add feature detection macros MP_DEFINED and MP_HAS #214

minad · 2019-04-08T21:50:13Z

additionally, make bn_mp_mul private

sjaeckel · 2019-04-09T05:14:12Z

@karel-m can you please try this out on your various compilers?

sjaeckel · 2019-04-09T06:16:10Z

@karel-m can you please try this out on your various compilers?

and most importantly, please also try it out with the following patch

diff --git a/tommath_superclass.h b/tommath_superclass.h
index c9090a8..8761f81 100644
--- a/tommath_superclass.h
+++ b/tommath_superclass.h
@@ -6,3 +6,5 @@
 /* default ... include all MPI */
+#ifndef LTM_NOT_ALL
 #define LTM_ALL
+#endif

compiling it with CFLAGS="-DLTM_NOT_ALL -DSC_RSA_1"

and check that the libraries created are different

@minad ~~can you please include that patch in the branch?~~ thx!

karel-m · 2019-04-09T11:50:15Z

1/ gcc 7.4.0 - Target: x86_64-pc-cygwin (the following warning does not appear with develop) do we really need -Wundef?

cc -c -I./ -Wall -Wsign-compare -Wextra -Wshadow -Wsystem-headers -Wdeclaration-after-statement -Wbad-function-cast -Wcast-align -Wstrict-prototypes -Wpointer-arith -Wundef -O3 -funroll-loops -fomit-frame-pointer bn_mp_add.c -o bn_mp_add.o
In file included from /usr/include/sys/config.h:5:0,
                 from /usr/include/_ansi.h:11,
                 from /usr/include/stdlib.h:11,
                 from tommath.h:7,
                 from tommath_private.h:7,
                 from bn_mp_add.c:1:
/usr/include/sys/features.h:274:36: warning: "_XOPEN_SOURCE" is not defined, evaluates to 0 [-Wundef]
 #if defined(_LARGEFILE_SOURCE) || (_XOPEN_SOURCE - 0) >= 500
                                    ^~~~~~~~~~~~~
/usr/include/sys/features.h:308:6: warning: "_XOPEN_SOURCE" is not defined, evaluates to 0 [-Wundef]
 #if (_XOPEN_SOURCE - 0) >= 700
      ^~~~~~~~~~~~~
/usr/include/sys/features.h:310:8: warning: "_XOPEN_SOURCE" is not defined, evaluates to 0 [-Wundef]
 #elif (_XOPEN_SOURCE - 0) >= 600
        ^~~~~~~~~~~~~
/usr/include/sys/features.h:312:8: warning: "_XOPEN_SOURCE" is not defined, evaluates to 0 [-Wundef]
 #elif (_XOPEN_SOURCE - 0) >= 500
        ^~~~~~~~~~~~~
/usr/include/sys/features.h:322:5: warning: "_FORTIFY_SOURCE" is not defined, evaluates to 0 [-Wundef]
 #if _FORTIFY_SOURCE > 0 && !defined(__cplusplus) && !defined(__lint__) && \
     ^~~~~~~~~~~~~~~

karel-m · 2019-04-09T11:53:46Z

2/ HP-UX + C compiler by HP

$ make -f makefile.unix
        [...]
        cc -I. -O2 -c bn_mp_div_d.c -o bn_mp_div_d.o
cpp: "bn_mp_div_d.c", line 60: warning 2006: Parameter holes filled with a null string.
        [...]
        cc -I. -O2 -c bn_mp_exptmod.c -o bn_mp_exptmod.o
cpp: "bn_mp_exptmod.c", line 25: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod.c", line 55: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod.c", line 55: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod.c", line 55: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod.c", line 61: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod.c", line 64: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod.c", line 69: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod.c", line 71: warning 2006: Parameter holes filled with a null string.
        cc -I. -O2 -c bn_mp_exptmod_fast.c -o bn_mp_exptmod_fast.o
cpp: "bn_mp_exptmod_fast.c", line 73: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod_fast.c", line 84: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod_fast.c", line 88: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod_fast.c", line 96: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod_fast.c", line 104: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod_fast.c", line 104: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_exptmod_fast.c", line 128: warning 2006: Parameter holes filled with a null string.
        [...]
        cc -I. -O2 -c bn_mp_invmod.c -o bn_mp_invmod.o
cpp: "bn_mp_invmod.c", line 15: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_invmod.c", line 19: warning 2006: Parameter holes filled with a null string.
        [...]
        cc -I. -O2 -c bn_mp_mul.c -o bn_mp_mul.o
cpp: "bn_mp_mul.c", line 15: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_mul.c", line 28: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_mul.c", line 31: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_mul.c", line 34: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_mul.c", line 45: warning 2006: Parameter holes filled with a null string.
        [...]
        cc -I. -O2 -c bn_mp_reduce.c -o bn_mp_reduce.o
cpp: "bn_mp_reduce.c", line 28: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_reduce.c", line 32: warning 2006: Parameter holes filled with a null string.
        [...]
        cc -I. -O2 -c bn_mp_set_double.c -o bn_mp_set_double.o
cpp: "bn_mp_set_double.c", line 46: warning 2029: "mp_set_double implementation is only available on platforms with IEEE754 floating point format"
        [...]
        cc -I. -O2 -c bn_mp_sqr.c -o bn_mp_sqr.o
cpp: "bn_mp_sqr.c", line 11: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_sqr.c", line 14: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_sqr.c", line 17: warning 2006: Parameter holes filled with a null string.
cpp: "bn_mp_sqr.c", line 22: warning 2006: Parameter holes filled with a null string.
        [...]
        cc -I. -O2 -c bn_s_mp_mul_high_digs.c -o bn_s_mp_mul_high_digs.o
cpp: "bn_s_mp_mul_high_digs.c", line 18: warning 2006: Parameter holes filled with a null string.

