring: add support for zone-append
Date: Thu, 30 Jul 2020 19:08:28 +0300	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>

On 27/07/2020 23:34, Jens Axboe wrote:
> On 7/27/20 1:16 PM, Kanchan Joshi wrote:
>> On Fri, Jul 24, 2020 at 10:00 PM Jens Axboe <[email protected]> wrote:
>>>
>>> On 7/24/20 9:49 AM, Kanchan Joshi wrote:
>>>> diff --git a/fs/io_uring.c b/fs/io_uring.c
>>>> index 7809ab2..6510cf5 100644
>>>> --- a/fs/io_uring.c
>>>> +++ b/fs/io_uring.c
>>>> @@ -1284,8 +1301,15 @@ static void __io_cqring_fill_event(struct io_kiocb *req, long res, long cflags)
>>>>       cqe = io_get_cqring(ctx);
>>>>       if (likely(cqe)) {
>>>>               WRITE_ONCE(cqe->user_data, req->user_data);
>>>> -             WRITE_ONCE(cqe->res, res);
>>>> -             WRITE_ONCE(cqe->flags, cflags);
>>>> +             if (unlikely(req->flags & REQ_F_ZONE_APPEND)) {
>>>> +                     if (likely(res > 0))
>>>> +                             WRITE_ONCE(cqe->res64, req->rw.append_offset);
>>>> +                     else
>>>> +                             WRITE_ONCE(cqe->res64, res);
>>>> +             } else {
>>>> +                     WRITE_ONCE(cqe->res, res);
>>>> +                     WRITE_ONCE(cqe->flags, cflags);
>>>> +             }
>>>
>>> This would be nice to keep out of the fast path, if possible.
>>
>> I was thinking of keeping a function-pointer (in io_kiocb) during
>> submission. That would have avoided this check......but argument count
>> differs, so it did not add up.
> 
> But that'd grow the io_kiocb just for this use case, which is arguably
> even worse. Unless you can keep it in the per-request private data,
> but there's no more room there for the regular read/write side.
> 
>>>> diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
>>>> index 92c2269..2580d93 100644
>>>> --- a/include/uapi/linux/io_uring.h
>>>> +++ b/include/uapi/linux/io_uring.h
>>>> @@ -156,8 +156,13 @@ enum {
>>>>   */
>>>>  struct io_uring_cqe {
>>>>       __u64   user_data;      /* sqe->data submission passed back */
>>>> -     __s32   res;            /* result code for this event */
>>>> -     __u32   flags;
>>>> +     union {
>>>> +             struct {
>>>> +                     __s32   res;    /* result code for this event */
>>>> +                     __u32   flags;
>>>> +             };
>>>> +             __s64   res64;  /* appending offset for zone append */
>>>> +     };
>>>>  };
>>>
>>> Is this a compatible change, both for now but also going forward? You
>>> could randomly have IORING_CQE_F_BUFFER set, or any other future flags.
>>
>> Sorry, I didn't quite understand the concern. CQE_F_BUFFER is not
>> used/set for write currently, so it looked compatible at this point.
> 
> Not worried about that, since we won't ever use that for writes. But it
> is a potential headache down the line for other flags, if they apply to
> normal writes.
> 
>> Yes, no room for future flags for this operation.
>> Do you see any other way to enable this support in io-uring?
> 
> Honestly I think the only viable option is as we discussed previously,
> pass in a pointer to a 64-bit type where we can copy the additional
> completion information to.

TBH, I hate the idea of such overhead/latency at times when SSDs can
serve writes in less than 10ms. Any chance you measured how long does it
take to drag through task_work?

> 
>>> Layout would also be different between big and little endian, so not
>>> even that easy to set aside a flag for this. But even if that was done,
>>> we'd still have this weird API where liburing or the app would need to
>>> distinguish this cqe from all others based on... the user_data? Hence
>>> liburing can't do it, only the app would be able to.
>>>
>>> Just seems like a hack to me.
>>
>> Yes, only user_data to distinguish. Do liburing helpers need to look
>> at cqe->res (and decide something) before returning the cqe to
>> application?
> 
> They generally don't, outside of the internal timeout. But it's an issue
> for the API, as it forces applications to handle the CQEs a certain way.
> Normally there's flexibility. This makes the append writes behave
> differently than everything else, which is never a good idea.
> 

-- 
Pavel Begunkov

  reply	other threads:[~2020-07-30 16:10 UTC|newest]

