io_uring: add support for fixed wait regions

Generally applications have 1 or a few waits of waiting, yet they pass
in a struct io_uring_getevents_arg every time. This needs to get copied
and, in turn, the timeout value needs to get copied.

Rather than do this for every invocation, allow the application to
register a fixed set of wait regions that can simply be indexed when
asking the kernel to wait on events.

At ring setup time, the application can register a number of these wait
regions and initialize region/index 0 upfront:

	struct io_uring_reg_wait *reg;

	reg = io_uring_setup_reg_wait(ring, nr_regions, &ret);

	/* set timeout and mark as set, sigmask/sigmask_sz as needed */
	reg->ts.tv_sec = 0;
	reg->ts.tv_nsec = 100000;
	reg->flags = IORING_REG_WAIT_TS;

where nr_regions >= 1 && nr_regions <= PAGE_SIZE / sizeof(*reg). The
above initializes index 0, but 63 other regions can be initialized,
if needed. Now, instead of doing:

	struct __kernel_timespec timeout = { .tv_nsec = 100000, };

	io_uring_submit_and_wait_timeout(ring, &cqe, nr, &t, NULL);

to wait for events for each submit_and_wait, or just wait, operation, it
can just reference the above region at offset 0 and do:

	io_uring_submit_and_wait_reg(ring, &cqe, nr, 0);

to achieve the same goal of waiting 100usec without needing to copy
both struct io_uring_getevents_arg (24b) and struct __kernel_timeout
(16b) for each invocation. Struct io_uring_reg_wait looks as follows:

struct io_uring_reg_wait {
	struct __kernel_timespec	ts;
	__u32				min_wait_usec;
	__u32				flags;
	__u64				sigmask;
	__u32				sigmask_sz;
	__u32				pad[3];
	__u64				pad2[2];
};

embedding the timeout itself in the region, rather than passing it as
a pointer as well. Note that the signal mask is still passed as a
pointer, both for compatability reasons, but also because there doesn't
seem to be a lot of high frequency waits scenarios that involve setting
and resetting the signal mask for each wait.

The application is free to modify any region before a wait call, or it
can use keep multiple regions with different settings to avoid needing to
modify the same one for wait calls. Up to a page size of regions is mapped
by default, allowing PAGE_SIZE / 64 available regions for use.

The registered region must fit within a page. On a 4kb page size system,
that allows for 64 wait regions if a full page is used, as the size of
struct io_uring_reg_wait is 64b. The region registered must be aligned
to io_uring_reg_wait in size. It's valid to register less than 64
entries.

In network performance testing with zero-copy, this reduced the time
spent waiting on the TX side from 3.12% to 0.3% and the RX side from 4.4%
to 0.3%.

Wait regions are fixed for the lifetime of the ring - once registered,
they are persistent until the ring is torn down. The regions support
minimum wait timeout as well as the regular waits.

Signed-off-by: Jens Axboe <axboe@kernel.dk>

1 files changed, 57 insertions, 11 deletions

diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c
index 4cd0ee52710d..2863b957e373 100644
--- a/io_uring/io_uring.c
+++ b/io_uring/io_uring.c
@@ -2736,6 +2736,7 @@ static __cold void io_ring_ctx_free(struct io_ring_ctx *ctx)
 	io_alloc_cache_free(&ctx->msg_cache, io_msg_cache_free);
 	io_futex_cache_free(ctx);
 	io_destroy_buffers(ctx);
+	io_unregister_cqwait_reg(ctx);
 	mutex_unlock(&ctx->uring_lock);
 	if (ctx->sq_creds)
 		put_cred(ctx->sq_creds);
@@ -3224,21 +3225,43 @@ void __io_uring_cancel(bool cancel_all)
 	io_uring_cancel_generic(cancel_all, NULL);
 }
 
-static int io_validate_ext_arg(unsigned flags, const void __user *argp, size_t argsz)
+static struct io_uring_reg_wait *io_get_ext_arg_reg(struct io_ring_ctx *ctx,
+			const struct io_uring_getevents_arg __user *uarg)
 {
-	if (flags & IORING_ENTER_EXT_ARG) {
-		struct io_uring_getevents_arg arg;
+	struct io_uring_reg_wait *arg = READ_ONCE(ctx->cq_wait_arg);
 
