From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tony Lindgren Subject: Re: [PATCH v5 1/4] TI816X: Update common omap platform files Date: Thu, 10 Feb 2011 10:25:46 -0800 Message-ID: <20110210182545.GN20795@atomide.com> References: <1297270555-32447-1-git-send-email-hemantp@ti.com> <20110210013910.GF20795@atomide.com> <2A3DCF3DA181AD40BDE86A3150B27B6B036ACA41C9@dbde02.ent.ti.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mho-02-ewr.mailhop.org ([204.13.248.72]:19473 "EHLO mho-02-ewr.mailhop.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751102Ab1BJS0W (ORCPT ); Thu, 10 Feb 2011 13:26:22 -0500 Content-Disposition: inline In-Reply-To: <2A3DCF3DA181AD40BDE86A3150B27B6B036ACA41C9@dbde02.ent.ti.com> Sender: linux-omap-owner@vger.kernel.org List-Id: linux-omap@vger.kernel.org To: "Pedanekar, Hemant" Cc: "linux-omap@vger.kernel.org" , "Hilman, Kevin" , "paul@pwsan.com" * Pedanekar, Hemant [110210 09:37]: > cvTony Lindgren wrote on Thursday, February 10, 2011 7:09 AM: > > I will take care of this using separate map_io as you mentioned on > the other mail on patch 2 comment. OK great. > Do you think the 2nd case restriction mentioned in this patch comment is OK? > > " 2) OMAP3 only build with CONFIG_SOC_OMAPTI816X: This will build TI816X > optimized kernel. May not boot on other OMAP3 SoCs." > > As I had mentioned on follow up on Paul's comment on v4, some OMAP3 specific > code may get skipped on non-multi-omap builds with CONFIG_SOC_OMAPTI816X > (cpu_is_ti816x() is true) such as: > > - if (omap_rev() >= OMAP3430_REV_ES2_0) > + if (!cpu_is_ti816x() && (omap_rev() >= OMAP3430_REV_ES2_0)) > omap3_clk_lock_dpll5(); I suggest that if CONFIG_ARCH_OMAP3 is selected, we do dynamic checks for cpu_is omap macros for 34xx/36xx/ti816x. We can use CONFIG_SOC_OMAP to optimize the size a bit if some features are not needed, like clock data etc. But in general the cpu_is checks should be only done during the init for most part, so at this point doing performance optimizations with the cpu_is macros should not really be needed. Regards, Tony