i.MX multi SoC patches

Shawn Guo shawn.guo at freescale.com
Wed Aug 31 10:05:28 EDT 2011


On Wed, Aug 31, 2011 at 02:02:58PM +0100, Hector Oron wrote:
> Hello,
> 
> 
> 2011/8/24 Sascha Hauer <s.hauer at pengutronix.de>:
> > This is hopefully the last iteration of the patches for compiling
> > all ARMv4/v5 based i.MXs together. CONFIG_ARM_PATCH_PHYS_VIRT
> > lost its experimental status and we can still build uImages
> > even with CONFIG_AUTO_ZRELADDR enabled. This series selects
> > these options and makes a multi SoC image the default.
> > Also, we can now build all i.MX5 SoCs into a single image.
> > This series also updates the defconfigs.
> 
> On a mx53 loco board I have rebased Richard's patchset for SATA and
> got it running, but I see this on mapping on iomem, board is more or
> less working but I was unsure if this is something significant or just
> some ugly cosmetic names for mx53 (i.e. imx21-uart.0, imx31-gpio, ..)
> 
This is expected.  The platform_device_id had been used to kill
cpu_is_mxXX() stuff from drivers.  And we decided to use soc prefix
to define IP block version in platform_device_id table, because
hardware does not define revisions, and soc is stable enough to
identify specific IP block version.

-- 
Regards,
Shawn




More information about the linux-arm-kernel mailing list