2 INSTALLATION ON THE UNIX PLATFORM
3 ---------------------------------
5 [Installation on DOS (with djgpp), Windows, OpenVMS, MacOS (before MacOS X)
6 and NetWare is described in INSTALL.DJGPP, INSTALL.W32, INSTALL.VMS,
7 INSTALL.MacOS and INSTALL.NW.
9 This document describes installation on operating systems in the Unix
12 To install OpenSSL, you will need:
17 * a development environment in form of development libraries and C
19 * a supported Unix operating system
24 If you want to just get on with it, do:
31 [If any of these steps fails, see section Installation in Detail below.]
33 This will build and install OpenSSL in the default location, which is (for
34 historical reasons) /usr/local/ssl. If you want to install it anywhere else,
37 $ ./config --prefix=/usr/local --openssldir=/usr/local/openssl
43 There are several options to ./config (or ./Configure) to customize
46 --prefix=DIR Install in DIR/bin, DIR/lib, DIR/include/openssl.
47 Configuration files used by OpenSSL will be in DIR/ssl
48 or the directory specified by --openssldir.
50 --openssldir=DIR Directory for OpenSSL files. If no prefix is specified,
51 the library files and binaries are also installed there.
53 no-threads Don't try to build with support for multi-threaded
56 threads Build with support for multi-threaded applications.
57 This will usually require additional system-dependent options!
58 See "Note on multi-threading" below.
60 no-zlib Don't try to build with support for zlib compression and
63 zlib Build with support for zlib compression/decompression.
65 zlib-dynamic Like "zlib", but has OpenSSL load the zlib library dynamically
66 when needed. This is only supported on systems where loading
67 of shared libraries is supported. This is the default choice.
69 no-shared Don't try to create shared libraries.
71 shared In addition to the usual static libraries, create shared
72 libraries on platforms where it's supported. See "Note on
73 shared libraries" below.
75 no-asm Do not use assembler code.
77 386 Use the 80386 instruction set only (the default x86 code is
78 more efficient, but requires at least a 486). Note: Use
79 compiler flags for any other CPU specific configuration,
80 e.g. "-m32" to build x86 code on an x64 system.
82 no-sse2 Exclude SSE2 code pathes. Normally SSE2 extention is
83 detected at run-time, but the decision whether or not the
84 machine code will be executed is taken solely on CPU
85 capability vector. This means that if you happen to run OS
86 kernel which does not support SSE2 extension on Intel P4
87 processor, then your application might be exposed to
88 "illegal instruction" exception. There might be a way
89 to enable support in kernel, e.g. FreeBSD kernel can be
90 compiled with CPU_ENABLE_SSE, and there is a way to
91 disengage SSE2 code pathes upon application start-up,
92 but if you aim for wider "audience" running such kernel,
93 consider no-sse2. Both 386 and no-asm options above imply
96 no-<cipher> Build without the specified cipher (bf, cast, des, dh, dsa,
97 hmac, md2, md5, mdc2, rc2, rc4, rc5, rsa, sha).
98 The crypto/<cipher> directory can be removed after running
101 -Dxxx, -lxxx, -Lxxx, -fxxx, -mxxx, -Kxxx These system specific options will
102 be passed through to the compiler to allow you to
103 define preprocessor symbols, specify additional libraries,
104 library directories or other compiler options.
107 Installation in Detail
108 ----------------------
110 1a. Configure OpenSSL for your operation system automatically:
114 This guesses at your operating system (and compiler, if necessary) and
115 configures OpenSSL based on this guess. Run ./config -t to see
116 if it guessed correctly. If you want to use a different compiler, you
117 are cross-compiling for another platform, or the ./config guess was
118 wrong for other reasons, go to step 1b. Otherwise go to step 2.
