update ltm to 1.1.0 and enable FIPS 186.4 compliant key-generation (#79)

* make key-generation compliant to FIPS 186.4

* fix includes in tommath_class.h

* update fuzzcorpus instead of error-out

* fixup fuzzing make-targets

* update Makefile.in

* apply necessary patches to ltm sources

* clean-up not required ltm files

* update to vanilla ltm 1.1.0

this already only contains the required files

* remove set/get double

.gitignore

@@ -14,6 +14,9 @@
 /dropbearconvert
 /dropbearkey
 /dropbearmulti
+/fuzzcorpus
+/fuzzer-*
+/fuzzer-*.options
 /scp
 /scp-progress
 Makefile

Makefile.in

@@ -278,27 +278,29 @@ fuzzstandalone: fuzz-harness.o fuzz-targets
 # exclude svr-main.o to avoid duplicate main
 svrfuzzobjs=$(subst svr-main.o, ,$(dropbearobjs))
 
+fuzz-harness.o: $(HEADERS) $(LIBTOM_DEPS) Makefile $(svrfuzzobjs) fuzz-common.o
+
 # build all the fuzzers. This will require fail to link unless built with
 # make fuzz-targets FUZZLIB=-lFuzzer.a 
 # or similar - the library provides main().
 fuzz-targets: $(FUZZ_TARGETS) $(FUZZER_OPTIONS)
 
-fuzzer-preauth: fuzzer-preauth.o $(HEADERS) $(LIBTOM_DEPS) Makefile $(svrfuzzobjs)
+fuzzer-preauth: fuzzer-preauth.o fuzz-harness.o
 	$(CXX) $(CXXFLAGS) $@.o $(LDFLAGS) $(svrfuzzobjs) -o $@$(EXEEXT) $(LIBTOM_LIBS) $(LIBS) $(FUZZLIB) @CRYPTLIB@
 
-fuzzer-preauth_nomaths: fuzzer-preauth_nomaths.o $(HEADERS) $(LIBTOM_DEPS) Makefile $(svrfuzzobjs)
+fuzzer-preauth_nomaths: fuzzer-preauth_nomaths.o fuzz-harness.o
 	$(CXX) $(CXXFLAGS) $@.o $(LDFLAGS) $(svrfuzzobjs) -o $@$(EXEEXT) $(LIBTOM_LIBS) $(LIBS) $(FUZZLIB) @CRYPTLIB@
 
-fuzzer-pubkey: fuzzer-pubkey.o $(HEADERS) $(LIBTOM_DEPS) Makefile $(svrfuzzobjs)
+fuzzer-pubkey: fuzzer-pubkey.o fuzz-harness.o
 	$(CXX) $(CXXFLAGS) $@.o $(LDFLAGS) $(svrfuzzobjs) -o $@$(EXEEXT) $(LIBTOM_LIBS) $(LIBS) $(FUZZLIB) @CRYPTLIB@
 
-fuzzer-verify: fuzzer-verify.o $(HEADERS) $(LIBTOM_DEPS) Makefile $(svrfuzzobjs)
+fuzzer-verify: fuzzer-verify.o fuzz-harness.o
 	$(CXX) $(CXXFLAGS) $@.o $(LDFLAGS) $(svrfuzzobjs) -o $@$(EXEEXT) $(LIBTOM_LIBS) $(LIBS) $(FUZZLIB) @CRYPTLIB@
 
-fuzzer-kexdh: fuzzer-kexdh.o $(HEADERS) $(LIBTOM_DEPS) Makefile $(svrfuzzobjs)
+fuzzer-kexdh: fuzzer-kexdh.o fuzz-harness.o
 	$(CXX) $(CXXFLAGS) $@.o $(LDFLAGS) $(svrfuzzobjs) -o $@$(EXEEXT) $(LIBTOM_LIBS) $(LIBS) $(FUZZLIB) @CRYPTLIB@
 
