1 files changed, 83 insertions, 67 deletions

diff --git a/fs/aio.c b/fs/aio.c
index f8e52a1854c1..435ca29eca31 100644
--- a/fs/aio.c
+++ b/fs/aio.c
@@ -151,6 +151,38 @@ struct kioctx {
 	unsigned		id;
 };
 
+/*
+ * 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))
+
+struct aio_kiocb {
+	struct kiocb		common;
+
+	struct kioctx		*ki_ctx;
+	kiocb_cancel_fn		*ki_cancel;
+
+	struct iocb __user	*ki_user_iocb;	/* user's aiocb */
+	__u64			ki_user_data;	/* user's data for completion */
+
+	struct list_head	ki_list;	/* the aio core uses this
+						 * for cancellation */
+
+	/*
+	 * If the aio_resfd field of the userspace iocb is not zero,
+	 * this is the underlying eventfd context to deliver events to.
+	 */
+	struct eventfd_ctx	*ki_eventfd;
+};
+
 /*------ sysctl variables----*/
 static DEFINE_SPINLOCK(aio_nr_lock);
 unsigned long aio_nr;		/* current system wide number of aio requests */
@@ -220,7 +252,7 @@ static int __init aio_setup(void)
 	if (IS_ERR(aio_mnt))
 		panic("Failed to create aio fs mount.");
 
-	kiocb_cachep = KMEM_CACHE(kiocb, SLAB_HWCACHE_ALIGN|SLAB_PANIC);
+	kiocb_cachep = KMEM_CACHE(aio_kiocb, SLAB_HWCACHE_ALIGN|SLAB_PANIC);
 	kioctx_cachep = KMEM_CACHE(kioctx,SLAB_HWCACHE_ALIGN|SLAB_PANIC);
 
 	pr_debug("sizeof(struct page) = %zu\n", sizeof(struct page));
@@ -480,8 +512,9 @@ static int aio_setup_ring(struct kioctx *ctx)
 #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)
+void kiocb_set_cancel_fn(struct kiocb *iocb, kiocb_cancel_fn *cancel)
 {
+	struct aio_kiocb *req = container_of(iocb, struct aio_kiocb, common);
 	struct kioctx *ctx = req->ki_ctx;
 	unsigned long flags;
 
@@ -496,7 +529,7 @@ void kiocb_set_cancel_fn(struct kiocb *req, kiocb_cancel_fn *cancel)
 }
 EXPORT_SYMBOL(kiocb_set_cancel_fn);
 
-static int kiocb_cancel(struct kiocb *kiocb)
+static int kiocb_cancel(struct aio_kiocb *kiocb)
 {
 	kiocb_cancel_fn *old, *cancel;
 
@@ -514,7 +547,7 @@ static int kiocb_cancel(struct kiocb *kiocb)
 		cancel = cmpxchg(&kiocb->ki_cancel, old, KIOCB_CANCELLED);
 	} while (cancel != old);
 
-	return cancel(kiocb);
+	return cancel(&kiocb->common);
 }
 
 static void free_ioctx(struct work_struct *work)
@@ -550,13 +583,13 @@ static void free_ioctx_reqs(struct percpu_ref *ref)
 static void free_ioctx_users(struct percpu_ref *ref)
 {
 	struct kioctx *ctx = container_of(ref, struct kioctx, users);
-	struct kiocb *req;
+	struct aio_kiocb *req;
 
 	spin_lock_irq(&ctx->ctx_lock);
 
 	while (!list_empty(&ctx->active_reqs)) {
 		req = list_first_entry(&ctx->active_reqs,
-				       struct kiocb, ki_list);
+				       struct aio_kiocb, ki_list);
 
 		list_del_init(&req->ki_list);
 		kiocb_cancel(req);
@@ -778,22 +811,6 @@ static int kill_ioctx(struct mm_struct *mm, struct kioctx *ctx,
 	return 0;
 }
 
