io_uring: restore back registered wait arguments

JIRA: https://issues.redhat.com/browse/RHEL-106845

commit d617b31
Author: Pavel Begunkov <asml.silence@gmail.com>
Date:   Fri Nov 15 16:54:43 2024 +0000

    io_uring: restore back registered wait arguments

    Now we've got a more generic region registration API, place
    IORING_ENTER_EXT_ARG_REG and re-enable it.

    First, the user has to register a region with the
    IORING_MEM_REGION_REG_WAIT_ARG flag set. It can only be done for a
    ring in a disabled state, aka IORING_SETUP_R_DISABLED, to avoid races
    with already running waiters. With that we should have stable constant
    values for ctx->cq_wait_{size,arg} in io_get_ext_arg_reg() and hence no
    READ_ONCE required.

    The other API difference is that we're now passing byte offsets instead
    of indexes. The user _must_ align all offsets / pointers to the native
    word size, failing to do so might but not necessarily has to lead to a
    failure usually returned as -EFAULT. liburing will be hiding this
    details from users.

    Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
    Link: https://lore.kernel.org/r/81822c1b4ffbe8ad391b4f9ad1564def0d26d990.1731689588.git.asml.silence@gmail.com
    Signed-off-by: Jens Axboe <axboe@kernel.dk>

Signed-off-by: Ming Lei <ming.lei@redhat.com>

Author: Ming Lei

include/linux/io_uring_types.h

@@ -324,6 +324,9 @@ struct io_ring_ctx {
 		unsigned		cq_entries;
 		struct io_ev_fd	__rcu	*io_ev_fd;
 		unsigned		cq_extra;
+
+		void			*cq_wait_arg;
+		size_t			cq_wait_size;
 	} ____cacheline_aligned_in_smp;
 
 	/*

include/uapi/linux/io_uring.h

@@ -679,6 +679,11 @@ struct io_uring_region_desc {
 	__u64 __resv[4];
 };
 
+enum {
+	/* expose the region as registered wait arguments */
+	IORING_MEM_REGION_REG_WAIT_ARG		= 1,
+};
+
 struct io_uring_mem_region_reg {
 	__u64 region_uptr; /* struct io_uring_region_desc * */
 	__u64 flags;

io_uring/io_uring.c

@@ -3206,7 +3206,19 @@ void __io_uring_cancel(bool cancel_all)
 static struct io_uring_reg_wait *io_get_ext_arg_reg(struct io_ring_ctx *ctx,
 			const struct io_uring_getevents_arg __user *uarg)
 {
-	return ERR_PTR(-EFAULT);
+	unsigned long size = sizeof(struct io_uring_reg_wait);
+	unsigned long offset = (uintptr_t)uarg;
+	unsigned long end;
+
+	if (unlikely(offset % sizeof(long)))
+		return ERR_PTR(-EFAULT);
+
+	/* also protects from NULL ->cq_wait_arg as the size would be 0 */
+	if (unlikely(check_add_overflow(offset, size, &end) ||
+		     end > ctx->cq_wait_size))
+		return ERR_PTR(-EFAULT);
+
+	return ctx->cq_wait_arg + offset;
 }
 
 static int io_validate_ext_arg(struct io_ring_ctx *ctx, unsigned flags,

io_uring/register.c

@@ -588,7 +588,16 @@ static int io_register_mem_region(struct io_ring_ctx *ctx, void __user *uarg)
 
 	if (memchr_inv(&reg.__resv, 0, sizeof(reg.__resv)))
 		return -EINVAL;
-	if (reg.flags)
+	if (reg.flags & ~IORING_MEM_REGION_REG_WAIT_ARG)
+		return -EINVAL;
+
+	/*
+	 * This ensures there are no waiters. Waiters are unlocked and it's
+	 * hard to synchronise with them, especially if we need to initialise
+	 * the region.
+	 */
+	if ((reg.flags & IORING_MEM_REGION_REG_WAIT_ARG) &&
+	    !(ctx->flags & IORING_SETUP_R_DISABLED))
 		return -EINVAL;
 
 	ret = io_create_region(ctx, &ctx->param_region, &rd);
@@ -598,6 +607,11 @@ static int io_register_mem_region(struct io_ring_ctx *ctx, void __user *uarg)
 		io_free_region(ctx, &ctx->param_region);
 		return -EFAULT;
 	}
+
+	if (reg.flags & IORING_MEM_REGION_REG_WAIT_ARG) {
+		ctx->cq_wait_arg = io_region_get_ptr(&ctx->param_region);
+		ctx->cq_wait_size = rd.size;
+	}
 	return 0;
 }