minad · 2019-04-09T11:57:02Z

@karel-m We can activate Wundef together with the Wconversion warnings in #204. We will only do it without -Wsystem-headers. Then this should be a non-issue. Besides that, its ok?

karel-m · 2019-04-09T11:58:57Z

3/ AIX + IBM XL C compiler

$ make -f makefile.unix
        [...]
        cc -I. -O2 -c bn_mp_exptmod_fast.c -o bn_mp_exptmod_fast.o
"bn_mp_exptmod_fast.c", line 50.8: 1506-041 (E) The invocation of macro _MP_DEFINED3 contains fewer arguments than are required by the macro definition.
"bn_mp_exptmod_fast.c", line 96.34: 1506-041 (E) The invocation of macro _MP_DEFINED3 contains fewer arguments than are required by the macro definition.
        [...]
        cc -I. -O2 -c bn_mp_set_double.c -o bn_mp_set_double.o
"bn_mp_set_double.c", line 46.6: 1506-948 (W) #warning "mp_set_double implementation is only available on platforms with IEEE754 floating point format"
        [...]
        cc -I. -O2 -c bn_s_mp_exptmod.c -o bn_s_mp_exptmod.o
"bn_s_mp_exptmod.c", line 37.8: 1506-041 (E) The invocation of macro _MP_DEFINED3 contains fewer arguments than are required by the macro definition.

karel-m · 2019-04-09T11:59:19Z

4/ pretty old gcc 3.4.4

$ make V=1 CC=gcc
[...]
gcc -c -I./ -Wall -Wsign-compare -Wextra -Wshadow -Wsystem-headers -Wdeclaration-after-statement -Wbad-function-cast -Wcast-align -Wstrict-prototypes -Wpointer-arith -Wundef -O3 -funroll-loops -fomit-frame-pointer bn_mp_exptmod_fast.c -o bn_mp_exptmod_fast.o
bn_mp_exptmod_fast.c: In function `mp_exptmod_fast':
bn_mp_exptmod_fast.c:30: warning: declaration of 'x' shadows a previous local
bn_mp_exptmod_fast.c:24: warning: shadowed declaration is here
[...]
gcc -c -I./ -Wall -Wsign-compare -Wextra -Wshadow -Wsystem-headers -Wdeclaration-after-statement -Wbad-function-cast -Wcast-align -Wstrict-prototypes -Wpointer-arith -Wundef -O3 -funroll-loops -fomit-frame-pointer bn_mp_prime_strong_lucas_selfridge.c -o bn_mp_prime_strong_lucas_selfridge.o
bn_mp_prime_strong_lucas_selfridge.c: In function `mp_prime_strong_lucas_selfridge':
bn_mp_prime_strong_lucas_selfridge.c:113: warning: passing arg 3 of `mp_kronecker' from incompatible pointer type
[...]
gcc -c -I./ -Wall -Wsign-compare -Wextra -Wshadow -Wsystem-headers -Wdeclaration-after-statement -Wbad-function-cast -Wcast-align -Wstrict-prototypes -Wpointer-arith -Wundef -O3 -funroll-loops -fomit-frame-pointer bn_mp_set_double.c -o bn_mp_set_double.o
bn_mp_set_double.c:46:6: warning: #warning "mp_set_double implementation is only available on platforms with IEEE754 floating point format"
[...]
gcc -c -I./ -Wall -Wsign-compare -Wextra -Wshadow -Wsystem-headers -Wdeclaration-after-statement -Wbad-function-cast -Wcast-align -Wstrict-prototypes -Wpointer-arith -Wundef -O3 -funroll-loops -fomit-frame-pointer bn_s_mp_exptmod.c -o bn_s_mp_exptmod.o
bn_s_mp_exptmod.c: In function `s_mp_exptmod':
bn_s_mp_exptmod.c:17: warning: declaration of 'x' shadows a previous local
bn_s_mp_exptmod.c:16: warning: shadowed declaration is here
bn_s_mp_exptmod.c:17: warning: declaration of 'mu' shadows a previous local
bn_s_mp_exptmod.c:14: warning: shadowed declaration is here

