home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!usc!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!eff!snorkelwacker.mit.edu!bloom-beacon!eru.mt.luth.se!lunic!sunic!news.funet.fi!fuug!ksvltd!jarmo
- From: jarmo@ksvltd.fi (Jarmo Raiha)
- Newsgroups: comp.dcom.modems
- Subject: Re: 16550A UART upgrade.
- Message-ID: <1992Aug25.112828.437@ksvltd.fi>
- Date: 25 Aug 92 11:28:28 GMT
- References: <Bt39ID.1z3@undergrad.math.waterloo.edu> <ZT2yPB1w165w@zswamp <1992Aug24.191332.12388@mnemosyne.cs.du.edu>
- Organization: KSV Instruments
- Lines: 23
-
- aduell@nyx.cs.du.edu (Tony Duell) writes:
-
- >Surely internal modems (I don't have one) miss out the last 2 uarts, and
- >directly transfer data from the modem CPU to the PC. In which case, how do
- >you use a 16550 with one? I can understand the Modem CPU doing buffering,
- >and it could be a lot more intellegent than a 16550.
-
- First of all, your modem software expects to see something that
- behaves like 8250. In many internal modems there really is a true 8250 chip,
- (or 16550) connected to PC bus just like an ordinary serial adapter.
-
- If that is the case, the modem's internal CPU can't do buffering
- because it can't see whether the 8250's buffer is full or empty.
- It might do the buffering if you could control RTS line fast enough,
- (Because modem cpu can read RTS line status)
- but without 16550 ,the interrupt latency problem still remains.
-
- *IF* the modem uses some proprietary chip that simulates 8250
- then it might implement some buffering.
-
- Maybe some modem manufacturer could tell something about their implementation.
-
- Jarmo Raiha
-