Subject: Re: [PATCH 3/5] ARM: davinci: da8xx: add cfgchip2 to resources
To: David Lechner <david@lechnology.com>, Sekhar Nori <nsekhar@ti.com>,
        Kevin Hilman <khilman@kernel.org>,
        Alan Stern <stern@rowland.harvard.edu>, Bin Liu <b-liu@ti.com>,
        Petr Kulhavy <petr@barix.com>
References: <1458081473-8223-1-git-send-email-david@lechnology.com>
 <1458081473-8223-3-git-send-email-david@lechnology.com>
 <56E8906E.7070800@cogentembedded.com> <56E8D720.7000508@lechnology.com>
 <56E8E7C7.70901@lechnology.com>
Cc: Russell King <linux@arm.linux.org.uk>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Felipe Balbi <balbi@kernel.org>, linux-arm-kernel@lists.infradead.org,
        linux-kernel@vger.kernel.org, linux-usb@vger.kernel.org
From: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Organization: Cogent Embedded
Message-ID: <56E99A16.9020709@cogentembedded.com>
Date: Wed, 16 Mar 2016 20:38:30 +0300
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:38.0) Gecko/20100101
 Thunderbird/38.6.0
MIME-Version: 1.0
In-Reply-To: <56E8E7C7.70901@lechnology.com>
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1127
Lines: 25

On 03/16/2016 07:57 AM, David Lechner wrote:

>>>     No, this register is shared b/w MUSB and OHCI. The proper thing to
>>> do is to write the PHY driver and let it control this shared register.
>>>
>> OK. I've started working on this. I am looking at using struct usb_phy,
>> however, enum usb_phy_type only has USB_PHY_TYPE_UNDEFINED,
>> USB_PHY_TYPE_USB2, and USB_PHY_TYPE_USB3. Would it be acceptable to use
>> USB_PHY_TYPE_UNDEFINED for the ohci since it is USB 1.1? Or perhaps I
>> should use the more generic struct phy for that one?
>>
> Also, I am not finding any existing data structure to pass the musb set_mode
> function to the phy in either usb_phy or usb_otg. Setting the mode
> (host/peripheral/otg) is done in the same PHY register, so it seems like it
> should be implemented in the new phy driver as well.

    Perhaps we'd have to sacrifice that functionality...

> I guess I could use a generic phy instead and use phy_set_drvdata() to share
> data between the phy driver and the musb driver. Does this sound like a
> reasonable thing to do?

    Not sure what you mean, could you elaborate?

MBR, Sergei