FWIW, the CUPS libusb-based backend only sets the alt setting if there is more than 1 alt setting in the descriptor.
> On Jan 18, 2021, at 11:31 AM, Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote:
>
> On Sun, Jan 17, 2021 at 11:44:16PM -0600, Pete Zaitcev wrote:
>> On Sun, 17 Jan 2021 15:36:39 -0600
>> Jeremy Figgins <kernel@xxxxxxxxxxxxxxxxx> wrote:
>>
>>> The naming is designed to mirror the existing
>>> USB_QUIRK_NO_SET_INTF flag, but that flag is
>>> not sufficient to make these devices work.
>>> + { 0x0416, 0x5011, USBLP_QUIRK_NO_SET_INTF }, /* Winbond Electronics Corp. Virtual Com Port */
>>
>> Jeremy, thanks for the patch. It looks mostly fine code-wise (quirk is
>> out of numerical order), but I have a question: did you consider keying
>> off usblp->dev->quirks instead?
>>
>> How about this:
>>
>> diff --git a/drivers/usb/class/usblp.c b/drivers/usb/class/usblp.c
>> index 37062130a03c..0c4a98f00797 100644
>> --- a/drivers/usb/class/usblp.c
>> +++ b/drivers/usb/class/usblp.c
>> @@ -1315,7 +1315,11 @@ static int usblp_set_protocol(struct usblp *usblp, int protocol)
>> alts = usblp->protocol[protocol].alt_setting;
>> if (alts < 0)
>> return -EINVAL;
>> - r = usb_set_interface(usblp->dev, usblp->ifnum, alts);
>> + if (usblp->dev->quirks & USB_QUIRK_NO_SET_INTF) {
>> + r = 0;
>> + } else {
>> + r = usb_set_interface(usblp->dev, usblp->ifnum, alts);
>> + }
>> if (r < 0) {
>> printk(KERN_ERR "usblp: can't set desired altsetting %d on interface %d\n",
>> alts, usblp->ifnum);
>
> Would it be practical simply to skip the usb_set_interface() call
> whenever alts is 0? After all, devices use altsetting 0 by default; it
> shouldn't be necessary to tell them to do so.
>
> Alan Stern
________________________
Michael Sweet