-fuzzer-kexecdh: fuzzer-kexecdh.o $(HEADERS) $(LIBTOM_DEPS) Makefile $(svrfuzzobjs)
+fuzzer-kexecdh: fuzzer-kexecdh.o fuzz-harness.o
 	$(CXX) $(CXXFLAGS) $@.o $(LDFLAGS) $(svrfuzzobjs) -o $@$(EXEEXT) $(LIBTOM_LIBS) $(LIBS) $(FUZZLIB) @CRYPTLIB@
 
 fuzzer-%.options: Makefile

fuzzer-verify.c

@@ -37,11 +37,13 @@ int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) {
 
 				if (type == DROPBEAR_SIGNKEY_DSS) {
 					/* So far have seen dss keys with bad p/q/g domain parameters */
-					int pprime, qprime;
-				    assert(mp_prime_is_prime(key->dsskey->p, 5, &pprime) == MP_OKAY);
-				    assert(mp_prime_is_prime(key->dsskey->q, 18, &qprime) == MP_OKAY);
-				    boguskey = !(pprime && qprime);
-				    /* Could also check g**q mod p == 1 */
+					int pprime, qprime, trials;
+					trials = mp_prime_rabin_miller_trials(mp_count_bits(key->dsskey->p));
+					assert(mp_prime_is_prime(key->dsskey->p, trials, &pprime) == MP_OKAY);
+					trials = mp_prime_rabin_miller_trials(mp_count_bits(key->dsskey->q));
+					assert(mp_prime_is_prime(key->dsskey->q, trials, &qprime) == MP_OKAY);
+					boguskey = !(pprime && qprime);
+					/* Could also check g**q mod p == 1 */
 				}
 
 				if (!boguskey) {

fuzzers_test.sh

@@ -4,7 +4,7 @@
 
 result=0
 
-hg clone https://secure.ucc.asn.au/hg/dropbear-fuzzcorpus fuzzcorpus || exit 1
+test -d fuzzcorpus && hg --repository fuzzcorpus/ pull || hg clone https://secure.ucc.asn.au/hg/dropbear-fuzzcorpus fuzzcorpus || exit 1
 for f in `make list-fuzz-targets`; do
     ./$f fuzzcorpus/$f/* || result=1
 done

gendss.c

@@ -68,6 +68,7 @@ dropbear_dss_key * gen_dss_priv_key(unsigned int size) {
 static void getq(const dropbear_dss_key *key) {
 
 	unsigned char buf[QSIZE];
+	int trials;
 
 	/* 160 bit prime */
 	genrandom(buf, QSIZE);
@@ -76,8 +77,9 @@ static void getq(const dropbear_dss_key *key) {
 
 	bytes_to_mp(key->q, buf, QSIZE);
 
-	/* 18 rounds are required according to HAC */
-	if (mp_prime_next_prime(key->q, 18, 0) != MP_OKAY) {
+	/* ask FIPS 186.4 how many Rabin-Miller trials are required */
+	trials = mp_prime_rabin_miller_trials(mp_count_bits(key->q));
+	if (mp_prime_next_prime(key->q, trials, 0) != MP_OKAY) {
 		fprintf(stderr, "DSS key generation failed\n");
 		exit(1);
 	}
@@ -89,7 +91,7 @@ static void getp(const dropbear_dss_key *key, unsigned int size) {
 	DEF_MP_INT(tempC);
 	DEF_MP_INT(tempP);
 	DEF_MP_INT(temp2q);
-	int result;
+	int result, trials;
 	unsigned char *buf;
 
 	m_mp_init_multi(&tempX, &tempC, &tempP, &temp2q, NULL);
@@ -129,9 +131,10 @@ static void getp(const dropbear_dss_key *key, unsigned int size) {
 			exit(1);
 		}
 
-		/* now check for prime, 5 rounds is enough according to HAC */
+		/* ask FIPS 186.4 how many Rabin-Miller trials are required */
+		trials = mp_prime_rabin_miller_trials(mp_count_bits(key->p));
 		/* result == 1  =>  p is prime */
-		if (mp_prime_is_prime(key->p, 5, &result) != MP_OKAY) {
+		if (mp_prime_is_prime(key->p, trials, &result) != MP_OKAY) {
 			fprintf(stderr, "DSS key generation failed\n");
 			exit(1);
 		}

genrsa.c

@@ -95,6 +95,7 @@ static void getrsaprime(mp_int* prime, mp_int *primeminus,
 		mp_int* rsa_e, unsigned int size_bytes) {
 
 	unsigned char *buf;
+	int trials;
 	DEF_MP_INT(temp_gcd);
 
 	buf = (unsigned char*)m_malloc(size_bytes);
@@ -108,8 +109,9 @@ static void getrsaprime(mp_int* prime, mp_int *primeminus,
 
 		bytes_to_mp(prime, buf, size_bytes);
 
