error: conflicting types for 'fmul' #9

pyperanger · 2018-10-17T01:55:46Z

I don't know if this error is part of the module, but i resolved this issue change variable name and the calls.

OS => ArchLinux

FIX
grep -rl "fmul" ./ | xargs sed -i 's/fmul/fixedvar/g'

ERROR
curve25519-donna-c64.c:99:1: error: conflicting types for 'fmul' fmull(felem output, const felem in2, const felem in) { ^~~~~ In file included from /usr/lib/perl5/5.28/core_perl/CORE/perl.h:2026, from Curve25519.xs:3: /usr/include/bits/mathcalls-narrow.h:30:20: note: previous declaration of 'fmul' was here __MATHCALL_NARROW (__MATHCALL_NAME (mul), __MATHCALL_REDIR_NAME (mul), 2); ^~~~~~~~~~~~~~~ make: *** [Makefile:334: Curve25519.o] Error 1

PS => I found this error during execution of Net::Perl::SSH module.

The text was updated successfully, but these errors were encountered:

pwr22 · 2018-12-11T11:06:03Z

I've also hit this on Manjaro - installing 0.05 works fine

andk · 2018-12-16T11:41:23Z

Also broken on Ubuntu cosmic (18.10).

pyperanger · 2018-12-17T13:30:09Z

Try this in the directory before compile: grep -rl "fmul" ./ | xargs sed -i 's/fmul/fixedvar/g'

I made a fork with the fix-> https://github.com/pyperanger/crypt-curve25519

nailara-technologies · 2018-12-19T13:11:24Z

same here on debian 686-pae, the conflict was in /usr/include/i386-linux-gnu/bits/mathcalls-narrow.h
and renaming fmul works as well

With thanks to Mike Benson for trawling the upstream bug reports: ajgb/crypt-curve25519#9 (comment) No revision bump as the package does not install without the fix. This uses a patch rather than a sed statement, as the latter could fail silently. Closes: https://bugs.gentoo.org/669436 Suggested-By: Mike Benson <mike@kambe.com.au> Signed-Off-By: Tony Vroon <chainsaw@gentoo.org> Package-Manager: Portage-2.3.66, Repoman-2.3.11

nailara-technologies · 2019-07-10T09:27:39Z

greetinx,

...now it no longer compiles on debian 64-bit as well... until this issue is fixed, the unmodified original has unfortunately become useless for direct installation from either CPAN or Github..

..is there any chance the author will fix the variable name, or can we consider Crypt::Curve25519 as abandoned / no longer maintained?

apologies for being so 'blunt' about this..

Taeki

pyperanger · 2019-07-12T19:03:09Z

Send here the new error code... ajgb is away, but the community is always here ;)

eserte · 2019-09-28T14:44:03Z

CPAN.pm users may use the following distropref file for automatic patching:
https://github.com/eserte/srezic-cpan-distroprefs/blob/master/Crypt-Curve25519.yml

gitcnd · 2020-03-03T03:38:47Z

See attachment "patch_Curve25519.patch.txt".


# cd /root/.cpan/build/Crypt-Curve25519-0.06-1
# patch < patch_Curve25519.patch.txt 
patching file curve25519-donna.c
patching file curve25519-donna-c64.c

# perl Makefile.PL 
Generating a Unix-style Makefile
Writing Makefile for Crypt::Curve25519
Writing MYMETA.yml and MYMETA.json

# make test

[snip - works now]

# make install
"/usr/bin/perl" -MExtUtils::Command::MM -e 'cp_nonempty' -- Curve25519.bs blib/arch/auto/Crypt/Curve25519/Curve25519.bs 644
Manifying 1 pod document
Files found in blib/arch: installing files in blib/lib into architecture dependent library tree
Installing /usr/local/lib64/perl5/auto/Crypt/Curve25519/Curve25519.so
Installing /usr/local/lib64/perl5/Crypt/Curve25519.pm
Installing /usr/local/share/man/man3/Crypt::Curve25519.3pm
Appending installation info to /usr/lib64/perl5/perllocal.pod

