From: Pavel Begunkov <asml.silence@gmail.com>
To: Jens Axboe <axboe@kernel.dk>, io-uring@vger.kernel.org
Subject: [PATCH 16/16] io_uring: return back safer resurrect
Date: Sun, 11 Apr 2021 01:46:40 +0100 [thread overview]
Message-ID: <7a080c20f686d026efade810b116b72f88abaff9.1618101759.git.asml.silence@gmail.com> (raw)
In-Reply-To: <cover.1618101759.git.asml.silence@gmail.com>
Revert of revert of "io_uring: wait potential ->release() on resurrect",
which adds a helper for resurrect not racing completion reinit, as was
removed because of a strange bug with no clear root or link to the
patch.
Was improved, instead of rcu_synchronize(), just wait_for_completion()
because we're at 0 refs and it will happen very shortly. Specifically
use non-interruptible version to ignore all pending signals that may
have ended prior interruptible wait.
This reverts commit cb5e1b81304e089ee3ca948db4d29f71902eb575.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
---
fs/io_uring.c | 18 ++++++++++++++----
1 file changed, 14 insertions(+), 4 deletions(-)
diff --git a/fs/io_uring.c b/fs/io_uring.c
index 2a465b6e90a4..257eddd4cd82 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -1083,6 +1083,18 @@ static inline void io_req_set_rsrc_node(struct io_kiocb *req)
}
}
+static void io_refs_resurrect(struct percpu_ref *ref, struct completion *compl)
+{
+ bool got = percpu_ref_tryget(ref);
+
+ /* already at zero, wait for ->release() */
+ if (!got)
+ wait_for_completion(compl);
+ percpu_ref_resurrect(ref);
+ if (got)
+ percpu_ref_put(ref);
+}
+
static bool io_match_task(struct io_kiocb *head,
struct task_struct *task,
struct files_struct *files)
@@ -9798,12 +9810,11 @@ static int __io_uring_register(struct io_ring_ctx *ctx, unsigned opcode,
if (ret < 0)
break;
} while (1);
-
mutex_lock(&ctx->uring_lock);
if (ret) {
- percpu_ref_resurrect(&ctx->refs);
- goto out_quiesce;
+ io_refs_resurrect(&ctx->refs, &ctx->ref_comp);
+ return ret;
}
}
@@ -9896,7 +9907,6 @@ static int __io_uring_register(struct io_ring_ctx *ctx, unsigned opcode,
if (io_register_op_must_quiesce(opcode)) {
/* bring the ctx back to life */
percpu_ref_reinit(&ctx->refs);
-out_quiesce:
reinit_completion(&ctx->ref_comp);
}
return ret;
--
2.24.0
next prev parent reply other threads:[~2021-04-11 0:51 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-04-11 0:46 [PATCH 5.13 00/16] random 5.13 patches Pavel Begunkov
2021-04-11 0:46 ` [PATCH 01/16] io_uring: unify task and files cancel loops Pavel Begunkov
2021-04-11 0:46 ` [PATCH 02/16] io_uring: track inflight requests through counter Pavel Begunkov
2021-04-11 0:46 ` [PATCH 03/16] io_uring: unify files and task cancel Pavel Begunkov
2021-04-11 0:46 ` [PATCH 04/16] io_uring: refactor io_close Pavel Begunkov
2021-04-11 0:46 ` [PATCH 05/16] io_uring: enable inline completion for more cases Pavel Begunkov
2021-04-11 0:46 ` [PATCH 06/16] io_uring: refactor compat_msghdr import Pavel Begunkov
2021-04-11 0:46 ` [PATCH 07/16] io_uring: optimise non-eventfd post-event Pavel Begunkov
2021-04-11 0:46 ` [PATCH 08/16] io_uring: always pass cflags into fill_event() Pavel Begunkov
2021-04-11 0:46 ` [PATCH 09/16] io_uring: optimise fill_event() by inlining Pavel Begunkov
2021-04-11 0:46 ` [PATCH 10/16] io_uring: simplify io_rsrc_data refcounting Pavel Begunkov
2021-04-11 0:46 ` [PATCH 11/16] io_uring: add buffer unmap helper Pavel Begunkov
2021-04-11 0:46 ` [PATCH 12/16] io_uring: cleanup buffer register Pavel Begunkov
2021-04-11 0:46 ` [PATCH 13/16] io_uring: split file table from rsrc nodes Pavel Begunkov
2021-04-11 0:46 ` [PATCH 14/16] io_uring: improve sqo stop Pavel Begunkov
2021-04-11 0:46 ` [PATCH 15/16] io_uring: improve hardlink code generation Pavel Begunkov
2021-04-11 0:46 ` Pavel Begunkov [this message]
2021-05-10 2:22 ` [PATCH 16/16] io_uring: return back safer resurrect yangerkun
2021-05-10 9:15 ` Pavel Begunkov
2021-05-11 1:11 ` yangerkun
2022-03-16 16:18 ` [PATCH] " Lee Jones
2022-03-16 16:38 ` Greg KH
2022-03-16 16:46 ` Lee Jones
2021-04-11 2:38 ` [PATCH 5.13 00/16] random 5.13 patches 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=7a080c20f686d026efade810b116b72f88abaff9.1618101759.git.asml.silence@gmail.com \
--to=asml.silence@gmail.com \
--cc=axboe@kernel.dk \
--cc=io-uring@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).