-		/* find the next integer which is prime, 8 round of miller-rabin */
-		if (mp_prime_next_prime(prime, 8, 0) != MP_OKAY) {
+		/* find the next integer which is prime */
+		trials = mp_prime_rabin_miller_trials(mp_count_bits(prime));
+		if (mp_prime_next_prime(prime, trials, 0) != MP_OKAY) {
 			fprintf(stderr, "RSA generation failed\n");
 			exit(1);
 		}

libtommath/LICENSE

@@ -1,29 +1,26 @@
-LibTomMath is licensed under DUAL licensing terms.
-
-Choose and use the license of your needs.
-
-[LICENSE #1]
-
-LibTomMath is public domain.  As should all quality software be.
-
-Tom St Denis
-
-[/LICENSE #1]
-
-[LICENSE #2]
-
-            DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE
-                    Version 2, December 2004
-
- Copyright (C) 2004 Sam Hocevar <sam@hocevar.net>
-
- Everyone is permitted to copy and distribute verbatim or modified
- copies of this license document, and changing it is allowed as long
- as the name is changed.
-
-            DO WHAT THE FUCK YOU WANT TO PUBLIC LICENSE
-   TERMS AND CONDITIONS FOR COPYING, DISTRIBUTION AND MODIFICATION
-
-  0. You just DO WHAT THE FUCK YOU WANT TO. 
-
-[/LICENSE #2]
+                          The LibTom license
+
+This is free and unencumbered software released into the public domain.
+
+Anyone is free to copy, modify, publish, use, compile, sell, or
+distribute this software, either in source code form or as a compiled
+binary, for any purpose, commercial or non-commercial, and by any
+means.
+
+In jurisdictions that recognize copyright laws, the author or authors
+of this software dedicate any and all copyright interest in the
+software to the public domain. We make this dedication for the benefit
+of the public at large and to the detriment of our heirs and
+successors. We intend this dedication to be an overt act of
+relinquishment in perpetuity of all present and future rights to this
+software under copyright law.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY CLAIM, DAMAGES OR
+OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+OTHER DEALINGS IN THE SOFTWARE.
+
+For more information, please refer to <http://unlicense.org/>

libtommath/Makefile.in

@@ -32,30 +32,32 @@ endif
 LCOV_ARGS=--directory .
 
