Re: may I commit this small umodem patch ?

On Thu, 3 Jul 2008 16:07:19 +0200
Luigi Rizzo <rizzo_at_icir.org> wrote:

> There was a discussion back in september about adding
> support for basic CDC tty devices in umodem.c.
> This lets you talk to a number of usb devices built around
> microcontrollers (e.g. Atmel), and puts us on par with
> Linux and Windows in terms of supporting these devices.
> 
> Because this simply requires the small patch below to the
> probe/attach routine, so if there are no objections I plan to add
> this to the system (CURRENT then RELENG_7 and RELENG_6) in the next
> few days.
> 
> > Index: umodem.c
> > ===================================================================
> > RCS file: /home/ncvs/src/sys/dev/usb/umodem.c,v
> > retrieving revision 1.57
> > diff -u -r1.57 umodem.c
> > --- umodem.c	31 Jan 2005 13:58:10 -0000	1.57
> > +++ umodem.c	20 Aug 2006 17:05:34 -0000
> > _at__at_ -256,6 +260,15 _at__at_
> >  	    id->bInterfaceProtocol == UIPROTO_CDC_AT)
> >  		ret = UMATCH_IFACECLASS_IFACESUBCLASS_IFACEPROTO;
> >  
> > +#if 1
> > +	if (ret == UMATCH_NONE &&
> > +	    id->bInterfaceClass == UICLASS_CDC_DATA &&
> > +	    id->bInterfaceSubClass == UISUBCLASS_DATA &&
> > +	    id->bInterfaceProtocol == 0x00)
> > +		ret = UMATCH_IFACECLASS_IFACESUBCLASS_IFACEPROTO;
> > +		return ret;
> > +#endif
> > +
> >  	if (ret == UMATCH_NONE)
> >  		return (ret);
> 

Is there any reason to keep the #if 1 ... #endif?  And why not just
directly return UMATCH_IFACECLASS_IFACESUBCLASS_IFACEPROTO rather than
assigning it to ret first?

---
Gary Jennejohn (gj_at_freebsd.org)