linux-kernel - [PATCH V6 07/10] io_uring: switch to user

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]
Message-Id: <20211129194707.5863-8-michael.christie@oracle.com>
Date:   Mon, 29 Nov 2021 13:47:04 -0600
From:   Mike Christie <michael.christie@...cle.com>
To:     geert@...ux-m68k.org, vverma@...italocean.com, hdanton@...a.com,
        hch@...radead.org, stefanha@...hat.com, jasowang@...hat.com,
        mst@...hat.com, sgarzare@...hat.com,
        virtualization@...ts.linux-foundation.org,
        christian.brauner@...ntu.com, axboe@...nel.dk,
        linux-kernel@...r.kernel.org
Cc:     Mike Christie <michael.christie@...cle.com>
Subject: [PATCH V6 07/10] io_uring: switch to user_worker

The user_worker_create/start helpers allow callers to create threads that
copy or inherit the caller's attributes like mm, cgroups, namespaces, etc,
and are accounted for under the callers rlimits nproc value similar to if
the caller did a clone() in userspace. However, instead of returning to
userspace the thread is usable in the kernel for modules like vhost or
layers like io_uring.

The user_worker_create function provides the same functionality as
create_io_thread while also allowing different signal and file handling
for users like vhost, so this patch converts io_uring and io-wq. The next
patch will then remove the create_io_thread code.

Signed-off-by: Mike Christie <michael.christie@...cle.com>
---
 fs/io-wq.c    | 15 ++++++++-------
 fs/io_uring.c | 11 +++++------
 2 files changed, 13 insertions(+), 13 deletions(-)

diff --git a/fs/io-wq.c b/fs/io-wq.c
index 88202de519f6..761796c14d3b 100644
--- a/fs/io-wq.c
+++ b/fs/io-wq.c
@@ -70,6 +70,9 @@ struct io_worker {
 
 #define IO_WQ_NR_HASH_BUCKETS	(1u << IO_WQ_HASH_ORDER)
 
+#define IO_WQ_CLONE_FLAGS	(CLONE_FS | CLONE_FILES | CLONE_SIGHAND | \
+				 CLONE_THREAD | CLONE_IO)
+
 struct io_wqe_acct {
 	unsigned nr_workers;
 	unsigned max_workers;
@@ -600,13 +603,9 @@ static int io_wqe_worker(void *data)
 	struct io_wqe *wqe = worker->wqe;
 	struct io_wq *wq = wqe->wq;
 	bool last_timeout = false;
-	char buf[TASK_COMM_LEN];
 
 	worker->flags |= (IO_WORKER_F_UP | IO_WORKER_F_RUNNING);
 
-	snprintf(buf, sizeof(buf), "iou-wrk-%d", wq->task->pid);
-	set_task_comm(current, buf);
-
 	audit_alloc_kernel(current);
 
 	while (!test_bit(IO_WQ_BIT_EXIT, &wq->state)) {
@@ -704,7 +703,7 @@ static void io_init_new_worker(struct io_wqe *wqe, struct io_worker *worker,
 	list_add_tail_rcu(&worker->all_list, &wqe->all_list);
 	worker->flags |= IO_WORKER_F_FREE;
 	raw_spin_unlock(&wqe->lock);
-	wake_up_new_task(tsk);
+	user_worker_start(tsk, "iou-wrk-%d", wqe->wq->task->pid);
 }
 
 static bool io_wq_work_match_all(struct io_wq_work *work, void *data)
@@ -734,7 +733,8 @@ static void create_worker_cont(struct callback_head *cb)
 	worker = container_of(cb, struct io_worker, create_work);
 	clear_bit_unlock(0, &worker->create_state);
 	wqe = worker->wqe;
-	tsk = create_io_thread(io_wqe_worker, worker, wqe->node);
+	tsk = user_worker_create(io_wqe_worker, worker, wqe->node,
+				 IO_WQ_CLONE_FLAGS, USER_WORKER_IO);
 	if (!IS_ERR(tsk)) {
 		io_init_new_worker(wqe, worker, tsk);
 		io_worker_release(worker);
@@ -801,7 +801,8 @@ static bool create_io_worker(struct io_wq *wq, struct io_wqe *wqe, int index)
 	if (index == IO_WQ_ACCT_BOUND)
 		worker->flags |= IO_WORKER_F_BOUND;
 
-	tsk = create_io_thread(io_wqe_worker, worker, wqe->node);
+	tsk = user_worker_create(io_wqe_worker, worker, wqe->node,
+				 IO_WQ_CLONE_FLAGS, USER_WORKER_IO);
 	if (!IS_ERR(tsk)) {
 		io_init_new_worker(wqe, worker, tsk);
 	} else if (!io_should_retry_thread(PTR_ERR(tsk))) {
diff --git a/fs/io_uring.c b/fs/io_uring.c
index c4f217613f56..d275f01c0828 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -7475,12 +7475,8 @@ static int io_sq_thread(void *data)
 	struct io_sq_data *sqd = data;
 	struct io_ring_ctx *ctx;
 	unsigned long timeout = 0;
-	char buf[TASK_COMM_LEN];
 	DEFINE_WAIT(wait);
 
-	snprintf(buf, sizeof(buf), "iou-sqp-%d", sqd->task_pid);
-	set_task_comm(current, buf);
-
 	if (sqd->sq_cpu != -1)
 		set_cpus_allowed_ptr(current, cpumask_of(sqd->sq_cpu));
 	else
@@ -8709,6 +8705,8 @@ static __cold int io_sq_offload_create(struct io_ring_ctx *ctx,
 		fdput(f);
 	}
 	if (ctx->flags & IORING_SETUP_SQPOLL) {
+		unsigned long flags = CLONE_FS | CLONE_FILES | CLONE_SIGHAND |
+					CLONE_THREAD | CLONE_IO;
 		struct task_struct *tsk;
 		struct io_sq_data *sqd;
 		bool attached;
@@ -8754,7 +8752,8 @@ static __cold int io_sq_offload_create(struct io_ring_ctx *ctx,
 
 		sqd->task_pid = current->pid;
 		sqd->task_tgid = current->tgid;
-		tsk = create_io_thread(io_sq_thread, sqd, NUMA_NO_NODE);
+		tsk = user_worker_create(io_sq_thread, sqd, NUMA_NO_NODE, flags,
+					 USER_WORKER_IO);
 		if (IS_ERR(tsk)) {
 			ret = PTR_ERR(tsk);
 			goto err_sqpoll;
@@ -8762,7 +8761,7 @@ static __cold int io_sq_offload_create(struct io_ring_ctx *ctx,
 
 		sqd->thread = tsk;
 		ret = io_uring_alloc_task_context(tsk, ctx);
-		wake_up_new_task(tsk);
+		user_worker_start(tsk, "iou-sqp-%d", sqd->task_pid);
 		if (ret)
 			goto err;
 	} else if (p->flags & IORING_SETUP_SQ_AFF) {
-- 
2.25.1