Annotation of www/porting.html, Revision 1.22
1.20 rohee 1: <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2 Final//EN">
1.1 marc 2: <html>
3: <head>
4: <meta http-equiv="Content-Type"
1.15 espie 5: content="text/html; charset=iso-8859-1">
1.1 marc 6: <meta name="resource-type"
1.15 espie 7: content="document">
1.1 marc 8: <meta name="description"
1.15 espie 9: CONTENT="How to make an OpenBSD port">
1.1 marc 10: <meta name="keywords"
1.15 espie 11: content="openbsd,ports">
1.1 marc 12: <meta name="distribution"
1.15 espie 13: content="global">
1.1 marc 14: <meta name="copyright"
1.20 rohee 15: content="This document copyright 1997-1999 by the OpenBSD project">
1.1 marc 16: <title>Building an OpenBSD port</title>
17: <link rev="made" HREF="mailto:www@openbsd.org">
18: </head>
19: <body text="#000000" bgcolor="#FFFFFF" link="#23238E">
1.6 pauls 20: <img height=30 width=141 src=images/smalltitle.gif alt="[OpenBSD]" >
1.1 marc 21:
1.20 rohee 22: <h2><font color="#e00000">Building an OpenBSD port</font></h2>
1.1 marc 23:
24: So you've just compiled your favorite software package on your
25: OpenBSD machine and you want to share your effort by turning
26: it into a standard port. What to do?
27: <p>
1.9 marc 28: First look at the porting information on this page. Then check
29: out the referenced documents, especially the OpenBSD porting
1.14 espie 30: <a href="checklist.html">checklist</a>.
1.1 marc 31: <p>
1.9 marc 32: Test, then re-test, and finally test again!
33: <p>
34: Submit the port. Create a gzipped tarball of the port directory.
35: You can then either place it on a public FTP or HTTP server, sending
1.20 rohee 36: its address to <a href="mailto:ports@openbsd.org">ports@openbsd.org</a>
1.9 marc 37: or send the port mime encoded to the same address. Pick whichever
38: method works best for you.
1.1 marc 39: <p>
1.20 rohee 40: <h3><font color="#0000e0">Available Porting Information</font></h3>
1.1 marc 41: <ul>
42: <li>The file <code>/usr/share/mk/bsd.port.mk</code>. This is the
43: system ports makefile included at the end of each individual
44: port makefile. Read the comments at the start of the makefile.
45: They do a good job of describing the available make options.
46: <li>The
1.13 art 47: <a href="http://www.netbsd.org/Documentation/software/packages.html">
48: NetBSD Package System</a> documentation. This document describes
49: NetBSD's version of the FreeBSD ports system and is quite helpful.
1.8 marc 50: <li>Section 19.2.5 of the
1.1 marc 51: <a href="http://www.freebsd.org/handbook/porting.html">FreeBSD
52: Handbook</a>. This is the FreeBSD porting bible.
1.9 marc 53: <li>OpenBSD porting <a href="checklist.html">checklist</a>.
1.15 espie 54: <li><a href="audio-port.html">Porting audio applications to OpenBSD</a>.
1.1 marc 55: <li>The OpenBSD ports mailing list,
56: <a href="mailto:ports@openbsd.org">ports@OpenBSD.ORG</a>.
57: </ul>
1.20 rohee 58: <h3><font color="#0000e0">OpenBSD Porting Policy</font></h3>
1.1 marc 59: <ul>
60: <li>OpenBSD does NOT use /usr/local/etc/rc.d.<br>
1.7 espie 61: <code>/usr/local</code> is often shared between several machines
1.15 espie 62: thanks to NFS. For this reason, configuration files that are specific
63: to a given machine can't be stored under <code>/usr/local</code>,
64: <code>/etc</code> is the central repository for per machine
65: configuration files. Moreover, OpenBSD policy is to never update
66: files under <code>/etc</code> automatically. Ports that need some
67: specific boot setup should advise the administrator about what to do
68: instead of blindly installing files.
1.1 marc 69: <li>OpenBSD does NOT compress man pages.