[patch_Curve25519.patch.txt](https://github.com/ajgb/crypt-curve25519/files/4279114/patch_Curve25519.patch.txt)

gitcnd · 2020-03-03T03:39:44Z

diff -Naur './curve25519-donna-200303.c' './curve25519-donna.c'
--- /home/cnd/keep/curve25519-donna-200303.c	2017-04-05 23:25:50.000000000 +0000
+++ ./curve25519-donna.c	2020-03-03 03:28:00.525177420 +0000
@@ -325,7 +325,7 @@
  * reduced coefficient.
  */
 static void
-fmul(limb *output, const limb *in, const limb *in2) {
+fmulcnd(limb *output, const limb *in, const limb *in2) {
   limb t[19];
   fproduct(t, in, in2);
   freduce_degree(t);
@@ -661,54 +661,54 @@
   /* 2 */ fsquare(z2,z);
   /* 4 */ fsquare(t1,z2);
   /* 8 */ fsquare(t0,t1);
-  /* 9 */ fmul(z9,t0,z);
-  /* 11 */ fmul(z11,z9,z2);
+  /* 9 */ fmulcnd(z9,t0,z);
+  /* 11 */ fmulcnd(z11,z9,z2);
   /* 22 */ fsquare(t0,z11);
-  /* 2^5 - 2^0 = 31 */ fmul(z2_5_0,t0,z9);
+  /* 2^5 - 2^0 = 31 */ fmulcnd(z2_5_0,t0,z9);
 
   /* 2^6 - 2^1 */ fsquare(t0,z2_5_0);
   /* 2^7 - 2^2 */ fsquare(t1,t0);
   /* 2^8 - 2^3 */ fsquare(t0,t1);
   /* 2^9 - 2^4 */ fsquare(t1,t0);
   /* 2^10 - 2^5 */ fsquare(t0,t1);
-  /* 2^10 - 2^0 */ fmul(z2_10_0,t0,z2_5_0);
+  /* 2^10 - 2^0 */ fmulcnd(z2_10_0,t0,z2_5_0);
 
   /* 2^11 - 2^1 */ fsquare(t0,z2_10_0);
   /* 2^12 - 2^2 */ fsquare(t1,t0);
   /* 2^20 - 2^10 */ for (i = 2;i < 10;i += 2) { fsquare(t0,t1); fsquare(t1,t0); }
-  /* 2^20 - 2^0 */ fmul(z2_20_0,t1,z2_10_0);
+  /* 2^20 - 2^0 */ fmulcnd(z2_20_0,t1,z2_10_0);
 
   /* 2^21 - 2^1 */ fsquare(t0,z2_20_0);
   /* 2^22 - 2^2 */ fsquare(t1,t0);
   /* 2^40 - 2^20 */ for (i = 2;i < 20;i += 2) { fsquare(t0,t1); fsquare(t1,t0); }
-  /* 2^40 - 2^0 */ fmul(t0,t1,z2_20_0);
+  /* 2^40 - 2^0 */ fmulcnd(t0,t1,z2_20_0);
 
   /* 2^41 - 2^1 */ fsquare(t1,t0);
   /* 2^42 - 2^2 */ fsquare(t0,t1);
   /* 2^50 - 2^10 */ for (i = 2;i < 10;i += 2) { fsquare(t1,t0); fsquare(t0,t1); }
-  /* 2^50 - 2^0 */ fmul(z2_50_0,t0,z2_10_0);
+  /* 2^50 - 2^0 */ fmulcnd(z2_50_0,t0,z2_10_0);
 
   /* 2^51 - 2^1 */ fsquare(t0,z2_50_0);
   /* 2^52 - 2^2 */ fsquare(t1,t0);
   /* 2^100 - 2^50 */ for (i = 2;i < 50;i += 2) { fsquare(t0,t1); fsquare(t1,t0); }
-  /* 2^100 - 2^0 */ fmul(z2_100_0,t1,z2_50_0);
+  /* 2^100 - 2^0 */ fmulcnd(z2_100_0,t1,z2_50_0);
 
   /* 2^101 - 2^1 */ fsquare(t1,z2_100_0);
   /* 2^102 - 2^2 */ fsquare(t0,t1);
   /* 2^200 - 2^100 */ for (i = 2;i < 100;i += 2) { fsquare(t1,t0); fsquare(t0,t1); }
-  /* 2^200 - 2^0 */ fmul(t1,t0,z2_100_0);
+  /* 2^200 - 2^0 */ fmulcnd(t1,t0,z2_100_0);
 
   /* 2^201 - 2^1 */ fsquare(t0,t1);
   /* 2^202 - 2^2 */ fsquare(t1,t0);
   /* 2^250 - 2^50 */ for (i = 2;i < 50;i += 2) { fsquare(t0,t1); fsquare(t1,t0); }
-  /* 2^250 - 2^0 */ fmul(t0,t1,z2_50_0);
+  /* 2^250 - 2^0 */ fmulcnd(t0,t1,z2_50_0);
 