 #START_INS
-OBJECTS=bncore.o bn_error.o bn_fast_mp_invmod.o bn_fast_mp_montgomery_reduce.o bn_fast_s_mp_mul_digs.o \
+OBJECTS=bn_error.o bn_fast_mp_invmod.o bn_fast_mp_montgomery_reduce.o bn_fast_s_mp_mul_digs.o \
 bn_fast_s_mp_mul_high_digs.o bn_fast_s_mp_sqr.o bn_mp_2expt.o bn_mp_abs.o bn_mp_add.o bn_mp_add_d.o \
 bn_mp_addmod.o bn_mp_and.o bn_mp_clamp.o bn_mp_clear.o bn_mp_clear_multi.o bn_mp_cmp.o bn_mp_cmp_d.o \
-bn_mp_cmp_mag.o bn_mp_cnt_lsb.o bn_mp_copy.o bn_mp_count_bits.o bn_mp_div_2.o bn_mp_div_2d.o bn_mp_div_3.o \
-bn_mp_div.o bn_mp_div_d.o bn_mp_dr_is_modulus.o bn_mp_dr_reduce.o bn_mp_dr_setup.o bn_mp_exch.o \
-bn_mp_export.o bn_mp_expt_d.o bn_mp_expt_d_ex.o bn_mp_exptmod.o bn_mp_exptmod_fast.o bn_mp_exteuclid.o \
-bn_mp_fread.o bn_mp_fwrite.o bn_mp_gcd.o bn_mp_get_int.o bn_mp_get_long.o bn_mp_get_long_long.o \
-bn_mp_grow.o bn_mp_import.o bn_mp_init.o bn_mp_init_copy.o bn_mp_init_multi.o bn_mp_init_set.o \
-bn_mp_init_set_int.o bn_mp_init_size.o bn_mp_invmod.o bn_mp_invmod_slow.o bn_mp_is_square.o \
-bn_mp_jacobi.o bn_mp_karatsuba_mul.o bn_mp_karatsuba_sqr.o bn_mp_lcm.o bn_mp_lshd.o bn_mp_mod_2d.o \
-bn_mp_mod.o bn_mp_mod_d.o bn_mp_montgomery_calc_normalization.o bn_mp_montgomery_reduce.o \
-bn_mp_montgomery_setup.o bn_mp_mul_2.o bn_mp_mul_2d.o bn_mp_mul.o bn_mp_mul_d.o bn_mp_mulmod.o bn_mp_neg.o \
-bn_mp_n_root.o bn_mp_n_root_ex.o bn_mp_or.o bn_mp_prime_fermat.o bn_mp_prime_is_divisible.o \
+bn_mp_cmp_mag.o bn_mp_cnt_lsb.o bn_mp_complement.o bn_mp_copy.o bn_mp_count_bits.o bn_mp_div.o \
+bn_mp_div_2.o bn_mp_div_2d.o bn_mp_div_3.o bn_mp_div_d.o bn_mp_dr_is_modulus.o bn_mp_dr_reduce.o \
+bn_mp_dr_setup.o bn_mp_exch.o bn_mp_export.o bn_mp_expt_d.o bn_mp_expt_d_ex.o bn_mp_exptmod.o \
+bn_mp_exptmod_fast.o bn_mp_exteuclid.o bn_mp_fread.o bn_mp_fwrite.o bn_mp_gcd.o bn_mp_get_bit.o \
+bn_mp_get_int.o bn_mp_get_long.o bn_mp_get_long_long.o bn_mp_grow.o bn_mp_import.o bn_mp_init.o \
+bn_mp_init_copy.o bn_mp_init_multi.o bn_mp_init_set.o bn_mp_init_set_int.o bn_mp_init_size.o \
+bn_mp_invmod.o bn_mp_invmod_slow.o bn_mp_is_square.o bn_mp_jacobi.o bn_mp_karatsuba_mul.o \
+bn_mp_karatsuba_sqr.o bn_mp_kronecker.o bn_mp_lcm.o bn_mp_lshd.o bn_mp_mod.o bn_mp_mod_2d.o bn_mp_mod_d.o \
+bn_mp_montgomery_calc_normalization.o bn_mp_montgomery_reduce.o bn_mp_montgomery_setup.o bn_mp_mul.o \
+bn_mp_mul_2.o bn_mp_mul_2d.o bn_mp_mul_d.o bn_mp_mulmod.o bn_mp_n_root.o bn_mp_n_root_ex.o bn_mp_neg.o \
+bn_mp_or.o bn_mp_prime_fermat.o bn_mp_prime_frobenius_underwood.o bn_mp_prime_is_divisible.o \
 bn_mp_prime_is_prime.o bn_mp_prime_miller_rabin.o bn_mp_prime_next_prime.o \
-bn_mp_prime_rabin_miller_trials.o bn_mp_prime_random_ex.o bn_mp_radix_size.o bn_mp_radix_smap.o \
-bn_mp_rand.o bn_mp_read_radix.o bn_mp_read_signed_bin.o bn_mp_read_unsigned_bin.o bn_mp_reduce_2k.o \
-bn_mp_reduce_2k_l.o bn_mp_reduce_2k_setup.o bn_mp_reduce_2k_setup_l.o bn_mp_reduce.o \
-bn_mp_reduce_is_2k.o bn_mp_reduce_is_2k_l.o bn_mp_reduce_setup.o bn_mp_rshd.o bn_mp_set.o bn_mp_set_int.o \
-bn_mp_set_long.o bn_mp_set_long_long.o bn_mp_shrink.o bn_mp_signed_bin_size.o bn_mp_sqr.o bn_mp_sqrmod.o \
-bn_mp_sqrt.o bn_mp_sqrtmod_prime.o bn_mp_sub.o bn_mp_sub_d.o bn_mp_submod.o bn_mp_toom_mul.o \
-bn_mp_toom_sqr.o bn_mp_toradix.o bn_mp_toradix_n.o bn_mp_to_signed_bin.o bn_mp_to_signed_bin_n.o \
-bn_mp_to_unsigned_bin.o bn_mp_to_unsigned_bin_n.o bn_mp_unsigned_bin_size.o bn_mp_xor.o bn_mp_zero.o \
+bn_mp_prime_rabin_miller_trials.o bn_mp_prime_random_ex.o bn_mp_prime_strong_lucas_selfridge.o \
+bn_mp_radix_size.o bn_mp_radix_smap.o bn_mp_rand.o bn_mp_read_radix.o bn_mp_read_signed_bin.o \
+bn_mp_read_unsigned_bin.o bn_mp_reduce.o bn_mp_reduce_2k.o bn_mp_reduce_2k_l.o bn_mp_reduce_2k_setup.o \
+bn_mp_reduce_2k_setup_l.o bn_mp_reduce_is_2k.o bn_mp_reduce_is_2k_l.o bn_mp_reduce_setup.o bn_mp_rshd.o \
+bn_mp_set.o bn_mp_set_int.o bn_mp_set_long.o bn_mp_set_long_long.o bn_mp_shrink.o bn_mp_signed_bin_size.o \
+bn_mp_sqr.o bn_mp_sqrmod.o bn_mp_sqrt.o bn_mp_sqrtmod_prime.o bn_mp_sub.o bn_mp_sub_d.o bn_mp_submod.o \
+bn_mp_tc_and.o bn_mp_tc_div_2d.o bn_mp_tc_or.o bn_mp_tc_xor.o bn_mp_to_signed_bin.o \
+bn_mp_to_signed_bin_n.o bn_mp_to_unsigned_bin.o bn_mp_to_unsigned_bin_n.o bn_mp_toom_mul.o \
+bn_mp_toom_sqr.o bn_mp_toradix.o bn_mp_toradix_n.o bn_mp_unsigned_bin_size.o bn_mp_xor.o bn_mp_zero.o \
 bn_prime_tab.o bn_reverse.o bn_s_mp_add.o bn_s_mp_exptmod.o bn_s_mp_mul_digs.o bn_s_mp_mul_high_digs.o \
-bn_s_mp_sqr.o bn_s_mp_sub.o
+bn_s_mp_sqr.o bn_s_mp_sub.o bncore.o
 