-/* wait_on_sync_kiocb:
- *	Waits on the given sync kiocb to complete.
- */
-ssize_t wait_on_sync_kiocb(struct kiocb *req)
-{
-	while (!req->ki_ctx) {
-		set_current_state(TASK_UNINTERRUPTIBLE);
-		if (req->ki_ctx)
-			break;
-		io_schedule();
-	}
-	__set_current_state(TASK_RUNNING);
-	return req->ki_user_data;
-}
-EXPORT_SYMBOL(wait_on_sync_kiocb);
-
 /*
  * exit_aio: called when the last user of mm goes away.  At this point, there is
  * no way for any new requests to be submited or any of the io_* syscalls to be
@@ -948,9 +965,9 @@ static void user_refill_reqs_available(struct kioctx *ctx)
  *	Allocate a slot for an aio request.
  * Returns NULL if no requests are free.
  */
-static inline struct kiocb *aio_get_req(struct kioctx *ctx)
+static inline struct aio_kiocb *aio_get_req(struct kioctx *ctx)
 {
-	struct kiocb *req;
+	struct aio_kiocb *req;
 
 	if (!get_reqs_available(ctx)) {
 		user_refill_reqs_available(ctx);
@@ -971,10 +988,10 @@ out_put:
 	return NULL;
 }
 
-static void kiocb_free(struct kiocb *req)
+static void kiocb_free(struct aio_kiocb *req)
 {
-	if (req->ki_filp)
-		fput(req->ki_filp);
+	if (req->common.ki_filp)
+		fput(req->common.ki_filp);
 	if (req->ki_eventfd != NULL)
 		eventfd_ctx_put(req->ki_eventfd);
 	kmem_cache_free(kiocb_cachep, req);
@@ -1010,8 +1027,9 @@ out:
 /* aio_complete
  *	Called when the io request on the given iocb is complete.
  */
-void aio_complete(struct kiocb *iocb, long res, long res2)
+static void aio_complete(struct kiocb *kiocb, long res, long res2)
 {
+	struct aio_kiocb *iocb = container_of(kiocb, struct aio_kiocb, common);
 	struct kioctx	*ctx = iocb->ki_ctx;
 	struct aio_ring	*ring;
 	struct io_event	*ev_page, *event;
@@ -1025,13 +1043,7 @@ void aio_complete(struct kiocb *iocb, long res, long res2)
 	 *    ref, no other paths have a way to get another ref
 	 *  - the sync task helpfully left a reference to itself in the iocb
 	 */
-	if (is_sync_kiocb(iocb)) {
-		iocb->ki_user_data = res;
-		smp_wmb();
-		iocb->ki_ctx = ERR_PTR(-EXDEV);
-		wake_up_process(iocb->ki_obj.tsk);
-		return;
-	}
+	BUG_ON(is_sync_kiocb(kiocb));
 
 	if (iocb->ki_list.next) {
 		unsigned long flags;
@@ -1057,7 +1069,7 @@ void aio_complete(struct kiocb *iocb, long res, long res2)
 	ev_page = kmap_atomic(ctx->ring_pages[pos / AIO_EVENTS_PER_PAGE]);
 	event = ev_page + pos % AIO_EVENTS_PER_PAGE;
 
-	event->obj = (u64)(unsigned long)iocb->ki_obj.user;
+	event->obj = (u64)(unsigned long)iocb->ki_user_iocb;
 	event->data = iocb->ki_user_data;
 	event->res = res;
 	event->res2 = res2;
@@ -1066,7 +1078,7 @@ void aio_complete(struct kiocb *iocb, long res, long res2)
 	flush_dcache_page(ctx->ring_pages[pos / AIO_EVENTS_PER_PAGE]);
 
 	pr_debug("%p[%u]: %p: %p %Lx %lx %lx\n",
-		 ctx, tail, iocb, iocb->ki_obj.user, iocb->ki_user_data,
+		 ctx, tail, iocb, iocb->ki_user_iocb, iocb->ki_user_data,
 		 res, res2);
 
