>>>>> "Chris" == Chris G Demetriou <Chris_G_Demetriou@UX2.SP.CS.CMU.EDU> writes:

[my pleas for reaonable com support deleted]

Chris> While i'd like to say, "I hope that problem is fixed before the
Chris> driver is committed," it's much more serious than that...  If
Chris> the problem isn't fixed before it's committed, then a whole
Chris> class of users will more or less lose the use of their serial
Chris> ports.  And, to go back to the original point, about it being
Chris> integrated now, for 1.2: changes which can seriously break
Chris> system operation shouldn't be committed after feature freezes.

	While I can respect your point, it's not like this has been a
sudden surprise to all of you.  From the looks of the volume of mail
this and similar threads have generated, a large number of people have
'lost' their serial ports since the 1.0 --> 1.1 upgrade.  By your
comments, this will be the second NetBSD release where slower x86
machines will have largely useless serial ports.  In fact, it seem to
have gotten worse from 1.1 to current.

	From my tests on my 386/40, using a 28.8 modem:

	Current-ppp	Current-uucico	Patched-ppp	Patched-uucico
9600	works		works		works		works
19200	works		almost		works		works
38400	almost		broken		works		works
57600	broken		broken		works		not tested

	Here, ppp performance seems to be different than that of the
uucico.  Uucico seems similar to all other user-land programs (like
seyon, tip and cu) in performance.

Dave.

-- 
----------------------------------------------------------------------------
|David Gilbert, PCI, Richmond Hill, Ontario.  | Two things can only be     |
|Mail:      dgilbert@jaywon.pci.on.ca         |  equal if and only if they |
|http://www.pci.on.ca/~dgilbert               |   are precisely opposite.  |
---------------------------------------------------------GLO----------------