This message is from the T13 list server.
On Fri, 02 Aug 2002 11:47:24 +0100, Simon Hildebrand wrote: >This message is from the T13 list server. >> We have never seen the hard drive abort the data transfer, so we cannot >> complete this part of our validation. Under what cicumstances will a >> "modern" hard drive abort a data transfer, has anyone ever seen a drive >> do this? Is there a way of coercing a healthy drive into this behaviour. What do you mean by 'abort the data transfer'? Do you mean transfer only some of the sectors for the DMA command and then end the command with some error such as ABRT or UNC or IDNF? If a drive did this then I would expect it to happen at a sector boundary (but nothing in ATA/ATAPI-x requires that because any error means the entire transfer is invalid). In any case, early termination of a DMA command by a drive would require the drive to have some kind of error condition. For the host side, especially from the DMA engine standpoint, the drive just terminates the last data burst and then sets BSY=0 DRQ=1 and asserts INTRQ. If you are trying to test a host side DMA engine then just set up the DMA engine to transfer some extra data - the result at the hardware level will the basically the same - that data will not be transferred by the drive in a read command - some may be transferred in a write command but ignored by the drive - either way the drive will end the command before the host DMA engine reaches the end of the PRD list. *** Hale Landis *** www.ata-atapi.com ***
