gdmr@cstvax.UUCP (George D M Ross) (07/23/86)
In article <131@pollux.UUCP> pedz@pollux.UUCP writes: > >I am having problems with my dmf driver. This is on a Vax 11/750 >running 4.2bsd unix and concerns lines connected to out going modems. >..... We are seeing much the same problem with our 750 attached to our Gandalf PACX 2000, viz lines stopping responding to the modem signals. Any pointers to a solution would be most appreciated. We've tried using Chris Maloney's (MDDC) dmf driver, but that just wedged up the processor whenever we tried talking to the board (it's an Emulex one, by the way); if anyone has a version which works for our configuration (750, SI9900, a DEUNA, a couple of DZ11s, an Emulex board looking like a couple of DMF32s, and a mag-tape (TS11 look-alike?)) we'd be glad to hear from you. -- George D M Ross, Dept. of Computer Science, Univ. of Edinburgh, Scotland Phone: +44 31-667 1081 x2730 JANET: gdmr@UK.AC.ed.cstvax --> ARPA: gdmr@cstvax.ed.AC.UK UUCP: <UK>!ukc!cstvax!gdmr
muller@sdcc7.ucsd.EDU (Keith Muller) (07/26/86)
Older cs21 emulex versions of the dmf had a microcode bug that will cause it to wedge the system by holding onto the unibus forever while doing dma output (which many drivers do). See your local emulex person about getting the proms upgraded. I have mods to the 4.3 (june 86) dmf driver that works on both cs11 and cs21 dmfs. It could be moved to 4.2 with a little work. Some of the dmf drivers accidently turn off modem signal interrupts in the section of code that handles a loss of carrier (in dmfrint()) That would cause the line to never respond to carrier after the first connection completed. Keith Muller muller@sdcsvax.ucsd.edu muller@nprdc.ARPA