70: <li>OpenBSD does NOT require <code>-lcrypt</code>.<br>
71: DES encryption is part of the standard <code>libc</code>.
1.10 espie 72: <li>OpenBSD is strongly security-oriented. You should read and understand
73: this page's <a href="#security">security section</a>.
74: <li>Be sure to add the <code>$</code><code>OpenBSD$</code> CVS tag to
75: the Makefile. If importing a port from another system be sure to
76: leave their tag in the Makefile, too. However, replace the FreeBSD
77: <code>$</code><code>Id$</code> tag with the
78: <code>$</code><code>FreeBSD$</code> tag.
79: <li>The goal is to get all ported applications to support OpenBSD. To
80: achieve this goal you <strong>must</strong> feed any OpenBSD patches
81: back to the application maintainer.
82: </ul>
1.20 rohee 83: <a name=security></a>
84: <h3><font color="#0000e0">Security recommendations</font></h3>
1.10 espie 85: There are many security problems to worry about. If
1.2 marc 86: you are not absolutely sure of what you are doing please request
1.1 marc 87: help from the <a href="mailto:ports@openbsd.org">ports</a> mailing
88: list.
1.10 espie 89:
90: <ul>
1.20 rohee 91: <li>Do <em>not</em> use alpha or beta code when preparing a port. Use the
1.10 espie 92: latest regular or patch release.
93:
1.1 marc 94: <li>Any software to be installed as a server should be scanned
95: for buffer overflows, especially unsafe use of
96: <code>strcat/strcpy/strcmp/sprintf</code>. In general,
97: <code>sprintf</code> should be replaced with <code>snprintf</code>.
1.10 espie 98:
1.17 espie 99: <li>Never use filenames instead of true security. There are numerous race
1.10 espie 100: conditions where you don't have proper control. For instance, an attacker
101: who already has user privileges on your machines may replace files in
102: <code>/tmp</code> with symbolic links to more strategic files, such as
1.19 rohee 103: <code>/etc/master.passwd</code>.
1.16 espie 104:
105: <li>For instance, both <code>fopen</code> and <code>freopen</code>
106: <strong>create a new file or open an existing file</strong> for
107: writing. An attacker may create a symbolic link from
1.19 rohee 108: <code>/etc/master.passwd</code> to <code>/tmp/addrpool_dump</code>. The
1.16 espie 109: instant you open it, your password file is hosed. Yes, even with
110: an <code>unlink</code> right before. You only narrow the window
111: of opportunity. Use <code>open</code> with
1.22 ! rohee 112: <code>O_CREAT|O_EXCL</code> and <code>fdopen</code> instead.
1.15 espie 113:
1.16 espie 114: <li>Another very common problem is the <code>mktemp</code>
1.17 espie 115: function. Heed the warnings of the bsd linker about its uses.
1.15 espie 116: <strong>These must be fixed</strong>.
117: This is not quite as simple as <code>s/mktemp/mkstemp/g</code>. <br>
118: Refer to the <code>mktemp(3)</code> man page of OpenBSD current
119: for more indications.
120: Correct code using <code>mkstemp</code> includes the source to
121: <code>ed</code> or <code>mail</code>.
122: A rare instance of code that uses <code>mktemp</code> correctly
123: can be found in the <code>rsync</code> port.
1.10 espie 124:
125: <li>Just because you can read it doesn't mean you should. A well-known hole
126: of this nature was the <code>startx</code> problem. As a setuid program,
127: you could launch startx with any file as a script. If the file was not
128: a valid shell script, a syntax error message would follow, along with the
129: first line of the offending file, without any further permission check.
130: Pretty handy to grab the first line of a shadow passwd file, considering
1.12 deraadt 131: these often start with root entry. Do not open your file, and then do
132: an <code>fstat</code> on the open descriptor to check if you should have
1.15 espie 133: been able to open it (or the attacker will play with /dev/rst0 and rewind
1.12 deraadt 134: your tape) -- open it with the correct uid/gid/grouplist set.
