From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tony Lindgren Subject: Re: [PATCH v4 4/4] ARM: OMAP2+: Add pdata quirk for sys_clkout2 for omap3 DBB056 Date: Wed, 5 Mar 2014 09:21:56 -0800 Message-ID: <20140305172156.GF4983@atomide.com> References: <1392387656-15186-1-git-send-email-chf.fritz@googlemail.com> <1392387656-15186-5-git-send-email-chf.fritz@googlemail.com> <20140228223741.GM13624@atomide.com> <53172175.5060603@ti.com> <1394025932.16566.44.camel@mars> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <1394025932.16566.44.camel@mars> Sender: linux-omap-owner@vger.kernel.org To: Christoph Fritz Cc: Tero Kristo , bcousson@baylibre.com, Tomi Valkeinen , Nishanth Menon , Javier Martinez Canillas , Daniel Mack , devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-omap@vger.kernel.org List-Id: devicetree@vger.kernel.org * Christoph Fritz [140305 05:28]: > On Wed, 2014-03-05 at 15:07 +0200, Tero Kristo wrote: > > On 03/01/2014 12:37 AM, Tony Lindgren wrote: > > > * Christoph Fritz [140214 06:24]: > > >> Full device tree support for clock control, especially to set frequencies, > > >> is not yet accomplished. Until then, configure the 24Mhz of sys_clkout2 to > > >> feed an USB-Hub here. > > > > > > Hmm would like to see Tero's comments on this, I wonder if we should > > > wait on this to avoid extra churn? > > > > Well, the set I posted that adds default clock parenting / default rate > > support to DT hasn't actually moved forward, so you might want to take > > this in as is now if it should be rushed. The clk-enable part isn't done > > currently by anything anyway. Just wondering though, should you create > > some sort of driver for your usb-hub which would control these clocks? > > The USB-Hub, a USB2512 is wired as non-configurable. Its pin > 'XTAL1/CLKIN' can be connected to either a crystal or an external clock > input. To economise the crystal, the clock-out of the omap is wired > here. > > So in my view, a driver is unnecessary. Yeah that makes sense. Let's try not to rely on pdata-quirks.c for the new boards. Regards, Tony