From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stefano Babic Date: Fri, 09 Sep 2011 12:29:01 +0200 Subject: [U-Boot] [PATCH 01/31] iMX28: Initial support for iMX28 CPU In-Reply-To: <201109091212.09015.marek.vasut@gmail.com> References: <1315514579-19215-1-git-send-email-marek.vasut@gmail.com> <1315514579-19215-2-git-send-email-marek.vasut@gmail.com> <4E69CCED.8050504@denx.de> <201109091212.09015.marek.vasut@gmail.com> Message-ID: <4E69EA6D.7060607@denx.de> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de On 09/09/2011 12:12 PM, Marek Vasut wrote: >> >> Another general remark here: we tend to have the same structure and the >> same files for all IMX SOC. This means that all IMX SOC have a >> imx-regs.h that contain the required register definitions (really a >> subset what we have in kernel). Is it really necessary to split the >> definitions in several small files ? > > Just like Heiko said ... it'd produce terribly big and messy file. I'd prefer to > avoid that. This is ok, and it is ok also that your imx-regs.h file contains only the reg-specific include files. What is not ok, is that you do not use imx-regs.h and you still includes each separate file. And as I see, they must be included in a specific order. This order must be defined only in imx-regs.h, as you have already done and then each driver/board requires to include only imx-regs.h, without having to care of the single files. > Honestly, I need to check if the file is used at all. We replaced the original > driver with pl011 driver. Ok >> Again, regarding file splitting for register. If you prefer to have >> several files, I think it is better that imx-regs.h include them. Then >> we have still a common header imx-regs.h that contains all needed >> register definitions. This is a better interface for a board maintainer. > > Won't it make the compiler slower at compile time if it has to go through all of > the included files instead of a subset ? Well, I do not know if on our Intel-PC this makes a so noticeable difference ;-) What I remark here, it is to have a clear and identical interface among the several SOCs. This is not only easier for board developers, but also removes some nasty #ifdef CONFIG_MX from the drivers. >> >> To understand: does a reset mean on the i.MX28 a power off ? Do you turn >> off the power ? If this is the case, some features are not possible (as >> PRAM, for example) on this SOC. > > This should just reset the chip. No poweroff. Ok - thanks for clarification. > > I'll probably wait for someone to clearly say how this should be to avoid > reworking it for the fourth time. Right. Best regards, Stefano Babic -- ===================================================================== DENX Software Engineering GmbH, MD: Wolfgang Denk & Detlev Zundel HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany Phone: +49-8142-66989-0 Fax: +49-8142-66989-80 Email: office at denx.de =====================================================================