Thread overview: 55+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CGME20200724155244epcas5p2902f57e36e490ee8772da19aa9408cdc@epcas5p2.samsung.com>
2020-07-24 15:49 ` [PATCH v4 0/6] zone-append support in io-uring and aio Kanchan Joshi
     [not found]   ` <CGME20200724155258epcas5p1a75b926950a18cd1e6c8e7a047e6c589@epcas5p1.samsung.com>
2020-07-24 15:49     ` [PATCH v4 1/6] fs: introduce FMODE_ZONE_APPEND and IOCB_ZONE_APPEND Kanchan Joshi
2020-07-24 16:34       ` Jens Axboe
2020-07-26 15:18       ` Christoph Hellwig
2020-07-28  1:49         ` Matthew Wilcox
2020-07-28  7:26           ` Christoph Hellwig
     [not found]   ` <CGME20200724155324epcas5p18e1d3b4402d1e4a8eca87d0b56a3fa9b@epcas5p1.samsung.com>
2020-07-24 15:49     ` [PATCH v4 2/6] fs: change ki_complete interface to support 64bit ret2 Kanchan Joshi
2020-07-26 15:18       ` Christoph Hellwig
     [not found]   ` <CGME20200724155329epcas5p345ba6bad0b8fe18056bb4bcd26c10019@epcas5p3.samsung.com>
2020-07-24 15:49     ` [PATCH v4 3/6] uio: return status with iov truncation Kanchan Joshi
     [not found]   ` <CGME20200724155341epcas5p15bfc55927f2abb60f19784270fe8e377@epcas5p1.samsung.com>
2020-07-24 15:49     ` [PATCH v4 4/6] block: add zone append handling for direct I/O path Kanchan Joshi
2020-07-26 15:19       ` Christoph Hellwig
     [not found]   ` <CGME20200724155346epcas5p2cfb383fe9904a45280c6145f4c13e1b4@epcas5p2.samsung.com>
2020-07-24 15:49     ` [PATCH v4 5/6] block: enable zone-append for iov_iter of bvec type Kanchan Joshi
2020-07-26 15:20       ` Christoph Hellwig
     [not found]   ` <CGME20200724155350epcas5p3b8f1d59eda7f8fbb38c828f692d42fd6@epcas5p3.samsung.com>
2020-07-24 15:49     ` [PATCH v4 6/6] io_uring: add support for zone-append Kanchan Joshi
2020-07-24 16:29       ` Jens Axboe
2020-07-27 19:16         ` Kanchan Joshi
2020-07-27 20:34           ` Jens Axboe
2020-07-30 16:08             ` Pavel Begunkov [this message]
2020-07-30 16:13               ` Jens Axboe
2020-07-30 16:26                 ` Pavel Begunkov
2020-07-30 17:16                   ` Jens Axboe
2020-07-30 17:38                     ` Pavel Begunkov
2020-07-30 17:51                       ` Kanchan Joshi
2020-07-30 17:54                         ` Jens Axboe
2020-07-30 18:25                           ` Kanchan Joshi
2020-07-31  6:42                             ` Damien Le Moal
2020-07-31  6:45                               ` hch
2020-07-31  6:59                                 ` Damien Le Moal
2020-07-31  7:58                                   ` Kanchan Joshi
2020-07-31  8:14                                     ` Damien Le Moal
2020-07-31  9:14                                       ` hch
2020-07-31  9:34                                         ` Damien Le Moal
2020-07-31  9:41                                           ` hch
2020-07-31 10:16                                             ` Damien Le Moal
2020-07-31 12:51                                               ` hch
2020-07-31 13:08                                                 ` hch
2020-07-31 15:07                                                   ` Kanchan Joshi
2022-03-02 20:47                                                   ` Luis Chamberlain
2020-08-05  7:35                                                 ` Damien Le Moal
2020-08-14  8:14                                                   ` hch
2020-08-14  8:27                                                     ` Damien Le Moal
2020-08-14 12:04                                                       ` hch
2020-08-14 12:20                                                         ` Damien Le Moal
2020-09-07  7:01                                                     ` Kanchan Joshi
2020-09-08 15:18                                                       ` hch
2020-09-24 17:19                                                         ` Kanchan Joshi
2020-09-25  2:52                                                           ` Damien Le Moal
2020-09-28 18:58                                                             ` Kanchan Joshi
2020-09-29  1:24                                                               ` Damien Le Moal
2020-09-29 18:49                                                                 ` Kanchan Joshi
2022-03-02 20:43                                                         ` Luis Chamberlain
2020-07-31  9:38                                       ` Kanchan Joshi
2022-03-02 20:51                                 ` Luis Chamberlain
2020-07-31  7:08                               ` Kanchan Joshi
2020-07-30 15:57       ` Pavel Begunkov

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox