BBB MMC / SD detection instability with U-Boot 2014.04 (CPU 1GHz)

[SAITOU Toshihide]

In message: <1400708280.1152.199.camel at revolution.hippie.lan>
            Ian Lepore <ian at FreeBSD.org> writes:
> On Tue, 2014-05-20 at 23:42 +0900, SAITOU Toshihide wrote:
>> In message: <537B62D1.4090901 at hot.ee>
>>             "Sulev-Madis Silber (ketas)" <madis555 at hot.ee> writes:
>> > On 2014-05-20 15:20, SAITOU Toshihide wrote:
>> >> In message: <20140520.191001.03109216.toshi at ruby.ocn.ne.jp>
>> >>             SAITOU Toshihide <toshi at ruby.ocn.ne.jp> writes:
>> >>> In message: <537ACDB2.9080808 at hot.ee>
>> >>>             "Sulev-Madis Silber (ketas)" <madis555 at hot.ee> writes:
>> >>>>
>> >>>> Actually I guess many people might think like me... "HELL, optimizing
>> >>>> boot time of 1min?! I have more important tasks to do than this".
>> >>>
>> >>> If you have ``device sdhci'' line in conf/BEAGLEBONE, is there any
>> >>> differences by changing mmchs to sdhci in am335x.dtsi and
>> >>> beaglebone-black.dts? And also remove ``status = "disabled"''
>> > 
>> > Don't edit am335x.dtsi
>> > 
>> > And beaglebone-black.dts already has proper config.
>> 
>> In this case, how does ``device sdhci'' driver know the register address?
>> I thought that there is an inconsistency in BEAGLEBONE config and
>> dts file for the SD/MMC driver.
> 
> The name at address tag in the dts[i] files doesn't have to match the
> driver name in the kernel config.  What matters for matching the driver
> to the right entry in the dts is the "compatible=" property.  The
> ti_sdhci driver looks for an entry with any one of these compatible
> strings:
> 
>   "ti,omap3-hsmmc", "ti,omap4-hsmmc", "ti,mmchs"

Ah! I see ti_sdhci.c has ofw_bus_is_compatible(dev, "ti,mmchs").
(Why I felt difference of eMMC detection... Hmm...)

On the other hand, u-boot's MMC/SD driver has problem when using eMMC.
Crochet has patch for this. The card detection and init problem is
exist not only for FreeBSD but also u-boot's MMC operation.

-- 
SAITOU Toshihide