-
Notifications
You must be signed in to change notification settings - Fork 62
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Apple Network Framework Socket Changes #662
base: grand_dispatch_queue
Are you sure you want to change the base?
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I haven't looked at nw_socket.c yet. I'm concerned about larger issues implied by the changes outside that file first.
|
||
Upon a connection being established, the new socket (either as the result of a `connect()` or `start_accept()` call) | ||
will not be attached to any event loops. It is your responsibility to register it with an event loop to begin receiving | ||
notifications. | ||
|
||
|
||
#### V-Table |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Why is this documentation necessary in the public readme? A giant list of functions feels like a distraction
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It would help with anybody who would like to extend the socket implementation. Also, we listed the vtable for other classes as well.
|
||
For LOCAL (Unix Domain Sockets or Named Pipes), the socket will be immediately ready for use upon a successful return. | ||
|
||
int aws_socket_bind(struct aws_socket *socket, struct aws_socket_endpoint *local_endpoint); | ||
|
||
Binds the socket to a local address. In UDP mode, the socket is ready for `aws_socket_read()` operations. In connection oriented | ||
modes, you still must call `aws_socket_listen()` and `aws_socket_start_accept()` before using the socket. | ||
Binds the socket to a local address. In UDP mode, the socket is ready for `aws_socket_read()` operations. In connection oriented |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This if-else conditional behavior bothers me. What happens if you call aws_socket_listen on a bound UDP socket?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
aws_socket_listen would fail with AWS_IO_SOCKET_ILLEGAL_OPERATION_FOR_STATE
on a bounded UDP socket.
return aws_event_loop_thread_is_callers_thread(channel->loop); | ||
// As aws_event_loop is not ref-counted, it is possible that the socket read/write callback get triggered after | ||
// the event loop is released, ignore the function call here. | ||
if (channel && channel->loop) { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
This is alarming. I would like to know a lot more about how this is possible: how can you have a channel still alive while its event loop is not?
@@ -200,6 +200,9 @@ static int s_tls_common_tester_clean_up(struct tls_common_tester *tester) { | |||
|
|||
aws_condition_variable_clean_up(&tester->condition_variable); | |||
aws_mutex_clean_up(&tester->mutex); | |||
// wait for socket ref count drop and released | |||
aws_thread_current_sleep(1000000000); |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Doing this (presumably to prevent memory-still-in-use test failures) is the sign of a much larger problem that could easily make life terrible for us: we don't understand, can't reason about, and/or can't tightly control an async cleanup/shutdown process. We should talk about this with Mike.
Apple Network Framework socket integration
By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.