[Bug 220027] USB 3.2 capture cards using incorrect USB speed

[bugzilla-daemon@kernel.org]

https://bugzilla.kernel.org/show_bug.cgi?id=220027

--- Comment #5 from Michał Pecio (michal.pecio@gmail.com) ---
Looks like a problem with those capture devices, then. Maybe they all use the
same chip? One name I have seen somewhere is Realtek RTD2801VPF.

It could be what Mathias Nyman suggested in the other bug: the device expects
some particular order of descriptor requests (as issued by Windows) and doesn't
know what to do when Linux does similar things in other order. It would be a
little weird that such a bug only occurs at 10G speed, but hardware can be all
sorts of buggy...

To test this theory it would be necessary to patch the kernel and observe what
happens. Can you build the kernel from source, are you familiar with C
programming by any chance?

Another possible approach is to sniff the USB bus with Wireshark on Windows and
on Linux and compare. Disclaimer: I have never really used Wireshark on Windows
before, so IDK how usable it is, and any changes to Linux behavior would still
require kernel patches of course.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.