Your message dated Wed, 11 Sep 2019 06:04:11 +0000
with message-id <[email protected]>
and subject line Bug#939559: fixed in gcc-mingw-w64 21.5
has caused the Debian Bug report #939559,
regarding x86_64-w64-mingw32-g++-win32: program compiled with -march=native
crashes on same CPU
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact [email protected]
immediately.)
--
939559: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=939559
Debian Bug Tracking System
Contact [email protected] with problems
--- Begin Message ---
Package: g++-mingw-w64-x86-64
Version: 8.3.0-19+21.4
Severity: normal
File: /usr/bin/x86_64-w64-mingw32-g++-win32
Dear Maintainer,
I wrote this C++ source code and saved it in `bug.cpp`:
#include <cstdint>
#include <cstdio>
#include <cstdlib>
typedef int64_t int2 __attribute__ ((vector_size (16)));
typedef int64_t int4 __attribute__ ((vector_size (32)));
typedef int64_t int8 __attribute__ ((vector_size (64)));
typedef int64_t int16 __attribute__ ((vector_size (128)));
__attribute__ ((noinline))
bool any(const int64_t &i) {
return i;
}
__attribute__ ((noinline))
bool any(const int2 &i) {
return (((i[0] || i[1])));
}
__attribute__ ((noinline))
bool any(const int4 &i) {
return (((i[0] || i[1]) || (i[2] || i[3])));
}
__attribute__ ((noinline))
bool any(const int8 &i) {
return (((i[0] || i[1]) || (i[2] || i[3])) ||
((i[4] || i[5]) || (i[6] || i[7])));
}
__attribute__ ((noinline))
bool any(const int16 &i) {
return (((i[0] || i[1]) || (i[2] || i[3])) ||
((i[4] || i[5]) || (i[6] || i[7]))) ||
(((i[8] || i[9]) || (i[10] || i[11])) ||
((i[12] || i[13]) || (i[14] || i[15])));
}
typedef double double2 __attribute__ ((vector_size (16)));
typedef double double4 __attribute__ ((vector_size (32)));
typedef double double8 __attribute__ ((vector_size (64)));
typedef double double16 __attribute__ ((vector_size (128)));
template <typename R, typename T>
__attribute__ ((noinline))
R broadcast(T x) { return R(x); }
template <>
__attribute__ ((noinline))
double2 broadcast<double2 ,double>(double x) { double2 r = { x, x
}; return r; }
template <>
__attribute__ ((noinline))
double4 broadcast<double4 ,double>(double x) { double4 r = { x, x,
x, x }; return r; }
template <>
__attribute__ ((noinline))
double8 broadcast<double8 ,double>(double x) { double8 r = { x, x,
x, x, x, x, x, x }; return r; }
template <>
__attribute__ ((noinline))
double16 broadcast<double16,double>(double x) { double16 r = { x,
x, x, x, x, x, x, x, x, x, x, x, x, x, x, x }; return r; }
int main(int argc, char **argv)
{
if (argc > 1)
{
int N = atoi(argv[1]);
switch (N)
{
case 1: { double x = broadcast<double ,double>(1.0);
return any(x != 1.0); }
case 2: { double2 x = broadcast<double2 ,double>(1.0);
return any(x != 1.0); }
case 4: { double4 x = broadcast<double4 ,double>(1.0);
return any(x != 1.0); }
case 8: { double8 x = broadcast<double8 ,double>(1.0);
return any(x != 1.0); }
case 16: { double16 x = broadcast<double16,double>(1.0);
return any(x != 1.0); }
}
}
return 1;
}
I compiled it like this:
$ x86_64-w64-mingw32-g++ bug.cpp -march=native -O3
bug.cpp: In function ‘R broadcast(T) [with R = __vector(8) double;
T = double]’:
bug.cpp:56:45: warning: AVX512F vector return without AVX512F
enabled changes the ABI [-Wpsabi]
double8 broadcast<double8 ,double>(double x) { double8 r = { x,
x, x, x, x, x, x, x }; return r; }
^
I ran it on the same host like this, and it crashed:
$ ./a.exe 4
wine: Unhandled page fault on read access to 0xffffffffffffffff at
address 0x4016a8 (thread 002a), starting debugger...
