Re: [hatari-devel] ACSI: READ/WRITE (10) incorrectly evaluate block number

[Uwe Seimet <Uwe.Seimet@xxxxxxxxx>]

Hi,

> > Yes, this works fine now, thank you. Reading beyond the last sector is
> > not yet rejected, though.
> 
> Oh, well, I (or whoever wrote that code) was assuming that doing
> an fseek() beyond the end of the file would return an error ... but it
> apparently does not. So I've added now additional checks for correct
> sector numbers, so that this problem should now be gone, too.

Yes, perfect. So two things remain that as far as I can tell are not yet as
they are supposed to be:

1. Commands like READ CAPACITY (16), READ (16) or REPORT LUNS are not
yet rejected with a Sense Key of $05 and an ASC of $20.
2. Trying to read a block with READ (6) from a non-existing LUN should
return SenseKey $05 and ASC $25 but returns $00 both times. There is a
special rule that applies when REQUEST SENSE is called with a (the same)
non-existing LUN in order to retrieve the error information for the command
that failed for this LUN. For this case SPC-4 says:

In response to a REQUEST SENSE command issued to a logical unit that
reports a peripheral qualifier of 011b in its standard INQUIRY data
(see 6.4.2) the device server shall:
1) Return parameter data containing sense data with the sense key set to
ILLEGAL REQUEST and the additional sense code shall be set to LOGICAL UNIT
NOT SUPPORTED; and
2) Complete the REQUEST SENSE command with GOOD status.

So in this case REQUEST SENSE does not return CHECK CONDITION even though
error information is available.

Take care

Uwe