 #END_INS
 
@@ -73,20 +75,20 @@ $(LIBNAME):  $(OBJECTS)
 # So far I've seen improvements in the MP math
 profiled:
 	make CFLAGS="$(CFLAGS) -fprofile-arcs -DTESTING" timing
-	./ltmtest
-	rm -f *.a *.o ltmtest
+	./timing
+	rm -f *.a *.o timing
 	make CFLAGS="$(CFLAGS) -fbranch-probabilities"
 
 #make a single object profiled library
 profiled_single:
 	perl gen.pl
 	$(CC) $(CFLAGS) -fprofile-arcs -DTESTING -c mpi.c -o mpi.o
-	$(CC) $(CFLAGS) -DTESTING -DTIMER demo/timing.c mpi.o -lgcov -o ltmtest
-	./ltmtest
-	rm -f *.o ltmtest
+	$(CC) $(CFLAGS) -DTESTING -DTIMER demo/timing.c mpi.o -lgcov -o timing
+	./timing
+	rm -f *.o timing
 	$(CC) $(CFLAGS) -fbranch-probabilities -DTESTING -c mpi.c -o mpi.o
 	$(AR) $(ARFLAGS) $(LIBNAME) mpi.o
-	$(RANLIB) $(LIBNAME)	
+	ranlib $(LIBNAME)
 
 install: $(LIBNAME)
 	install -d $(DESTDIR)$(LIBPATH)
@@ -108,12 +110,8 @@ test_standalone: $(LIBNAME) demo/demo.o
 mtest:
 	cd mtest ; $(CC) $(CFLAGS) -O0 mtest.c $(LFLAGS) -o mtest
 
-travis_mtest: test mtest
-	@ for i in `seq 1 10` ; do sleep 500 && echo alive; done &
-	./mtest/mtest 666666 | ./test > test.log
-
-timing: $(LIBNAME)
-	$(CC) $(CFLAGS) -DTIMER demo/timing.c $(LIBNAME) $(LFLAGS) -o ltmtest
+timing: $(LIBNAME) demo/timing.c
+	$(CC) $(CFLAGS) -DTIMER demo/timing.c $(LIBNAME) $(LFLAGS) -o timing
 
 # You have to create a file .coveralls.yml with the content "repo_token: <the token>"
 # in the base folder to be able to submit to coveralls
@@ -128,11 +126,12 @@ pretty:
 
