summaryrefslogtreecommitdiffstats
path: root/fs/io-wq.c
diff options
context:
space:
mode:
authorJens Axboe <axboe@kernel.dk>2019-11-13 06:31:31 +0100
committerJens Axboe <axboe@kernel.dk>2019-11-13 19:37:54 +0100
commit7d7230652e7c788ef908536fd79f4cca077f269f (patch)
tree111768c5fef3d709f65a336ba42aaefd772c1efe /fs/io-wq.c
parentio_uring: check for validity of ->rings in teardown (diff)
downloadlinux-7d7230652e7c788ef908536fd79f4cca077f269f.tar.xz
linux-7d7230652e7c788ef908536fd79f4cca077f269f.zip
io_wq: add get/put_work handlers to io_wq_create()
For cancellation, we need to ensure that the work item stays valid for as long as ->cur_work is valid. Right now we can't safely dereference the work item even under the wqe->lock, because while the ->cur_work pointer will remain valid, the work could be completing and be freed in parallel. Only invoke ->get/put_work() on items we know that the caller queued themselves. Add IO_WQ_WORK_INTERNAL for io-wq to use, which is needed when we're queueing a flush item, for instance. Signed-off-by: Jens Axboe <axboe@kernel.dk>
Diffstat (limited to 'fs/io-wq.c')
-rw-r--r--fs/io-wq.c25
1 files changed, 23 insertions, 2 deletions
diff --git a/fs/io-wq.c b/fs/io-wq.c
index 33b14b85752b..26d81540c1fc 100644
--- a/fs/io-wq.c
+++ b/fs/io-wq.c
@@ -106,6 +106,9 @@ struct io_wq {
unsigned long state;
unsigned nr_wqes;
+ get_work_fn *get_work;
+ put_work_fn *put_work;
+
struct task_struct *manager;
struct user_struct *user;
struct mm_struct *mm;
@@ -392,7 +395,7 @@ static struct io_wq_work *io_get_next_work(struct io_wqe *wqe, unsigned *hash)
static void io_worker_handle_work(struct io_worker *worker)
__releases(wqe->lock)
{
- struct io_wq_work *work, *old_work;
+ struct io_wq_work *work, *old_work = NULL, *put_work = NULL;
struct io_wqe *wqe = worker->wqe;
struct io_wq *wq = wqe->wq;
@@ -424,6 +427,8 @@ static void io_worker_handle_work(struct io_worker *worker)
wqe->flags |= IO_WQE_FLAG_STALLED;
spin_unlock_irq(&wqe->lock);
+ if (put_work && wq->put_work)
+ wq->put_work(old_work);
if (!work)
break;
next:
@@ -444,6 +449,11 @@ next:
if (worker->mm)
work->flags |= IO_WQ_WORK_HAS_MM;
+ if (wq->get_work && !(work->flags & IO_WQ_WORK_INTERNAL)) {
+ put_work = work;
+ wq->get_work(work);
+ }
+
old_work = work;
work->func(&work);
@@ -455,6 +465,12 @@ next:
}
if (work && work != old_work) {
spin_unlock_irq(&wqe->lock);
+
+ if (put_work && wq->put_work) {
+ wq->put_work(put_work);
+ put_work = NULL;
+ }
+
/* dependent work not hashed */
hash = -1U;
goto next;
@@ -950,13 +966,15 @@ void io_wq_flush(struct io_wq *wq)
init_completion(&data.done);
INIT_IO_WORK(&data.work, io_wq_flush_func);
+ data.work.flags |= IO_WQ_WORK_INTERNAL;
io_wqe_enqueue(wqe, &data.work);
wait_for_completion(&data.done);
}
}
struct io_wq *io_wq_create(unsigned bounded, struct mm_struct *mm,
- struct user_struct *user)
+ struct user_struct *user, get_work_fn *get_work,
+ put_work_fn *put_work)
{
int ret = -ENOMEM, i, node;
struct io_wq *wq;
@@ -972,6 +990,9 @@ struct io_wq *io_wq_create(unsigned bounded, struct mm_struct *mm,
return ERR_PTR(-ENOMEM);
}
+ wq->get_work = get_work;
+ wq->put_work = put_work;
+
/* caller must already hold a reference to this */
wq->user = user;