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

From: Alan Stern <stern@rowland.harvard.edu>
To: "Bjørn Mork" <bjorn@mork.no>
Cc: Oliver Neukum <oneukum@suse.com>,
	Marcel Holtmann <marcel@holtmann.org>,
	Jose Ignacio Tornos Martinez <jtornosm@redhat.com>,
	gregkh@linuxfoundation.org, linux-usb@vger.kernel.org
Subject: Re: [PATCH v4] USB: core: skip unconfiguration if device doesn't support it
Date: Thu, 5 May 2022 13:04:33 -0400	[thread overview]
Message-ID: <YnQDoc0JETR3KA8A@rowland.harvard.edu> (raw)
In-Reply-To: <875ymkknqy.fsf@miraculix.mork.no>

On Thu, May 05, 2022 at 05:22:45PM +0200, Bjørn Mork wrote:
> Alan Stern <stern@rowland.harvard.edu> writes:
> 
> > Please pardon me for butting in, but I don't see how this tests the 
> > condition that Jose is worried about.
> ..
> > And presumably the device is working again.  However, none of this shows 
> > what happens when the device is unconfigured.  To test that, you would 
> > have to do:
> >
> > 	echo 0 >/sys/bus/usb/devices/1-8/bConfigurationValue
> > 	echo 1 >/sys/bus/usb/devices/1-8/bConfigurationValue
> >
> > and then see if the device continues to work.
> 
> Ah, sorry. Scanned briefly, saw "bind", and assumed too much.  Making an
> ass out of... you know.
> 
> Actually I didn't understand the part about unconfiguration since I
> can't see how that would happen during normal usage?  Anyway,

Jose was exporting his device over usbip, which perhaps does not count 
as "normal" usage.

> unconfiguring also works for me:
> 
> canardo:/tmp# hciconfig hci0
> hci0:   Type: Primary  Bus: USB
>         BD Address: 00:1A:7D:DA:71:15  ACL MTU: 310:10  SCO MTU: 64:8
>         UP RUNNING 
>         RX bytes:660 acl:0 sco:0 events:43 errors:0
>         TX bytes:2178 acl:0 sco:0 commands:43 errors:0
> 
> canardo:/tmp# echo 0 >/sys/bus/usb/devices/1-8/bConfigurationValue
> canardo:/tmp# hciconfig hci0
> Can't get device info: No such device
> canardo:/tmp# echo 1 >/sys/bus/usb/devices/1-8/bConfigurationValue
> 
> 
> Not entirely sure how to validate that "everything" works at this point?

If it weren't, you would know.  Error messages would have shown up at 
this point.

> I can use the rfcomm session the adapter usually handles and also run
> lescan after this.  So I guess both BLE and BDR works.
> 
> And the counters count something:
> 
> canardo:/tmp# hciconfig hci0
> hci0:   Type: Primary  Bus: USB
>         BD Address: 00:1A:7D:DA:71:15  ACL MTU: 310:10  SCO MTU: 64:8
>         UP RUNNING 
>         RX bytes:3883 acl:40 sco:0 events:81 errors:0
>         TX bytes:2518 acl:19 sco:0 commands:50 errors:0
> 
> 
> I can see an error like this logged every time I unconfigure the device:
> 
> Bluetooth: hci0: urb 00000000e66a2492 failed to resubmit (2)

That's normal.  The driver probably could avoid giving that error 
message under these circumstances, but it's not a big deal.

> There is nothing else logged in kernel log

Okay, thanks for testing.

Alan Stern