As part of an ongoing scsi problem on one of our netbsd fileservers
I was looking at sys/dev/ic/bha.c and came across the following chunk
of code:

        /*
         * If it has been through before, then
         * a previous abort has failed, don't
         * try abort again
         */
        if (ccb->flags & CCB_ABORT) {
                /* abort timed out */
                printf(" AGAIN\n");
                /* XXX Must reset! */
        } else {
                /* abort the operation that has timed out */
                printf("\n");
                ccb->xs->error = XS_TIMEOUT;
                ccb->timeout = BHA_ABORT_TIMEOUT;
                ccb->flags |= CCB_ABORT;
                bha_queue_ccb(sc, ccb); 
        }  


So does the XXX comment mean that this is something that needs to be
implemented but hasn't yet?  When the scsi devices start misbehaiving we
usually see:

sd0(bha0:0:0): timed out
sd0(bha0:0:0): timed out AGAIN

type of messages.  

					--Greg