tools: Rename IPC functions to reduce misleading

[liuqun]

Hi dear developers of WireGuard!

Recently, I've been reading the source code of WireGuard. The function name like ipc_get_device() looks to be a little bit misleading. To make the code more readable for new developers, here are some suggestions:

• Rename ipc_get_device(&dev) to ipc_fetch_conf(&conf) because it only fetches the runtime configuations rather than a device instance or device handler.
• Funtion prototype ipc_set_device(struct wgdevice *dev) is changed to ipc_set_device(const struct wgdevice *newconf).

Change the following IPC function names and parameters which used to be confusing:

- ipc_set_device(dev)                  => ipc_set_device(newconf)
- ipc_get_device(dev, interface)       => ipc_fetch_conf(conf, interface)
- kernel_set_device(dev)               => kernel_set_device(newconf)
- kernel_get_device(device, interface) => kernel_fetch_conf(conf, interface)
- userspace_set_device(dev)            => userspace_set_device(newconf)
- userspace_get_device(out, interface) => userspace_fetch_conf(out, interface)

Renamed local variable/parameter symbols:

• In showconf_main(), show_main():
  device => conf
• In ipc_fetch_conf(), kernel_fetch_conf(), userspace_fetch_conf():
  dev => conf
• In ipc_set_device(), kernel_set_device(), userspace_set_device():
  dev => newconf

Note:
A lot of braces are added to the if-else code blocks to keep ourselves from making unexpected mistakes in the future.

Changed function prototypes:

• int ipc_set_device(struct wgdevice *) => int ipc_set_device(const struct wgdevice *newconf)
• int kernel_set_device(struct wgdevice *) => int kernel_set_device(const struct wgdevice *newconf)
• int userspace_set_device(struct wgdevice *) => int userspace_set_device(const struct wgdevice *newconf)

[liuqun]

• Contributing Guide: https://www.wireguard.com/#contributing
• Mail list: [email protected]

[thinrope]

While I like the ideas in this, they are a bit intrusive and change the coding style in a way. And do it in one single big patch.
I think it will be a bit more easy to digest if the "adding extra curly braces" part was a separate diff.
Also, while at it, using a standard C-beautifyer (astyle, uncrustify, etc.) for the whole project and selecting a coding style might be a better step.

[liuqun]

Thanks @thinrope !

Now I've removed those curly braces and created a new PR #6

[liuqun]

I've found the on-line document Linux kernel coding style here from Intel's opensource web site 01.org:
https://01.org/linuxgraphics/gfx-docs/drm/process/coding-style.html#placing-braces-and-spaces

In the document:

Do not unnecessarily use braces where a single statement will do.

	if (condition)
		action();

and

	if (condition)
		do_this();
	else
		do_that();

This does not apply if only one branch of a conditional statement is a single
statement; in the latter case use braces in both branches:

	if (condition) {
		do_this();
		do_that();
	} else {
		otherwise();
	}

If we look into WireGuard/src/tools, we might found many if-else blocks didn‘t follow the last rule above, for example:
https://github.com/WireGuard/WireGuard/blob/033dc8974de052c1f1ab5d735ec2ee921127e352/src/tools/show.c#L447-L452