Re: [PATCH v4] USB: core: skip unconfiguration if device doesn't support it

From: Oliver Neukum <oneukum@suse.com>
To: Jose Ignacio Tornos Martinez <jtornosm@redhat.com>,
	Oliver Neukum <oneukum@suse.com>
Cc: Marcel Holtmann <marcel@holtmann.org>,
	Greg KH <gregkh@linuxfoundation.org>,
	Alan Stern <stern@rowland.harvard.edu>,
	linux-usb@vger.kernel.org
Subject: Re: [PATCH v4] USB: core: skip unconfiguration if device doesn't support it
Date: Thu, 5 May 2022 14:06:57 +0200	[thread overview]
Message-ID: <50d48f9e-d7b1-a05a-5fbf-2c0461b540ef@suse.com> (raw)
In-Reply-To: <CABk-BGt8dFKkRvFZHpJnkHyObZ2G3D102zvyEFjVYjaFVRiQdg@mail.gmail.com>

On 05.05.22 13:15, Jose Ignacio Tornos Martinez wrote:
> Hello Oliver,
>
> Of course, I am working with usbip to remotize usb devices, that is
> the reason why unbind/bind is needed and with the btusb devices that I
> have, it was not working.
>
>

Hi,

sorry for being unclear. I was not referring to positive
knowledge about the devices you are testing with.
I was having dark thoughts about the other devices
they are sharing an ID with.
But Bjorn's testing has resolved that. In that case
we can indeed not penalize the compliant devices
for the broken ones.

One question, though. Your approach of simply doing
nothing if config 0 is to be selected again is a bit
brutal. Have you considered resetting the device
and stopping the reenumeration right as a config
is supposed to be chosen?

    Regards
        Oliver