[LinuxPPS] kernel does not see PPS on serial port

[Rodolfo Giometti]

On Fri, Mar 07, 2008 at 09:29:53PM +1100, Cirilo Bernardo wrote:
> On Fri, Mar 7, 2008 at 7:20 PM, LICHTENBERGER Janos <lityi at sas.elte.hu> wrote:
> [snip]
> >  > If you look into /proc/interrupts do you see that the interrupts
> >  > counter of the serial line is increasing?
> >  >
> >  No, not at all. I booted with the old 2.6.15.7 kernel and in this case the
> >  interrupt counter steadily increased. May it mean that something is wrong
> >  with serial port? The port itself is working, I tried with an app reading
> >  and writing data through the port (it does not use pin 1 howvever).
> >
> [snip]
> 
> Do you mean you booted your new computer with the older kernel to
> check that PPS works and it does work?
> 
> Rodolfo suggested the "cat /dev/ttyS0" because the serial port drivers
> are designed to disable the UART interrupts unless someone opens the
> device for communication.  This makes sense for the usual case -
> unless you are expecting data, why even watch the serial port.  The
> result of course is that you will not get the DCD interrupt unless you
> open the serial port - even if you never actually use the serial port
> data.
> 
> I'm asking if you used the older kernel on the same hardware because
> some hardware only implements the Rx/Tx lines and not the DCD and
> other lines, so if you tested on two different machines then that is
> not a valid test.
> 
> Another problem I encounter is that the 8250 driver in particular
> tries to work around faults in numerous UART hardware which claims to
> be compatible with the old 8250.  Sometimes hardware is not identified
> correctly and the wrong driver is loaded.  In such a case, you can see
> what driver is associated with your serial port in your old kernel and
> in your new kernel.
> 
> Yet another problem I encounter is with power management features on
> some boards and in some BIOSes.  Recently I found that with one board
> I have, I need to check the BIOS settings and turn off the "power
> management features active on boot" feature.  If it is active, then
> some peripherals are shut down on boot and Linux does not realize this
> or have any obvious mechanism for turning these peripherals on.  Well,
> maybe it is possible to turn on via the ACPI or APM interface, but for
> my purpose I don't even see the point in disabling these peripherals.

Good analysis! Can you please turn this letter into the first draft of
LinuxPPS-howto? :)

If you have no time to do it, can I add this letter the the wiki FAQs?

Ciao,

Rodolfo

-- 

GNU/Linux Solutions                  e-mail:    giometti at enneenne.com
Linux Device Driver                             giometti at gnudd.com
Embedded Systems                     		giometti at linux.it
UNIX programming                     phone:     +39 349 2432127