Comment # 69 on bug 922529 from Takashi Iwai

(In reply to Oliver Neukum from comment #54)
> The drive first errors out when you access an rpmb partition. It looks like
> we are running blkid on this partition. We must not do that. RPMB partitions
> cannot have a filesystem. The best option is to ignore it.
> Now it would be best if the driver were able to recover, but failing that,
> don't crash it.
> 
> [   28.005490] mmc1: Got data interrupt 0x00000002 even though no data
> operation was in progress.
> [   28.021864] mmcblk0rpmb: error -110 transferring data, sector 8064, nr 8,
> cmd response 0x900, card status 0xb00
> [   28.029502] mmcblk0rpmb: retrying using single block read
> [   28.031787] mmcblk0rpmb: timed out sending r/w cmd command, card status
> 0x400900
> [   28.041820] mmcblk0rpmb: timed out sending r/w cmd command, card status
> 0x400900
> [   28.051618] mmcblk0rpmb: timed out sending r/w cmd command, card status
> 0x400900
> [   28.061314] mmcblk0rpmb: timed out sending r/w cmd command, card status
> 0x400900
> [   28.070912] mmcblk0rpmb: timed out sending r/w cmd command, card status
> 0x400900
> [   28.080383] mmcblk0rpmb: timed out sending r/w cmd command, card status
> 0x400900
> [   28.087579] end_request: I/O error, dev mmcblk0rpmb, sector 8064

FYI, there is a recent upstream fix (in 4.1) regarding RPMB partition:
4e93b9a6abc0d028daf3c8a00cb77b679d8a4df4
    mmc: card: Don't access RPMB partitions for normal read/write

So, it's anyway good to test the latest Tumbleweed indeed.  The relevant dracut
fix has been already merged in TW.