Setting Power Condition with Start/Stop Unit Command

[George Penokie]

* From the SCSI Reflector, posted by:
* "George Penokie" <GOP at RCHVMP3.VNET.IBM.COM>
*
      All,

      Tony Cappellini from Quantum (see below) points out an area of the
      SBC having to do with the power conditions and the start/stop unit
      command that is confusing.  I agree with him that the way it is
      written today could lead to an improper implementation.  I propose
      the following EDITORIAL change be made to the SBC draft to correct
      this problem:

      In section 5.1.4 change:

      "If the START/STOP UNIT command is issued with the Power Conditions
      field set to 1, 2, 3, or 5 the logical unit shall:"

      to:

      "If the START/STOP UNIT command is issued with the Power Conditions
      field set to 1, 2, or 3 the logical unit shall:"

      and add the following after the first abc list:

      "If the START/STOP UNIT command is issued with the Power Conditions
      field set to 5 the logical unit shall:

      a) change power conditions only on receipt of a hard reset

      b) suspend any Power Condition timers that are active on receipt
      of the START/STOP UNIT command until a hard reset occurs

      c) not respond to a selection service until a hard reset occurs."

  >   Hi George
  >
  >   I have some additional concerns regarding sleep mode.
  >
  >   The spec states ...
  >
  >   "The lowest power consumption occurs in the Sleep condition.   When
  >   in sleep condition a unit needs a bus reset to be activated"
  >
  >   - SBC rev 1, section 5.1.4
  >
  >   "If the START/STOP UNIT command is issued with the Power Conditions
  >   field set to 1, 2, 3, or 5 <note: 5 is sleep mode> the logical unit
  >   shall:
  >
  >   a) change power conditions only on receipt of another START/STOP
  >   UNIT command or hard reset
  >
  >   b) suspend any Power Condition timers that are active on receipt
  >   of the START/STOP UNIT command until another START/STOP UNIT command
  >   is received which returns control of the power condition to the
  >   logical unit or a hard reset occurs
  >
  >   c) terminate any command received which requires more power than
  >   allowed by the START/STOP UNIT commands *most recent* power condition
  >   setting with a CHECK CONDITION status.  The sense key shall be set to
  >   ILLEGAL REQUEST and the additional sense code shall be set to LOW
  >   POWER CONDITION ACTIVE"
  >
  >   - SBC Rev 1, section 6.1.17
  >
  >   "In the sleep condition the device server shall only respond to a bus
  >   reset condition."
  >
  >   - SBC Rev 1, section 6.1.17
  >
  >   In the second excerpt, it sounds as if a device can be changed from
  >   sleep mode to another mode on receipt of another start/stop command.
  >   Additionally, it sounds like a command issued to a target in sleep
  >   mode should return a check condition status.
  >
  >   In the other excerpts, it sounds as if once a target is in sleep mode,
  >   only a bus reset can bring it out.  And they do not address whether or
  >   not a check condition should be returned and status generated, while
  >   the second excerpt seems to say that it should.  The spec does include
  >   sense codes and ASQs for a logical unit in standby and idle mode, but
  >   not for sleep mode.
  >
  >
  >   George, what are your views on the above ?
  >
  >
  >   thanks
  >
  >
  >   Tony Cappellini
  >   Quantum Corporation