From mboxrd@z Thu Jan 1 00:00:00 1970 From: Brian Swetland Subject: Re: Building kenel error Date: Fri, 18 Aug 2006 13:15:17 -0700 Message-ID: <20060818201517.GA4066@localhost.localdomain> References: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: linux-omap-open-source-bounces@linux.omap.com Errors-To: linux-omap-open-source-bounces@linux.omap.com To: "Woodruff, Richard" Cc: Linux-omap-open-source@linux.omap.com List-Id: linux-omap@vger.kernel.org ["Woodruff, Richard" ] > > > I never had success of using gcc-4.0.2 and later on 2.6.16/17 kernel. > So > > I just stick with the older gcc-3.4.1. Which version of tool chain are > > you using for 2.6.16 or 2.6.17 kernel? Any advantages of using the > later > > version of the compiler? > > Thanks, > > kwan > > We have to use several different versions. Currently I use 3.4.0, > 3.4.3, and 3.3.1 on older kernels. On the latest kernels we will use > 4.1.0. We have built these, gotten them from a vendor, or gotten from > www.codesourcery.com . > > One of the main reasons for using CodeSourcery based tool chains is they > provide newer ARM support though ARMv7 (including VFP & VFP2). If you > are doing gaming with the OMAP2 3d core it can provide a nice boost. What is codegen and stability like? I suffered through a number of pre-3.x versions while working at Danger and have found the 3.3.x and 3.4.x toolchains to be pretty solid. I had heard that the changes in 4.x resulted in big improvements in x86 codegen but other platforms suffered a bit -- any truth to this? New instruction set support is very cool, but I'm curious how the baseline arm and thumb support has held up. Brian