linux-kernel - Re: [PATCH 1/3] usb: USB Type-C Connector Class

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]

Message-ID: <20160218130916.GL1859@kuha.fi.intel.com>
Date:	Thu, 18 Feb 2016 15:09:16 +0200
From:	Heikki Krogerus <heikki.krogerus@...ux.intel.com>
To:	Oliver Neukum <oneukum@...e.com>
Cc:	Felipe Balbi <balbi@...nel.org>,
	Mathias Nyman <mathias.nyman@...ux.intel.com>,
	Greg KH <gregkh@...uxfoundation.org>,
	linux-kernel@...r.kernel.org, linux-usb@...r.kernel.org
Subject: Re: [PATCH 1/3] usb: USB Type-C Connector Class

On Thu, Feb 18, 2016 at 10:21:48AM +0100, Oliver Neukum wrote:
> On Thu, 2016-02-18 at 10:47 +0200, Heikki Krogerus wrote:
> 
> Hi,
> 
> > The modes that can actually be selected have to be supported by both
> > the connector and the partner, and this is where I'm putting the ball
> > on the userspace at the moment. I'm not offering a list of
> > "possible_alternate_modes" where I list the combination, but instead
> > expect the userspace to be figure out that on it's own.
> > 
> > Do you think we should add "possible_alternate_modes" file?
> 
> No, what do we answer to the DFP if we recieve "Discover SVIDs"?
> I don't think that we always should answer with all we physically
> can. If, for example, the hardware could do Thunderbolt, but the OS
> is not prepared to handle it, we shouldn't offer it. So this is
> a policy decision to be made in user space. Hence we need
> an API to tell it to the kernel.

OK. Makes sense.

> > P.S. That reminds me, here's my current draft for the
> > Documentation/ABI/. Could you take a look?
> 
> OK
> 
> Here are my comments:
> 
> What:		/sys/class/type-c/usbcN/connected
> 
> 		Connection status of the USB Type-C connector usbcN. "yes" when
> 		connected, otherwise "no".
> 
> Unnecessarily wordy. 0 and 1 would do

That works for me.

> What:		/sys/class/type-c/usbcN/current_data_role
> 
> Again, 0 and 1 would do

I disagree with this one. What would 0 mean and what would 1? It would
require us to make an agreement about the "index" of the role, which
creates a small risk of somebody getting it wrong, but for what
purpose?

Why couldn't it be human readable "host" or "device" so there is never
no confusion about it.

> What:		/sys/class/type-c/usbcN/partner_alternate_modes
> 
> You should say in which number base the values are given.
> 
> What:		/sys/class/type-c/usbcN/partner_type
> 
> That could be combined with "connected"

Hmm, so in practice getting rid of "connected" completely.. I guess
it's OK.

> What:		/sys/class/type-c/usbcN/supported_data_roles
> 
> A connector can be both. How is that expressed?

"host, device".

> What:		/sys/class/type-c/usbcN/supported_power_roles
> 
> Again, what if it can do both?

"source, sink".

So these last two are now listing the values that can be entered to
the current_data_role and current_power_role.


Thanks,

-- 
heikki