From mboxrd@z Thu Jan 1 00:00:00 1970 From: Kevin Hilman Subject: Re: [PATCH 2/2] OMAP3: PM: Configure CPUidle latencies/thresholds from board files Date: Tue, 30 Jun 2009 15:58:45 -0700 Message-ID: <87ab3p49cq.fsf@deeprootsystems.com> References: <1246286113-21197-1-git-send-email-rnayak@ti.com> <1246286113-21197-2-git-send-email-rnayak@ti.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from wf-out-1314.google.com ([209.85.200.175]:17884 "EHLO wf-out-1314.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753512AbZF3W6o (ORCPT ); Tue, 30 Jun 2009 18:58:44 -0400 Received: by wf-out-1314.google.com with SMTP id 26so177366wfd.4 for ; Tue, 30 Jun 2009 15:58:47 -0700 (PDT) In-Reply-To: (Sanjeev Premi's message of "Tue\, 30 Jun 2009 17\:11\:18 +0530") Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: "Premi, Sanjeev" Cc: "Nayak, Rajendra" , "linux-omap@vger.kernel.org" "Premi, Sanjeev" writes: >> extern void omap1_map_common_io(void); >> extern void omap1_init_common_hw(void); >> @@ -234,7 +235,8 @@ extern void omap2_init_common_hw(struct >> omap_sdrc_params *sp, >> struct omap_opp *mpu_opps, >> struct omap_opp *dsp_opps, >> struct omap_opp *l3_opps, >> - struct prm_setup_vc *setup_times); >> + struct prm_setup_vc *setup_times, >> + struct cpuidle_params *idle_params); > > Rajendra, all, > > Aren't we 'overloading' omap2_init_common_hw() each time we add another > argument to it? > > Shouldn't we defined define a function specific for PM initialization? > Some of the initialization actually doesn't even touch the HW. > Completely agree. We are absolutely overloading init_common_hw() and this is on the list of things I'd like to see changed before those change go upstream. I just haven't got to that yet. I would happily welcome any proposals or cleanups to this. My current feeling is that I think these board-specific settings should just be using platform_driver/platform_device model and then board-specifics could be passed via platform_data. If not settings are given, conservative defaults can be used. But honestly, I haven't given it much more thought, other than I *really* don't like the current overloading of that function. Kevin