   /* 2^251 - 2^1 */ fsquare(t1,t0);
   /* 2^252 - 2^2 */ fsquare(t0,t1);
   /* 2^253 - 2^3 */ fsquare(t1,t0);
   /* 2^254 - 2^4 */ fsquare(t0,t1);
   /* 2^255 - 2^5 */ fsquare(t1,t0);
-  /* 2^255 - 21 */ fmul(out,t1,z11);
+  /* 2^255 - 21 */ fmulcnd(out,t1,z11);
 }
 
 int curve25519_donna(u8 *, const u8 *, const u8 *);
@@ -727,7 +727,7 @@
   fexpand(bp, basepoint);
   cmult(x, z, e, bp);
   crecip(zmone, z);
-  fmul(z, x, zmone);
+  fmulcnd(z, x, zmone);
   freduce_coefficients(z);
   fcontract(mypublic, z);
   return 0;
diff -Naur './curve25519-donna-c64-200303.c' './curve25519-donna-c64.c'
--- /home/cnd/keep/curve25519-donna-c64-200303.c	2017-04-05 23:25:50.000000000 +0000
+++ ./curve25519-donna-c64.c	2020-03-03 03:28:21.942742166 +0000
@@ -96,7 +96,7 @@
  * On return, output[i] < 2**52
  */
 static inline void force_inline
-fmul(felem output, const felem in2, const felem in) {
+fmulcnd(felem output, const felem in2, const felem in) {
   uint128_t t[5];
   limb r0,r1,r2,r3,r4,s0,s1,s2,s3,s4,c;
 
@@ -305,22 +305,22 @@
   memcpy(origxprime, xprime, sizeof(limb) * 5);
   fsum(xprime, zprime);
   fdifference_backwards(zprime, origxprime);
-  fmul(xxprime, xprime, z);
-  fmul(zzprime, x, zprime);
+  fmulcnd(xxprime, xprime, z);
+  fmulcnd(zzprime, x, zprime);
   memcpy(origxprime, xxprime, sizeof(limb) * 5);
   fsum(xxprime, zzprime);
   fdifference_backwards(zzprime, origxprime);
   fsquare_times(x3, xxprime, 1);
   fsquare_times(zzzprime, zzprime, 1);
-  fmul(z3, zzzprime, qmqp);
+  fmulcnd(z3, zzzprime, qmqp);
 
   fsquare_times(xx, x, 1);
   fsquare_times(zz, z, 1);
-  fmul(x2, xx, zz);
+  fmulcnd(x2, xx, zz);
   fdifference_backwards(zz, xx);  // does zz = xx - zz
   fscalar_product(zzz, zz, 121665);
   fsum(zzz, xx);
-  fmul(z2, zz, zzz);
+  fmulcnd(z2, zz, zzz);
 }
 
 // -----------------------------------------------------------------------------
@@ -405,26 +405,26 @@
 
   /* 2 */ fsquare_times(a, z, 1); // a = 2
   /* 8 */ fsquare_times(t0, a, 2);
-  /* 9 */ fmul(b, t0, z); // b = 9
-  /* 11 */ fmul(a, b, a); // a = 11
+  /* 9 */ fmulcnd(b, t0, z); // b = 9
+  /* 11 */ fmulcnd(a, b, a); // a = 11
   /* 22 */ fsquare_times(t0, a, 1);
-  /* 2^5 - 2^0 = 31 */ fmul(b, t0, b);
+  /* 2^5 - 2^0 = 31 */ fmulcnd(b, t0, b);
   /* 2^10 - 2^5 */ fsquare_times(t0, b, 5);
-  /* 2^10 - 2^0 */ fmul(b, t0, b);
+  /* 2^10 - 2^0 */ fmulcnd(b, t0, b);
   /* 2^20 - 2^10 */ fsquare_times(t0, b, 10);
-  /* 2^20 - 2^0 */ fmul(c, t0, b);
+  /* 2^20 - 2^0 */ fmulcnd(c, t0, b);
   /* 2^40 - 2^20 */ fsquare_times(t0, c, 20);
-  /* 2^40 - 2^0 */ fmul(t0, t0, c);
+  /* 2^40 - 2^0 */ fmulcnd(t0, t0, c);
   /* 2^50 - 2^10 */ fsquare_times(t0, t0, 10);
-  /* 2^50 - 2^0 */ fmul(b, t0, b);
+  /* 2^50 - 2^0 */ fmulcnd(b, t0, b);
   /* 2^100 - 2^50 */ fsquare_times(t0, b, 50);
-  /* 2^100 - 2^0 */ fmul(c, t0, b);
+  /* 2^100 - 2^0 */ fmulcnd(c, t0, b);
   /* 2^200 - 2^100 */ fsquare_times(t0, c, 100);
-  /* 2^200 - 2^0 */ fmul(t0, t0, c);
+  /* 2^200 - 2^0 */ fmulcnd(t0, t0, c);
   /* 2^250 - 2^50 */ fsquare_times(t0, t0, 50);
-  /* 2^250 - 2^0 */ fmul(t0, t0, b);
+  /* 2^250 - 2^0 */ fmulcnd(t0, t0, b);
   /* 2^255 - 2^5 */ fsquare_times(t0, t0, 5);
-  /* 2^255 - 21 */ fmul(out, t0, a);
+  /* 2^255 - 21 */ fmulcnd(out, t0, a);
 }
 
