1 $Id: README.aix,v 1.5 2002/05/31 10:22:51 guy Exp $
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 Some earlier notes: the notes about libpcap may not apply, with libpcap
10 0.7.1, but they're preserved here for historical reasons.
12 The glib, gtk+, and Ethereal notes still apply.
14 After much work and toil, Craig Rodrigues was able to compile libpcap
15 and Ethereal on AIX 4.3.2. His odyssey is document in various e-mails
16 at http://www.ethereal.com/lists/ethereal-dev/199911/
18 Here are a few excerpts. Note that, to configure "libpcap" to use DLPI
19 rather than BPF (which it'll apparently use by default on AIX),
24 to the "configure" script for "libpcap" should do the trick.
26 The source code changes to Ethereal mentioned below should be in the
27 current source tree. The changes to the GLib configure script is in
28 GLib 1.2.7; the changes for the "-lgdk" problem are probably still
29 necessary in the current version of GTK+.
31 Subject: Re: [ethereal-dev] Re: [ethereal-users] Problems compiling 0.7.7 under AIX 4.3.2
32 From: Gilbert Ramirez <gram@xiexie.org>
33 Date: Fri, 5 Nov 1999 16:58:17 -0600
34 To: Guy Harris <guy@netapp.com>
35 Cc: Craig Rodrigues <rodrigc@mediaone.net>, ethereal-dev@zing.org
38 On Fri, Nov 05, 1999 at 01:42:44PM -0600, Guy Harris wrote:
43 > Looks suspiciously similar to the previous error; have you tried
44 > recompiling GTK+ with "xlc_r"?
46 I believe glib and gtk+ should both be compiled with xlc_r. I haven't
47 compiled on AIX in a long time, but I think it's because glib is including
48 pthread stuff, so the re-entrant C library, libc_r, is needed.
53 When compiling a multi-threaded program, you should invoke the C compiler
54 using one of the following commands:
57 Invokes the compiler with default language level of ansi.
59 Invokes the compiler with default language level of extended.
62 These commands ensure that the adequate options and libraries are used to be
63 compliant with the X/Open Version 5 Standard. The POSIX Threads
64 Specification 1003.1c is a subset of the X/Open Specification.
66 The following libraries are automatically linked with your program when using these commands:
74 For example, the following command compiles the foo.c multi-threaded C source file and produces the foo executable file:
78 See the cc command for more information about C For AIX.
84 To: ethereal-users@zing.org
85 Subject: [ethereal-dev] AIX: gtk problem solved, now an ethereal problem
86 From: Craig Rodrigues <rodrigc@mediaone.net>
87 Date: Mon, 8 Nov 1999 10:46:25 -0500
88 Cc: ethereal-dev@zing.org
93 After much sweat and toil, I have managed to get gtk 1.2.6 to
94 compile and not dump core under AIX. The solutions were to
95 (1) apply the attached patch to the configure.in in the glib-1.2.6
98 (2) In the file gtk+-1.2.6/gtk/Makefile, add a link flag -lgdk to link
101 I have submitted (1) to the gtk-devel mailing list where it has been
102 accepted. (2) is an uglier problem, but for now, adding -lgdk by hand
105 Now I have a problem....I compiled gtk, and that works.
106 I compiled ethereal (after some minor mods), and it starts,
107 but when I click on Capture -> Start, I get:
109 "There are no network interfaces that can be opened."
