[lug] ppp dial up server with RH7.2

[Hobbs: Nobuki Matsui]

I am sorry.  I forgot to mention that those are external modems connected
through serial ports.  Then I wonder where the mismatches happen?

N MATSUI
CIRES NOAA SRRB

On Wed, 30 Jan 2002, D. Stimits wrote:

> Hobbs: Nobuki Matsui wrote:
> > 
> > Thank you for your suggestion; I did not doubt the serial port.
> > These are my current settings below...
> > 
> > localhost:nobuki {109} sudo setserial -a /dev/modem
> > /dev/modem, Line 0, UART: 16550A, Port: 0x03f8, IRQ: 4
> >         Baud_base: 115200, close_delay: 50, divisor: 0
> >         closing_wait: 3000
> >         Flags: spd_normal skip_test
> 
> This looks right, but the trick is that the system view of this serial
> port must match in other places. Now take a close look at /proc/pci,
> find the modem, compare parameters that are visible/obvious. Also
> consider lspci -vvv, look for modem settings as seen by the pci bus (I
> am assuming this modem is PCI; if not, perhaps ISA jumpers apply). Also,
> if this is PCI, be sure your bios is set to the equivalent of "O/S is
> NOT plug-n-play aware". This will cause the bios to do initial setup of
> PCI hardware (but linux can still change it). PCI setup will be
> consistent across reboots, unless you move cards in their PCI slots.
> 
> D. Stimits, stimits at idcomm.com
> 
> > 
> > N Matsui
> > CIRES NOAA SRRB
> > 
> > On Tue, 29 Jan 2002, D. Stimits wrote:
> > 
> > > Hobbs: Nobuki Matsui wrote:
> > > >
> > > > Dear all,
> > > >         I am trying to set up a ppp dial-up server with RH7.2.  I had a
> > > > success with RH6.1 before, however it does not work with RH7.2 with the
> > > > same settings.  I searched the webs for a solution and I started wondering
> > > > if it is a 2.4 kernel issue...  There are some bugs as to ppp in the 2.4
> > > > kernel as far as I found, using google.  All the garbage in the log file
> > > > and LCP Timeout suggests the computers are having a hard time
> > > > negotiating.  Apending 'EXPECT ~' line to the chat.scr would give me a
> > > > normal login.
> > > >         I wonder if anybody has any suggestions.
> > >
> > > The last couple of times I had LCP timeout, I had to run setserial to
> > > make sure the serial port and the rest of the system had the same idea
> > > of what was there. In particular, I used almost the same as the default
> > > parameters, but added the spd_normal and skip_test, plus specified
> > > specifically the 16550A uart. Note that "setserial -a
> > > /dev/whatver_the_modem_is" will give you current settings.
> > >
> > > D. Stimits, stimits at idcomm.com
> >