From mboxrd@z Thu Jan 1 00:00:00 1970 From: u.kleine-koenig@pengutronix.de (Uwe =?iso-8859-1?Q?Kleine-K=F6nig?=) Date: Thu, 25 Nov 2010 09:14:06 +0100 Subject: [PATCH 1/1] arm: imx: share imx5x_register_gpios for mach-mx5 In-Reply-To: References: <1290695312-30031-1-git-send-email-linuxzsc@gmail.com> <20101125062924.GD4693@pengutronix.de> <20101125074244.GE4693@pengutronix.de> Message-ID: <20101125081406.GG4693@pengutronix.de> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Thu, Nov 25, 2010 at 04:01:04PM +0800, Richard Zhao wrote: > 2010/11/25 Uwe Kleine-K?nig : > > Hey Richard, > > > > On Thu, Nov 25, 2010 at 02:44:48PM +0800, Richard Zhao wrote: > >> > IMO the machine files shouldn't need to know how many gpio banks are > >> > available. ?This is a per-SoC thing and so should be set in a SoC function. > >> But we don't have SoC level device file. > > Check how I did it for imx{1,21,27,25} in arch/arm/plat-mxc/gpio.c. > I considered put it there too. but we can not add new defconfig, and > even can not create new items in Kconfig. So we can not use ifdefs. Why not add new Kconfig symbols? Up to now I was unfettered by scruples adding new ones. > And per my understanding, mx5x is going to use one zImage, correct? It should be possible to have a single image for more than one SoC, yes. > >> > How do you want to continue this change? ?mx51 and mx53 have the same > >> > number of ports? ?Does mx50 have more or less? ?The addresses are the > >> > same? > >> mx53 and mx50 have similar memory map. but mx51 is different. ?Sorry, > >> I forgot upstream version IO_ADDR can not handler differnt SoC base > >> addr offset. > > It might even work in this case, still I prefer a more explicit way. > > > >> ? ? ? ? ? ? ?It can not be shared with mx51, but it can be shared with > >> mx50/53, correct? May plan is mx53/50 share > >> arch/arm/plat-mxc/include/mach/mx5x.h file. > > mx5x.h is for mx50 and mx53 but not mx51? ?No please. > For memory map, mx51 is a special case. From mx53 on , IC tuned memory > map to support large memory. > If we use things like mx53-50.h, symbols is hard to define. > MX53_50_XXX_XXX is wired. Any possibly, we need to add new SoC with > similar memory map in the future. I still think defining symbols for a single SoC is fine most of the time. Best regards Uwe -- Pengutronix e.K. | Uwe Kleine-K?nig | Industrial Linux Solutions | http://www.pengutronix.de/ |