RE: write failed with errno 11

Terribly sorry - my bad. major oops on my side, totally stupid - happens to me all the time. I knocked my had on the walls a couple of times after I've seen this : hopefully that would cause the lesson to be learned this time around :-(

Here's the new patch against today's CVS :

--
Oded Arbel
m-Wise mobile solutions
[email protected]

+972-9-9581711 (116)
+972-67-340014

::..
"The best way for civilians to understand military culture is to read 'Starship 
Troopers', 'The Forever War', and 'Bill the Galactic Hero' all at one sitting." 




> -----Original Message-----
> From: Oded Arbel 
> Sent: Thursday, August 08, 2002 4:40 PM
> To: Stephane Pointu; Kannel-devel (E-mail)
> Subject: RE: write failed with errno 11
> 
> 
> 
> 
> 
> Yes - CMS error 500 is an unknown error, but I think I know 
> why it happened and it's my fault. I'm not sure yet why it 
> did what it did, but I'm going to fix it. 
> 
> --
> Oded Arbel
> m-Wise mobile solutions
> [email protected]
> 
> +972-9-9581711 (116)
> +972-67-340014
> 
> ::..
> "So mutable is really the prophecised anti-Const, yes?" 
> 	-- Doug Schmidt 
> 
> > -----Original Message-----
> > From: Stephane Pointu [mailto:[email protected]]
> > Sent: Thursday, August 08, 2002 4:32 PM
> > To: 'Paul Keogh'; Kannel-devel (E-mail)
> > Cc: Oded Arbel
> > Subject: RE: write failed with errno 11
> > 
> > 
> > I did a make distclean and recompiled... Now i have another 
> > error message:
> > 
> > 2002-08-08 16:22:29 [5] DEBUG: AT2[T68]: international 
> > starting with 00
> > (0033666666666)
> > 2002-08-08 16:22:29 [5] DEBUG: AT2[T68]: TP-Validity-Period: 
> > 24.0 hours
> > 2002-08-08 16:22:29 [5] DEBUG: AT2[T68]: --> AT+CMGS=21^M
> > 2002-08-08 16:22:29 [5] DEBUG: AT2[T68]: <-- >
> > 2002-08-08 16:22:29 [5] DEBUG: AT2[T68]: send command status: 1
> > 2002-08-08 16:22:29 [5] DEBUG: AT2[T68]: -->
> > 0000000B913366567902F60000A707D4F29CEE72B900
> > 2002-08-08 16:22:29 [5] DEBUG: AT2[T68]: --> ^Z
> > 2002-08-08 16:22:30 [5] DEBUG: AT2[T68]: <-- >
> > 2002-08-08 16:22:30 [5] DEBUG: AT2[T68]: <-- +CMS ERROR: 500
> > 2002-08-08 16:22:30 [5] ERROR: AT2[T68]: CMS ERROR: +CMS ERROR: 500
> > 2002-08-08 16:22:30 [5] DEBUG: AT2[T68]: send command status: 1
> > 
> > CMS ERROR 500 is an unknown error! Is it really unknown?
> > 
> > Stephane.
> > 
> > 
> > -----Original Message-----
> > From: Paul Keogh [mailto:[email protected]]
> > Sent: 08 August 2002 16:06
> > To: Kannel-devel (E-mail)
> > Subject: RE: write failed with errno 11
> > 
> > 
> > > > > 2002-08-08 15:25:32 [14] ERROR: Packet too short while 
> > > > unpacking Msg.
> > > > > 2002-08-08 15:25:32 [14] ERROR: Msg packet was invalid.
> > > > > 2002-08-08 15:25:32 [14] ERROR: Failed to unpack data!
> > > > > 
> > 
> > This is symptomatic of a change to the Msg struct without all the
> > affected modules being recompiled. I would guess that either the
> > wap or sms box is out of sync with the bearer box in this respect.
> > 
> 
>

--- gw/smsc_at2.c	2002-08-07 10:56:44.000000000 +0300
+++ gw/smsc_at2.c	2002-08-08 18:10:43.000000000 +0300
@@ -241,6 +241,7 @@
 {
     int count;
     int s = 0;
+    int write_count = 0;
     Octstr *linestr = NULL;
 
     linestr = octstr_format("%s\r", line);
@@ -248,7 +249,15 @@
     debug("bb.smsc.at2", 0, "AT2[%s]: --> %s^M", octstr_get_cstr(privdata->name), line);
 
     count = octstr_len(linestr);
-    s = write(privdata->fd, octstr_get_cstr(linestr), count);
+    while (1) {
+	errno = 0;
+	s = write(privdata->fd, octstr_get_cstr(linestr), count);
+	if (s < 0 && errno == EAGAIN && write_count < RETRY_SEND) {
+	    gwthread_sleep(1);
+	    ++write_count;
+	} else
+	    break;
+    };
     O_DESTROY(linestr);
     if (s < 0) {
         debug("bb.smsc.at2", 0, "AT2[%s]: write failed with errno %d", 
@@ -266,9 +275,18 @@
 {
     int s;
     char *ctrlz = "\032" ;
-
+    int write_count = 0;
+    
     debug("bb.smsc.at2", 0, "AT2[%s]: --> ^Z", octstr_get_cstr(privdata->name));
-    s = write(privdata->fd, ctrlz, 1);
+    while (1) {
+	errno = 0;
+	s = write(privdata->fd, ctrlz, 1);
+	if (s < 0 && errno == EAGAIN && write_count < RETRY_SEND) {
+	    gwthread_sleep(1);
+	    ++write_count;
+	} else
+	    break;
+    };
     if (s < 0) {
         debug("bb.smsc.at2", 0, "AT2[%s]: write failed with errno %d", 
               octstr_get_cstr(privdata->name), errno);
@@ -1517,14 +1612,18 @@
      */
 
     /* 
-     * message type SUBMIT
-     *    01010001 = 0x51 indicating add. UDH, TP-VP(Rel) & MSG_SUBMIT
-     * or 00010001 = 0x11 for just TP-VP(Rel) & MSG_SUBMIT 
+     * message type SUBMIT , bit mapped :
+     * bit7                            ..                                    bit0
+     * TP-RP , TP-UDHI, TP-SRR, TP-VPF(4), TP-VPF(3), TP-RD, TP-MTI(1), TP-MTI(0)
      */
-
-    pdu[pos] = octstr_len(msg->sms.udhdata) ? at2_numtext(5) : at2_numtext(1);
+    i  = (msg->sms.rpi ? 1 : 0) << 7 /* TP-RP */
+	| (octstr_len(msg->sms.udhdata)  ? 1 : 0) << 6 /* TP-UDHI */
+	| ((msg->sms.dlr_mask & (DLR_SUCCESS | DLR_FAIL | DLR_BUFFERED)) ? 1 : 0) << 5 /* TP-SRR */
+	| 16 /* TP-VP(Rel)*/
+	| 1; /* TP-MTI: SUBMIT_SM */
+    pdu[pos] = at2_numtext(i >> 4);
     pos++;
-    pdu[pos] = at2_numtext(AT_SUBMIT_SM);
+    pdu[pos] = at2_numtext(i & 15);
     pos++;
 
     /* message reference (0 for now) */