chuck@felix.UUCP (12/03/87)
Reply-Path: I got the note from Curt on the DCA 207 / DMF emulator problems, and moore's response. Just to give a brief synopsis of what was actually found, the jammed processes had the TS_STOP and TS_FLUSH bits set in the tty structure's state word. In the Ultrix 1.2 DMF32 driver, this occurs only in the ttystop() routine, if characters are currently being output but XOFF has not been seen from the terminal. (There may be one more condition: I don't have the code in front of me.) These are cleared iff a transmit interrupt occurs for that line, causing a subsequent ttstart(). This latter appears not to be happening. WHY it is not happening can only be determined from better documentation than DCA has seen fit to issue, or from DCA themselves, or from extensive hardware hours. As Curt said, there was negligible difference (in this place) with the BSD4.2 DMF32 driver. I have not yet looked at the BSD4.3 or Ux2.0 DMF drivers. Joe Yao