002c:fixme:dbghelp:elf_search_auxv can't find symbol in module
002c:fixme:dbghelp:elf_search_auxv can't find symbol in module
002c:fixme:dbghelp:elf_search_auxv can't find symbol in module
002c:fixme:dbghelp:elf_search_auxv can't find symbol in module
002c:fixme:dbghelp:elf_search_auxv can't find symbol in module
Unhandled exception: page fault on read access to
0xffffffffffffffff in 64-bit code (0x00000000004016a8).
002c:fixme:dbghelp:elf_search_auxv can't find symbol in module
002c:fixme:dbghelp:interpret_function_table_entry PUSH_MACHFRAME 6
Register dump:
rip:00000000004016a8 rsp:000000000032f928 rbp:00000000004619e0
eflags:00010203 ( R- -- I - - -C)
rax:000000000032fa10 rbx:000000000032fa80 rcx:000000000032fa10
rdx:0000000000405000
rsi:0000000000000002 rdi:00000000004619e0 r8:00000000ffffffd0
r9:0000000000000000 r10:0000000000000002
r11:0000000000461a68 r12:0000000000000018 r13:0000000000000010
r14:0000000000000000 r15:0000000000000000
Stack dump:
0x000000000032f928: 00000000004031e8 0000000000000000
0x000000000032f938: 0000000000000000 0000000000000000
0x000000000032f948: 0000000000000000 0000000000202020
0x000000000032f958: 0000000020000000 0000000000ffffff
0x000000000032f968: ffffffff00ffffff 0000000000202020
0x000000000032f978: 0000000000000000 0000000000000000
0x000000000032f988: 0000000000000000 0000000000000000
0x000000000032f998: 0000000000000000 0000000000000000
0x000000000032f9a8: 0000000000000000 0000000000000000
0x000000000032f9b8: 0000000000000000 0000000000000000
0x000000000032f9c8: 0000000000000000 0000000000000000
0x000000000032f9d8: 0000000000000000 0000000000000000
Backtrace:
=>0 0x00000000004016a8 in a (+0x16a8) (0x00000000004619e0)
1 0x00000000004031e8 in a (+0x31e7) (0x00000000004619e0)
0x00000000004016a8: ldsl %ebp,%edi
Modules:
Module Address Debug info Name (20 modules)
PE 400000- 458000 Dwarf a
ELF 7b400000- 7b831000 Deferred kernel32<elf>
\-PE 7b420000- 7b831000 \ kernel32
ELF 7bc00000- 7bd2c000 Deferred ntdll<elf>
\-PE 7bc20000- 7bd2c000 \ ntdll
ELF 7c000000- 7c005000 Deferred <wine-loader>
ELF 7f8e5120b000- 7f8e51239000 Deferred libtinfo.so.6
ELF 7f8e51239000- 7f8e51262000 Deferred libncurses.so.6
ELF 7f8e51362000- 7f8e51445000 Deferred msvcrt<elf>
\-PE 7f8e51380000- 7f8e51445000 \ msvcrt
ELF 7f8e51445000- 7f8e5145a000 Deferred libnss_files.so.2
ELF 7f8e5145a000- 7f8e51473000 Deferred libnsl.so.1
ELF 7f8e51473000- 7f8e51480000 Deferred libnss_nis.so.2
ELF 7f8e51480000- 7f8e5148b000 Deferred libnss_compat.so.2
ELF 7f8e51970000- 7f8e5198a000 Deferred libgcc_s.so.1
ELF 7f8e5198a000- 7f8e51b0d000 Deferred libm.so.6
ELF 7f8e51b0f000- 7f8e51b14000 Deferred libdl.so.2
ELF 7f8e51b14000- 7f8e51cd5000 Deferred libc.so.6
ELF 7f8e51cd5000- 7f8e51cf6000 Deferred libpthread.so.0
ELF 7f8e51edf000- 7f8e51f09000 Deferred
ld-linux-x86-64.so.2
Threads:
process tid prio (all id:s are in hex)
0000000e services.exe
00000020 0
0000001b 0
00000015 0
00000014 0
00000013 0
00000010 0
0000000f 0
00000011 winedevice.exe
00000018 0
00000017 0
00000016 0
00000012 0
00000019 plugplay.exe
0000001d 0
0000001c 0
0000001a 0
0000001e winedevice.exe
00000025 0
00000022 0
00000021 0
0000001f 0
00000023 explorer.exe
00000028 0
00000027 0
00000026 0
00000024 0
00000029 (D) Z:\home\claude\a.exe
0000002a 0 <==
System information:
Wine build: wine-4.0 (Debian 4.0-2)
Platform: x86_64
Version: Windows 7
Host system: Linux
Host version: 4.19.0-5-amd64
Running it with other arguments does not crash.
