All of lore.kernel.org
 help / color / mirror / Atom feed
From: Pavel Begunkov <asml.silence@gmail.com>
To: Jens Axboe <axboe@kernel.dk>, io-uring@vger.kernel.org
Cc: stable@vger.kernel.org
Subject: [PATCH 1/9] io_uring: fix leaking reg files on exit
Date: Tue, 13 Apr 2021 02:58:38 +0100	[thread overview]
Message-ID: <e696e9eade571b51997d0dc1d01f144c6d685c05.1618278933.git.asml.silence@gmail.com> (raw)
In-Reply-To: <cover.1618278933.git.asml.silence@gmail.com>

If io_sqe_files_unregister() faults on io_rsrc_ref_quiesce(), it will
fail to do unregister leaving files referenced. And that may well happen
because of a strayed signal or just because it does allocations inside.

In io_ring_ctx_free() do an unsafe version of unregister, as it's
guaranteed to not have requests by that point and so quiesce is useless.

Cc: stable@vger.kernel.org
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
---
 fs/io_uring.c | 29 +++++++++++++++--------------
 1 file changed, 15 insertions(+), 14 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 257eddd4cd82..44342ff5c4e1 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -7094,6 +7094,10 @@ static void __io_sqe_files_unregister(struct io_ring_ctx *ctx)
 			fput(file);
 	}
 #endif
+	io_free_file_tables(&ctx->file_table, ctx->nr_user_files);
+	kfree(ctx->file_data);
+	ctx->file_data = NULL;
+	ctx->nr_user_files = 0;
 }
 
 static inline void io_rsrc_ref_lock(struct io_ring_ctx *ctx)
@@ -7200,21 +7204,14 @@ static struct io_rsrc_data *io_rsrc_data_alloc(struct io_ring_ctx *ctx,
 
 static int io_sqe_files_unregister(struct io_ring_ctx *ctx)
 {
-	struct io_rsrc_data *data = ctx->file_data;
 	int ret;
 
-	if (!data)
+	if (!ctx->file_data)
 		return -ENXIO;
-	ret = io_rsrc_ref_quiesce(data, ctx);
-	if (ret)
-		return ret;
-
-	__io_sqe_files_unregister(ctx);
-	io_free_file_tables(&ctx->file_table, ctx->nr_user_files);
-	kfree(data);
-	ctx->file_data = NULL;
-	ctx->nr_user_files = 0;
-	return 0;
+	ret = io_rsrc_ref_quiesce(ctx->file_data, ctx);
+	if (!ret)
+		__io_sqe_files_unregister(ctx);
+	return ret;
 }
 
 static void io_sq_thread_unpark(struct io_sq_data *sqd)
@@ -7664,7 +7661,7 @@ static int io_sqe_files_register(struct io_ring_ctx *ctx, void __user *arg,
 
 	ret = io_sqe_files_scm(ctx);
 	if (ret) {
-		io_sqe_files_unregister(ctx);
+		__io_sqe_files_unregister(ctx);
 		return ret;
 	}
 
@@ -8465,7 +8462,11 @@ static void io_ring_ctx_free(struct io_ring_ctx *ctx)
 	}
 
 	mutex_lock(&ctx->uring_lock);
-	io_sqe_files_unregister(ctx);
+	if (ctx->file_data) {
+		if (!atomic_dec_and_test(&ctx->file_data->refs))
+			wait_for_completion(&ctx->file_data->done);
+		__io_sqe_files_unregister(ctx);
+	}
 	if (ctx->rings)
 		__io_cqring_overflow_flush(ctx, true);
 	mutex_unlock(&ctx->uring_lock);
-- 
2.24.0


  reply	other threads:[~2021-04-13  2:03 UTC|newest]

Thread overview: 12+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-04-13  1:58 [PATCH 5.13 0/9] another 5.13 pack Pavel Begunkov
2021-04-13  1:58 ` Pavel Begunkov [this message]
2021-04-13  1:58 ` [PATCH 2/9] io_uring: fix uninit old data for poll event upd Pavel Begunkov
2021-04-13  1:58 ` [PATCH 3/9] io_uring: split poll and poll update structures Pavel Begunkov
2021-04-13 17:14   ` Pavel Begunkov
2021-04-13  1:58 ` [PATCH 4/9] io_uring: add timeout completion_lock annotation Pavel Begunkov
2021-04-13  1:58 ` [PATCH 5/9] io_uring: refactor hrtimer_try_to_cancel uses Pavel Begunkov
2021-04-13  1:58 ` [PATCH 6/9] io_uring: clean up io_poll_remove_waitqs() Pavel Begunkov
2021-04-13  1:58 ` [PATCH 7/9] io_uring: don't fail overflow on in_idle Pavel Begunkov
2021-04-13  1:58 ` [PATCH 8/9] io_uring: skip futile iopoll iterations Pavel Begunkov
2021-04-13  1:58 ` [PATCH 9/9] io_uring: inline io_iopoll_getevents() Pavel Begunkov
2021-04-13 15:38 ` [PATCH 5.13 0/9] another 5.13 pack 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=e696e9eade571b51997d0dc1d01f144c6d685c05.1618278933.git.asml.silence@gmail.com \
    --to=asml.silence@gmail.com \
    --cc=axboe@kernel.dk \
    --cc=io-uring@vger.kernel.org \
    --cc=stable@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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.