> > However, there are side effects to it as the mmc devices will be enumerated > > in a different order (due to -EPROBE_DEFER), this means this patch would likely > > break existing platforms. > > However this doesn't. The user daemon should never relies on the > sequency of probing the mmc block part. We neither support to provide > alias for the device id, nor guarantee the first probed controller get > the first block id. Please use PARTUUID for rootfs, etc. So this > shouldn't be a big deal. That's exactly what I thought as well... Thanks for the patch! Will test it the next days, but it looks good already.
Attachment:
signature.asc
Description: PGP signature
