mario@hpindda.cup.hp.com (Mario Arias) (04/18/91)
A question concerning 3270 bisynch. Does anyone know if an RVI followed by status sense response of "DEVICE BUSY" is a legal response to a host device selection? The 3274 Control Unit description manual indicates that it is, but it is not clear. If a device is busy when a host tries to select it, should a WACK be used to respond or can the above mentioned status sense message of "DEVICE BUSY" be used? Any help or comments would be appreciated. Thanks in advance, Mario Arias
lstowell@pyrnova.pyramid.com (Lon Stowell) (04/20/91)
In article <5160006@hpindda.cup.hp.com> mario@hpindda.cup.hp.com (Mario Arias) writes: > > >A question concerning 3270 bisynch. Does anyone know if an >RVI followed by status sense response of "DEVICE BUSY" is >a legal response to a host device selection? The 3274 >Control Unit description manual indicates that it is, >but it is not clear. > The manual actually notes that WACK is the legal response according to Notes 2 and 3. Actually, as long as the ABSOLUTE NEXT outbound sequence to the device (after the RVI) is an EOT, then an Specific Poll, this is an error. Note 2 of the BSC Selection Addressing diagram specifically states that the device has pending status....EXCLUDING either Device End or Busy. In other words, a device with ONLY Busy or Device End should NOT issue an RVI. If the status is only Device End, the device should send an ACK 0, reset it's DE status internally, and take the following command. If the status is Busy (which would obviously be exclusive of Device End), the device should return a WACK. >If a device is busy when a host tries to select it, should >a WACK be used to respond or can the above mentioned >status sense message of "DEVICE BUSY" be used? As noted, if the ONLY status is Busy, the WACK is the proper response. (Page 4.21 in my version of the Desc, Manual....but these haven't changed since the older 3271...)