111 I am running as root, so I don't think permissions are a problem.
118 http://www.gis.net/~craigr
121 *** configure.in.old Thu Oct 7 17:27:43 1999
122 --- configure.in Sun Nov 7 19:34:36 1999
126 if test "$ac_cv_func_getpwuid_r" = "yes"; then
127 AC_MSG_CHECKING(whether getpwuid_r is posix like)
128 ! # getpwuid_r(0, NULL, NULL, 0) is the signature on
129 ! # solaris, if that is not found, the prog below won't
130 ! # compile, then the posix signature is assumed as
132 ! AC_TRY_COMPILE([#include <pwd.h>],
133 ! [getpwuid_r(0, NULL, NULL, 0);],
134 ! [AC_MSG_RESULT(no)],
135 ! [AC_MSG_RESULT(yes)
136 ! AC_DEFINE(HAVE_GETPWUID_R_POSIX)])
139 if test x"$have_threads" = xposix; then
142 if test "$ac_cv_func_getpwuid_r" = "yes"; then
143 AC_MSG_CHECKING(whether getpwuid_r is posix like)
144 ! # The signature for the POSIX version is:
145 ! # int getpwuid_r(uid_t, struct passwd *, char *, size_t, struct passwd **)
146 ! AC_TRY_COMPILE([#include <pwd.h>
147 ! #include <sys/types.h>
148 ! #include <stdlib.h>],
149 ! [getpwuid_r((uid_t)0, NULL, NULL, (size_t)0, NULL);],
150 ! [AC_DEFINE(HAVE_GETPWUID_R_POSIX)
151 ! AC_MSG_RESULT(yes)],
152 ! [AC_MSG_RESULT(no)])
155 if test x"$have_threads" = xposix; then
159 To: ethereal-dev@zing.org
160 Subject: Re: [ethereal-dev] AIX: gtk problem solved, now an ethereal problem
161 From: Craig Rodrigues <rodrigc@mediaone.net>
162 Date: Wed, 10 Nov 1999 12:18:47 -0500
168 OK, I'm getting closer and closer to this working on AIX.
172 (1) In a bunch of places in the code I removed '//' style C++ comments
173 which the IBM C compiler didn't like.
175 (2) I also found some places in the code like:
177 enum some_enum { FOO, BAR, };
179 IBM C did not like the trailing "," after BAR.
181 (3) In packet-ipv6.h, IPV6_VERSION is defined, but that is already
182 defined in <netinet/in.h> on AIX 4.3, so for now I just commented that out.
184 (4) in packet-afs.c, when it sucks in <netinet/in.h>, in.h sucks in
185 <sys/machine.h> which defines LITTLE_ENDIAN. This conflicts with
186 LITTLE_ENDIAN in globals.h. So what I did was, in globals.h, I added:
188 #ifdef HAVE_NETINET_IN_H
189 #include <netinet/in.h>
192 So after doing all these things, I can compile ethereal and run it.
194 correct network interfaces on my system: lo0 and en0. However,
195 when I start capturing packets on en0, they are all of the protocol type
196 "TRMAC" and "TR". The only problem is, I'm not on a Token Ring network.
200 No. Time Source Destination Protocol Info
201 1 0.000000 0a:30:a1:08:00:45 06:74:60:08:00:5a TR Token-Ring Unknown
202 2 0.210304 0a:30:a1:08:00:45 06:74:60:08:00:5a TR Token-Ring Unknown
203 3 0.926080 0a:30:a1:08:00:45 06:74:60:08:00:5a TR Token-Ring Unknown
204 4 0.4236416 0a:30:a1:08:00:45 06:74:60:08:00:5a TR Token-Ring Unknown
205 5 0.4712064 6f:06:74:60:08:00 5a:8a:30:a1:00:00 TR MAC Unknown Major Vector: 127
208 ---------------------
209 It turns out that libpcap was using IFT_* numbers instead of DLT_* numbers for
210 link types. That has been fixed
211 ---------------------
214 To: tcpdump-workers@tcpdump.org
215 Subject: [ethereal-dev] Sucess with libpcap under AIX
216 From: Craig Rodrigues <rodrigc@mediaone.net>
217 Date: Sat, 20 Nov 1999 03:34:50 -0500
218 Cc: ethereal-dev@zing.org
223 I have managed to successfully compile and use the latest
224 snapshot of libpcap under AIX using DLPI. bpf is majorly
225 brain-dead under AIX, and very unsupported. Rather than
226 find all the bugs in AIX's bpf, I decided to try using
227 dlpi, which is officially supported.
229 The first step is to get the setup right. To determine if
230 you have the dlpi driver loaded correctly, type:
236 then you are ready to use dlpi.
