Mouse focus lost

[7s3b]

Please fill in this template when posting!
XMBC Version: 2.21 beta52
Installed or Portable version: Portable
Windows Version: Win10 21H2 19044.3086
Mouse Information (brand/model): Logitech G305
Relevant Computer Information (CPU, RAM etc): i5 12400/16G
Did the problem occur after an upgrade of XMBC? (If so, from what version?): tried from 2.20 to newest beta, same problem
Did the problem occur after a Windows update/upgrade? (If so, from what version?): Never update my Windows
How long have you used XMBC?: 1 year
What language and keyboard layout do you use in Windows?: EN(US)

Clear description of the problem - try and include as much information as possible, including what button and mappings you are having problems with (if applicable).:

Mouse focus get stucked in some specific window randomly, sometimes focus on nothing, keyboard works fine. Tried increasing the hook timeout from 200ms to 1000ms but got no luck. When it happens, most times i I can only log off my Windows account and log in again, sometimes I can solve it by terminating XMBC and explorer.exe then run explorer via task manager. It seems as same as this thread viewtopic.php?t=4837 .

I have a log file saved at the exact time the problem occurred but don't know how to upload.

[phil]

I'm not sure what you mean by mouse focus lost, or stuck in a particular window.
Do you mean the mouse stops moving, is locked to a window (There is nothing in XMBC to do that intentionally!), stops responding to clicks or something else?

If when it happens you press CTRL+ALT+DEL to open the secure desktop, then cancel, does it fix the problem? That seemed to be the case in the other thread you mentioned. I still haven't been able to reproduce that on any of my machines (all now on Windows 11 24H2 since Windows 10 is dead - thanks MS!)

The log file might help, but only if you had debug logging when you had the problem, and only if this is something XMBC is causing (I need more detail before I can figure that out). You can upload it either by sending me a PM and attaching it, attaching it here, or simply pasting it into a code snippet.

NOTE: Usually you need 5 posts before uploads and links are allowed - this is to stop spam bots filling up the forums with rubbish. I have removed this restriction on your account already so no need to wait until 5 posts.

[7s3b]

Yes, I meant it stops responding to clicks, sometimes only a particular window responds.

If when it happens you press CTRL+ALT+DEL to open the secure desktop, then cancel, does it fix the problem?

Yes it works.

Log file attached.

[phil]

Thanks for the log, I should have also asked for some context... What window was failing to respond to clicks (and if a window did work, which one was that). More importantly, roughly what time did the problem start - so I can narrow down which bit of that log I need to focus on! I presume around 23:30 when you enabled debug logging but would be good to confirm.

This line is weird (and repeated often)...

Code: Select all

07-10-2025 22:10:07.136> Attempting to activate layer 12 which is greater than Max layers (1)
07-10-2025 22:10:07.142> Attempting to activate layer 13 which is greater than Max layers (1)

XMBC only supports up to 10 layers so that's a bit strange, I will need to look into that - but I don't think that's at all relevant here.

Did all buttons stop working, or just one? There don't seem to be any obvious mismatched down/up messages in that log. What Im trying to establish is if XMBC is seeing the messages or not. If the hook disappearance, it would not be logging messages. But then Ctrl+Alt+Del would not reinstall the hook, so if that fixed it, it suggests more that the Windows button state has become messed up (probably (but not guaranteed) by something XMBC did). These are very difficult to diagnose, especially after the fact, even when it happens to me (and it does every now and then - usually when Im least able to investigate it of course, but it's very rare for me which makes it even more difficult to reproduce under a debug environment).

What is your configuration - this seems to be chording related, but how is that setup?

[7s3b]

When the problem starts, there is at most one window that works normally, not an exact window, sometimes Centbrowser, sometimes WinRAR, and sometimes nothing. The focus could even get lost when Windows starts, so I checked all the auto start program but found nothing. Finally, I noticed that it might be XMBC because this problem occurred for the first time a year ago, I started using XMBC from that time on.

For this time, yes I enabled logging after the problem occurs at about 23:30. And I'm sure these lines are related to the problem, I once checked them right after the problem occured, there were once bigger numbers such as 20.

Code: Select all

07-10-2025 22:10:07.136> Attempting to activate layer 12 which is greater than Max layers (1)
07-10-2025 22:10:07.142> Attempting to activate layer 13 which is greater than Max layers (1)

But it's still strange because the log of today still contains these layer 12&13 lines, but no problems so far, maybe when this layer error is solved, so does the problem?

