Hello everyone.

This post is intended as a testimony to my latest adventure in the USB wolrd of STM32F1xx device, and I suppose, any usb device implementig an HID class.
I hope it will serve someone.

See, I love using VMWare : it's the perfect sandbox, and when 64-bits versions of windows refuses to connect to some Mathusalem-dated usb device, VMware makes it easy to have it plugged to a virtual 32-bits XP, with its old-timey drivers.
That was then... a new version of VMWare server-side usb driver appeared...sadly.

I spent TWO DAYS wondering why my HID-class Cortext-M3 project was "accepted" by Win7 x64 (this devices works properly, no error code, not even code 10), but its name, VID & PID values, shown in MS device manager, where completely wrong. As well as anything in its descriptor, on the host's side had nothing in common with the device's descriptor. But :
-The device was connecting properly...
-The host's usb port was perfectly identified, main usb controller, root hub 0, port 7.
-The device's USB clock was perfect to the slightest tick... no garbage.
-Windows 7 was derecting it as if it where another device... With different HID & VID. All of its device description was erroneous ! IMPOSSIBLE !!!

Then it struck me : amongst all those drivers, which one can hijack some unknow device at hotplug time, without even requiring knowledge of its descriptor ???? Well VMWare's server-side usb driver, can !
The worst part of all : windows device manager was swearing the device's driver was...rumble...rumble...Microsoft's...not a single lead on VMWare's. *applause*

So, the only solution here : select the device, and uninstall it. Unplug, replug.

Tadaaa...two days lost try to debug a perfectly sane HID implementation. Thank you Microsoft, thank you VMWare.