Re: HELP: can send but cannot receive (retrain negative)

[Kadlecsik Jozsi <kadlec@sunserv.kfki.hu>]

Hello,

I wrote:
> Here is the configuration:
> 
> HylaFAX v4.0pl1 on Linux (Slackware 3.1.0, kernel 2.0.27) compiled 
> by gcc 2.7.2 and a Smart One 2834 internal faxmodem (V1.510-V34_DP).
> 
> There is no problem with sending faxes, but we were so far *unable* to receive
> any fax. Even when quality checking is disabled, HylaFAX sends back
> retrain negative messages to the sender.

Matthias Apitz suggested to set the following
ModemRecvFillOrder: MSB2LSB

I modified another line in config.cua1:
Class2RecvDataTrigger:  "\022"

Now we can receive facsimile from machine B (another faxmodem), 
but not from machine A (a real faxmachine), from which we originally
started to test the configuration and have never been able to receive
facsimile.

The setting of ModemRecvFillOrder has no effect on the connection with
machine A.

Here follows excerpts from the logs:

Machine B, facsimile receives successfully:
....
Mar 04 14:20:32.22: [ 9962]: --> [22:+FDCS: 1,3,0,2,0,0,0,0]
....
Mar 04 14:20:47.54: [ 9962]: RECV: begin page
Mar 04 14:20:47.54: [ 9962]: RECV: send trigger 022
Mar 04 14:21:14.86: [ 9962]: RECV: 32767 bytes of data, 1232 total lines
Mar 04 14:21:38.70: [ 9962]: RECV/CQ: Bad 1D pixel count, row 2423, got 0, expected 1728
Mar 04 14:21:38.70: [ 9962]: RECV/CQ: Bad 1D pixel count, row 2424, got 0, expected 1728
Mar 04 14:21:38.70: [ 9962]: RECV/CQ: Bad 1D pixel count, row 2425, got 0, expected 1728
Mar 04 14:21:38.70: [ 9962]: RECV/CQ: Bad 1D pixel count, row 2426, got 0, expected 1728
Mar 04 14:21:38.70: [ 9962]: RECV/CQ: Bad 1D pixel count, row 2427, got 0, expected 1728
Mar 04 14:21:38.70: [ 9962]: RECV: 28612 bytes of data, 2428 total lines
Mar 04 14:21:38.70: [ 9962]: --> [17:+FPTS: 1,2219,0,0]
Mar 04 14:21:40.30: [ 9962]: --> [7:+FET: 2]
Mar 04 14:21:40.30: [ 9962]: RECV recv EOP (no more pages or documents)
Mar 04 14:21:40.30: [ 9962]: --> [2:OK]
Mar 04 14:21:40.30: [ 9962]: RECV send MCF (message confirmation)

Machine A, facsimile not received:
...
Mar 04 14:31:18.87: [ 9962]: --> [22:+FDCS: 0,0,0,2,1,0,0,0]
...
Mar 04 14:31:18.87: [ 9962]: RECV: begin page
Mar 04 14:31:18.87: [ 9962]: RECV: send trigger 022
Mar 04 14:33:08.34: [ 9962]: RECV: 32730 bytes of data, 926 total lines
Mar 04 14:33:25.62: [ 9962]: RECV/CQ: Bad 1D pixel count, row 1149, got 0, expected 1728
Mar 04 14:33:25.62: [ 9962]: RECV/CQ: Bad 1D pixel count, row 1150, got 0, expected 1728
Mar 04 14:33:25.62: [ 9962]: RECV/CQ: Bad 1D pixel count, row 1151, got 0, expected 1728
Mar 04 14:33:25.62: [ 9962]: RECV/CQ: Bad 1D pixel count, row 1152, got 0, expected 1728
Mar 04 14:33:25.62: [ 9962]: RECV/CQ: Bad 1D pixel count, row 1153, got 0, expected 1728
Mar 04 14:33:25.77: [ 9962]: RECV: 5305 bytes of data, 1154 total lines
Mar 04 14:33:25.77: [ 9962]: RECV/CQ: Adjusting for RTC found at row 1148
Mar 04 14:33:25.77: [ 9962]: --> [17:+FPTS: 2,2219,0,0]
Mar 04 14:33:27.20: [ 9962]: --> [7:+FET: 2]
Mar 04 14:33:27.20: [ 9962]: RECV recv EOP (no more pages or documents)
Mar 04 14:33:27.20: [ 9962]: --> [2:OK]
Mar 04 14:33:27.20: [ 9962]: RECV send RTN (retrain negative)

Why on the earth sends the modem retrain negative messages back in the
case A???

The only difference between the two sessions (as far as I can see) is:

Mar 04 14:31:18.87: [ 9962]: --> [22:+FDCS: 0,0,0,2,1,0,0,0]
                                                    ^
i.e machine A wants data compression. This Smart One modem supports both
V.42bis and MNP5 (AT\N3%C3). I added the AT command as ModemResetCmds
to the config file - no effect whatsoever in the connection.

What else could we check to find out the reason of the RTN messages?

How could we fix the configuration??

Thank you any suggestion,
Jozsef
--
E-mail  : kadlec@sunserv.kfki.hu, kadlec@blackhole.kfki.hu
WWW-Home: http://www.kfki.hu/~kadlec
Address : KFKI Research Institute for Particle and Nuclear Physics
          P.O.B 49 Budapest, 1525 Hungary