Comment 136 for bug 438136

To get a definitive answer to whether the drive is actually remapping defects due
to bad blocks you have to examine the "defect lists". It's easy to do on SCSI, I
expected libata to swizzle the cmd into ATA'ez but it apparently fails on my dell.

sg_reassign --grow /dev/sda

I don't know how to do it in ATA off the top of my head but I could find out. It's likely
accessible only through some SMART log page.

I agree with the assertion mentioned here:
http://thread.gmane.org/gmane.linux.utilities.smartmontools/5252

The reason being the drive firmware will remap to another block on write should
there be an error. Smart may be trying to be "too smart" by observing the block
re-allocation and coming to the conclusion that this is possible iif a write
error occured.

Here's a decent defintion of how defect tracking works:
http://www.dataclinic.co.uk/hard-drive-defects-table.htm