Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751849AbaKTFPo (ORCPT ); Thu, 20 Nov 2014 00:15:44 -0500 Received: from mail-vc0-f175.google.com ([209.85.220.175]:56552 "EHLO mail-vc0-f175.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750780AbaKTFPm (ORCPT ); Thu, 20 Nov 2014 00:15:42 -0500 MIME-Version: 1.0 In-Reply-To: <20141120023050.12298.26964@quantum> References: <1415408818-28028-1-git-send-email-dianders@chromium.org> <20141120023050.12298.26964@quantum> Date: Wed, 19 Nov 2014 21:15:41 -0800 X-Google-Sender-Auth: ETyRomP5HhYU68lv5WeZmTmQ_oo Message-ID: Subject: Re: [PATCH v2] clk: Propagate prepare and enable when reparenting orphans From: Doug Anderson To: Mike Turquette Cc: Heiko Stuebner , Dmitry Torokhov , "linux-arm-kernel@lists.infradead.org" , "open list:ARM/Rockchip SoC..." , Russell King - ARM Linux , "linux-kernel@vger.kernel.org" Content-Type: text/plain; charset=UTF-8 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Mike, On Wed, Nov 19, 2014 at 6:30 PM, Mike Turquette wrote: > Quoting Doug Anderson (2014-11-07 17:06:58) >> With the existing code, if you find a parent for an orhpan that has >> already been prepared / enabled, you won't enable the parent. That >> can cause later problems since the clock tree isn't in a consistent >> state. Fix by propagating the prepare and enable. >> >> NOTE: this does bring up the question about whether the enable of the >> orphan actually made sense. If the orphan's parent wasn't enabled by >> default (by the bootloader or the default state of the hardware) then >> the original enable of the orphan probably didn't do what the caller >> though it would. Some users of the orphan might have preferred an >> EPROBE_DEFER be returned until we had a full path to a root clock. >> This patch doesn't address those concerns and really just syncs up the >> state. > > -ECANOFWORMS > > I'm thinking about this patch but I haven't quite made up my mind. It is > reasonable, but also some nice kind of error might be preferable when > preparing/enabling an orphaned clock. > > Under what conditions might a clock be orphaned? An obvious example is > just bad luck during the thundering herd of clock registrations from a > driver. In this case deferring the probe might be a good idea. > > However what about the case where a loadable module provides the parent > clock? It might be a long time before the orphan clocks gets picked up > from the orphanage. Is deferring probe the right thing here as well? I will defer to your wisdom here. I agree that these are the two primary solutions and I've picked one, but I have no idea which will be more of a PITA in the long run. Note: I'm not sure that anyone expects EPROBE_DEFER to be returned from a clk_enable() (do they?). It almost seems like the right answer is to fail to allow anyone to clk_get() any clock that doesn't have a path to root. I will say that without this patch or the EPROBE_DEFER solution we have a clear bug. You can get into a situation where a clock is enabled/prepared but its parent isn't. -Doug -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/