public inbox for [email protected]
 help / color / mirror / Atom feed
From: Pavel Begunkov <[email protected]>
To: Xiaoguang Wang <[email protected]>,
	Joseph Qi <[email protected]>,
	Jens Axboe <[email protected]>,
	[email protected]
Subject: Re: [PATCH] io_uring: leave clean req to be done in flush overflow
Date: Thu, 21 Jan 2021 02:11:44 +0000	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>

On 21/01/2021 01:54, Xiaoguang Wang wrote:
> hi Pavel,
> 
>> On 20/01/2021 08:11, Joseph Qi wrote:
>>> Abaci reported the following BUG:
>>>
>>> [   27.629441] BUG: sleeping function called from invalid context at fs/file.c:402
>>> [   27.631317] in_atomic(): 1, irqs_disabled(): 1, non_block: 0, pid: 1012, name: io_wqe_worker-0
>>> [   27.633220] 1 lock held by io_wqe_worker-0/1012:
>>> [   27.634286]  #0: ffff888105e26c98 (&ctx->completion_lock){....}-{2:2}, at: __io_req_complete.part.102+0x30/0x70
>>> [   27.636487] irq event stamp: 66658
>>> [   27.637302] hardirqs last  enabled at (66657): [<ffffffff8144ba02>] kmem_cache_free+0x1f2/0x3b0
>>> [   27.639211] hardirqs last disabled at (66658): [<ffffffff82003a77>] _raw_spin_lock_irqsave+0x17/0x50
>>> [   27.641196] softirqs last  enabled at (64686): [<ffffffff824003c5>] __do_softirq+0x3c5/0x5aa
>>> [   27.643062] softirqs last disabled at (64681): [<ffffffff8220108f>] asm_call_irq_on_stack+0xf/0x20
>>> [   27.645029] CPU: 1 PID: 1012 Comm: io_wqe_worker-0 Not tainted 5.11.0-rc4+ #68
>>> [   27.646651] Hardware name: Alibaba Cloud Alibaba Cloud ECS, BIOS rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
>>> [   27.649249] Call Trace:
>>> [   27.649874]  dump_stack+0xac/0xe3
>>> [   27.650666]  ___might_sleep+0x284/0x2c0
>>> [   27.651566]  put_files_struct+0xb8/0x120
>>> [   27.652481]  __io_clean_op+0x10c/0x2a0
>>> [   27.653362]  __io_cqring_fill_event+0x2c1/0x350
>>> [   27.654399]  __io_req_complete.part.102+0x41/0x70
>>> [   27.655464]  io_openat2+0x151/0x300
>>> [   27.656297]  io_issue_sqe+0x6c/0x14e0
>>> [   27.657170]  ? lock_acquire+0x31a/0x440
>>> [   27.658068]  ? io_worker_handle_work+0x24e/0x8a0
>>> [   27.659119]  ? find_held_lock+0x28/0xb0
>>> [   27.660026]  ? io_wq_submit_work+0x7f/0x240
>>> [   27.660991]  io_wq_submit_work+0x7f/0x240
>>> [   27.661915]  ? trace_hardirqs_on+0x46/0x110
>>> [   27.662890]  io_worker_handle_work+0x501/0x8a0
>>> [   27.663917]  ? io_wqe_worker+0x135/0x520
>>> [   27.664836]  io_wqe_worker+0x158/0x520
>>> [   27.665719]  ? __kthread_parkme+0x96/0xc0
>>> [   27.666663]  ? io_worker_handle_work+0x8a0/0x8a0
>>> [   27.667726]  kthread+0x134/0x180
>>> [   27.668506]  ? kthread_create_worker_on_cpu+0x90/0x90
>>> [   27.669641]  ret_from_fork+0x1f/0x30
>>>
>>> It blames we call cond_resched() with completion_lock when clean
>>> request. In fact we will do it during flush overflow and it seems we
>>> have no reason to do it before. So just remove io_clean_op() in
>>> __io_cqring_fill_event() to fix this BUG.
>>
>> Nope, it would be broken. You may override, e.g. iov pointer
>> that is dynamically allocated, and the function makes sure all
>> those are deleted and freed. Most probably there will be problems
>> on flush side as well.
> Could you please explain more why this is a problem?
> io_clean_op justs does some clean work, free allocated memory, put file, etc,
> and these jobs should can be done in __io_cqring_overflow_flush():

struct io_kiocb {
	union {
		struct file		*file;
		struct io_rw		rw;
		...
		/* use only after cleaning per-op data, see io_clean_op() */
		struct io_completion	compl;
	};
};

io_clean_op() cleans everything in first 64B (not only), and that space
is used for overflow lists, etc.

io_clean_op(req);
req->compl.cflags = cflags;
     -----
list_add_tail(&req->compl.list, &ctx->cq_overflow_list);
                    -----

That's the reason why we need to call it. A bit different story is why
it does drop_files(). One time it was in io_req_clean_work(), which is
called without locks held, but there were nasty races with cancellations
of overflowed reqs, so it was much easier to move into io_clean_op(),
so we just don't ever have requests with ->files in overflowed lists.

As we just changed that cancellation scheme, those races are not
existent anymore, and it could be moved back as in the diff. 


>     while (!list_empty(&list)) {
>         req = list_first_entry(&list, struct io_kiocb, compl.list);
>         list_del(&req->compl.list);
>         io_put_req(req); // will call io_clean_op
>     }

-- 
Pavel Begunkov

      reply	other threads:[~2021-01-21  3:42 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-01-20  8:11 [PATCH] io_uring: leave clean req to be done in flush overflow Joseph Qi
2021-01-20 12:35 ` Pavel Begunkov
2021-01-21  1:37   ` Joseph Qi
2021-01-21  2:00     ` Pavel Begunkov
2021-01-21  1:54   ` Xiaoguang Wang
2021-01-21  2:11     ` Pavel Begunkov [this message]

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] \
    /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