From mboxrd@z Thu Jan 1 00:00:00 1970 From: Nishanth Menon Subject: Re: [PATCH 3/6] ARM: OMAP: voltage: move voltdm_reset to platform_data header Date: Tue, 6 Nov 2012 19:28:38 -0600 Message-ID: <20121107012838.GA13630@kahuna> References: <1352127734-30103-1-git-send-email-nm@ti.com> <1352127734-30103-5-git-send-email-nm@ti.com> <20121106184935.GL6801@atomide.com> <20121106214819.GA16246@kahuna> <20121107011808.GZ6801@atomide.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Return-path: Received: from devils.ext.ti.com ([198.47.26.153]:57996 "EHLO devils.ext.ti.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753314Ab2KGB2m (ORCPT ); Tue, 6 Nov 2012 20:28:42 -0500 Content-Disposition: inline In-Reply-To: <20121107011808.GZ6801@atomide.com> Sender: linux-pm-owner@vger.kernel.org List-Id: linux-pm@vger.kernel.org To: Tony Lindgren Cc: linux-omap , Rafael , Kevin Hilman , Anton Vorontsov , linux-pm On 17:18-20121106, Tony Lindgren wrote: > * Nishanth Menon [121106 13:50]: > > On 10:49-20121106, Tony Lindgren wrote: > > > > > > Looks like there are other things there too that's not platform data: > > > > > > struct voltagedomain *voltdm_lookup(const char *name); > > > int voltdm_scale(struct voltagedomain *voltdm, unsigned long target_volt); > > > unsigned long voltdm_get_voltage(struct voltagedomain *voltdm); > > > struct omap_volt_data *omap_voltage_get_voltdata(struct voltagedomain *voltdm, > > > unsigned long volt); > > > > > > Can you please add a patch fixing that ASAP? > > > > Agreed include/linux/platform_data/voltage-omap.h has more functions as well. > > Considering it did: > > rename arch/arm/plat-omap/include/plat/voltage.h => > > include/linux/platform_data/voltage-omap.h > > > > Where do we move these functions to? > > > > drivers/power/avs/smartreflex.c needs: > > omap_voltage_get_voltdata > > and > > drivers/power/avs/smartreflex-class3.c > > will need voltdm_reset and voltdm_get_voltage > > How about something local drivers/power/avs/smartreflex.h? These APIs are exposed by voltage layer, not smartreflex :( stuff like voltdm_scale will have to be used by regulator logic eventually, so moving to AVS driver header is probably not right. -- Regards, Nishanth Menon