Hi Benoit, On Tue, Sep 21, 2010 at 2:26 AM, Cousson, Benoit <b-cousson@xxxxxx> wrote: > Hi Kishore > > On 9/18/2010 6:34 PM, Kadiyala, Kishore wrote: >> >> The offset handling implementation of omap4 mmc registers which >> was already present can't be reused once hwmod modifications are done >> for mmc driver. >> Since hwmod data file for OMAP4 is an auto generated the base >> address for MMC will remain same as OMAP3 and thus the offset deviation >> of registers should be updated in the hsmmc driver. > > That explanation does not sound very accurate to me. > The fact that the hwmods are auto-generated has nothing to do with your > offset problem. Agree the change log needs to be updated as said but when coming to implementation this how it is: With current mainline kernel, offset appending is already implemented in mach-omap2/devices.c and works fine [here bases are updated with 0x100 and then passed to driver which has same register offset's as OMAP3]. When hwmod modifications are done to the MMC driver, the base address are picked from omap_hwmod_44xx_data.c and omap_device is build. Coming to the driver it still has the same old register offset's as OMAP3 and there should be a way to add 0x100 some where. The current patch moves the offset appending to the driver file which is updated in mach-omap2/hsmmc.c which will fit for current implementation and even hwmod adaptation. >The real cause is due to the update of the IP to support the > new PM programming model. > Because of that, and to keep a certain level of compatibility, the legacy > registers were all shifted by 0x100 and the new one were added from offset 0 > to 0x10. > > Regards, > Benoit > <snip> Regards, Kishore -- To unsubscribe from this list: send the line "unsubscribe linux-mmc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
