struct aio_ring_info ring_info;
+ spinlock_t completion_lock;
+
struct rcu_head rcu_head;
struct work_struct rcu_work;
};
#define AIO_EVENTS_FIRST_PAGE ((PAGE_SIZE - sizeof(struct aio_ring)) / sizeof(struct io_event))
#define AIO_EVENTS_OFFSET (AIO_EVENTS_PER_PAGE - AIO_EVENTS_FIRST_PAGE)
+void kiocb_set_cancel_fn(struct kiocb *req, kiocb_cancel_fn *cancel)
+{
+ struct kioctx *ctx = req->ki_ctx;
+ unsigned long flags;
+
+ spin_lock_irqsave(&ctx->ctx_lock, flags);
+
+ if (!req->ki_list.next)
+ list_add(&req->ki_list, &ctx->active_reqs);
+
+ req->ki_cancel = cancel;
+
+ spin_unlock_irqrestore(&ctx->ctx_lock, flags);
+}
+EXPORT_SYMBOL(kiocb_set_cancel_fn);
+
static int kiocb_cancel(struct kioctx *ctx, struct kiocb *kiocb,
struct io_event *res)
{
- int (*cancel)(struct kiocb *, struct io_event *);
+ kiocb_cancel_fn *old, *cancel;
int ret = -EINVAL;
- cancel = kiocb->ki_cancel;
- kiocbSetCancelled(kiocb);
- if (cancel) {
- atomic_inc(&kiocb->ki_users);
- spin_unlock_irq(&ctx->ctx_lock);
+ /*
+ * Don't want to set kiocb->ki_cancel = KIOCB_CANCELLED unless it
+ * actually has a cancel function, hence the cmpxchg()
+ */
+
+ cancel = ACCESS_ONCE(kiocb->ki_cancel);
+ do {
+ if (!cancel || cancel == KIOCB_CANCELLED)
+ return ret;
- memset(res, 0, sizeof(*res));
- res->obj = (u64)(unsigned long)kiocb->ki_obj.user;
- res->data = kiocb->ki_user_data;
- ret = cancel(kiocb, res);
+ old = cancel;
+ cancel = cmpxchg(&kiocb->ki_cancel, old, KIOCB_CANCELLED);
+ } while (cancel != old);
- spin_lock_irq(&ctx->ctx_lock);
- }
+ atomic_inc(&kiocb->ki_users);
+ spin_unlock_irq(&ctx->ctx_lock);
+
+ memset(res, 0, sizeof(*res));
+ res->obj = (u64)(unsigned long)kiocb->ki_obj.user;
+ res->data = kiocb->ki_user_data;
+ ret = cancel(kiocb, res);
+
+ spin_lock_irq(&ctx->ctx_lock);
return ret;
}
atomic_set(&ctx->users, 2);
atomic_set(&ctx->dead, 0);
spin_lock_init(&ctx->ctx_lock);
+ spin_lock_init(&ctx->completion_lock);
mutex_init(&ctx->ring_info.ring_lock);
init_waitqueue_head(&ctx->wait);
{
struct kiocb *req = NULL;
- req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL);
+ req = kmem_cache_alloc(kiocb_cachep, GFP_KERNEL|__GFP_ZERO);
if (unlikely(!req))
return NULL;
- req->ki_flags = 0;
atomic_set(&req->ki_users, 2);
- req->ki_key = 0;
req->ki_ctx = ctx;
- req->ki_cancel = NULL;
- req->ki_retry = NULL;
- req->ki_dtor = NULL;
- req->private = NULL;
- req->ki_iovec = NULL;
- req->ki_eventfd = NULL;
return req;
}
spin_lock_irq(&ctx->ctx_lock);
list_for_each_entry_safe(req, n, &batch->head, ki_batch) {
list_del(&req->ki_batch);
- list_del(&req->ki_list);
kmem_cache_free(kiocb_cachep, req);
atomic_dec(&ctx->reqs_active);
}
}
batch->count -= allocated;
- list_for_each_entry(req, &batch->head, ki_batch) {
- list_add(&req->ki_list, &ctx->active_reqs);
- atomic_inc(&ctx->reqs_active);
- }
+ atomic_add(allocated, &ctx->reqs_active);
kunmap_atomic(ring);
spin_unlock_irq(&ctx->ctx_lock);
info = &ctx->ring_info;
/*
- * Add a completion event to the ring buffer. Must be done holding
- * ctx->ctx_lock to prevent other code from messing with the tail
- * pointer since we might be called from irq context.
- *
* Take rcu_read_lock() in case the kioctx is being destroyed, as we
* need to issue a wakeup after decrementing reqs_active.
