4k sector drive reporting 512 byte sectors to OS - why?


Hi all, I have just purchased two Western Digital WD20EARS disks and discovered that they incorrectly report their physical sector size to the OS as 512 bytes, even though the specs suggest they are using 4k sectors internally:

/sys/block/sdh/device/model: WDC WD20EARS-00M
/sys/block/sdh/queue/physical_block_size: 512
/sys/block/sdh/queue/logical_block_size: 512
/sys/block/sdh/alignment_offset: 0

I am worried that the OS kernel is going to think the drive writes only need to be aligned on 512-byte boundaries instead of 4k. Is there some way I can make the drive correctly report its physical sector size, as apparently some other WD drives (and many non-WD drives) do?


it is ok, that’s ‘advanced format’, the firmware emulates a 512 byte sector disk.

thw WD20EARS is a 512 byte logical / 4K physical harware sector drive, that’s it.

just do not forget to align the partitions in order to avoid the performance penality of having

the partitions not aligned.


Yes, that’s my problem - it *emulates* a 512-byte sector disk. How do I turn that off, and make it behave like a real 4k sector disk? (i.e. with a lower maximum LBA)


A drive reports its Advanced Format capabilities via the ATA Identify Device command.

See these threads for examples:


The following thread explains how to access your drive’s Identify Device data in Linux:


Thanks for the info! I used smartctl under Linux and viewed the IDENTIFY DEVICE hex data. The model number correctly appears in words 27-46 as per the ATA8 spec, so it would seem I am looking at the correct data.

However word 106 (physical/logical sector size) is zero, as are words 117-118 (words per logical sector).

You mentioned in the linked thread that this information is optional, so I guess WD chose to omit it from these drives for some reason. This is a shame, as one of the reasons why I chose this drive was because I wanted to try out these 4k sectors to see how my system behaved. Maybe I should just return the drive? Seems a shame when there’s nothing mechanically wrong with it, but I’m guessing there’s not much likelihood of getting a firmware update any time soon to fix this… :frowning: