acb 4000 bad block behavior

[Al Kossow]

I just read a quantum 540 that was attached to an adaptec 4000
What is the behavior when the scsi host reads the disk? Does it just
reduce the number of logical blocks available skipping over the flagged ones
or does it reserve a track on the last cylinder and try to remap them?

[David Gesswein]

Did it report as Adaptec? That seems likely for that controller.
If it Looks like it skips bad sectors.
LBA address has MSB set for bad/spare sectors. Good sectors get consecutive
LBA addresses.
if (lba_addr & 0x800000) {
msg(MSG_DEBUG, "Sector marked bad/spare flag %x on cyl %d head %d physical sector %d\n",

Extracted data file should only contain the good sectors.

Manual has some hints. Looks like it skips bad sectors when formatting.

http://www.bitsavers.org/pdf/adaptec/ACB-4000/500504-00_ACB-4000_OEM_Manual_Mar84.pdf

3) The ACB-4000 handles defects on a sector level transparent to the host. Since
no alternate tracks
...
If, in time, other defects appear on a drive, the contents of the entire drive should be
backed up and a new format operation performed. To identify the physical locations
of the troublesome blocks use the TRANSLATE command. The new defect locations
must then be added and sorted into the complete list.
The defect skipping technique implemented by the ACB-4000 is at the sector level
and does not require time-consuming seeks to spare track locations