I expect a program compiled with -march=native to work on the machine
on which it was compiled.
This seems to be a mingw issue (or possibly a wine issue, also occurs with
wine-development), as compiling with regular x86-64 g++ does not exhibit
the problem.
$ g++ --version
g++ (Debian 8.3.0-6) 8.3.0
Copyright (C) 2018 Free Software Foundation, Inc.
This is free software; see the source for copying conditions. There
is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR
PURPOSE.
$ apt-cache policy g++
g++:
Installed: 4:8.3.0-1
Candidate: 4:8.3.0-1
Version table:
4:9.2.1-3 500
500 http://ftp.uk.debian.org/debian unstable/main amd64
Packages
*** 4:8.3.0-1 990
990 http://ftp.uk.debian.org/debian buster/main amd64 Packages
100 /var/lib/dpkg/status
$ apt-cache policy wine
wine:
Installed: 4.0-2
Candidate: 4.0-2
Version table:
4.0.2-1 500
500 http://ftp.uk.debian.org/debian unstable/main amd64
Packages
500 http://ftp.uk.debian.org/debian unstable/main i386 Packages
*** 4.0-2 990
990 http://ftp.uk.debian.org/debian buster/main amd64 Packages
990 http://ftp.uk.debian.org/debian buster/main i386 Packages
100 /var/lib/dpkg/status
$ apt-cache policy wine-development
wine-development:
Installed: 4.7-1
Candidate: 4.7-1
Version table:
*** 4.7-1 500
500 http://ftp.uk.debian.org/debian unstable/main amd64
Packages
500 http://ftp.uk.debian.org/debian unstable/main i386 Packages
100 /var/lib/dpkg/status
4.2-4 990
990 http://ftp.uk.debian.org/debian buster/main amd64 Packages
990 http://ftp.uk.debian.org/debian buster/main i386 Packages
Here's my CPU:
$ cat /proc/cpuinfo
processor : 0
vendor_id : AuthenticAMD
cpu family : 23
model : 8
model name : AMD Ryzen 7 2700X Eight-Core Processor
stepping : 2
microcode : 0x800820b
cpu MHz : 1985.533
cache size : 512 KB
physical id : 0
siblings : 16
core id : 0
cpu cores : 8
apicid : 0
initial apicid : 0
fpu : yes
fpu_exception : yes
cpuid level : 13
wp : yes
flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge
mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext
fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid
extd_apicid aperfmperf pni pclmulqdq monitor ssse3 fma cx16 sse4_1
sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm
extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw skinit wdt
tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb
hw_pstate sme ssbd sev ibpb vmmcall fsgsbase bmi1 avx2 smep bmi2 rdseed
adx smap clflushopt sha_ni xsaveopt xsavec xgetbv1 xsaves clzero irperf
xsaveerptr arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean
flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload
vgif overflow_recov succor smca
bugs : sysret_ss_attrs null_seg spectre_v1 spectre_v2
spec_store_bypass
bogomips : 7399.05
TLB size : 2560 4K pages
clflush size : 64
cache_alignment : 64
address sizes : 43 bits physical, 48 bits virtual
power management: ts ttp tm hwpstate cpb eff_freq_ro [13] [14]
[snip 15 more instances]
Thanks,
Claude
-- System Information:
Debian Release: 10.0
APT prefers stable
APT policy: (990, 'stable'), (500, 'stable-debug'), (500,
'unstable'), (1, 'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386
Kernel: Linux 4.19.0-5-amd64 (SMP w/16 CPU cores)
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8),
LANGUAGE=en_GB:en (charmap=UTF-8)
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)
LSM: AppArmor: enabled
Versions of packages g++-mingw-w64-x86-64 depends on:
ii gcc-mingw-w64-base 8.3.0-19+21.4
ii gcc-mingw-w64-x86-64 8.3.0-19+21.4
ii libc6 2.28-10
ii libgcc1 1:8.3.0-6
ii libgmp10 2:6.1.2+dfsg-4
ii libisl19 0.20-2
ii libmpc3 1.1.0-1
ii libmpfr6 4.0.2-1
ii libstdc++6 8.3.0-6
ii zlib1g 1:1.2.11.dfsg-1
g++-mingw-w64-x86-64 recommends no packages.
