netdev - [PATCH 11/33] aio: try to complete poll iocbs without context switch

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 23 May 2018 21:20:00 +0200
From:   Christoph Hellwig <hch@....de>
To:     viro@...iv.linux.org.uk
Cc:     Avi Kivity <avi@...lladb.com>, linux-aio@...ck.org,
        linux-fsdevel@...r.kernel.org, netdev@...r.kernel.org,
        linux-api@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: [PATCH 11/33] aio: try to complete poll iocbs without context switch

If we can acquire ctx_lock without spinning we can just remove our
iocb from the active_reqs list, and thus complete the iocbs from the
wakeup context.

Signed-off-by: Christoph Hellwig <hch@....de>
---
 fs/aio.c | 20 +++++++++++++++++---
 1 file changed, 17 insertions(+), 3 deletions(-)

diff --git a/fs/aio.c b/fs/aio.c
index bd711cfa4a1f..8274d09d44a2 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -1633,6 +1633,7 @@ static int aio_poll_wake(struct wait_queue_entry *wait, unsigned mode, int sync,
 		void *key)
 {
 	struct poll_iocb *req = container_of(wait, struct poll_iocb, wait);
+	struct aio_kiocb *iocb = container_of(req, struct aio_kiocb, poll);
 	struct file *file = req->file;
 	__poll_t mask = key_to_poll(key);
 
@@ -1648,9 +1649,22 @@ static int aio_poll_wake(struct wait_queue_entry *wait, unsigned mode, int sync,
 
 	__aio_poll_remove(req);
 
-	req->events = mask;
-	INIT_WORK(&req->work, aio_poll_work);
-	schedule_work(&req->work);
+	/*
+	 * Try completing without a context switch if we can acquire ctx_lock
+	 * without spinning.  Otherwise we need to defer to a workqueue to
+	 * avoid a deadlock due to the lock order.
+	 */
+	if (spin_trylock(&iocb->ki_ctx->ctx_lock)) {
+		list_del_init(&iocb->ki_list);
+		spin_unlock(&iocb->ki_ctx->ctx_lock);
+
+		__aio_poll_complete(req, mask);
+	} else {
+		req->events = mask;
+		INIT_WORK(&req->work, aio_poll_work);
+		schedule_work(&req->work);
+	}
+
 	return 1;
 }
 
-- 
2.17.0