 	/* after flagging the request as done, we
@@ -1113,7 +1125,6 @@ void aio_complete(struct kiocb *iocb, long res, long res2)
 
 	percpu_ref_put(&ctx->reqs);
 }
-EXPORT_SYMBOL(aio_complete);
 
 /* aio_read_events_ring
  *	Pull an event off of the ioctx's event ring.  Returns the number of
@@ -1344,12 +1355,13 @@ typedef ssize_t (rw_iter_op)(struct kiocb *, struct iov_iter *);
 static ssize_t aio_setup_vectored_rw(struct kiocb *kiocb,
 				     int rw, char __user *buf,
 				     unsigned long *nr_segs,
+				     size_t *len,
 				     struct iovec **iovec,
 				     bool compat)
 {
 	ssize_t ret;
 
-	*nr_segs = kiocb->ki_nbytes;
+	*nr_segs = *len;
 
 #ifdef CONFIG_COMPAT
 	if (compat)
@@ -1364,21 +1376,22 @@ static ssize_t aio_setup_vectored_rw(struct kiocb *kiocb,
 	if (ret < 0)
 		return ret;
 
-	/* ki_nbytes now reflect bytes instead of segs */
-	kiocb->ki_nbytes = ret;
+	/* len now reflect bytes instead of segs */
+	*len = ret;
 	return 0;
 }
 
 static ssize_t aio_setup_single_vector(struct kiocb *kiocb,
 				       int rw, char __user *buf,
 				       unsigned long *nr_segs,
+				       size_t len,
 				       struct iovec *iovec)
 {
-	if (unlikely(!access_ok(!rw, buf, kiocb->ki_nbytes)))
+	if (unlikely(!access_ok(!rw, buf, len)))
 		return -EFAULT;
 
 	iovec->iov_base = buf;
-	iovec->iov_len = kiocb->ki_nbytes;
+	iovec->iov_len = len;
 	*nr_segs = 1;
 	return 0;
 }
@@ -1388,7 +1401,7 @@ static ssize_t aio_setup_single_vector(struct kiocb *kiocb,
  *	Performs the initial checks and io submission.
  */
 static ssize_t aio_run_iocb(struct kiocb *req, unsigned opcode,
-			    char __user *buf, bool compat)
+			    char __user *buf, size_t len, bool compat)
 {
 	struct file *file = req->ki_filp;
 	ssize_t ret;
@@ -1423,21 +1436,21 @@ rw_common:
 		if (!rw_op && !iter_op)
 			return -EINVAL;
 
-		ret = (opcode == IOCB_CMD_PREADV ||
-		       opcode == IOCB_CMD_PWRITEV)
-			? aio_setup_vectored_rw(req, rw, buf, &nr_segs,
-						&iovec, compat)
-			: aio_setup_single_vector(req, rw, buf, &nr_segs,
-						  iovec);
+		if (opcode == IOCB_CMD_PREADV || opcode == IOCB_CMD_PWRITEV)
+			ret = aio_setup_vectored_rw(req, rw, buf, &nr_segs,
+						&len, &iovec, compat);
+		else
+			ret = aio_setup_single_vector(req, rw, buf, &nr_segs,
+						  len, iovec);
 		if (!ret)
-			ret = rw_verify_area(rw, file, &req->ki_pos, req->ki_nbytes);
+			ret = rw_verify_area(rw, file, &req->ki_pos, len);
 		if (ret < 0) {
 			if (iovec != inline_vecs)
 				kfree(iovec);
 			return ret;
 		}
 
-		req->ki_nbytes = ret;
+		len = ret;
 