*/
rcu_read_lock();
- spin_lock_irqsave(&ctx->ctx_lock, flags);
- list_del(&iocb->ki_list); /* remove from active_reqs */
+ if (iocb->ki_list.next) {
+ unsigned long flags;
+
+ spin_lock_irqsave(&ctx->ctx_lock, flags);
+ list_del(&iocb->ki_list);
+ spin_unlock_irqrestore(&ctx->ctx_lock, flags);
+ }
/*
* cancelled requests don't get events, userland was given one
* when the event got cancelled.
*/
- if (kiocbIsCancelled(iocb))
+ if (unlikely(xchg(&iocb->ki_cancel,
+ KIOCB_CANCELLED) == KIOCB_CANCELLED))
goto put_rq;
+ /*
+ * Add a completion event to the ring buffer. Must be done holding
+ * ctx->ctx_lock to prevent other code from messing with the tail
+ * pointer since we might be called from irq context.
+ */
+ spin_lock_irqsave(&ctx->completion_lock, flags);
+
tail = info->tail;
pos = tail + AIO_EVENTS_OFFSET;
kunmap_atomic(ring);
flush_dcache_page(info->ring_pages[0]);
+ spin_unlock_irqrestore(&ctx->completion_lock, flags);
+
pr_debug("added to ring %p at [%u]\n", iocb, tail);
/*
if (waitqueue_active(&ctx->wait))
wake_up(&ctx->wait);
- spin_unlock_irqrestore(&ctx->ctx_lock, flags);
rcu_read_unlock();
}
EXPORT_SYMBOL(aio_complete);
req->ki_opcode = iocb->aio_lio_opcode;
ret = aio_setup_iocb(req, compat);
-
if (ret)
goto out_put_req;
- if (unlikely(kiocbIsCancelled(req)))
- ret = -EINTR;
- else
- ret = req->ki_retry(req);
-
+ ret = req->ki_retry(req);
if (ret != -EIOCBQUEUED) {
/*
* There's no easy way to restart the syscall since other AIO's
return 0;
out_put_req:
- spin_lock_irq(&ctx->ctx_lock);
- list_del(&req->ki_list);
- spin_unlock_irq(&ctx->ctx_lock);
-
atomic_dec(&ctx->reqs_active);
aio_put_req(req); /* drop extra ref to req */
aio_put_req(req); /* drop i/o ref to req */
#include <linux/atomic.h>
struct kioctx;
+struct kiocb;
#define KIOCB_SYNC_KEY (~0U)
-/* ki_flags bits */
-#define KIF_CANCELLED 2
-
-#define kiocbSetCancelled(iocb) set_bit(KIF_CANCELLED, &(iocb)->ki_flags)
-
-#define kiocbClearCancelled(iocb) clear_bit(KIF_CANCELLED, &(iocb)->ki_flags)
+/*
+ * We use ki_cancel == KIOCB_CANCELLED to indicate that a kiocb has been either
+ * cancelled or completed (this makes a certain amount of sense because
+ * successful cancellation - io_cancel() - does deliver the completion to
+ * userspace).
+ *
+ * And since most things don't implement kiocb cancellation and we'd really like
+ * kiocb completion to be lockless when possible, we use ki_cancel to
+ * synchronize cancellation and completion - we only set it to KIOCB_CANCELLED
+ * with xchg() or cmpxchg(), see batch_complete_aio() and kiocb_cancel().
+ */
+#define KIOCB_CANCELLED ((void *) (~0ULL))
-#define kiocbIsCancelled(iocb) test_bit(KIF_CANCELLED, &(iocb)->ki_flags)
+typedef int (kiocb_cancel_fn)(struct kiocb *, struct io_event *);
/* is there a better place to document function pointer methods? */
/**
* calls may result in undefined behaviour.
*/
struct kiocb {
- unsigned long ki_flags;
atomic_t ki_users;
unsigned ki_key; /* id of this request */
struct file *ki_filp;
struct kioctx *ki_ctx; /* may be NULL for sync ops */
- int (*ki_cancel)(struct kiocb *, struct io_event *);
+ kiocb_cancel_fn *ki_cancel;
ssize_t (*ki_retry)(struct kiocb *);
void (*ki_dtor)(struct kiocb *);
extern void exit_aio(struct mm_struct *mm);
extern long do_io_submit(aio_context_t ctx_id, long nr,
struct iocb __user *__user *iocbpp, bool compat);
+void kiocb_set_cancel_fn(struct kiocb *req, kiocb_cancel_fn *cancel);
#else
static inline ssize_t wait_on_sync_kiocb(struct kiocb *iocb) { return 0; }
static inline void aio_put_req(struct kiocb *iocb) { }
static inline long do_io_submit(aio_context_t ctx_id, long nr,
struct iocb __user * __user *iocbpp,
bool compat) { return 0; }
+static inline void kiocb_set_cancel_fn(struct kiocb *req,
+ kiocb_cancel_fn *cancel) { }
#endif /* CONFIG_AIO */
static inline struct kiocb *list_kiocb(struct list_head *h)
projects / karo-tx-linux.git / commitdiff