linux-kernel - Re: bad endpoint address, MOTU FastLane

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Tue, 12 May 2009 10:08:01 -0400 (EDT)
From:	Alan Stern <stern@...land.harvard.edu>
To:	Clemens Ladisch <clemens@...isch.de>
cc:	David Fries <david@...es.net>,
	David Griffith <dgriffi@...csubak.edu>,
	Andrew Morton <akpm@...ux-foundation.org>,
	<linux-kernel@...r.kernel.org>, <linux-usb@...r.kernel.org>,
	<todd@...es.net>
Subject: Re: bad endpoint address, MOTU FastLane

On Tue, 12 May 2009, Clemens Ladisch wrote:

> David Fries wrote:
> > usb_submit_urb returns -22 EINVAL invalid argument as it is trying to
> > submit an interrupt packet to what is registered as an isoc endpoint.

How does one know which type the endpoint really is?  Through prior 
experience with the device?

> > The second interface has the same end point addresses as the first.
> > Any suggestions on how to deal with this class of broken hardware?
> > 
> > I:* If#= 0 Alt= 0 ...
> > E:  Ad=81(I) Atr=03(Int.) MxPS=   4 Ivl=1ms
> > I:* If#= 1 Alt= 0 ...
> > E:  Ad=81(I) Atr=01(Isoc) MxPS=   4 Ivl=1ms
> 
> AFAICS the USB core memorizes the descriptor belonging to a specific
> endpoint number whenever usb_set_interface() is called.  In this case,
> the second interface is initialized later, so the second set of
> descriptors wins.
> 
> The driver could call usb_set_interface() again for the first interface
> so that the USB core takes notice of the first set of endpoints again.
> Please try the patch below.
> 
> I guess I'll have to write another quirk for this.
> 
> 
> --- linux/sound/usb/usbmidi.c
> +++ linux/sound/usb/usbmidi.c
> @@ -1779,6 +1779,7 @@
>  		err = snd_usbmidi_detect_per_port_endpoints(umidi, endpoints);
>  		break;
>  	case QUIRK_MIDI_RAW:
> +		usb_set_interface(umidi->chip->dev, 0, 0);
>  		umidi->usb_protocol_ops = &snd_usbmidi_raw_ops;
>  		err = snd_usbmidi_detect_per_port_endpoints(umidi, endpoints);
>  		break;

This is only a bandaid.  The underlying problem is still there.  
Basically, the second interface shouldn't exist at all.  The only
proper way to deal with it is to pretend there's only one interface --
which would mean changing the descriptors before the kernel parses
them.

A much easier approach is to throw the device away and replace it with 
something that _is_ compliant with the USB specification.

Alan Stern

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/