3 libpcap 0.7.1 and later appear to work on AIX when using AIX's native
4 BPF; that appears to work better than DLPI does. Note that you may have
5 to run AIX's tcpdump, as root, before configuring, building, and
6 installing libpcap, in order to create the "/dev/bpf" devices and load
9 However, libpcap 0.7.1 doesn't work perfectly with AIX's BPF - it
10 appears that AIX's BPF devices inform their user that packets were
11 dropped since the last successful read by returning -1 and setting
12 "errno" to EFAULT, which libpcap 0.7.1 treats as an error. The current
13 CVS version of libpcap ignores EFAULT on AIX; it appears that this fixes
16 Some earlier notes: the notes about libpcap may not apply, with libpcap
17 0.7.1, but they're preserved here for historical reasons.
19 The glib, gtk+, and Ethereal notes still apply.
21 After much work and toil, Craig Rodrigues was able to compile libpcap
22 and Ethereal on AIX 4.3.2. His odyssey is document in various e-mails
23 at http://www.ethereal.com/lists/ethereal-dev/199911/
25 Here are a few excerpts. Note that, to configure "libpcap" to use DLPI
26 rather than BPF (which it'll apparently use by default on AIX),
31 to the "configure" script for "libpcap" should do the trick.
33 The source code changes to Ethereal mentioned below should be in the
34 current source tree. The changes to the GLib configure script is in
35 GLib 1.2.7; the changes for the "-lgdk" problem are probably still
36 necessary in the current version of GTK+.
38 Subject: Re: [ethereal-dev] Re: [ethereal-users] Problems compiling 0.7.7 under AIX 4.3.2
39 From: Gilbert Ramirez <gram@xiexie.org>
40 Date: Fri, 5 Nov 1999 16:58:17 -0600
41 To: Guy Harris <guy@netapp.com>
42 Cc: Craig Rodrigues <rodrigc@mediaone.net>, ethereal-dev@zing.org
45 On Fri, Nov 05, 1999 at 01:42:44PM -0600, Guy Harris wrote:
50 > Looks suspiciously similar to the previous error; have you tried
51 > recompiling GTK+ with "xlc_r"?
53 I believe glib and gtk+ should both be compiled with xlc_r. I haven't
54 compiled on AIX in a long time, but I think it's because glib is including
55 pthread stuff, so the re-entrant C library, libc_r, is needed.
60 When compiling a multi-threaded program, you should invoke the C compiler
61 using one of the following commands:
64 Invokes the compiler with default language level of ansi.
66 Invokes the compiler with default language level of extended.
69 These commands ensure that the adequate options and libraries are used to be
70 compliant with the X/Open Version 5 Standard. The POSIX Threads
71 Specification 1003.1c is a subset of the X/Open Specification.
73 The following libraries are automatically linked with your program when using these commands:
81 For example, the following command compiles the foo.c multi-threaded C source file and produces the foo executable file:
85 See the cc command for more information about C For AIX.
91 To: ethereal-users@zing.org
92 Subject: [ethereal-dev] AIX: gtk problem solved, now an ethereal problem
93 From: Craig Rodrigues <rodrigc@mediaone.net>
94 Date: Mon, 8 Nov 1999 10:46:25 -0500
95 Cc: ethereal-dev@zing.org
100 After much sweat and toil, I have managed to get gtk 1.2.6 to
101 compile and not dump core under AIX. The solutions were to
102 (1) apply the attached patch to the configure.in in the glib-1.2.6
105 (2) In the file gtk+-1.2.6/gtk/Makefile, add a link flag -lgdk to link
108 I have submitted (1) to the gtk-devel mailing list where it has been
109 accepted. (2) is an uglier problem, but for now, adding -lgdk by hand
112 Now I have a problem....I compiled gtk, and that works.
113 I compiled ethereal (after some minor mods), and it starts,
114 but when I click on Capture -> Start, I get:
116 "There are no network interfaces that can be opened."