 		/* XXX: move/kill - rw_verify_area()? */
 		/* This matches the pread()/pwrite() logic */
@@ -1450,7 +1463,7 @@ rw_common:
 			file_start_write(file);
 
 		if (iter_op) {
-			iov_iter_init(&iter, rw, iovec, nr_segs, req->ki_nbytes);
+			iov_iter_init(&iter, rw, iovec, nr_segs, len);
 			ret = iter_op(req, &iter);
 		} else {
 			ret = rw_op(req, iovec, nr_segs, req->ki_pos);
@@ -1500,7 +1513,7 @@ rw_common:
 static int io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb,
 			 struct iocb *iocb, bool compat)
 {
-	struct kiocb *req;
+	struct aio_kiocb *req;
 	ssize_t ret;
 
 	/* enforce forwards compatibility on users */
@@ -1523,11 +1536,14 @@ static int io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb,
 	if (unlikely(!req))
 		return -EAGAIN;
 
-	req->ki_filp = fget(iocb->aio_fildes);
-	if (unlikely(!req->ki_filp)) {
+	req->common.ki_filp = fget(iocb->aio_fildes);
+	if (unlikely(!req->common.ki_filp)) {
 		ret = -EBADF;
 		goto out_put_req;
 	}
+	req->common.ki_pos = iocb->aio_offset;
+	req->common.ki_complete = aio_complete;
+	req->common.ki_flags = 0;
 
 	if (iocb->aio_flags & IOCB_FLAG_RESFD) {
 		/*
@@ -1542,6 +1558,8 @@ static int io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb,
 			req->ki_eventfd = NULL;
 			goto out_put_req;
 		}
+
+		req->common.ki_flags |= IOCB_EVENTFD;
 	}
 
 	ret = put_user(KIOCB_KEY, &user_iocb->aio_key);
@@ -1550,13 +1568,12 @@ static int io_submit_one(struct kioctx *ctx, struct iocb __user *user_iocb,
 		goto out_put_req;
 	}
 
-	req->ki_obj.user = user_iocb;
+	req->ki_user_iocb = user_iocb;
 	req->ki_user_data = iocb->aio_data;
-	req->ki_pos = iocb->aio_offset;
-	req->ki_nbytes = iocb->aio_nbytes;
 
-	ret = aio_run_iocb(req, iocb->aio_lio_opcode,
+	ret = aio_run_iocb(&req->common, iocb->aio_lio_opcode,
 			   (char __user *)(unsigned long)iocb->aio_buf,
+			   iocb->aio_nbytes,
 			   compat);
 	if (ret)
 		goto out_put_req;
@@ -1643,10 +1660,10 @@ SYSCALL_DEFINE3(io_submit, aio_context_t, ctx_id, long, nr,
 /* lookup_kiocb
  *	Finds a given iocb for cancellation.
  */
-static struct kiocb *lookup_kiocb(struct kioctx *ctx, struct iocb __user *iocb,
-				  u32 key)
+static struct aio_kiocb *
+lookup_kiocb(struct kioctx *ctx, struct iocb __user *iocb, u32 key)
 {
-	struct list_head *pos;
+	struct aio_kiocb *kiocb;
 
 	assert_spin_locked(&ctx->ctx_lock);
 
@@ -1654,9 +1671,8 @@ static struct kiocb *lookup_kiocb(struct kioctx *ctx, struct iocb __user *iocb,
 		return NULL;
 
 	/* TODO: use a hash or array, this sucks. */
-	list_for_each(pos, &ctx->active_reqs) {
-		struct kiocb *kiocb = list_kiocb(pos);
-		if (kiocb->ki_obj.user == iocb)
+	list_for_each_entry(kiocb, &ctx->active_reqs, ki_list) {
+		if (kiocb->ki_user_iocb == iocb)
 			return kiocb;
 	}
 	return NULL;
@@ -1676,7 +1692,7 @@ SYSCALL_DEFINE3(io_cancel, aio_context_t, ctx_id, struct iocb __user *, iocb,
 		struct io_event __user *, result)
 {
 	struct kioctx *ctx;
-	struct kiocb *kiocb;
+	struct aio_kiocb *kiocb;
 	u32 key;
 	int ret;