channels: fix memory ordering violation in iterate (#52407)

Channel `iterate` calls might miss trailing items without this patch. I
have not seen proof of this reaching a failure, but we do appear to be
missing this ordering specification in visual review.

Should not make a difference to generated code on x86, which already has
TSO guaranteed, but may alter optimizations and other CPUs with weaker
memory orderings.

(cherry picked from commit 856e112)

Author: vtjnash

base/channels.jl

@@ -211,7 +211,12 @@ function close(c::Channel, @nospecialize(excp::Exception))
     end
     nothing
 end
-isopen(c::Channel) = ((@atomic :monotonic c.state) === :open)
+
+# Use acquire here to pair with release store in `close`, so that subsequent `isready` calls
+# are forced to see `isready == true` if they see `isopen == false`. This means users must
+# call `isopen` before `isready` if you are using the race-y APIs (or call `iterate`, which
+# does this right for you).
+isopen(c::Channel) = ((@atomic :acquire c.state) === :open)
 
 """
     bind(chnl::Channel, task::Task)