SCSI-2 Parity Errors
monia at am.shrmsg.shr.mts.dec.com
Tue Mar 21 17:59:13 PST 1995
In his reply to Denis, Kurt Chan stated:
>| What should happen if parity errors are detected in COMMAND and/or DATA OUT
>| phase? How should retries be performed in each of these situations and
>| what should happen after retries have been exhausted?
>Denis asks an interesting question. What about the initiator response
>if it detects bad parity on a DATA IN phase?
>More specifically, what would happen if an error occurred on a command
>sent with NACA=1, and a parity error occurred on DATA IN during the
>subsequent Request Sense data transfer? Would the sense data be lost
>permanently regardless of whether ACA is in effect?
My take is that sense data may be lost but so what? I and apparently a lot of
others are under the impression that SPI parity errors are symptoms of a fatal
illness. If that's true, trying to recover sounds like rearranging the deck
chairs on the Titanic.
As someone else pointed out, attempted recovery may even make the problem
worse, since the parity errors you see may indicate the presence of double-bit
errors that you don't see.
More information about the T10