| Age | Commit message (Collapse) | Author |
|---|
|
|
|
Make CheckDeviceAccess() more flat by inverting checks, reducing
indentation and removing dead code.
|
|
|
|
|
|
|
|
|
|
This is only ever queried and not set outside of the Core.cpp, so this
should just be hidden internally and just have a function exposed that
allows querying it.
|
|
This reverts commit c8a6dc6c23a5efe99b4bf937003ba3f2f7f8f200.
libusb on Windows isn't really safe to use from different threads
with a single context.
|
|
Places all of the SI code under the SerialInterface namespace instead of
only the main source file. This keeps all SI code under a common name,
as well as out of the global namespace
|
|
Given this constant is related to the serial interface, it shouldn't be
outside of the namespace.
|
|
This prevents libusb warnings from showing up even when the user is
not using Bluetooth or USB passthrough, or the Wii U GC adapter.
|
|
IOS: USB support (OH0, USB_VEN, USB_HID)
|
|
|
|
libusb on Windows is limited to only a single context. Trying to open
more than one can cause device enumerations to fail randomly.
libusb is thread-safe and we don't use the manual polling support (with
`poll()`) so this should be safe.
|
|
|
|
|
|
Most of this commits changes performance decreasing logs from info to debug and also cleans up innacurate levels.
|
|
|
|
There's no reason this only should apply to netplay.
|
|
|
|
GCAdapter: improve thread safety
|
|
Reset() and Setup() are not used outside of this namespace
|
|
make sure Reset() can’t be run concurrently with AddGCAdapter() or
ResetRumble() (which is called on other threads) which can cause
crashes (issue #9462)
|
|
In order to avoid desyncs
|
|
where the payload size would end up being crazy values on the other
thread. Mirroring the similar commit made for the android side:
https://github.com/Sonicadvance1/dolphin/commit/2833685fbfd782b78bbf1813f30cbb0a4ecd801d
|
|
Previously we would iterate through every GC adapter plugged in to the PC and steal ownership of it.
This causes issues all over the place in the implementation if this happens.
Break on the first adapter we can get access to.
|
|
This cleans up a bunch of #ifdef checks in places.
|
|
|
|
The Wii U Gamecube controller adapter setup has always been a bit weird. It tries to be as automatic as possible to make the user experience as easy
as possible.
The problem with this approach is that it brings a large disconnect in the user experience because you have the Gamecube controller setup with regular
gamepads and then for some reason below that you have a "direct connect" option which will cause the Gamecube Adapter to overwrite the regular inputs
if something was connected.
While this works and allows the user to only click one checkbox to get the device working, it breaks the user's experience because they don't really
know what "direct connect" means and won't look it up to figure out what it is. Just expecting the device to work (At least one occurence of this in
the IRC channel in the last week).
This way around also had the terrible nature of making the code more filthy than it needed to be. The GCAdapter namespace was parasitic and hooked in
to the regular GC Controller SI class to overwrite the data that it was getting from the default configuration.
Now instead we have a specific SIDevice class for the Wii U Gamecube adapter. This class is fairly simple and is a child of the regular SI Gamecube
Pad device and only reimplements what it needs to.
This also gives the ability to configure controllers individually, which allows the user to configure rumble individually per pad input.
Overall the code is cleaner, and it fits more in line with how the rest of Dolphin works.
|
|
|
