Re: [hylafax-users] 2.6 kernel and Multiport boards (MainPine)

["andrew rinaldi" <andrew.rinaldi@xxxxxxxxxxxx>]

Hi Robert.

I don't have the authority to give you an 'official statement' but I can
give you some background information.

The Agere/Lucent Venus chipset has a bug which can result in a interrupt
register read failure.  This condition is extremely rare but effects boards
from Multitech, Perle, Zoom and other Venus based products that we have
tested.  As a company we focus on providing reliable fax solutions and
therefore implemented a workaround in our hardware for the interrupt
register read bug - this is the multiport feature.  Kernels prior to 2.6 had
native support for this feature.  However, major changes in the Linux serial
driver removed this from 2.6 and it can therefore no longer be used.  The
good news is that these changes also modified the behaviour of the interrupt
service routine and we see the bug very rarely today.  Bottom line, you can
ignore this feature if you are running 2.6 kernels.

I will get the website updated with this information to save confusion in
the future.... The only excuse that I can offer is that we have a new
website in development and they are focused on that more than updating the
current one.

You may also like to know that we are about to extend our sideband features
to Linux.  This will allow us to automatically recover when the modems
'Wedge' following a Venus Firmware/DSP crash (unfortunately, despite many
years of effort on our part we are still unable to cure all of the Venus
bugs).

We have thousands of customers running our products in a production
environment and I have no doubt that we have the most reliable fax hardware
in the industry.  We regularly test fax server software which is unique in
this industry (http://www.mainpine.com/corporate/press/2006/isvlab.html).
We also see a lot of customers dump Brooktrout and Eicon hardware to move to
our solutions because of the better reliability and price point
(http://www.mainpine.com/downloads/WP-Lobar.pdf).

Please let me know if I can offer you any further clarification.

Regards

ANDREW RINALDI
Mainpine Ltd. Support
USA +1 718 701 2422 | Asia/Europe +44 1225 436137 
andrew.rinaldi@xxxxxxxxxxxx | www.mainpine.com

-----Original Message-----
From: hylafax-users-bounce@xxxxxxxxxxx
[mailto:hylafax-users-bounce@xxxxxxxxxxx] On Behalf Of Robert
Sent: 10 May 2006 02:53
To: hylafax-users@xxxxxxxxxxx
Subject: [hylafax-users] 2.6 kernel and Multiport boards (MainPine)

I've been messing around for the past several hours trying to get the
RockForce Octo+ from Mainpine working on a 2.6 kernel.

All is working well except, the multiport command comes back invalid.

After 'Several' recompiles of my kernel I decided to search the group.

I found a post (quoted below) by Darren Nickerson of iFax, stating basically
in 2.6 multiport support is different, you no longer need the mutliport
statement...
it should just work. He also mentioned MainPain would make a official
statement (Which I do not see on their site) is regards to a recommended
installation procedure.

What I would like to know is:

Is anyone running this setup in a production environment?
If so what are your experiences, is it stable?


Thanks,
Robert Silvia
DEXIS, LLC
www.dexray.com



-- Quote Mentioned above --


"Lee Howard" <fax... 
<http://groups.google.com/groups/unlock?msg=dfdc066ed46d2b8f&_done=/group/fa
.hylafax/browse_thread/thread/a6707a5a079e3009/dfdc066ed46d2b8f%3Fq%3Dmultip
ort%26rnum%3D1>@howardsilvan.com>: 


 > Chris Giesel wrote:

 >>All:

 >>I am having difficulty installing my mainpine rockforce octo+ and
>>configuring Fedora kernel 2.6.5-1.358.i386 serial ports/ttyS latches. I
>>have  >>the following information:

 >>We have recompiled the fedora core 2 kernel with the options for
*multiport*
 >>serial devices enabled as suggested earlier.

 >>We ran the setserial command:
 >>setserial /dev/ttyS14 port 0xdc80 irq 9 autoconfig ^fourport substituting
>>apropriate io addresses and ttyS where apropriate

 >>we are still getting "Cannot get *multiport* config: Invalid argument"
after  >>running  setserial /dev/ttyS14 set_multiport port1 0xdce0 mask1
0xFF
 >>match1
 >>00x00

 > You may have to recompile your kernel with the *multiport* options
enabled.

Chris/Lee,

In kernel-2.4, the *multiport* 'latch' is required. Most major distros
default kernels are compiled with *multiport* support, and so the
set_multiport line would execute properly (at least when there's no ioport
interrupts) and a kernel recompile would not be required.

This has completely changed in kernel-2.6 ... they've reworked the
*multiport*
serial support quite a bit.

We've been working with Mainpine to help them develop a recommended
configuration procedure for kernel-2.6, and we each have 2.6-based fedora's
in our labs in testing. So far, it seems like you can run all the setserial
commands _except_ for the *multiport* latch (which is not relevant to
kernel-2.6 and as far as we can tell will never work) and it should 'just
work'.

Look for an update on Mainpine's website once they've finalized their
official recommendations.

-Darren

--
Darren Nickerson
Senior Sales & Support Engineer
iFAX Solutions, Inc. www.ifax.com <http://www.ifax.com> darren.nicker... 
<http://groups.google.com/groups/unlock?msg=dfdc066ed46d2b8f&_done=/group/fa
.hylafax/browse_thread/thread/a6707a5a079e3009/dfdc066ed46d2b8f%3Fq%3Dmultip
ort%26rnum%3D1>@ifax.com 

+1.215.438.4638 x8106
+1.215.243.8335 (fax)

____________________ HylaFAX(tm) Users Mailing List _______________________
  To subscribe/unsubscribe, click http://lists.hylafax.org/cgi-bin/lsg2.cgi
 On UNIX: mail -s unsubscribe hylafax-users-request@xxxxxxxxxxx < /dev/null
  *To learn about commercial HylaFAX(tm) support, mail sales@xxxxxxxxx*




____________________ HylaFAX(tm) Users Mailing List _______________________
  To subscribe/unsubscribe, click http://lists.hylafax.org/cgi-bin/lsg2.cgi
 On UNIX: mail -s unsubscribe hylafax-users-request@xxxxxxxxxxx < /dev/null
  *To learn about commercial HylaFAX(tm) support, mail sales@xxxxxxxxx*