From: Pavel Begunkov <[email protected]>
To: Jens Axboe <[email protected]>, [email protected]
Subject: [PATCH v2 3/5] io_uring: fix POLL_REMOVE removing apoll
Date: Wed, 14 Apr 2021 13:38:35 +0100 [thread overview]
Message-ID: <50808409506c69dcaaa8aefdc49689e6b25c1522.1618403742.git.asml.silence@gmail.com> (raw)
In-Reply-To: <[email protected]>
Don't allow REQ_OP_POLL_REMOVE to kill apoll requests, users should not
know about it. Also, remove weird -EACCESS in io_poll_update(), it
shouldn't know anything about apoll, and have to work even if happened
to have a poll and an async poll'ed request with same user_data.
Signed-off-by: Pavel Begunkov <[email protected]>
---
fs/io_uring.c | 23 +++++++++++------------
1 file changed, 11 insertions(+), 12 deletions(-)
diff --git a/fs/io_uring.c b/fs/io_uring.c
index 6a70bf455c49..ce75a859a376 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -5258,7 +5258,8 @@ static bool io_poll_remove_all(struct io_ring_ctx *ctx, struct task_struct *tsk,
return posted != 0;
}
-static struct io_kiocb *io_poll_find(struct io_ring_ctx *ctx, __u64 sqe_addr)
+static struct io_kiocb *io_poll_find(struct io_ring_ctx *ctx, __u64 sqe_addr,
+ bool poll_only)
__must_hold(&ctx->completion_lock)
{
struct hlist_head *list;
@@ -5268,18 +5269,20 @@ static struct io_kiocb *io_poll_find(struct io_ring_ctx *ctx, __u64 sqe_addr)
hlist_for_each_entry(req, list, hash_node) {
if (sqe_addr != req->user_data)
continue;
+ if (poll_only && req->opcode != IORING_OP_POLL_ADD)
+ continue;
return req;
}
-
return NULL;
}
-static int io_poll_cancel(struct io_ring_ctx *ctx, __u64 sqe_addr)
+static int io_poll_cancel(struct io_ring_ctx *ctx, __u64 sqe_addr,
+ bool poll_only)
__must_hold(&ctx->completion_lock)
{
struct io_kiocb *req;
- req = io_poll_find(ctx, sqe_addr);
+ req = io_poll_find(ctx, sqe_addr, poll_only);
if (!req)
return -ENOENT;
if (io_poll_remove_one(req))
@@ -5311,7 +5314,7 @@ static int io_poll_remove(struct io_kiocb *req, unsigned int issue_flags)
int ret;
spin_lock_irq(&ctx->completion_lock);
- ret = io_poll_cancel(ctx, req->poll_remove.addr);
+ ret = io_poll_cancel(ctx, req->poll_remove.addr, true);
spin_unlock_irq(&ctx->completion_lock);
if (ret < 0)
@@ -5412,14 +5415,10 @@ static int io_poll_update(struct io_kiocb *req)
int ret;
spin_lock_irq(&ctx->completion_lock);
- preq = io_poll_find(ctx, req->poll_update.old_user_data);
+ preq = io_poll_find(ctx, req->poll_update.old_user_data, true);
if (!preq) {
ret = -ENOENT;
goto err;
- } else if (preq->opcode != IORING_OP_POLL_ADD) {
- /* don't allow internal poll updates */
- ret = -EACCES;
- goto err;
}
/*
@@ -5748,7 +5747,7 @@ static void io_async_find_and_cancel(struct io_ring_ctx *ctx,
ret = io_timeout_cancel(ctx, sqe_addr);
if (ret != -ENOENT)
goto done;
- ret = io_poll_cancel(ctx, sqe_addr);
+ ret = io_poll_cancel(ctx, sqe_addr, false);
done:
if (!ret)
ret = success_ret;
@@ -5790,7 +5789,7 @@ static int io_async_cancel(struct io_kiocb *req, unsigned int issue_flags)
ret = io_timeout_cancel(ctx, sqe_addr);
if (ret != -ENOENT)
goto done;
- ret = io_poll_cancel(ctx, sqe_addr);
+ ret = io_poll_cancel(ctx, sqe_addr, false);
if (ret != -ENOENT)
goto done;
spin_unlock_irq(&ctx->completion_lock);
--
2.24.0
next prev parent reply other threads:[~2021-04-14 12:43 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-04-14 12:38 [PATCH v2 5.13 0/5] poll update into poll remove Pavel Begunkov
2021-04-14 12:38 ` [PATCH v2 1/5] io_uring: improve sqpoll event/state handling Pavel Begunkov
2021-04-14 16:41 ` Jens Axboe
2021-04-14 19:27 ` Pavel Begunkov
2021-04-14 12:38 ` [PATCH v2 2/5] io_uring: refactor io_ring_exit_work() Pavel Begunkov
2021-04-14 12:38 ` Pavel Begunkov [this message]
2021-04-14 12:38 ` [PATCH v2 4/5] io_uring: add helper for parsing poll events Pavel Begunkov
2021-04-14 12:38 ` [PATCH v2 5/5] io_uring: move poll update into remove not add Pavel Begunkov
2021-04-14 16:20 ` [PATCH v2 5.13 0/5] poll update into poll remove Jens Axboe
2021-04-14 19:32 ` 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 \
--in-reply-to=50808409506c69dcaaa8aefdc49689e6b25c1522.1618403742.git.asml.silence@gmail.com \
[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