From mboxrd@z Thu Jan 1 00:00:00 1970 From: Albert ARIBAUD Date: Sun, 18 Sep 2011 19:58:14 +0200 Subject: [U-Boot] [PATCH v7] dreamplug: initial board support. In-Reply-To: References: Message-ID: <4E763136.4010709@aribaud.net> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de (Please do not use my @free.fr address to copy me on U-boot matters...) Le 14/09/2011 08:39, Prafulla Wadaskar a ?crit : > > >> -----Original Message----- >> From: Jason [mailto:u-boot at lakedaemon.net] >> Sent: Tuesday, September 13, 2011 7:47 PM >> To: Wolfgang Denk >> Cc: Prafulla Wadaskar; Albert Aribaud; u-boot at lists.denx.de >> Subject: Re: [U-Boot] [PATCH v7] dreamplug: initial board support. >> >> Albert, >> >> On Tue, Sep 13, 2011 at 03:00:59PM +0200, Wolfgang Denk wrote: >>> Dear Prafulla, >>> >>> In message> VEXCH4.marvell.com> you wrote: >>>> >>>> Let's have Wolfgang's opinion on this, since this not aligned with >> current >>>> u-boot development strategy. >>>> >>>> May be we can create a separate header file for tracking >> (unsupported/tobe >>>> supported) arm machine-types. >>> >>> Actually this is for Albert to comment. He is the ARM custodian and >>> has to live with the results. >> >> Wolfgang, Marek Vasut, and I discussed this here [1]. To summarize, by >> declaring non-mainlined mach_types in the respective board config, an >> error will be thrown at compile time after mach-types.h is updated to >> include the mach_type. >> >> The other idea is to have a separate file, say mach-types-local.h where >> all non-mainlined mach-types would be defined. > > I will vote for this second approach so that it becomes independent change and anyone can update it in future. > > Let's get Albert's opinion on this. My opinion on the whole mach-type question is "if a board needs a mach-type it's because it will run Linux, so its mach-type should eventually be in the Linux mach-type list". As I understand it, the only case when is not there is because U-Boot support is submitted before Linux mainline support. Thus I second the idea of defining it in the board config header file, possibly even testing for it first and if it already exists, throwing a #error to remind the board maintainer to remove the now useless define from the config file. > Regards.. > Prafulla . . Amicalement, -- Albert.