== DETAILS
So, it turns out that there *is* a autoconfig disconnect handler. Took digging
through tasks/task_autodetect.c to find it!
So, I added a call to the handler when the pad gets disconnected.
This seems to solve the problem of the pad not disappearing from the menu.
(At the very least, the user's pad index reverts to "none" which is still
an improvement)
== TESTING
Tested manually, made sure it didn't crash or leak slots.
== DETAILS
The Wii U GC adapter doesn't seem to like doing async reads if it is connected
via a USB hub. It seems to be device-specific, though, because my DS3 works
just fine through the same hub.
I tried creating a fallback to synchronous reads, but it resulted in a hard
lock of the system. So, for the time being, it's going to be a known
limitation. Might be solved by using a powered USB hub.
Learned that the cache alignment is 64, not 32, so the alignment math has
been updated. Thanks, @aliaspider for that info.
This should fix the issue where R/L buttons didn't register when doing
input detection.
This also brings the GC pad in line with the rest of the gamepads in
input_autodetect_builtin.c.
Also fixed a really stupid bug that was part of why analog inputs aren't
being read. Analog still isn't working, mind, but it's a lot closer to
working now that it's actually getting down into the pad driver level!
=== DETAILS
So, the GCA has 2 USB connections; one is the data connection, and the
second is used to drive rumble.
Due to a driver bug, if the second cable wasn't attached, the pads wouldn't
get detected.
I fixed that bug.
== DETAILS
Hooray for conditional compile directives.
Moving things around broke things in unexpected ways on non-WiiU builds.
Well, not *completely* unexpected. But still.
Changes:
- Move some typedefs around to avoid circular include dependencies
- Include the file where the HID driver definition got moved to
== TESTING
- verified build for Wii U still runs successfully
- did a local build without any errors (some weird warnings, but since they
happen in code I didn't change, I'm assuming they're pre-existing?)
== DETAILS
I think this will fix the problem with duplicate pads--pads weren't properly
de-initializing and registering as disconnected. When a pad is disconnected,
the slot should properly release now.
retro_bits_t turned into input_bits_t and there were parts of my
code that needed to update.
== TESTING
No idea if upstream changes broke anything, but it compiles cleanly
now.
== DETAILS
TIL that it's bad to call synchronization code from callbacks.
To avoid that, I made the following changes:
- Implemented an atomic swap (see previous commit) to avoid explicit
locking when working with the event list
- ensure locks are only acquired in either the main thread or the
I/O polling thread
- use an explicit polling loop; we still use async reads, but the
read doesn't immediately re-invoke itself.
- remove the sleep in the polling thread.
- remove unnecessary locking in the thread cleanup call--verified that
the list can't be modified while it is being executed.
== TESTING
I tested locally, and was able to disconnect/reconnect USB devices several times without the worker thread getting deadlocked.
Fortunately, the gcc port implements the builtins and, from basic
testing, they seem to work.
This is only really useful on Wii U--other platforms have more
robust atomic operations, or aren't using gcc to build.
new icons: freeintv
redrawn icons: 81, nekop2, np2kai
tweaked icons: fceumm, mednafen_pce_fast, nestopia, o2em, picodrive
new banners: freeintv
tweaked banners: 81
all images have also been optimised
