pulls in x86 headers on every arch:
--- src/smack1.c.orig 2017-06-06 03:59:39 UTC
@@ -115,6 +115,8 @@
static __inline__ unsigned long long __rdtsc(void)
see also: ports r533887 (which is wrong imo)
In file included from src/smack1.c:119:
In file included from /usr/lib/clang/11.0.0/include/x86intrin.h:15:
In file included from /usr/lib/clang/11.0.0/include/immintrin.h:15:
/usr/lib/clang/11.0.0/include/mmintrin.h:50:12: error: invalid conversion between vector type '__m64' (vector of 1 'long long' value) and integer type 'int' of different size
return (__m64)__builtin_ia32_vec_init_v2si(__i, 0);
/usr/lib/clang/11.0.0/include/mmintrin.h:129:12: error: invalid conversion between vector type '__m64' (vector of 1 'long long' value) and integer type 'int' of different size
return (__m64)__builtin_ia32_packsswb((__v4hi)__m1, (__v4hi)__m2);
broken commit: https://svnweb.freebsd.org/ports/head/security/masscan/files/patch-src_smack1.c?view=markup&pathrev=443915
Thanks for your notice.
I will look it.
(In reply to Mikael Urankar from comment #0)
That's interesting. I've just rebuilt all my ports which includes security/masscan which were successfully built on
FreeBSD 12.2-STABLE #0 r367125M: Fri Oct 30 02:52:58 AEDT 2020 i386 and amd64
We build for each of: c3-2, pentium3 and prescott; core2, haswell. masscan is functional on the c3-2 and haswell (which I tested just tested for this PR)
So for the brave, my cflags (used with clang 10.0.1, the default on 12.2 i386):
-O2 -pipe -g0 -ggdb0 -DSTRIP_FBSDID -UDEBUG -fno-math-errno -DOPENSSL_NO_SSL2 -DOPENSSL_NO_SSL3 -Wl,--strip-debug -Wl,--build-id=md5 -Wl,--hash-style=sysv -Wl,-m,elf_i386_fbsd -fPIC -DPIC -fPIE -pie -fomit-frame-pointer -fno-asynchronous-unwind-tables -Wl,-z,relro -Wl,-z,now -fno-common -march=c3-2 -Wno-format
-Wl,--strip-debug -Wl,--build-id=md5 -Wl,--hash-style=sysv -pie -Wl,-z,notext -z notext -z relro -z now
I should note, that I've made no change for masscan, these are my default options for all ports for each cpu_type. (Though some ports need hand-holding.)
(In reply to dewayne from comment #3)
I haven't tried on x86, only aarch64 and armv7, it fails here.
Apologies, I misinterpreted your !x86.
^Triage: Reporter is committer, assign accordingly.
A commit references this bug:
Date: Sat Nov 21 13:37:33 UTC 2020
New revision: 555980
security/masscan: fix build on !x86
Only include x86 header on x86
Approved by: firstname.lastname@example.org (maintainer timeout)