Re: [hylafax-users] modem (or faxgetty?) hangs "receiving from ..." STRACE AND LOG

[Lee Howard <faxguy@xxxxxxxxxxxxxxxx>]

On 2003.12.08 08:11 Aidan Van Dyk wrote:

> It did hang.  It hung for an hour, until he put the strace on it,
>
> 	Dec 08 15:02:31.30: [ 5284]: RECV: send trigger 022
> 	Dec 08 15:02:31.30: [ 5284]: <-- data [1]
> 	Dec 08 15:02:39.82: [ 5284]: RECV: 14569 bytes of data, 554
> total lines

Ah, I didn't notice that.  I was looking for the error to be within 
getModemLine().

Okay, so the problem isn't with ModemPageDoneTimeout nor with 
getModemLine().

We appear to be hung, then, within FaxModem::recvPageDLEData().  It 
also appears that copy quality checking is not being done (because 
flushRawData() is being called - "x bytes of data, y total lines").

In that function there is a for loop that will loop forever until an 
RTCraised is true (which triggers a siglongjmp back out of the loop and 
will subseqently not enter the loop and will move past it).  I don't 
think that we're getting stuck within that loop because flushRawData() 
is being called without the total received number of bytes of data 
being larger than RCVBUFSIZ.

Since we seem to be stuck on a syscall and not within a loop (due to 
strace "kicking" the process back into action), I'd venture to guess 
that we're hung within recvEndPage(), specifically:

    TIFFSetField(tif, TIFFTAG_FAXRECVTIME,
        (uint32) server.getPageTransferTime());

Which (in faxd/FaxServer.c++) makes a call to Sys::now().  Notice the 
strace...

16:01:56.727457 rt_sigaction(SIGALRM, {0x400a0b10, [],
SA_RESTORER|SA_INTERRUPT, 0x4029a5c8}, NULL, 8) = 0
16:01:56.727711 setitimer(ITIMER_REAL, {it_interval={0, 0},
it_value={180, 0}}, NULL) = 0
16:01:56.727809 gettimeofday({1070895716, 727838}, NULL) = 0
16:01:56.727896 write(9, "Dec 08 16:01:56.72: [ 5284]: awa"..., 92) = 92
16:01:56.728001 gettimeofday({1070895716, 728030}, NULL) = 0
16:01:56.728080 write(9, "Dec 08 16:01:56.72: [ 5284]: no "..., 63) = 63
16:01:56.728160 setitimer(ITIMER_REAL, {it_interval={0, 0}, it_value={0,
0}}, NULL) = 0

calls on the real-time clock.

What would make the process hang when calling on Sys::now()?

Patrick, what are those log messages that occurred at Dec 08 
16:01:56.72 beginning "awa..." and "no ..." ?

Lee.

____________________ 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@xxxxxxxxxxxx*