FTDI Com ports "just stop"

[Karl Pielorz]

--On 31 January 2011 08:36 +0100 Hans Petter Selasky <hselasky at c2i.net> 
wrote:

> On Sunday 30 January 2011 19:57:07 Karl Pielorz wrote:
>> Any hints how I can try and debug this? I think it might be related to
>> the  2nd set of 4 FTDI ports, as I didn't notice this issue for the few
>> weeks  the system ran with just 4 ports connected.
>
> Try connecting a null-modem cable and see if you can loop any data.

Nope, no data is sent or received. I also connected an RS232 breakout 
connector (gee, remember those?!) - and no lights flash either... Looking 
at the output of 'ps' shows processes using the ports at the time either 
hang in 'ttyin' or 'ttyout' (I guess depending on whether they were 
sending, or receiving data).

> Also you can try a "usbconfig -d X.Y reset"

I'll give that a go, thanks!

I've also noticed - if you boot the machine and leave it (it has 4-5 
programs running using the serial ports) - it seems to be fine. However if 
you start opening, and closing ports (e.g. "tip modem", run some commands - 
quit tip, re-run it) - or stop/start the other software (causing the serial 
port to close, then open) - that seems to pretty reliably trigger the 
problem.

When the problem has happened - anything that 'touches' the ports seems to 
just hang (e.g. if you run 'tip modem' when it's in this state - you get a 
'connected.' line, but no data is sent/received - tip then gets very 
sluggish to quit, you can quit/kill it but it takes around 20 seconds).

Some of the other software just blocks immediately it tries to read from 
the port, other software must use a background timeout as 'tries' to use 
the port, but then starts issuing timeouts itself.

All the ports exist in /dev at the time, and having double checked - 
there's nothing logged from the kernel anywhere when it happens :(

The rest of the machine seems fine as well when this happens (i.e. it 
doesn't affect anything else, the box is a simple NAT router in addition to 
looking after the serial ports). Next time it happens I'll try the reset 
about - and double check that the single FT232R on the system does still 
work (it seemed to be last time).

-Karl