Code: Select all

*******************************************************************************************************************************************
17-10-2025 19:13:24.083> X-Mouse Button Control v2.21 Beta 52 (x64) Startup. Commandline ''
17-10-2025 19:13:24.086> Running in high integrity mode (as Administrator) (0x00003000)
17-10-2025 19:13:24.134> Loaded 8 application specific profiles (8 normal and 0 custom window).
17-10-2025 19:13:24.134> NewApplicationSettings: SPIThread is not initialised.
17-10-2025 19:13:24.135> Running on Microsoft Windows 10 Professional Edition (build 19044), 64-bit, Hook Timeout: 1000 ms
17-10-2025 19:13:24.135> Startup folder: 'C:\Program Files\X-Mouse Button Control\'
17-10-2025 19:13:24.135> Settings (and log) folder: 'C:\Users\StormBolt\AppData\Roaming\Highresolution Enterprises\XMouseButtonControl\'
17-10-2025 19:13:24.190> SetMouseHook: Starting timer thread [0x570]...
17-10-2025 19:13:24.200> SetMouseHook: Started timer thread [0x0]...
17-10-2025 19:13:24.200> SetMouseHook: Starting hook thread [0x54C]...
17-10-2025 19:13:24.201> HookThreadProc[0x3224]: Hook thread starting...
17-10-2025 19:13:24.201> HookThreadProc[0x3224]: Current mouse cursor speed=10, system default cursor speed=10
17-10-2025 19:13:24.201> HookThreadProc[0x3224]: Creating new SPI object...
17-10-2025 19:13:24.201> SPIThread constructor...
17-10-2025 19:13:24.225> Start SPIThread...
17-10-2025 19:13:24.235> SPI Thread [0x00000554] has started.
17-10-2025 19:13:24.235> HookThreadProc[0x3224]: Creating new SendInput queue...
17-10-2025 19:13:24.236> SPIThread[0x00003108]: Running thread...
17-10-2025 19:13:24.236> SendInput Thread [0x5BC] has started.
17-10-2025 19:13:24.236> HookThreadProc[0x3224]: Creating new brightness OSD...
17-10-2025 19:13:24.248> SendInputThread: Running thread ID 0x3584.
17-10-2025 19:13:24.249> CVolumeManager(SendInput Thread 13700)::Initialise: Initialising volume manager...
17-10-2025 19:13:24.286> HookThreadProc[0x3224]: Secure attention sequence (SAS) is not available (0): ??????? 
17-10-2025 19:13:24.287> UIAccess is disabled. Unable to interact with elevated processes.
17-10-2025 19:13:24.287> HookThreadProc[0x3224]: Using keyboard language 0x04090409
17-10-2025 19:13:24.287> HookThreadProc[0x3224]: Installing low level mouse hook
17-10-2025 19:13:24.288> HookThreadProc[0x3224]: Installing low level key hook (for layer modifier keys)
17-10-2025 19:13:24.288> SetMouseHook: Started hook thread [0x54C]...
17-10-2025 19:13:24.292> CVolumeManager(SendInput Thread 13700)::Initialise: Initialised volume manager.
17-10-2025 19:13:24.316> SetHotkeys installed 1 global hotkeys...
17-10-2025 19:15:40.330> Attempting to activate layer 12 which is greater than Max layers (1)
17-10-2025 19:15:40.343> Attempting to activate layer 13 which is greater than Max layers (1)

At least MB1/2/3 stop working, not sure about MB4/5. Actually, I have no impression that Ctrl+Alt+Del works every time because I would remember that if it did, but it did work yesterday.

I do use chording but the problem occurs before I started using them. Here is my config file.

7s3b's XMBC Settings.zip

[7s3b]

There is one thing, maybe I missunderstood this "What language and keyboard layout do you use in Windows?"
I'm using Chinese language and the layout is EN(US), we input chinese charaters via imput method (IM), these IMs belong to the Chinese language in Windows.

zz.jpg

So do I have 2 languages in my Windows? I don't use a single language because I have more than one IM, Windows can remember
the last IM when switching between languages, but can only switch squentially between IMs in the same language.

Others in China rarely use Windows with IM disabled, they keep the IM enabled all the time and only imput Chinese, so I'm kind of unique, and if this is the orgin of the problem, it's also the reason why others don't report.

[phil]

I don't think this will be related to the keyboard layout (its worth me asking about it because things like simulated keystrokes may well matter, but not in this case).

Its more likely that XMBC is getting windows confused by sending a button down, but not sending a button up, thus Windows thinks the button is stuck down. This of course shouldn't happen but if it does, its really tricky to figure out why (and often this sort of thing can be because of timing (threading) issues, so related to specific hardware/CPU speeds).

Thanks for the updated information - I will have a further look into it when I get a moment, but probably not this week as its a busy week at work and I wont get any time at home for XMBC.

[7s3b]

Oh…my CPU is overclocked by increasing BCLK to 130.15Mhz, so it could be the reason? but it's not easy to verify this since the problem may not occur for more than one week, using stock clock for such a long time is boring.

So I was tangled to tell you this, after all making XMBC compatible for OC is not your duty, and that take times, what if there is no solution objectively? But... in the other thread I mentioned before, the author was using a moblie CPU which can't be overclocked, that seems to indicate the possibllity and justice of solving this problem.

It's up to you, but if you decided not digging into it, please let me know, I may consider updating my hardware.

[phil]

I'm pretty sure its not due to being overclocked.