linux-block.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Jens Axboe <axboe@kernel.dk>
To: Bart Van Assche <bvanassche@acm.org>,
	linux-block@vger.kernel.org, linux-fsdevel@vger.kernel.org,
	linux-aio@kvack.org
Cc: hch@lst.de
Subject: Re: [PATCH 05/27] block: ensure that async polled IO is marked REQ_NOWAIT
Date: Fri, 30 Nov 2018 10:17:49 -0700	[thread overview]
Message-ID: <6044f9e1-0175-823b-b923-998094bdca72@kernel.dk> (raw)
In-Reply-To: <1543597960.4347.3.camel@acm.org>

On 11/30/18 10:12 AM, Bart Van Assche wrote:
> On Fri, 2018-11-30 at 09:56 -0700, Jens Axboe wrote:
>> We can't wait for polled events to complete, as they may require active
>> polling from whoever submitted it. If that is the same task that is
>> submitting new IO, we could deadlock waiting for IO to complete that
>> this task is supposed to be completing itself.
>>
>> Signed-off-by: Jens Axboe <axboe@kernel.dk>
>> ---
>>  fs/block_dev.c | 10 +++++++++-
>>  1 file changed, 9 insertions(+), 1 deletion(-)
>>
>> diff --git a/fs/block_dev.c b/fs/block_dev.c
>> index 6de8d35f6e41..ebc3d5a0f424 100644
>> --- a/fs/block_dev.c
>> +++ b/fs/block_dev.c
>> @@ -402,8 +402,16 @@ __blkdev_direct_IO(struct kiocb *iocb, struct iov_iter *iter, int nr_pages)
>>  
>>  		nr_pages = iov_iter_npages(iter, BIO_MAX_PAGES);
>>  		if (!nr_pages) {
>> -			if (iocb->ki_flags & IOCB_HIPRI)
>> +			if (iocb->ki_flags & IOCB_HIPRI) {
>>  				bio->bi_opf |= REQ_HIPRI;
>> +				/*
>> +				 * For async polled IO, we can't wait for
>> +				 * requests to complete, as they may also be
>> +				 * polled and require active reaping.
>> +				 */
>> +				if (!is_sync)
>> +					bio->bi_opf |= REQ_NOWAIT;
>> +			}
>>  
>>  			qc = submit_bio(bio);
>>  			WRITE_ONCE(iocb->ki_cookie, qc);
> 
> Setting REQ_NOWAIT from inside the block layer will make the code that
> submits requests harder to review. Have you considered to make this code
> fail I/O if REQ_NOWAIT has not been set and to require that the context
> that submits I/O sets REQ_NOWAIT?

It's technically still feasible to do for sync polled IO, it's only
the async case that makes it a potential deadlock.

-- 
Jens Axboe


  reply	other threads:[~2018-11-30 17:17 UTC|newest]

