public inbox for [email protected]
 help / color / mirror / Atom feed
From: Jens Axboe <[email protected]>
To: Pavel Begunkov <[email protected]>,
	Josef <[email protected]>,
	[email protected]
Cc: [email protected]
Subject: Re: io_uring process termination/killing is not working
Date: Wed, 12 Aug 2020 17:32:02 -0600	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>

On 8/12/20 12:28 PM, Pavel Begunkov wrote:
> On 12/08/2020 21:22, Pavel Begunkov wrote:
>> On 12/08/2020 21:20, Pavel Begunkov wrote:
>>> On 12/08/2020 21:05, Jens Axboe wrote:
>>>> On 8/12/20 11:58 AM, Josef wrote:
>>>>> Hi,
>>>>>
>>>>> I have a weird issue on kernel 5.8.0/5.8.1, SIGINT even SIGKILL
>>>>> doesn't work to kill this process(always state D or D+), literally I
>>>>> have to terminate my VM because even the kernel can't kill the process
>>>>> and no issue on 5.7.12-201, however if IOSQE_IO_LINK is not set, it
>>>>> works
>>>>>
>>>>> I've attached a file to reproduce it
>>>>> or here
>>>>> https://gist.github.com/1Jo1/15cb3c63439d0c08e3589cfa98418b2c
>>>>
>>>> Thanks, I'll take a look at this. It's stuck in uninterruptible
>>>> state, which is why you can't kill it.
>>>
>>> It looks like one of the hangs I've been talking about a few days ago,
>>> an accept is inflight but can't be found by cancel_files() because it's
>>> in a link.
>>
>> BTW, I described it a month ago, there were more details.
> 
> https://lore.kernel.org/io-uring/[email protected]

Yeah I think you're right. How about something like the below? That'll
potentially cancel more than just the one we're looking for, but seems
kind of silly to only cancel from the file table holding request and to
the end.


diff --git a/fs/io_uring.c b/fs/io_uring.c
index 8a2afd8c33c9..0630a9622baa 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -4937,6 +5003,7 @@ static bool io_poll_remove_one(struct io_kiocb *req)
 		io_cqring_fill_event(req, -ECANCELED);
 		io_commit_cqring(req->ctx);
 		req->flags |= REQ_F_COMP_LOCKED;
+		req_set_fail_links(req);
 		io_put_req(req);
 	}
 
@@ -7935,6 +8002,47 @@ static bool io_wq_files_match(struct io_wq_work *work, void *data)
 	return work->files == files;
 }
 
+static bool __io_poll_remove_link(struct io_kiocb *preq, struct io_kiocb *req)
+{
+	struct io_kiocb *link;
+
+	if (!(preq->flags & REQ_F_LINK_HEAD))
+		return false;
+
+	list_for_each_entry(link, &preq->link_list, link_list) {
+		if (link != req)
+			break;
+		io_poll_remove_one(preq);
+		return true;
+	}
+
+	return false;
+}
+
+/*
+ * We're looking to cancel 'req' because it's holding on to our files, but
+ * 'req' could be a link to another request. See if it is, and cancel that
+ * parent request if so.
+ */
+static void io_poll_remove_link(struct io_ring_ctx *ctx, struct io_kiocb *req)
+{
+	struct hlist_node *tmp;
+	struct io_kiocb *preq;
+	int i;
+
+	spin_lock_irq(&ctx->completion_lock);
+	for (i = 0; i < (1U << ctx->cancel_hash_bits); i++) {
+		struct hlist_head *list;
+
+		list = &ctx->cancel_hash[i];
+		hlist_for_each_entry_safe(preq, tmp, list, hash_node) {
+			if (__io_poll_remove_link(preq, req))
+				break;
+		}
+	}
+	spin_unlock_irq(&ctx->completion_lock);
+}
+
 static void io_uring_cancel_files(struct io_ring_ctx *ctx,
 				  struct files_struct *files)
 {
@@ -7989,6 +8097,8 @@ static void io_uring_cancel_files(struct io_ring_ctx *ctx,
 			}
 		} else {
 			io_wq_cancel_work(ctx->io_wq, &cancel_req->work);
+			/* could be a link, check and remove if it is */
+			io_poll_remove_link(ctx, cancel_req);
 			io_put_req(cancel_req);
 		}
 

-- 
Jens Axboe


  reply	other threads:[~2020-08-12 23:32 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-08-12 17:58 io_uring process termination/killing is not working Josef
2020-08-12 18:05 ` Jens Axboe
2020-08-12 18:20   ` Pavel Begunkov
2020-08-12 18:22     ` Pavel Begunkov
2020-08-12 18:28       ` Pavel Begunkov
2020-08-12 23:32         ` Jens Axboe [this message]
2020-08-13 16:07           ` Josef
2020-08-13 16:09             ` Jens Axboe
2020-08-15  7:45           ` Pavel Begunkov
2020-08-15 15:12             ` Jens Axboe
2020-08-15 16:48               ` Pavel Begunkov
2020-08-15 21:43                 ` Josef
2020-08-15 22:35                   ` Jens Axboe
2020-08-15 23:21                     ` Josef
2020-08-15 23:31                     ` Jens Axboe
2020-08-16  0:36                       ` Josef
2020-08-16  0:41                         ` Jens Axboe
2020-08-16  1:21                           ` Jens Axboe
2020-08-16  3:14                             ` Josef
2020-08-16  3:20                               ` Jens Axboe
2020-08-16 17:30                                 ` Jens Axboe
2020-08-16 21:09                                   ` Josef
2020-08-16 22:17                                     ` Jens Axboe
2020-08-17  8:58                                       ` Josef
2020-08-17 10:08                                         ` Pavel Begunkov
2020-08-16 13:45                 ` Jens Axboe
2020-08-16 14:53                   ` Jens Axboe
2020-08-16 15:22                     ` Jens Axboe
2020-08-17 10:16                       ` 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] \
    /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