Disk retries on error

Hi,

I'm new to the Linux SCSI community. I got a few questions on how SCSI drivers handle read/write failures.

When there is a latent sector fault (unable to read/write into a sector), or if the data is corrupted (CRC failed), does the SCSI drivers reissue the command ?

Does this differ from one SCSI driver (say, for IBM disk) to another (say, for Seagate) ?

of them are:

1. MEDIUM_ERROR
What types of errors are classified as MEDIUM_ERROR ? Is this same as data corruption ?

2. ABORTED_ERROR
Again, what types of errors are classified as this type ?

Any information regarding this will be greatly helpful.

Thanks,
Vijayan