Dolphin, the GameCube and Wii emulator - Forums

Full Version: Bluetooth Passthrough Bug Introduced between 5.0-2501 and 5.0-2505
You're currently viewing a stripped down version of our content. View the full version with proper formatting.
Pages: 1 2 3 4
I've noticed an odd problem with Bluetooth Passthrough on my system.  Everything worked normally in 5.0-2501, but 5.0-2505 and beyond cause Dolphin to hang on my system when closed, eating all of a CPU core's processing time.  I have to end task on the Dolphin process to get it to close.  This problem doesn't occur with BT passthrough turned off.  I'm using a Sena UD100 BT adapter on Windows 10 x64.

Another strange thing I've noticed that may be related, in 2501, when I press the close X with a game running, the game window fades to black, but dolphin doesn't close.  It takes a second press of the X to actually shut Dolphin down, and the wiimote stays powered on.  When I reopen Dolphin and start a game, the wiimote powers off, but I can press a button and it connects back up just fine.  In 2505, when I press the close X, the Dolphin windows close and the wiimote turns off, but the process stays running in the background until I kill the task.
The change you've bisected to makes no sense to me. It's fully possible something is messed up thought.

The delay when you close the game is actually Dolphin sending a command to the Wii game to shut down, just like you would on a real Wii. If it's shutting down instantly, that could mean something else is wrong.
I've swapped back and forth between the two versions several times, and it is consistently broken in 2505 and beyond.  I'm happy to provide any additional information.  I just don't know what to submit.

Here's a log from 2501 that works:

http://pastebin.com/PvFGjGZP

And here's a log 2569 when it fails:

http://pastebin.com/6yA1Ence

If you compare the end of the two logs, you can see in the failed log that it hangs after "GC Adapter scanning thread stopped", but before "GC Adapter detached" which follows the former in the good log.  I guess I should mention that I have the Mayflash Wii U GC controller adapter in passthrough mode as well.  With the GC adapter unplugged, Dolphin doesn't hang on exit.  It does close on the first press of the close X, though.

Well, now that I connect the GC adapter back up, the first click of the X goes to a black screen, and the second click closes Dolphin without hanging!  The WiiMote does not power off, though.

So, it's kinda fixed, I guess?  I don't understand what the issue was, ultimately.  The WiiMote not powering off is not desirable, but overall it's an improvement.
Actually, it's still doing it. All new versions get to where they won't shut down properly. Sometimes replugging the GCN adapter will make it work a few times, but the problem always returns.
I'm not exactly sure what's going on here...
I even nuked my Dolphin folder and started over, keeping only my savegames.  It still crashes on exit.

So, I started fresh with 2501, and I discovered something interesting.  It wasn't shutting down properly, either, but because I had Panic Handlers turned off I didn't know it.  Since the fresh copy had it on, I was able to see a warning saying a shutdown was already in progress and asking if I wanted to force stop.  I look in the log at this point and see the following:

Code:
59:39:205 HW\EXI\EXI_DeviceIPL.cpp:340 N[OSREPORT]: BTU HCI command timeout - cmd opcode = 0xc1a


59:39:205 HW\EXI\EXI_DeviceIPL.cpp:340 N[OSREPORT]: ---- WARNING: USB FATAL ERROR! ----

Does this shed any light on the issue?
The question now is whether it has always been crashing, or if that crash is actually a regression and it used to work somewhen. Please test more builds.
It doesn't. It just means the game is waiting for the adapter to reply to a command, and that never happens. I can't see what the issue would be with adding back code that was there in the first place… Does the issue happen with 5.0-1886?

If it does, then that means the USB KBD stuff is just broken and possibly affects other parts of IOS HLE.
5.0-1886 works just fine!

So, what does that mean?
No idea. Sounds like USB KBD is affecting Bluetooth somehow, and the fact that I accidentally broke USB KBD masked the issue.

Okay… so please bisect and find out which version caused the issue to happen! You don't have to try every single build; just do a binary search. Before you try other builds though: what about 5.0-1905?
Pages: 1 2 3 4