241 Then you need to type:
242 strload -f /etc/dlpi.conf
244 Check again with strload -q -d dlpi that the dlpi driver is loaded.
246 I had to make one minor code change to pcap-dlpi.c. Maybe someone
247 can explain it to me, because I am not familiar with dlpi or
248 streams programming. It took me hours to figure this out, because
249 I'm not familiar with dlpi.
251 In pcap-dlpi.c, lines 316-320:
252 #if !defined(HAVE_HPUX9) && !defined(HAVE_HPUX10_20) && !defined(sinix)
253 if (dlbindreq(p->fd, 0, ebuf) < 0 ||
254 dlbindack(p->fd, (char *)buf, ebuf) < 0)
259 #if !defined(HAVE_HPUX9) && !defined(HAVE_HPUX10_20) && !defined(sinix)
260 if (dlbindreq(p->fd, 1620, ebuf) < 0 ||
261 dlbindack(p->fd, (char *)buf, ebuf) < 0)
265 I picked the number 1620 out of thin air. The second parameter
266 to dlbindreq() sets the value of dl_sap. This dl_sap
267 value is then passed along to the DLPI driver through
268 the DL_BIND_REQ primitive. I guess that it cannot be 0 under
269 AIX, but I'm not sure.
271 If someone knows anything about DLPI, I'd appreciate a clarification.
272 Basically, I am just using the DLPI specification at:
273 http://www.opengroup.org/onlinepubs/009638599/ which is pretty good.
274 The AIX documentation is not so well written.
276 But basically, after I fixed up pcap-dlpi.c, I managed to get libpcap
277 working under AIX. This enabled me to successfully run Ethereal,
278 ie. all the packets on my Ethernet network correctly showed up
279 as Ethernet and not Token Ring in the Ethereal screen.
284 http://www.gis.net/~craigr
287 Date: Thu, 11 Nov 1999 23:47:02 -0500
288 From: Craig Rodrigues <rodrigc@mediaone.net>
289 To: ethereal-dev@zing.org
290 Subject: Re: [ethereal-dev] AIX: gtk problem solved, now an ethereal problem
292 On Thu, Nov 11, 1999 at 11:50:23AM -0800, Guy Harris wrote:
293 > > The only differences between gtkclist.c in the gtk distribution and
294 > > gtkclist.c in the ethereal distribution relate to the ROW_ELEMENT
295 > > macro. It looks like an optimization for retrieving the GList item
296 > > when the requested row is the last row in the list.
298 > Yup - as per my other mail, Ethereal does that rather a lot when
299 > building the CList, and the optimization changes quadratic behavior to
302 > > Any ideas why this causes trouble?
304 > Mismatches between the layouts of data structures as declared in the
305 > "gtk/gtk*.h" files in the Ethereal source tree and the layouts as
306 > declared in the header files in the GTK+ source (either due to header
307 > file differences - although the header files appear to be identical to
308 > the GTK+ 1.2.6 ones - or due to compiler behavior differences)?
310 I tried stepping things through the debugger, and constantly
311 hit the same segfault inside gdk_string_width(), line 308 of gdkfont.c
313 Fails on line: switch(font->type),
314 where *font is: (type = -1, ascent = -1, descent = -1)
317 gdk_string_width(font = 0x7caf01a4, string = "../"), line 308 in "gdkfont.c"
318 gtk_file_selection_populate(fs = 0x20094468, rel_path = "", try_complete = 0), line 1341 in "gtkfilesel.c"
319 gtk_file_selection_init(filesel = 0x20094468), line 513 in "gtkfilesel.c"
320 gtk_type_new(0xc315), line 403 in "gtktypeutils.c"
321 gtk_file_selection_new(title = "Ethereal: Open Capture File"), line 524 in "gtkfilesel.c"
322 file_open_cmd_cb(0x200640f4, 0x0), line 79 in "file_dlg.c"
324 Removing gtkclist.o from libui.a and recompiling removed this problem.
326 Any ideas? I'm stumped.
330 http://www.gis.net/~craigr