Re: [PATCH] USB: serial: simple: narrow OWON match to avoid breaking VDS1022

[Huang Wei]

Hi leStoppe,

Thanks a lot for the dump - this settles it. Johan's hypothesis was
exactly right: the florentbr tool talks to the VDS1022 over libusb,
and the kernel driver binding is what blocks it from claiming the
interface. And the point that the VDS1022 is a headless scope where
the PC tool is the primary interface makes it clear it should stay
unclaimed.

The descriptor data also settles the match question, in a slightly
unexpected way. The two families share interface class 5, but:

  HDS272S (dump in the commit message of 4cc01410e1c1):
    class 5, subclass 0, protocol 0
  VDS1022 (your dump):
    class 5, subclass 6, protocol 0x50

So a class-only match - the USB_DEVICE_INTERFACE_CLASS() form that
came up in the August review - would still claim the VDS1022. The
full interface tuple does discriminate, which is what my original v1
used. I'll follow up with a v2 that keeps the full-tuple match and
reworks the commit message around the libusb breakage, with both
descriptor dumps as the justification.

Craig - if you happen to know whether subclass 0 / protocol 0 holds
across the HDS200 series and firmware versions, that would be a
useful confirmation; the dump in your commit message is the only
HDS200 data point I have.

Thanks,
Huang Wei