summaryrefslogtreecommitdiff
path: root/geography/gpsbabel/patches/patch-af
blob: 16e10cd9294a6da092aac5cd930dd5af34e9b6b0 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
$NetBSD: patch-af,v 1.4 2013/03/16 18:52:58 gdt Exp $

From Giles Lean, who places it in the public domain.  Workarounds for
problems with buggy USB serial adapators.  Not applied by upstream
2009-02-21 because it's too dangerous to work around problems without
understanding them.

--- jeeps/gpsread.c.orig	2011-07-26 01:56:33.000000000 +0000
+++ jeeps/gpsread.c
@@ -98,6 +98,36 @@ int32 GPS_Serial_Packet_Read(gpsdevh* fd
       GPS_Diag("%02x ", u);
 
       if (!len) {
+                /*
+                 * Missed DLE characters have been observed with Geko
+                 * 201 and Legend GPSRs with Prolific USB-serial
+                 * cables.  The following kludge seems to help.
+                 *
+                 * It has been tested so far with the following
+                 * combinations of software and hardware:
+                 *
+                 * OS X 10.4.x and 10.5.x:
+                 *
+                 * - a Geko 201 (firmware version 2.70) and a third
+                 *   party cable using a Prolific USB-serial converter
+                 * - a Legend using a serial cable with a non-integral
+                 *   Prolific USB-serial converter.
+                 *
+                 * NetBSD-4.0/i386 with the Geko and cable as above.
+                 *
+                 * REVISIT GFL Should this be a switch in a .ini file?
+                 *
+                 * + it's a kludge, so leaving it always on is ugly
+                 * - if it's harmless to properly working hardware,
+                 *   then it's a better user experience to leave it on
+                 */
+                if (u == 0x06 || u == 0x15)
+                {
+                    ++len;
+		    (void) fprintf(stderr,"GPS_Packet_Read: inserted DLE due to 0x%02x.\n", u);
+                    goto dle_missed;
+                }
+
         if (u != DLE) {
           (void) fprintf(stderr,"GPS_Packet_Read: No DLE.  Data received, but probably not a garmin packet.\n");
           (void) fflush(stderr);
@@ -107,6 +137,7 @@ int32 GPS_Serial_Packet_Read(gpsdevh* fd
         continue;
       }
 
+dle_missed:
       if (len==1) {
         (*packet)->type = u;
         ++len;
@@ -130,6 +161,20 @@ int32 GPS_Serial_Packet_Read(gpsdevh* fd
       if (u == ETX)
         if (isDLE) {
           if (p-(*packet)->data-2 != (*packet)->n) {
+	    /*
+	     * When used with a buggy Prolific USB-serial converter the
+	     * calling sequence GPS_A000() -> GPS_Get_Ack() sometimes
+	     * returns data != type causing this routine to fail and the
+	     * following error message to be emitted:
+	     *
+	     *   GARMIN:Can't init /dev/cu.usbserial
+	     *
+	     * Manually retrying usually works, and subsequent
+	     * investigation shows that the call that fails is made from
+	     * GPS_Init() which in this case is GPS_Serial_Init().  Simply
+	     * retrying the call a workaround: see note and retry loop in
+	     * gpsapp.c:GPS_Init().
+	     */
             GPS_Error("GPS_Packet_Read: Bad count");
             gps_errno = FRAMING_ERROR;
             return 0;