-		if (argsz != sizeof(arg))
+	if (arg) {
+		unsigned int index = (unsigned int) (uintptr_t) uarg;
+
+		if (index <= ctx->cq_wait_index)
+			return arg + index;
+	}
+
+	return ERR_PTR(-EFAULT);
+}
+
+static int io_validate_ext_arg(struct io_ring_ctx *ctx, unsigned flags,
+			       const void __user *argp, size_t argsz)
+{
+	struct io_uring_getevents_arg arg;
+
+	if (!(flags & IORING_ENTER_EXT_ARG))
+		return 0;
+
+	if (flags & IORING_ENTER_EXT_ARG_REG) {
+		if (argsz != sizeof(struct io_uring_reg_wait))
 			return -EINVAL;
-		if (copy_from_user(&arg, argp, sizeof(arg)))
-			return -EFAULT;
+		return PTR_ERR(io_get_ext_arg_reg(ctx, argp));
 	}
+	if (argsz != sizeof(arg))
+		return -EINVAL;
+	if (copy_from_user(&arg, argp, sizeof(arg)))
+		return -EFAULT;
 	return 0;
 }
 
-static int io_get_ext_arg(unsigned flags, const void __user *argp,
-			  struct ext_arg *ext_arg)
+static int io_get_ext_arg(struct io_ring_ctx *ctx, unsigned flags,
+			  const void __user *argp, struct ext_arg *ext_arg)
 {
 	const struct io_uring_getevents_arg __user *uarg = argp;
 	struct io_uring_getevents_arg arg;
@@ -3252,6 +3275,28 @@ static int io_get_ext_arg(unsigned flags, const void __user *argp,
 		return 0;
 	}
 
+	if (flags & IORING_ENTER_EXT_ARG_REG) {
+		struct io_uring_reg_wait *w;
+
+		if (ext_arg->argsz != sizeof(struct io_uring_reg_wait))
+			return -EINVAL;
+		w = io_get_ext_arg_reg(ctx, argp);
+		if (IS_ERR(w))
+			return PTR_ERR(w);
+
+		if (w->flags & ~IORING_REG_WAIT_TS)
+			return -EINVAL;
+		ext_arg->min_time = READ_ONCE(w->min_wait_usec) * NSEC_PER_USEC;
+		ext_arg->sig = u64_to_user_ptr(READ_ONCE(w->sigmask));
+		ext_arg->argsz = READ_ONCE(w->sigmask_sz);
+		if (w->flags & IORING_REG_WAIT_TS) {
+			ext_arg->ts.tv_sec = READ_ONCE(w->ts.tv_sec);
+			ext_arg->ts.tv_nsec = READ_ONCE(w->ts.tv_nsec);
+			ext_arg->ts_set = true;
+		}
+		return 0;
+	}
+
 	/*
 	 * EXT_ARG is set - ensure we agree on the size of it and copy in our
 	 * timespec and sigset_t pointers if good.
@@ -3297,7 +3342,8 @@ SYSCALL_DEFINE6(io_uring_enter, unsigned int, fd, u32, to_submit,
 	if (unlikely(flags & ~(IORING_ENTER_GETEVENTS | IORING_ENTER_SQ_WAKEUP |
 			       IORING_ENTER_SQ_WAIT | IORING_ENTER_EXT_ARG |
 			       IORING_ENTER_REGISTERED_RING |
-			       IORING_ENTER_ABS_TIMER)))
+			       IORING_ENTER_ABS_TIMER |
+			       IORING_ENTER_EXT_ARG_REG)))
 		return -EINVAL;
 
 	/*
@@ -3380,7 +3426,7 @@ SYSCALL_DEFINE6(io_uring_enter, unsigned int, fd, u32, to_submit,
 			 */
 			mutex_lock(&ctx->uring_lock);
 iopoll_locked:
-			ret2 = io_validate_ext_arg(flags, argp, argsz);
+			ret2 = io_validate_ext_arg(ctx, flags, argp, argsz);
 			if (likely(!ret2)) {
 				min_complete = min(min_complete,
 						   ctx->cq_entries);
@@ -3390,7 +3436,7 @@ iopoll_locked:
 		} else {
 			struct ext_arg ext_arg = { .argsz = argsz };
 
-			ret2 = io_get_ext_arg(flags, argp, &ext_arg);
+			ret2 = io_get_ext_arg(ctx, flags, argp, &ext_arg);
 			if (likely(!ret2)) {
 				min_complete = min(min_complete,
 						   ctx->cq_entries);