118 I am running as root, so I don't think permissions are a problem.
125 http://www.gis.net/~craigr
128 *** configure.in.old Thu Oct 7 17:27:43 1999
129 --- configure.in Sun Nov 7 19:34:36 1999
133 if test "$ac_cv_func_getpwuid_r" = "yes"; then
134 AC_MSG_CHECKING(whether getpwuid_r is posix like)
135 ! # getpwuid_r(0, NULL, NULL, 0) is the signature on
136 ! # solaris, if that is not found, the prog below won't
137 ! # compile, then the posix signature is assumed as
139 ! AC_TRY_COMPILE([#include <pwd.h>],
140 ! [getpwuid_r(0, NULL, NULL, 0);],
141 ! [AC_MSG_RESULT(no)],
142 ! [AC_MSG_RESULT(yes)
143 ! AC_DEFINE(HAVE_GETPWUID_R_POSIX)])
146 if test x"$have_threads" = xposix; then
149 if test "$ac_cv_func_getpwuid_r" = "yes"; then
150 AC_MSG_CHECKING(whether getpwuid_r is posix like)
151 ! # The signature for the POSIX version is:
152 ! # int getpwuid_r(uid_t, struct passwd *, char *, size_t, struct passwd **)
153 ! AC_TRY_COMPILE([#include <pwd.h>
154 ! #include <sys/types.h>
155 ! #include <stdlib.h>],
156 ! [getpwuid_r((uid_t)0, NULL, NULL, (size_t)0, NULL);],
157 ! [AC_DEFINE(HAVE_GETPWUID_R_POSIX)
158 ! AC_MSG_RESULT(yes)],
159 ! [AC_MSG_RESULT(no)])
162 if test x"$have_threads" = xposix; then
166 To: ethereal-dev@zing.org
167 Subject: Re: [ethereal-dev] AIX: gtk problem solved, now an ethereal problem
168 From: Craig Rodrigues <rodrigc@mediaone.net>
169 Date: Wed, 10 Nov 1999 12:18:47 -0500
175 OK, I'm getting closer and closer to this working on AIX.
179 (1) In a bunch of places in the code I removed '//' style C++ comments
180 which the IBM C compiler didn't like.
182 (2) I also found some places in the code like:
184 enum some_enum { FOO, BAR, };
186 IBM C did not like the trailing "," after BAR.
188 (3) In packet-ipv6.h, IPV6_VERSION is defined, but that is already
189 defined in <netinet/in.h> on AIX 4.3, so for now I just commented that out.
191 (4) in packet-afs.c, when it sucks in <netinet/in.h>, in.h sucks in
192 <sys/machine.h> which defines LITTLE_ENDIAN. This conflicts with
193 LITTLE_ENDIAN in globals.h. So what I did was, in globals.h, I added:
195 #ifdef HAVE_NETINET_IN_H
196 #include <netinet/in.h>
199 So after doing all these things, I can compile ethereal and run it.
201 correct network interfaces on my system: lo0 and en0. However,
202 when I start capturing packets on en0, they are all of the protocol type
203 "TRMAC" and "TR". The only problem is, I'm not on a Token Ring network.
207 No. Time Source Destination Protocol Info
208 1 0.000000 0a:30:a1:08:00:45 06:74:60:08:00:5a TR Token-Ring Unknown
209 2 0.210304 0a:30:a1:08:00:45 06:74:60:08:00:5a TR Token-Ring Unknown
210 3 0.926080 0a:30:a1:08:00:45 06:74:60:08:00:5a TR Token-Ring Unknown
211 4 0.4236416 0a:30:a1:08:00:45 06:74:60:08:00:5a TR Token-Ring Unknown
212 5 0.4712064 6f:06:74:60:08:00 5a:8a:30:a1:00:00 TR MAC Unknown Major Vector: 127
215 ---------------------
216 It turns out that libpcap was using IFT_* numbers instead of DLT_* numbers for
217 link types. That has been fixed
218 ---------------------
221 To: tcpdump-workers@tcpdump.org
222 Subject: [ethereal-dev] Sucess with libpcap under AIX
223 From: Craig Rodrigues <rodrigc@mediaone.net>
224 Date: Sat, 20 Nov 1999 03:34:50 -0500
225 Cc: ethereal-dev@zing.org
230 I have managed to successfully compile and use the latest
231 snapshot of libpcap under AIX using DLPI. bpf is majorly
232 brain-dead under AIX, and very unsupported. Rather than
233 find all the bugs in AIX's bpf, I decided to try using
234 dlpi, which is officially supported.