karel-m · 2019-04-09T12:05:14Z

5/ Visual Studio 2017

1>------ Build started: Project: tommath, Configuration: Debug x64 ------
[...]
1>bn_s_mp_mul_high_digs.c
1>bn_s_mp_mul_high_digs.c(18): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
[...]
1>bn_mp_sqr.c
1>bn_mp_sqr.c(11): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_sqr.c(14): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_sqr.c(17): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_sqr.c(22): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
[...]
1>bn_mp_set_double.c
1>mp_set_double implementation is only available on platforms with IEEE754 floating point format
[...]
1>bn_mp_reduce.c
1>bn_mp_reduce.c(28): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_reduce.c(32): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
[...]
1>bn_mp_mul.c
1>bn_mp_mul.c(15): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_mul.c(28): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_mul.c(31): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_mul.c(34): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_mul.c(45): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
[...]
1>bn_mp_invmod.c
1>bn_mp_invmod.c(15): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_invmod.c(19): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
[...]
1>bn_mp_exptmod_fast.c
1>bn_mp_exptmod_fast.c(73): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_exptmod_fast.c(84): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_exptmod_fast.c(88): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_exptmod_fast.c(96): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_exptmod_fast.c(104): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_exptmod_fast.c(128): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_exptmod.c
1>bn_mp_exptmod.c(25): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_exptmod.c(55): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_exptmod.c(61): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_exptmod.c(64): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_exptmod.c(69): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
1>bn_mp_exptmod.c(71): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
[...]
1>bn_mp_div_d.c
1>bn_mp_div_d.c(60): warning C4003: not enough actual parameters for macro '_MP_DEFINED1'
[...]

1>libtommath_VS2008.vcxproj -> C:\cygwin64\home\bob\libtommath\MSVC_x64_Debug\tommath.lib
1>Done building project "libtommath_VS2008.vcxproj".
========== Build: 1 succeeded, 0 failed, 0 up-to-date, 0 skipped ==========

karel-m · 2019-04-09T12:06:38Z

All from me for now.

minad · 2019-04-09T12:09:43Z

Hmm, I think this requires C99. One could also try the GCC variadic argument extension args....

Enable together with Wconversion warnings in #204 with -Wno-system-headers!

minad · 2019-04-12T09:25:39Z

@sjaeckel What is your opinion about this?

minad mentioned this pull request Apr 8, 2019

do not expose mp_balance_mul, add MP_ENABLED feature detection #213

Closed

minad force-pushed the feature-detection branch from bf44351 to 96af296 Compare April 8, 2019 22:02

minad requested review from czurnieden and sjaeckel April 8, 2019 22:02

minad force-pushed the feature-detection branch from 96af296 to a1e9b6f Compare April 8, 2019 22:53

minad added 2 commits April 9, 2019 01:10

add feature detection macros MP_DEFINED and MP_HAS

d0d7a6d

additionally, make bn_mp_mul private

activate Wundef

6e7027a

minad force-pushed the feature-detection branch from a1e9b6f to 6e7027a Compare April 8, 2019 23:11

sjaeckel requested a review from karel-m April 9, 2019 09:00

add LTM_NOT_ALL

9501556

minad mentioned this pull request Apr 9, 2019

Rename internals #172

Merged

deactivate -Wundef

1566606

Enable together with Wconversion warnings in #204 with -Wno-system-headers!

minad force-pushed the feature-detection branch from f7136e0 to 1566606 Compare April 9, 2019 12:20

minad mentioned this pull request Apr 18, 2019

more linting #218

Merged

minad mentioned this pull request May 14, 2019

add feature detection macro MP_HAS (2) #262

Merged

minad closed this May 14, 2019

minad deleted the feature-detection branch May 14, 2019 15:55

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

add feature detection macros MP_DEFINED and MP_HAS #214

add feature detection macros MP_DEFINED and MP_HAS #214

minad commented Apr 8, 2019

sjaeckel commented Apr 9, 2019

sjaeckel commented Apr 9, 2019 •

edited

Loading

karel-m commented Apr 9, 2019

karel-m commented Apr 9, 2019

minad commented Apr 9, 2019 •

edited

Loading

karel-m commented Apr 9, 2019

karel-m commented Apr 9, 2019

karel-m commented Apr 9, 2019

karel-m commented Apr 9, 2019

minad commented Apr 9, 2019

minad commented Apr 12, 2019

add feature detection macros MP_DEFINED and MP_HAS #214

add feature detection macros MP_DEFINED and MP_HAS #214

Conversation

minad commented Apr 8, 2019

sjaeckel commented Apr 9, 2019

sjaeckel commented Apr 9, 2019 • edited Loading

karel-m commented Apr 9, 2019

karel-m commented Apr 9, 2019

minad commented Apr 9, 2019 • edited Loading

karel-m commented Apr 9, 2019

karel-m commented Apr 9, 2019

karel-m commented Apr 9, 2019

karel-m commented Apr 9, 2019

minad commented Apr 9, 2019

minad commented Apr 12, 2019

sjaeckel commented Apr 9, 2019 •

edited

Loading

minad commented Apr 9, 2019 •

edited

Loading