From mboxrd@z Thu Jan 1 00:00:00 1970 From: Rob Herring Subject: Re: [RFC PATCH 3/5] of: add clock providers Date: Tue, 01 Feb 2011 16:42:34 -0600 Message-ID: <4D488C5A.7020107@gmail.com> References: <20110126043338.16410.75019.stgit@localhost6.localdomain6> <20110126044416.16410.68942.stgit@localhost6.localdomain6> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20110126044416.16410.68942.stgit@localhost6.localdomain6> Sender: linux-kernel-owner@vger.kernel.org To: Grant Likely Cc: devicetree-discuss@lists.ozlabs.org, jeremy.kerr@canonical.com, linux-kernel@vger.kernel.org, linux-arm-kernel@lists.infradead.org, nicolas.pitre@linaro.org List-Id: devicetree@vger.kernel.org Grant, On 01/25/2011 10:44 PM, Grant Likely wrote: > +struct clk *of_clk_get(struct device *dev, const char *id) It would help if this took a struct device_node rather than struct device. This would allow using of_clk_get directly for cases where you don't have a struct device. This is fairly common in core platform core which are not full drivers. The case I have run into is the timer init code. To make clk_get work, I would have to create a dummy struct device. Alternatively, an of_clk_get_sys function is needed. > +{ > + struct device_node *provnode; > + u32 provhandle; > + int sz; > + struct clk *clk; > + char prop_name[32]; /* 32 is max size of property name */ > + const void *prop; > + > + dev_dbg(dev, "Looking up %s-clock from device tree\n", id); > + > + snprintf(prop_name, 32, "%s-clock", id ? id : "bus"); Many times a module just has 1 clock and id will be NULL. Is "bus-clock" really the best choice of name? A bus clock may exist, but be transparent to s/w. How about allowing "clock" or "%s-clock"? If bus-clock is the default, then you should add this to the clock binding wiki page. Rob