Peter Dons Tychsen

@sonatique we will get there, do not worry. What is really needed is sort of a "mode" switch. There are 2 scenarios: 1) The user wants to use async APIs....

> When you're writing about "async API" vs "sync API" which API are you referring to? The one from libusb0 or libusb-1.0? Both libraries support both mode. 1) If you...

OK I fixed the problem. So now, if another request is queued for the same endpoint, the IRP cannot be resubmitted as it would cause the order to go bonkers,...

@sonatique: Please test: https://github.com/mcuee/libusb-win32/releases/tag/test_release_async_order

@sonatique thanks for testing. 1) I think the high CPU comes from my use of spinlocks during read/write requests. I have come up with a better idea, which I will...

@sonatique I re-fixed this and also fixed up some things that where not safe on multi-core setups. 1) I have removed the spin-lock and replaced it with some simple interlocked...

New test release here: https://github.com/mcuee/libusb-win32/releases/tag/test_release_async_order2

@sonatique I can see that USBPcap appears in your stack trace. USBPcap is notorious for causing problems. You probably installed it along with Wireshark I guess. Could you uninstall USBPcap...

Hi @sonatique 1) There cold easily be a bug in USBPcap that causes this. I have a had all kinds of side-effects from running with USBCap attached. It has definitely...

> Regarding the high CPU usage: as you expected I guess: there is no improvement after I removed USBPcap.. 1) Yes this sounds plausable. 2) Did you try running the...