 .PHONY: pre_gen
 pre_gen:
+	mkdir -p pre_gen
 	perl gen.pl
 	sed -e 's/[[:blank:]]*$$//' mpi.c > pre_gen/mpi.c
 	rm mpi.c
 
-zipup: clean pre_gen new_file manual poster docs
+zipup: clean astyle new_file manual poster docs
 	@# Update the index, so diff-index won't fail in case the pdf has been created.
 	@#   As the pdf creation modifies the tex files, git sometimes detects the
 	@#   modified files, but misses that it's put back to its original version.
@@ -141,10 +140,15 @@ zipup: clean pre_gen new_file manual poster docs
 	rm -rf libtommath-$(VERSION) ltm-$(VERSION).*
 	@# files/dirs excluded from "git archive" are defined in .gitattributes
 	git archive --format=tar --prefix=libtommath-$(VERSION)/ HEAD | tar x
+	@echo 'fixme check'
+	-@(find libtommath-$(VERSION)/ -type f | xargs grep 'FIXM[E]') && echo '############## BEWARE: the "fixme" marker was found !!! ##############' || true
 	mkdir -p libtommath-$(VERSION)/doc
 	cp doc/bn.pdf doc/tommath.pdf doc/poster.pdf libtommath-$(VERSION)/doc/
+	$(MAKE) -C libtommath-$(VERSION)/ pre_gen
 	tar -c libtommath-$(VERSION)/ | xz -6e -c - > ltm-$(VERSION).tar.xz
 	zip -9rq ltm-$(VERSION).zip libtommath-$(VERSION)
+	cp doc/bn.pdf bn-$(VERSION).pdf
+	cp doc/tommath.pdf tommath-$(VERSION).pdf
 	rm -rf libtommath-$(VERSION)
 	gpg -b -a ltm-$(VERSION).tar.xz
 	gpg -b -a ltm-$(VERSION).zip
@@ -154,4 +158,7 @@ new_file:
 	perl dep.pl
 
 perlcritic:
-	perlcritic *.pl
+	perlcritic *.pl doc/*.pl
+
+astyle:
+	astyle --options=astylerc $(OBJECTS:.o=.c) tommath*.h demo/*.c etc/*.c mtest/mtest.c

libtommath/README.md

@@ -1,15 +1,25 @@
-[![Build Status - master](https://travis-ci.org/libtom/libtommath.png?branch=master)](https://travis-ci.org/libtom/libtommath)
+# libtommath
 
-[![Build Status - develop](https://travis-ci.org/libtom/libtommath.png?branch=develop)](https://travis-ci.org/libtom/libtommath)
+This is the git repository for [LibTomMath](http://www.libtom.net/LibTomMath/), a free open source portable number theoretic multiple-precision integer (MPI) library written entirely in C.
 
-This is the git repository for [LibTomMath](http://www.libtom.org/), a free open source portable number theoretic multiple-precision integer (MPI) library written entirely in C.
+## Build Status
+
+master: [![Build Status](https://api.travis-ci.org/libtom/libtommath.png?branch=master)](https://travis-ci.org/libtom/libtommath)
+
+develop: [![Build Status](https://api.travis-ci.org/libtom/libtommath.png?branch=develop)](https://travis-ci.org/libtom/libtommath)
+
+API/ABI changes: [check here](https://abi-laboratory.pro/tracker/timeline/libtommath/)
+
+## Summary
 
 The `develop` branch contains the in-development version. Stable releases are tagged.
 
 Documentation is built from the LaTeX file `bn.tex`. There is also limited documentation in `tommath.h`. There is also a document, `tommath.pdf`, which describes the goals of the project and many of the algorithms used.
 
 The project can be build by using `make`. Along with the usual `make`, `make clean` and `make install`, there are several other build targets, see the makefile for details. There are also makefiles for certain specific platforms.
 
+## Testing
+
 Tests are located in `demo/` and can be built in two flavors.
 * `make test` creates a test binary that is intended to be run against `mtest`. `mtest` can be built with `make mtest` and test execution is done like `./mtest/mtest | ./test`. `mtest` is creating test vectors using an alternative MPI library and `test` is consuming these vectors to verify correct behavior of ltm
 * `make test_standalone` creates a stand-alone test binary that executes several test routines.