From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Warren Date: Wed, 02 Jul 2014 15:53:51 -0600 Subject: [U-Boot] [PATCH] ARM: tegra: Use mem size from MC rather than ODMDATA In-Reply-To: <20140702211821.3EFF6380185@gemini.denx.de> References: <1404331950-4916-1-git-send-email-swarren@wwwdotorg.org> <20140702211821.3EFF6380185@gemini.denx.de> Message-ID: <53B47F6F.1090405@wwwdotorg.org> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de On 07/02/2014 03:18 PM, Wolfgang Denk wrote: > Dear Stephen Warren, > > In message <1404331950-4916-1-git-send-email-swarren@wwwdotorg.org> you wrote: >> >> In at least Tegra124, the Tegra memory controller (MC) has a register >> that controls the memory size. Read this to determine the memory size >> rather than requiring this to be redundantly encoded into the ODMDATA. >> This way, changes to the BCT (i.e. MC configuration) automatically >> updated SW's view of the memory size, without requiring manual changes >> to the ODMDATA. > > Is there a specific reason for not using get_ram_size()? Since we know the exact RAM size, we may as well simply use it directly rather than "probing" for it. I know that if non-existent peripheral addresses are accessed by the CPU, the CPU or some bus hangs. I'm not sure if the same applies to addresses within the memory window where there is not actually RAM present on a particular board, but I'd rather not risk it by touching them during probing. BTW, I'm out on vacation starting tomorrow, so may not respond soon.