Thread overview: 59+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-11-30 16:56 [PATCHSET v4] Support for polled aio Jens Axboe
2018-11-30 16:56 ` [PATCH 01/27] aio: fix failure to put the file pointer Jens Axboe
2018-11-30 17:07   ` Bart Van Assche
2018-11-30 17:08     ` Jens Axboe
2018-11-30 17:24       ` Bart Van Assche
2018-11-30 16:56 ` [PATCH 02/27] aio: clear IOCB_HIPRI Jens Axboe
2018-11-30 17:13   ` Christoph Hellwig
2018-11-30 17:14     ` Jens Axboe
2018-12-04 14:46       ` Christoph Hellwig
2018-12-04 16:40         ` Jens Axboe
2018-11-30 16:56 ` [PATCH 03/27] fs: add an iopoll method to struct file_operations Jens Axboe
2018-11-30 16:56 ` [PATCH 04/27] block: wire up block device iopoll method Jens Axboe
2018-11-30 16:56 ` [PATCH 05/27] block: ensure that async polled IO is marked REQ_NOWAIT Jens Axboe
2018-11-30 17:12   ` Bart Van Assche
2018-11-30 17:17     ` Jens Axboe [this message]
2018-12-04 14:48       ` Christoph Hellwig
2018-12-04 18:13         ` Jens Axboe
2018-11-30 16:56 ` [PATCH 06/27] iomap: wire up the iopoll method Jens Axboe
2018-11-30 16:56 ` [PATCH 07/27] iomap: ensure that async polled IO is marked REQ_NOWAIT Jens Axboe
2018-11-30 16:56 ` [PATCH 08/27] aio: use assigned completion handler Jens Axboe
2018-11-30 16:56 ` [PATCH 09/27] aio: separate out ring reservation from req allocation Jens Axboe
2018-11-30 16:56 ` [PATCH 10/27] aio: don't zero entire aio_kiocb aio_get_req() Jens Axboe
2018-12-04 14:49   ` Christoph Hellwig
2018-12-04 15:27     ` Jens Axboe
2018-11-30 16:56 ` [PATCH 11/27] aio: only use blk plugs for > 2 depth submissions Jens Axboe
2018-12-04 14:50   ` Christoph Hellwig
2018-11-30 16:56 ` [PATCH 12/27] aio: use iocb_put() instead of open coding it Jens Axboe
2018-12-04 14:50   ` Christoph Hellwig
2018-11-30 16:56 ` [PATCH 13/27] aio: split out iocb copy from io_submit_one() Jens Axboe
2018-11-30 16:56 ` [PATCH 14/27] aio: abstract out io_event filler helper Jens Axboe
2018-11-30 16:56 ` [PATCH 15/27] aio: add io_setup2() system call Jens Axboe
2018-11-30 16:56 ` [PATCH 16/27] aio: add support for having user mapped iocbs Jens Axboe
2018-11-30 16:56 ` [PATCH 17/27] aio: support for IO polling Jens Axboe
2018-11-30 16:56 ` [PATCH 18/27] aio: add submission side request cache Jens Axboe
2018-11-30 16:56 ` [PATCH 19/27] fs: add fget_many() and fput_many() Jens Axboe
2018-11-30 16:56 ` [PATCH 20/27] aio: use fget/fput_many() for file references Jens Axboe
2018-11-30 16:56 ` [PATCH 21/27] aio: split iocb init from allocation Jens Axboe
2018-11-30 16:56 ` [PATCH 22/27] aio: batch aio_kiocb allocation Jens Axboe
2018-11-30 16:56 ` [PATCH 23/27] block: add BIO_HOLD_PAGES flag Jens Axboe
2018-11-30 16:56 ` [PATCH 24/27] block: implement bio helper to add iter kvec pages to bio Jens Axboe
2018-11-30 19:21   ` Al Viro
2018-11-30 20:15     ` Jens Axboe
2018-11-30 20:32       ` Jens Axboe
2018-11-30 21:11         ` Al Viro
2018-11-30 21:16           ` Jens Axboe
2018-11-30 21:25             ` Al Viro
2018-11-30 21:34               ` Jens Axboe
2018-11-30 22:06                 ` Jens Axboe
2018-12-04 14:55     ` Christoph Hellwig
2018-12-04 15:25       ` Jens Axboe
2018-11-30 16:56 ` [PATCH 25/27] fs: add support for mapping an ITER_KVEC for O_DIRECT Jens Axboe
2018-11-30 16:56 ` [PATCH 26/27] iov_iter: add import_kvec() Jens Axboe
2018-11-30 19:17   ` Al Viro
2018-11-30 20:15     ` Jens Axboe
2018-11-30 16:56 ` [PATCH 27/27] aio: add support for pre-mapped user IO buffers Jens Axboe
2018-11-30 21:44   ` Jeff Moyer
2018-11-30 21:57     ` Jens Axboe
2018-11-30 22:04       ` Jeff Moyer
2018-11-30 22:11         ` Jens Axboe

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 \
    --in-reply-to=6044f9e1-0175-823b-b923-998094bdca72@kernel.dk \
    --to=axboe@kernel.dk \
    --cc=bvanassche@acm.org \
    --cc=hch@lst.de \
    --cc=linux-aio@kvack.org \
    --cc=linux-block@vger.kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    /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;
as well as URLs for NNTP newsgroup(s).