From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Nelson Date: Sun, 10 Mar 2013 07:23:17 -0700 Subject: [U-Boot] [PATCH v2] mx6dl: Add initial support for Wandboard dual lite In-Reply-To: <20130310075140.55FE4200642@gemini.denx.de> References: <1362873704-23598-1-git-send-email-festevam@gmail.com> <20130310075140.55FE4200642@gemini.denx.de> Message-ID: <513C9755.3040802@boundarydevices.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de Hi Wolfgang, On 03/10/2013 12:51 AM, Wolfgang Denk wrote: > Dear Fabio Estevam, > > In message <1362873704-23598-1-git-send-email-festevam@gmail.com> you wrote: >> >> Add initial support for Wandboard dual lite. > ... > >> Changes since v1: >> - Use get_ram_size >> - Remove ARP_TIMEOUT >> - Remove unneeded "1u" > > Thanks. I have still a question: > >> +int dram_init(void) >> +{ >> + gd->ram_size = get_ram_size((void *)PHYS_SDRAM, PHYS_SDRAM_SIZE); >> + >> + return 0; >> +} > > I see this new code does not reference imx_ddr_size() any more. If > it's so easy to switch to get_ram_size(), should we not then also > convert the existing MX53 and MX6 boards to using get_ram_size(), and > remove the then unneeded imx_ddr_size() code? > > Or would there be any drawbacks? > In the specific case of Nitrogen6X, we're defining different configurations for each of the memory arrangements, so we could skip the call to imx_ddr_size(). For that matter, we could skip the call to get_ram_size() entirely because we're configuring each of the DDR settings very explicitly. Since there are timing difference between the x128M and x256M DDR chips we're using, we've split them off into separate files: 1066mhz_4x128mx16.cfg 1066mhz_4x256mx16.cfg and defined two board configurations to select between the two. Regards, Eric