1.10 espie 135:
136: <li>Don't use anything that forks a shell in setuid programs before dropping
1.15 espie 137: your privileges. This includes <code>popen</code> and
138: <code>system</code>.
1.10 espie 139: Use <code>fork</code>, <code>pipe</code> and <code>execve</code> instead.
140:
1.15 espie 141: <li>Pass open descriptors instead of filenames to other programs to
142: avoid race conditions. Even if a program does not accept file
143: descriptors, you can still use <code>/dev/fd/0</code>.
1.10 espie 144:
1.15 espie 145: <li>Access rights are attached to file descriptors. If you need setuid rights
1.10 espie 146: to open a file, open that file, then drop your privileges. You can still
147: access the open descriptor, but you have less to worry about. This is
148: double-edged: even after dropping privileges, you should still watch out
149: for those descriptors.
150:
151: <li>Avoid root setuid as much as you can. Basically, root can do anything,
1.15 espie 152: but root rights are very rarely needed, except maybe to create
153: socket ports with a number under 1024. It is arguably better to
154: keep that under <code>inetd</code>
1.10 espie 155: control and just add the relevant entries to <code>inetd.conf</code>.
156: You must know the appropriate magic for writing daemons to achieve that.
1.15 espie 157: It could be argued that you have no business writing setuid programs
158: if you don't know how to do that.
1.10 espie 159:
1.15 espie 160: <li>Use setgid instead of setuid. Apart from those specific files needed
161: by setgid programs, most files are not group-writable. Hence, a
162: security problem in a setgid program won't compromise your system as
163: much: with only group rights, the worst an intruder will be able to
164: do is hack a games score table or some such.
1.10 espie 165: See the <code>xkobo</code> port for an instance of such a change.
166:
1.15 espie 167: <li>Don't trust group-writable files. Even though they are audited,
168: setgid programs are not perceived as important potential security
169: holes. Hence stuff they can tamper with shouldn't be considered
170: sensitive information.
171:
172: <li>Don't trust your environment ! This involves simple things such as
173: your <code>PATH</code> (never use <code>system</code> with an
174: unqualified name, avoid <code>execvp</code>), but also more subtle
175: items such as your locale, timezone, termcap, and so on.
176: Be aware of transitivity: even though you're taking full precautions,
177: programs you call directly won't necessarily. <strong>Never</strong>
178: use <code>system</code> in privileged programs, build your command
179: line, a controlled environment, and call <code>execve</code> directly.
180: The <code>perlsec</code> man page is a good tutorial on such problems.
181:
182: <li>Never used setuid shell-scripts. These are inherently insecure.
183: Wrap them into some C code that ensures a proper environment.
184: On the other hand, OpenBSD features secure perl scripts.
185:
186: <li>Beware the dynamic loader. If you are running setuid, it will only
187: use trusted libraries that were scanned with ldconfig.
188: Setgid is not enough.
189:
190: <li>Dynamic libraries are tricky. C++ code sets a similar problem.
191: Basically, libraries may take some action based upon your environment
192: before your main program even gets to check its setuid status.
193: OpenBSD <code>issetugid</code> addresses this problem, from the
194: library writer point of view. Don't try to port libraries unless you
195: understand this issue thoroughly.
1.10 espie 196: </ul>
1.20 rohee 197: <h3><font color="#0000e0">Generic porting hints</font></h3>
1.10 espie 198: <ul>
199: <li><code>__OpenBSD__</code> should be used sparingly, if at all.
200: Constructs that look like
201: <pre>
202: #if defined(__NetBSD__) || defined(__FreeBSD__)
203: </pre>
204: are often inappropriate. Don't add blindly <code>__OpenBSD__</code>
205: to it. Instead, try to figure out what's going on, and what actual
206: feature is needed. Manual pages are often useful, as they include
207: historic comments, stating when a particular feature was incorporated
208: into BSD. Checking the numeric value of <code>BSD</code> against known
209: releases is often the right way. See
1.17 espie 210: <a href="ftp://ftp.netbsd.org/pub/NetBSD/packages/pkgsrc/Packages.txt">the NetBSD package guide</a>
1.10 espie 211: for more information.
