From mboxrd@z Thu Jan  1 00:00:00 1970
Received: from duke.felloff.net ([216.126.196.34]) by ewsd; Thu Jun 11 13:10:41 EDT 2020
Message-ID: <46EFE19EBB1B74CCB387820F084840C6@felloff.net>
Date: Thu, 11 Jun 2020 19:10:22 +0200
From: cinap_lenrek@felloff.net
To: 9front@9front.org
Subject: Re: [9front] new nusb/serial driver
In-Reply-To: <CAHwi9bz-SAHQUbG8ftSNeBcZu19eXkQM_8YOfUCOWPq_CKCeiQ@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset="US-ASCII"
Content-Transfer-Encoding: 7bit
List-ID: <9front.9front.org>
List-Help: <http://lists.9front.org>
X-Glyph: ➈
X-Bullshit: responsive shared SOAP over SQL browser-scale realtime-java generator 

> I only did that because the other drivers did it.

yeah, it could be that they'r also wrong. or the devices they support
have broken descriptors... or they did the same as you and just copied
it from another driver. :)

> It seems to work just as well without that

good. then i'd suggest we leave it out. given that cdc is a generic
standard, we should use the value the device advertises as we'r not
targeting specific chips where we know which packet sizes will work
or not.

another issue:

> 	if(epin == -1 || epout == -1 || (ser->hasepintr && epintr == -1)){
> 		ifc++;
> 		if(ifc > 1)
> 			return -1;
> 		goto TOPFINDEPS;
> 	}

i think this should not be done here. rather, we should do that in
the caller of findendendpoints(), and have it continue if it cant
open a interface. and only fail in the end when all interfaces
failed to be usable.

the reason is the caller already iterates over the interfaces,
and the task of this function is to bring the interface up that
was passed to it.

otherwise, everything looks good. you are almost there! :-)

--
cinap