This message is from the T13 list server.

On Tue, 11 Nov 2003 20:50:40 +0900, [EMAIL PROTECTED]
wrote:
>This message is from the T13 list server.
>About 1)
>ATA/ATAPI specification does not allow any time-out of device interface. So
>if data transfer direction is incorrect, it hangs up.

It *may* hang. In MW DMA many devices see any IOR-/IOW- assertion as
a valid transfer of data without paying attention to what is on the
Data bus. This means the command may complete with neither the host
or device knowing that there was a direction mismatch. I am happy to
say that U-DMA does not have this problem.

>Unfortunately, OS driver/software shows incorrect data transfer direction
>to PATA in some cases. This causes problem in P1394/USB-ATAPI bridge board
>actually. SATA-PATA bridge should have some knowledge of SCSI ope-code too.

I would say far more ATAPI devices set IO and CD wrong than there are
hosts that use the wrong transfer direction for a command.

>This protocol is valid between OS driver/software and SATA-PATA bridge.
>Current ATAPI ODD never mind DMADIR bit. This means SATA-PATA bridge should
>modify the ATAPI Identify Device data. ? How does it work now?

I was going to ask this question too... Thanks for asking it. This
proposal should be written such that it will work for a bridge that
modifies the ID data on-the-fly. This also means the bridge must turn
off the DMADIR bit in the Features register so an older device
doesn't see it. However, I really would like to know how the bridge
will know the difference between the FR value used with SET FEATURES
from the FR value used with a PACKET command - this means the bridge
must look at the command code and command parameters before sending
them on to the device.

This isn't a kludge - it is a HUGE KLUDGE!!!

Hale



*** Hale Landis *** www.ata-atapi.com ***



Reply via email to