212: <li>Defining <code>BSD</code> is a bad idea. Try to include <code>sys/param.h</code>.
213: This not only defines <code>BSD</code>, it also gives it a proper value.
214: The right code fragment should look like:
215: <pre>
216: #if (defined(__unix__) || defined(unix)) && !defined(USG)
217: #include <sys/param.h>
218: #endif
219: </pre>
220: <li>Test for features, not for specific OSes. In the long run, it is much
1.15 espie 221: better to test whether <code>tcgetattr</code> works than whether
222: you're running under BSD 4.3 or later, or SystemVR4. These kind of
223: tests just confuse the issue. The way to go about it is, for instance,
224: to test for one particular system, set up a slew of
225: <code>HAVE_TCGETATTR</code> defines, then proceed to the next system.
226: This technique separates features tests from specific OSes.
227: In a hurry, another porter can just add the whole set of
228: <code>-DHAVE_XXX</code> defines to the Makefile. One may also write
229: or add to a configure script to check for that feature and add it
230: automatically. As an example not to follow, check nethack 3.2.2
231: source: it assumes loads of things based on the system type. Most
232: of these assumptions are obsolete and no longer reflect reality:
233: POSIX functions are more useful than older BSD versus SystemV
234: differences, to the point that some traditional bsd functions are
1.10 espie 235: now only supported through compatibility libraries.
236:
237: <li>Avoid include files that include other includes that... First, because
1.15 espie 238: this is inefficient. Your project will end up including a file that
239: includes everything. Also, because it makes some problems difficult
240: to fix. It becomes harder to <em>not</em> include one particular file
241: at a given point.
1.10 espie 242:
243: <li>Avoid bizarre macro tricks. Undefining a macro that was defined by a
244: header file is a bad idea. Defining macros to get some specific behavior
1.15 espie 245: from an include file is also a bad idea: compilation modes should be
246: global. If you want POSIX behavior, say so, and
247: <code>#define POSIX_C_SOURCE</code>
1.10 espie 248: throughout the whole project, not when you feel like it.
249:
250: <li>Don't ever write system function prototypes. All modern systems,
251: OpenBSD included, have a standard location for these prototypes. Likely
1.15 espie 252: places include <code>unistd.h</code>, <code>fcntl.h</code> or
253: <code>termios.h</code>.
254: The man page frequently states where the prototype can be found.
255: You might need another slew of <code>HAVE_XXX</code> macros to
256: procure the right file. Don't worry about including the same file
257: twice, include files have guards that prevent all kinds of nastiness.<br>
1.10 espie 258: If some broken system needs you to write the prototype, don't impose
1.15 espie 259: on all other systems. Roll-your-own prototypes will break because they
260: may use types that are equivalent on your system, but are not portable
261: to other systems (<code>unsigned long</code> instead of
262: <code>size_t</code>), or get some <code>const</code> status wrong.
263: Also, some compilers, such as OpenBSD's own gcc,
1.10 espie 264: are able to do a better job with some very frequent functions such as
265: <code>strlen</code> if you include the right header file.
266:
267: <li>Don't use the name of a standard system function for anything else.
1.15 espie 268: If you want to write your own function, give it its own name, and
269: call that function everywhere. If you wish to revert to the
270: default system function, you just need to comment your code out,
271: and define your own name to the system function. Don't do it the
272: other way round. Code should look like this
1.10 espie 273: <pre>
274: /* prototype part */
275: #ifdef USE_OWN_GCVT
276: char *foo_gcvt(double number, size_t ndigit, char *buf);
277: #else
278: /* include correct file */
279: #include <stdlib.h>
280: /* use system function */
281: #define foo_gcvt gcvt
282: #endif
283:
284: /* definition part */
285: #ifdef USE_OWN_GCVT
286: char *foo_gcvt(double number, size_t ndigit, char *buf)
287: {
1.15 espie 288: /* proper definition */
289: }
1.10 espie 290:
291: /* typical use */
292: s = foo_gcvt(n, 15, b);
293: </pre>
1.1 marc 294: </ul>
1.20 rohee 295: <h3><font color="#0000e0">Other Helpful Hints</font></h3>
1.1 marc 296: <ul>
1.15 espie 297: <li>Recent versions of <code>bsd.port.mk</code> set the installers
298: path. Specifically, they set <code>/usr/bin</code> and
299: <code>/bin</code> to be searched <em>before</em>
300: <code>/usr/local/bin</code> and <code>/usr/X11R6/bin</code>.
