Lines Matching refs:we
52 * After starting the selection hardware, we check for reconnecting targets
54 * bus arbitration. The problem with this is that we must keep track of the
55 * SCB that we've already pulled from the QINFIFO and started the selection
56 * on just in case the reselection wins so that we can retry the selection at
104 * We have at least one queued SCB now and we don't have any
124 * before we completed the DMA operation. If it was,
211 /* The Target ID we were selected at */
239 * Watch ATN closely now as we pull in messages from the
285 * we've got a failed selection and must transition to bus
333 * Reselection has been initiated by a target. Make a note that we've been
362 * when we have outstanding transactions, so we can safely
364 * we start sending out transactions again.
412 * message byte we receive so it can be checked prior to
467 * If we are the result of a tagged command, send
486 * we properly identified ourselves.
541 * and we believe this selection to be valid. If the ring
572 * the total for the command since we've already fetched
615 /* MSG_CMDCMPLT is 0, but we can't do an immediate of 0 */
647 * As soon as we get a successful selection, the target
648 * should go into the message out phase since we have ATN
698 * We don't know the target we will connect to,
731 * Do we need any more segments for this transfer?
735 /* Did we just finish fetching segs? */
738 /* Are we actively fetching segments? */
742 * Do we have any prefetch left???
747 * Need to fetch segments, but we can only do that
749 * sure we don't have an SCB prefetch going on.
755 * so we don't end up referencing a non-existent page.
791 * XXX Can we optimize this for PCI writes only???
839 * If we re-enter the data phase after going through another
855 * modify the saved values in the SCB until we see a save
921 * As a target, we control the phases,
948 * flush process going as we complete the
957 * allows us to update our notion of where we are
960 * If, by chance, we stopped before being able
967 * If we happened to stop on the last segment, then
974 /* Record if we've consumed all S/G entries */
990 * the SG_CACEPTR register so we can restore the
1008 * In the future, we'll want to handle end of
1026 * and for FIFOEMP to de-assert. Here we
1076 * Here we fixup the HSHADDR stored in the high byte
1078 * we can batch the clearing of HADDR with the fixup.
1079 * If we halted on the last segment, the residual is
1080 * already correct. If we are not on the last
1093 /* If we are the last SG block, tell the hardware. */
1121 * PHASEMIS is active). If we are doing a SCSI->Host transfer,
1154 /* See if we have completed this last segment */
1188 * into the DMA engine. If we don't have hardware
1189 * to perform a prefetch, we'll have to fetch the
1258 * we should only loop if there is a data overrun. For
1259 * all other adapters, we'll loop after each S/G element
1299 * lost by some other type of SCSI I/O (e.g. PIO). If we have
1329 * Since we've been through a data phase, the SCB_RESID* fields
1335 /* Clear the channel in case we return to data phase later */
1407 * we aren't going to touch host memory.
1480 * associated with this target is set, we will also interrupt the host,
1569 * we perform the phase_lock before checking to see
1570 * if we should exit the loop and skip the phase_lock
1605 * In either of these conditions, we upload the SCB back to the host so it can
1606 * process this information. In the case of a non zero status byte, we
1616 * If ATN is raised, we still want to give the target a message.
1621 * for this command and setting ATN while we are still processing
1627 * If we are identified and have successfully sent the CDB,
1640 * If we recevied good status but never successfully sent the
1648 * See if we attempted to deliver a message but the target ingnored us.
1700 * If ATN is raised, we still want to give the target a message.
1702 * or we want to abort this command. Either way, the target
1723 * only if we've actually been into a data phase to change them. This
1725 * since they are only initialized when we go into data_in or data_out.
1727 * we can only ack the message after SHADDR has been saved. On these
1739 * If we are asked to save our position at the end of the
1754 * The SCB_SGPTR becomes the next one we'll download,
1775 * SCB anytime we enter a data phase for the first time, so all
1776 * we need to do is clear the DPHASE flag and let the data phase
1778 * sure we have a clean start for the next data or command phase.
1806 * clearing the "disconnected" bit so we don't "find" it by accident later.
1864 * disconnected list, so we must replace
1873 * Here we "snoop" the bus looking for a SIMPLE QUEUE TAG message.
1874 * If we get one, we use the tag returned to find the proper
1875 * SCB. With SCB paging, we must search for non-tagged
1876 * transactions since the SCB may exist in any slot. If we're not
1877 * using SCB paging, we can use the tag as the direct index to the
1962 * the target until SCSIDATL is read from. So we wait until SCSIDATL is
1964 * using SCSIBUSL. When we have pulled the ATN line, or we just want to
1965 * acknowledge the byte, then we do a dummy read from SCISDATL. The SCSI
1967 * we send our ACK.
1998 * Change to a new phase. If we are changing the state of the I/O signal,
2013 * we must wait at least a data release delay plus
2252 * transfer, we must empty the fifo of all
2257 * When the hang occurs, we know that a multiple