/etc/fstab: no entry relatime. But the partitions are mounted with option relatime. But this causes an error in dmesg. NCQ is disabled for a SSD and a HDD. ---------------------- [ 5.640926] ata2.00: ATA-9: WDC WD4003FZEX-00Z4SA0, 01.01A01, max UDMA/133 [ 5.640931] ata2.00: 7814037168 sectors, multi 16: LBA48 NCQ (depth 31/32), AA [ 5.641470] ata2.00: configured for UDMA/133 [ 5.642179] ata3.00: ATAPI: HL-DT-ST BD-RE BH16NS55, 1.01, max UDMA/133 [ 5.642252] ata1.00: supports DRM functions and may not be fully accessible [ 5.642296] ata1.00: READ LOG DMA EXT failed, trying unqueued [ 5.642320] ata1.00: failed to get NCQ Send/Recv Log Emask 0x1 [ 5.642323] ata1.00: ATA-9: Samsung SSD 850 PRO 512GB, EXM01B6Q, max UDMA/133 [ 5.642326] ata1.00: 1000215216 sectors, multi 1: LBA48 NCQ (depth 31/32), AA [ 5.642542] ata1.00: supports DRM functions and may not be fully accessible [ 5.642580] ata1.00: failed to get NCQ Send/Recv Log Emask 0x1 [ 5.642632] ata1.00: configured for UDMA/133 -------------------------------------
These errors have nothing to do with relatime. You can control whether the last access time is updated on every single read (which is a performance disaster, but required by POSIX) using the mount option strictatime. You can also set the last access time to never be updated using the mount option noatime. The default, relatime will update atime if it is more than 24 hours out of date, or if atime < mtime. None of this has anything to do with the NCQ error, which is coming from the device, and not the file system.
I have disabled relatime in fstab. And this error did not occur.
So number one this is a distro kernel, so I suggest you ask the SuSE help desk for support. Number 2, I suggest you give exact reproduction details --- e.g., what was in your fstab before and after. And then try reproducing this on an upstream kernel if you want support on bugzilla.kernel.org.