236 The first step is to get the setup right. To determine if
237 you have the dlpi driver loaded correctly, type:
243 then you are ready to use dlpi.
248 Then you need to type:
249 strload -f /etc/dlpi.conf
251 Check again with strload -q -d dlpi that the dlpi driver is loaded.
253 I had to make one minor code change to pcap-dlpi.c. Maybe someone
254 can explain it to me, because I am not familiar with dlpi or
255 streams programming. It took me hours to figure this out, because
256 I'm not familiar with dlpi.
258 In pcap-dlpi.c, lines 316-320:
259 #if !defined(HAVE_HPUX9) && !defined(HAVE_HPUX10_20) && !defined(sinix)
260 if (dlbindreq(p->fd, 0, ebuf) < 0 ||
261 dlbindack(p->fd, (char *)buf, ebuf) < 0)
266 #if !defined(HAVE_HPUX9) && !defined(HAVE_HPUX10_20) && !defined(sinix)
267 if (dlbindreq(p->fd, 1620, ebuf) < 0 ||
268 dlbindack(p->fd, (char *)buf, ebuf) < 0)
272 I picked the number 1620 out of thin air. The second parameter
273 to dlbindreq() sets the value of dl_sap. This dl_sap
274 value is then passed along to the DLPI driver through
275 the DL_BIND_REQ primitive. I guess that it cannot be 0 under
276 AIX, but I'm not sure.
278 If someone knows anything about DLPI, I'd appreciate a clarification.
279 Basically, I am just using the DLPI specification at:
280 http://www.opengroup.org/onlinepubs/009638599/ which is pretty good.
281 The AIX documentation is not so well written.
283 But basically, after I fixed up pcap-dlpi.c, I managed to get libpcap
284 working under AIX. This enabled me to successfully run Ethereal,
285 ie. all the packets on my Ethernet network correctly showed up
286 as Ethernet and not Token Ring in the Ethereal screen.
291 http://www.gis.net/~craigr
294 Date: Thu, 11 Nov 1999 23:47:02 -0500
295 From: Craig Rodrigues <rodrigc@mediaone.net>
296 To: ethereal-dev@zing.org
297 Subject: Re: [ethereal-dev] AIX: gtk problem solved, now an ethereal problem
299 On Thu, Nov 11, 1999 at 11:50:23AM -0800, Guy Harris wrote:
300 > > The only differences between gtkclist.c in the gtk distribution and
301 > > gtkclist.c in the ethereal distribution relate to the ROW_ELEMENT
302 > > macro. It looks like an optimization for retrieving the GList item
303 > > when the requested row is the last row in the list.
305 > Yup - as per my other mail, Ethereal does that rather a lot when
306 > building the CList, and the optimization changes quadratic behavior to
309 > > Any ideas why this causes trouble?
311 > Mismatches between the layouts of data structures as declared in the
312 > "gtk/gtk*.h" files in the Ethereal source tree and the layouts as
313 > declared in the header files in the GTK+ source (either due to header
314 > file differences - although the header files appear to be identical to
315 > the GTK+ 1.2.6 ones - or due to compiler behavior differences)?
317 I tried stepping things through the debugger, and constantly
318 hit the same segfault inside gdk_string_width(), line 308 of gdkfont.c
320 Fails on line: switch(font->type),
321 where *font is: (type = -1, ascent = -1, descent = -1)
324 gdk_string_width(font = 0x7caf01a4, string = "../"), line 308 in "gdkfont.c"
325 gtk_file_selection_populate(fs = 0x20094468, rel_path = "", try_complete = 0), line 1341 in "gtkfilesel.c"
326 gtk_file_selection_init(filesel = 0x20094468), line 513 in "gtkfilesel.c"
327 gtk_type_new(0xc315), line 403 in "gtktypeutils.c"
328 gtk_file_selection_new(title = "Ethereal: Open Capture File"), line 524 in "gtkfilesel.c"
329 file_open_cmd_cb(0x200640f4, 0x0), line 79 in "file_dlg.c"
331 Removing gtkclist.o from libui.a and recompiling removed this problem.
333 Any ideas? I'm stumped.
337 http://www.gis.net/~craigr