From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from muru.com ([72.249.23.125]:46294 "EHLO muru.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750837AbeFHGWC (ORCPT ); Fri, 8 Jun 2018 02:22:02 -0400 Date: Thu, 7 Jun 2018 23:21:58 -0700 From: Tony Lindgren To: Faiz Abbas Cc: linux-kernel@vger.kernel.org, devicetree@vger.kernel.org, linux-omap@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-clk@vger.kernel.org, mark.rutland@arm.com, paul@pwsan.com, t-kristo@ti.com, robh+dt@kernel.org, bcousson@baylibre.com Subject: Re: [PATCH v3 4/6] bus: ti-sysc: Add support for software reset Message-ID: <20180608062158.GI5738@atomide.com> References: <20180606060826.14671-1-faiz_abbas@ti.com> <20180606060826.14671-5-faiz_abbas@ti.com> <20180607073530.GH5738@atomide.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: Sender: linux-clk-owner@vger.kernel.org List-ID: * Faiz Abbas [180607 10:24]: > Hi, > > On Thursday 07 June 2018 01:05 PM, Tony Lindgren wrote: > > * Faiz Abbas [180606 06:14]: > >> +static int sysc_reset(struct sysc *ddata) > >> +{ > >> + int offset = ddata->offsets[SYSC_SYSCONFIG]; > >> + int val = sysc_read(ddata, offset); > >> + > >> + val |= (0x1 << ddata->cap->regbits->srst_shift); > >> + sysc_write(ddata, offset, val); > >> + > >> + /* Poll on reset status */ > >> + if (ddata->cfg.quirks & SYSC_QUIRK_RESET_STATUS) { > >> + offset = ddata->offsets[SYSC_SYSSTATUS]; > >> + > >> + return readl_poll_timeout(ddata->module_va + offset, val, > >> + (val & ddata->cfg.syss_mask) == 0x0, > >> + 100, MAX_MODULE_SOFTRESET_WAIT); > >> + } > >> + > >> + return 0; > >> +} > > > > I wonder if we should also add SYSS_QUIRK_RESET_STATUS in > > addition to SYSC_QUIRK_RESET status to make it easy to > > read the right register? > > I assumed SYSC_QUIRK is the prefix to indicate the ti-sysc driver not > the register. Are there layouts in which the reset status bit is in the > sysconfig register rather than the sysstatus register? Yes we can have reset status bit in either syss or syssconfig register. We detect that in sysc_init_syss_mask() but we should also set something at that point to make it clear which reset to use. But as we don't need the quirk flag, it's probably set a function pointer after the detection. So how about let's have two functions sysc_reset() and sysc_syss_reset() and then we can implement sysc_syss_reset() in a separate patch after testing it when we have a non-platform data using example for sysc_syss_reset(). Regards, Tony