 int curve25519_donna(u8 *, const u8 *, const u8 *);
@@ -443,7 +443,7 @@
   fexpand(bp, basepoint);
   cmult(x, z, e, bp);
   crecip(zmone, z);
-  fmul(z, x, zmone);
+  fmulcnd(z, x, zmone);
   fcontract(mypublic, z);
   return 0;
 }

nicomen · 2020-04-19T19:29:33Z

I got this error now on Ubuntu 19.10 – needed to build Net::SFTP – is there some way I can help getting this fix to CPAN?

eserte · 2020-04-19T19:46:24Z

@pyperanger : Would you like to take over the module?
@nicomen : as I wrote some comments above --- for CPAN.pm users there's a solution which works out of the box without manual patching, so maybe it's worth to switch back from cpanminus to good old CPAN.pm...

nicomen · 2020-04-19T20:08:04Z

@eserte Thanks, I investigated the distroprefs feature, but it sems to not be implemented in cpanminus (nor CPANPLUS), so it's not viable for our deployments, but I will do a manual patch with cpan locally.

Tux · 2020-06-21T16:22:50Z

FWIW the posted patch works for me

Linux 5.7.1-1-default [openSUSE Tumbleweed 20200618]
This is perl 5, version 32, subversion 0 (v5.32.0) built for x86_64-linux-thread-multi-ld
gcc (SUSE Linux) 10.1.1 20200507 [revision dd38686d9c810cecbaa80bb82ed91caaa58ad635]

duttaANI · 2022-03-06T19:29:28Z

same here on debian 686-pae, the conflict was in /usr/include/i386-linux-gnu/bits/mathcalls-narrow.h and renaming fmul works as well

Thanks a lot, this approach worked . The issue was resolved by opening the file mathcalls-narrow.h and removing the below line -

__MATHCALL_NARROW (__MATHCALL_NAME (mul), __MATHCALL_REDIR_NAME (mul), 2)

pyperanger closed this as completed Mar 3, 2020

willsheppard mentioned this issue Feb 1, 2021

New release w/compile fixes wanted #13

Open

duttaANI mentioned this issue Mar 6, 2022

Issue while using Net::SSH::Perl lkinley/Net-SSH-Perl#18

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

error: conflicting types for 'fmul' #9

error: conflicting types for 'fmul' #9

pyperanger commented Oct 17, 2018 •

edited

Loading

pwr22 commented Dec 11, 2018

andk commented Dec 16, 2018

pyperanger commented Dec 17, 2018 •

edited

Loading

nailara-technologies commented Dec 19, 2018 •

edited

Loading

nailara-technologies commented Jul 10, 2019

pyperanger commented Jul 12, 2019

eserte commented Sep 28, 2019

gitcnd commented Mar 3, 2020

gitcnd commented Mar 3, 2020

nicomen commented Apr 19, 2020

eserte commented Apr 19, 2020

nicomen commented Apr 19, 2020

Tux commented Jun 21, 2020

duttaANI commented Mar 6, 2022

error: conflicting types for 'fmul' #9

error: conflicting types for 'fmul' #9

Comments

pyperanger commented Oct 17, 2018 • edited Loading

pwr22 commented Dec 11, 2018

andk commented Dec 16, 2018

pyperanger commented Dec 17, 2018 • edited Loading

nailara-technologies commented Dec 19, 2018 • edited Loading

nailara-technologies commented Jul 10, 2019

pyperanger commented Jul 12, 2019

eserte commented Sep 28, 2019

gitcnd commented Mar 3, 2020

gitcnd commented Mar 3, 2020

nicomen commented Apr 19, 2020

eserte commented Apr 19, 2020

nicomen commented Apr 19, 2020

Tux commented Jun 21, 2020

duttaANI commented Mar 6, 2022

pyperanger commented Oct 17, 2018 •

edited

Loading

pyperanger commented Dec 17, 2018 •

edited

Loading

nailara-technologies commented Dec 19, 2018 •

edited

Loading