1.19 rohee 301: <li>Do <em>NOT</em> generate shared libraries if
1.21 rohee 302: <code>${NO_SHARED_LIBS}</code> is defined (beware, it can be defined
303: only after inclusion of <code>bsd.port.mk</code>). If your port has
304: a GNU configure simply add the line
305: <code>CONFIGURE_ARGS += ${CONFIGURE_SHARED}</code> to the Makefile.
1.19 rohee 306: <li>If you rely on a feature that appeared in a recent version of
307: <code>bsd.port.mk</code> don't forget to add a line
308: <code>NEED_VERSION = x.yy</code> in the Makefile.
1.1 marc 309: <li>In OpenBSD <code>curses.h/libcurses/libtermlib</code> are the
310: ``new curses''. Change:<br>
1.15 espie 311: <code>ncurses.h ==> curses.h</code><br>
312: <code>-lncurses ==> -lcurses</code><br>
313: ``old (BSD) curses'' is available by defining
314: <code>_USE_OLD_CURSES_</code>
1.11 millert 315: before including <code>curses.h</code> (usually in a Makefile) and
1.15 espie 316: linking with <code>-locurses</code>.
1.11 millert 317: <li>In OpenBSD, terminal discipline has been upgraded from the older BSD
318: <code>sgtty</code> to the newer POSIX <code>tcgetattr</code> family.
1.15 espie 319: Avoid the older style in new code. Some code may define
320: <code>tcgetattr</code> to be a synonym for the older
321: <code>sgtty</code>, but this is at best a stopgap measure on OpenBSD.
322: The <code>xterm</code> source code is a very good example of
323: what not to do. Try to get your system functionality right: you
324: want a type that remembers the state of your terminal
325: (possible typedef), you want a function that extracts the current
326: state, and a function that sets the new state.
327: Functions that modify this state are more difficult, as they tend
328: to vary depending upon the system. Also, don't forget that you will
329: have to handle cases where you are not connected to a terminal,
330: and that you need to handle signals: not only termination, but
331: also background (<code>SIGTSTP</code>). You should always leave
332: the terminal in a sane state. Do your tests under an older shell,
333: such as sh, which does not reset the terminal in any way at
1.10 espie 334: program's termination.
1.15 espie 335: <li>The newer termcap/terminfo and curses, as included with OpenBSD,
336: include standard sequences for controlling color terminals. You
337: should use these if possible, reverting to standard ANSI color
338: sequences on other systems. However, some terminal descriptions
339: have not been updated yet, and you may need to be able to specify
340: these sequences manually. This is the way vim handles it. This is
341: not strictly necessary. Except for privileged programs, it is
342: generally possible to override a termcap definition through the
1.10 espie 343: <code>TERMCAP</code> variable and get it to work properly.
1.15 espie 344: <li>Signal semantics are tricky, and vary from one system to another.
345: Use <code>sigaction</code> to ensure a specific semantics, along
346: with other system calls referenced in the corresponding manpage.
1.1 marc 347: </ul>
348: <hr>
1.6 pauls 349: <a href="index.html"><img height=24 width=24 src=back.gif border=0 alt=OpenBSD></a>
1.20 rohee 350: <a href="mailto:www@openbsd.org">www@openbsd.org</a>
1.22 ! rohee 351: <br><small>$OpenBSD: porting.html,v 1.21 1999/03/13 18:15:43 rohee Exp $</small>
1.1 marc 352: </body>
353: </html>