120 On some systems, you can include debugging information as follows:
122 $ ./config -d [options]
124 1b. Configure OpenSSL for your operating system manually
126 OpenSSL knows about a range of different operating system, hardware and
127 compiler combinations. To see the ones it knows about, run
131 Pick a suitable name from the list that matches your system. For most
132 operating systems there is a choice between using "cc" or "gcc". When
133 you have identified your system (and if necessary compiler) use this name
134 as the argument to ./Configure. For example, a "linux-elf" user would
137 $ ./Configure linux-elf [options]
139 If your system is not available, you will have to edit the Configure
140 program and add the correct configuration for your system. The
141 generic configurations "cc" or "gcc" should usually work on 32 bit
144 Configure creates the file Makefile.ssl from Makefile.org and
145 defines various macros in crypto/opensslconf.h (generated from
146 crypto/opensslconf.h.in).
148 2. Build OpenSSL by running:
152 This will build the OpenSSL libraries (libcrypto.a and libssl.a) and the
153 OpenSSL binary ("openssl"). The libraries will be built in the top-level
154 directory, and the binary will be in the "apps" directory.
156 If "make" fails, look at the output. There may be reasons for
157 the failure that aren't problems in OpenSSL itself (like missing
158 standard headers). If it is a problem with OpenSSL itself, please
159 report the problem to <openssl-bugs@openssl.org> (note that your
160 message will be recorded in the request tracker publicly readable
161 via http://www.openssl.org/support/rt2.html and will be forwarded to a
162 public mailing list). Include the output of "make report" in your message.
163 Please check out the request tracker. Maybe the bug was already
164 reported or has already been fixed.
166 [If you encounter assembler error messages, try the "no-asm"
167 configuration option as an immediate fix.]
169 Compiling parts of OpenSSL with gcc and others with the system
170 compiler will result in unresolved symbols on some systems.
172 3. After a successful build, the libraries should be tested. Run:
176 If a test fails, look at the output. There may be reasons for
177 the failure that isn't a problem in OpenSSL itself (like a missing
178 or malfunctioning bc). If it is a problem with OpenSSL itself,
179 try removing any compiler optimization flags from the CFLAG line
180 in Makefile.ssl and run "make clean; make". Please send a bug
181 report to <openssl-bugs@openssl.org>, including the output of
182 "make report" in order to be added to the request tracker at
183 http://www.openssl.org/support/rt2.html.
185 4. If everything tests ok, install OpenSSL with
189 This will create the installation directory (if it does not exist) and
190 then the following subdirectories:
192 certs Initially empty, this is the default location
193 for certificate files.
194 man/man1 Manual pages for the 'openssl' command line tool
195 man/man3 Manual pages for the libraries (very incomplete)
196 misc Various scripts.
197 private Initially empty, this is the default location
198 for private key files.
200 If you didn't choose a different installation prefix, the
201 following additional subdirectories will be created:
203 bin Contains the openssl binary and a few other
205 include/openssl Contains the header files needed if you want to
206 compile programs with libcrypto or libssl.
207 lib Contains the OpenSSL library files themselves.
209 Package builders who want to configure the library for standard
210 locations, but have the package installed somewhere else so that
211 it can easily be packaged, can use
213 $ make INSTALL_PREFIX=/tmp/package-root install
215 (or specify "--install_prefix=/tmp/package-root" as a configure
216 option). The specified prefix will be prepended to all
217 installation target filenames.
220 NOTE: The header files used to reside directly in the include
221 directory, but have now been moved to include/openssl so that
222 OpenSSL can co-exist with other libraries which use some of the
223 same filenames. This means that applications that use OpenSSL
224 should now use C preprocessor directives of the form
226 #include <openssl/ssl.h>
228 instead of "#include <ssl.h>", which was used with library versions
229 up to OpenSSL 0.9.2b.
231 If you install a new version of OpenSSL over an old library version,
232 you should delete the old header files in the include directory.
234 Compatibility issues:
236 * COMPILING existing applications
238 To compile an application that uses old filenames -- e.g.
239 "#include <ssl.h>" --, it will usually be enough to find
240 the CFLAGS definition in the application's Makefile and
241 add a C option such as
243 -I/usr/local/ssl/include/openssl
247 But don't delete the existing -I option that points to
248 the ..../include directory! Otherwise, OpenSSL header files
249 could not #include each other.
