Windows Cursor Acceleration Fix is a kernel-mode program for Windows 2000 and above
to disable Windows' cursor acceleration by patching win32k.sys of the Win32 subsystem in volatile memory.
It is meant to load with every system start, modify system memory, and then unload.

It supersedes the ambiguously-named Windows Mouse Acceleration Fix.

The 64-bit build has not been tested.

Installation
0. If running Windows Vista "x64", enable test signing: Enabling Test Signing
1. Download the attached archive and extract its content to a local directory.
2. Run install.cmd with administrative privileges.

A restart should not be necessary.

Although it is designed to provide a high level of safety, I cannot guarantee failure-free operation.
In case you are unable to access the Windows installation because of a reoccurring system crash,
you can boot into Safe Mode and disable service "wcafix".

Well the fact is games use mouse quite extensively nowadays Anyway Microsoft could make the mouse much better but they choose not to. Win2k had the best mouse, in WinXP it was worse but in Vista its just awful. Thanks for this, really.

Now all there is left to do is to get 1000hz usbport.sys hack working in Vista with (KB932596), (KB938979) and (KB941649) windows updates. Well the kernel protection patch is useless, but the other patches are important...

Pull that off and you'll be no small deity in my book   

This is expected behaviour.
The driver intentionally fails the initialization routine in order to enforce its immediate unload.

Edit: This has changed with the current version

No, test-signing wasn't difficult (makecert & signtool).
The documentation, however, is indeed poor and sometimes even erroneous. For example, you don't have to import the root certificate at all - any signature will work in Test Mode.

Tested with Windows Vista x64 Ultimate.

The patch has no effect; it does not work. Properly enabled Test Signing (evident by "Test Mode" on desktop); acceleration still present in Windows as well as in all games.

I applied that patch. Here's what it did:

After reboot and logging back on, the mouse cursor did not move at all, then after a few seconds, started working again.

Is that what it's supposed to do? And if it didn't move, that means that acceleration is still enabled?

Just to be sure: The cursor worked again without you changing any settings?

Yes, it is supposed to work that way. The moment the cursor worked, the patch had been successfully applied. The driver is loaded in a relatively late initialization phase, so you may still notice cursor acceleration (or in this case the "deadlock") for a short time. The driver itself should usually take less than a second to complete.


The previous fix patched win32k.sys on disk, this one patches it in memory.
As long as the known byte patterns keep working, you won't have to care about Windows updates installing newer versions of win32k.sys. However, this fix has to load with every system start, but it will unload once it has finished and won't waste CPU or RAM resources.

If the "deadlock" has vanished, Windows' own cursor acceleration has been disabled. If you do not agree, you will have to prove the contrary. I don't know what software you have installed, but maybe the cursor/mouse acceleration you experience has a different cause.