Versions of packages g++-mingw-w64-x86-64 suggests:
pn gcc-8-locales <none>
-- no debconf information
--- End Message ---
--- Begin Message ---
Source: gcc-mingw-w64
Source-Version: 21.5
We believe that the bug you reported is fixed in the latest version of
gcc-mingw-w64, which is due to be installed in the Debian FTP archive.
A summary of the changes between this version and the previous one is
attached.
Thank you for reporting the bug, which will now be closed. If you
have further comments please address them to [email protected],
and the maintainer will reopen the bug report if appropriate.
Debian distribution maintenance software
pp.
Stephen Kitt <[email protected]> (supplier of updated gcc-mingw-w64 package)
(This message was generated automatically at their request; if you
believe that there is a problem with it please contact the archive
administrators by mailing [email protected])
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512
Format: 1.8
Date: Wed, 11 Sep 2019 07:39:39 +0200
Source: gcc-mingw-w64
Architecture: source
Version: 21.5
Distribution: unstable
Urgency: medium
Maintainer: Stephen Kitt <[email protected]>
Changed-By: Stephen Kitt <[email protected]>
Closes: 939559
Changes:
gcc-mingw-w64 (21.5) unstable; urgency=medium
.
* Use unaligned VMOV instructions to avoid crashes caused by mis-aligned
values; thanks to Claude Heiland-Allen for the patch! Closes: #939559.
* Standards-Version 4.4.0, no change required.
Checksums-Sha1:
9d98c20e3973d3177278fc7f6f5be18d170ed748 3440 gcc-mingw-w64_21.5.dsc
a9071040b094e316c0dd0abbdefd118adaa3e5b8 58240 gcc-mingw-w64_21.5.tar.xz
24ca6f38bc7b90ad0e99c795e5b90dd2b8bda1b8 7263
gcc-mingw-w64_21.5_source.buildinfo
Checksums-Sha256:
e9c7aa5b03dc43a6786f10a1db7ef5c6756547579a66ba572ff830d161f6e59c 3440
gcc-mingw-w64_21.5.dsc
626d95c51d35a2e3ec57d7c2ec8779ce160850311826018551817e960e14db28 58240
gcc-mingw-w64_21.5.tar.xz
e50e76d372d3c6c9a8316ef10502ff0feb22b6a6ee87ea8fdc9860095afc43da 7263
gcc-mingw-w64_21.5_source.buildinfo
Files:
00035b8766dc01a1912dd50e6fb63fe6 3440 devel optional gcc-mingw-w64_21.5.dsc
0c097bec648c2d01ea0332d77cd72cf8 58240 devel optional gcc-mingw-w64_21.5.tar.xz
6ae9cb27831ed6c1ed7a5725b0a8fd83 7263 devel optional
gcc-mingw-w64_21.5_source.buildinfo
-----BEGIN PGP SIGNATURE-----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=fAls
-----END PGP SIGNATURE-----
--- End Message ---