251 * WRITING applications
253 To write an application that is able to handle both the new
254 and the old directory layout, so that it can still be compiled
255 with library versions up to OpenSSL 0.9.2b without bothering
256 the user, you can proceed as follows:
258 - Always use the new filename of OpenSSL header files,
259 e.g. #include <openssl/ssl.h>.
261 - Create a directory "incl" that contains only a symbolic
262 link named "openssl", which points to the "include" directory
264 For example, your application's Makefile might contain the
265 following rule, if OPENSSLDIR is a pathname (absolute or
266 relative) of the directory where OpenSSL resides:
270 cd $(OPENSSLDIR) # Check whether the directory really exists
271 -ln -s `cd $(OPENSSLDIR); pwd`/include incl/openssl
273 You will have to add "incl/openssl" to the dependencies
274 of those C files that include some OpenSSL header file.
276 - Add "-Iincl" to your CFLAGS.
278 With these additions, the OpenSSL header files will be available
279 under both name variants if an old library version is used:
280 Your application can reach them under names like <openssl/foo.h>,
281 while the header files still are able to #include each other
282 with names of the form <foo.h>.
285 Note on multi-threading
286 -----------------------
288 For some systems, the OpenSSL Configure script knows what compiler options
289 are needed to generate a library that is suitable for multi-threaded
290 applications. On these systems, support for multi-threading is enabled
291 by default; use the "no-threads" option to disable (this should never be
294 On other systems, to enable support for multi-threading, you will have
295 to specify at least two options: "threads", and a system-dependent option.
296 (The latter is "-D_REENTRANT" on various systems.) The default in this
297 case, obviously, is not to include support for multi-threading (but
298 you can still use "no-threads" to suppress an annoying warning message
299 from the Configure script.)
302 Note on shared libraries
303 ------------------------
305 Shared libraries have certain caveats. Binary backward compatibility
306 can't be guaranteed before OpenSSL version 1.0. The only reason to
307 use them would be to conserve memory on systems where several programs
310 For some systems, the OpenSSL Configure script knows what is needed to
311 build shared libraries for libcrypto and libssl. On these systems,
312 the shared libraries are currently not created by default, but giving
313 the option "shared" will get them created. This method supports Makefile
314 targets for shared library creation, like linux-shared. Those targets
315 can currently be used on their own just as well, but this is expected
316 to change in future versions of OpenSSL.
318 Note on random number generation
319 --------------------------------
321 Availability of cryptographically secure random numbers is required for
322 secret key generation. OpenSSL provides several options to seed the
323 internal PRNG. If not properly seeded, the internal PRNG will refuse
324 to deliver random bytes and a "PRNG not seeded error" will occur.
325 On systems without /dev/urandom (or similar) device, it may be necessary
326 to install additional support software to obtain random seed.
327 Please check out the manual pages for RAND_add(), RAND_bytes(), RAND_egd(),
328 and the FAQ for more information.
330 Note on support for multiple builds
331 -----------------------------------
333 OpenSSL is usually built in its source tree. Unfortunately, this doesn't
334 support building for multiple platforms from the same source tree very well.
335 It is however possible to build in a separate tree through the use of lots
336 of symbolic links, which should be prepared like this:
338 mkdir -p objtree/"`uname -s`-`uname -r`-`uname -m`"
339 cd objtree/"`uname -s`-`uname -r`-`uname -m`"
340 (cd $OPENSSL_SOURCE; find . -type f) | while read F; do
341 mkdir -p `dirname $F`
342 rm -f $F; ln -s $OPENSSL_SOURCE/$F $F
343 echo $F '->' $OPENSSL_SOURCE/$F
345 make -f Makefile.org clean
347 OPENSSL_SOURCE is an environment variable that contains the absolute (this
348 is important!) path to the OpenSSL source tree.
350 Also, operations like 'make update' should still be made in the source tree.