[Bug c++/96634] New: missing documentation for __builtin_addressof()
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96634 Bug ID: 96634 Summary: missing documentation for __builtin_addressof() Product: gcc Version: unknown Status: UNCONFIRMED Severity: normal Priority: P3 Component: c++ Assignee: unassigned at gcc dot gnu.org Reporter: mirq-gccboogs at rere dot qmqm.pl Target Milestone: --- The __builtin_addressof() function is not documented. It is available in gcc since r240873 (git commit be845b04a8e1).
[Bug tree-optimization/51513] Only partially optimizes away __builtin_unreachable switch default case
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=51513 Michał Mirosław changed: What|Removed |Added CC||mirq-gccboogs at rere dot qmqm.pl --- Comment #12 from Michał Mirosław --- Created attachment 48236 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=48236=edit testcase This bug reoccurred in gcc-8. gcc-7 and gcc-9+ seem not affected. $ gcc-8 -O3 -S b.c -o - [...] foo: .LFB0: .cfi_startproc cmpl$1, %edi jne .L10 movslq %esi, %rsi leaqfoz(%rip), %rax movl(%rax,%rsi,4), %eax movl%eax, (%rdx) xorl%eax, %eax ret .p2align 4,,10 .p2align 3 .L10: cmpl$2, %edi jne .L11 .L3: movslq %esi, %rsi leaqbaz(%rip), %rax movl(%rax,%rsi,4), %eax movl%eax, (%rdx) xorl%eax, %eax ret .p2align 4,,10 .p2align 3 .L11: testl %edi, %edi je .L3 .cfi_endproc [...] $ gcc-8 -O -S b.c -o - [...] foo: .LFB0: .cfi_startproc cmpl$1, %edi je .L2 cmpl$2, %edi je .L3 testl %edi, %edi jne .L4 .L3: movslq %esi, %rsi leaqbaz(%rip), %rax movl(%rax,%rsi,4), %eax movl%eax, (%rdx) .L5: movl$0, %eax ret .L2: movslq %esi, %rsi leaqfoz(%rip), %rax movl(%rax,%rsi,4), %eax movl%eax, (%rdx) jmp .L5 .L4: .cfi_endproc
[Bug c/92507] False positives with -Wsign-compare
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92507 Michał Mirosław changed: What|Removed |Added CC||mirq-gccboogs at rere dot qmqm.pl --- Comment #4 from Michał Mirosław --- Created attachment 48152 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=48152=edit testcase A simple testcase.
[Bug c++/66586] Template backtrace is truncated/absent after 'template argument deduction/substitution failed:'
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66586 Michał Mirosław changed: What|Removed |Added CC||mirq-gccboogs at rere dot qmqm.pl --- Comment #3 from Michał Mirosław --- Created attachment 44379 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=44379=edit Testcase with enable_if_t Happens also in 6.3.0, 7.3.0, 8.1.0 (same messages generated for those versions). Slightly different testcase, with enable_if_t. $ g++ -c -std=c++17 a.cc a.cc: In function ‘void bar()’: a.cc:22:7: error: no matching function for call to ‘foo(A&)’ foo(a); ^ a.cc:6:6: note: candidate: template void foo(S&) void foo(S&); ^~~ a.cc:6:6: note: template argument deduction/substitution failed: a.cc:9:6: note: candidate: template void foo(S&) void foo(S&); ^~~ a.cc:9:6: note: template argument deduction/substitution failed: a.cc:8:23: error: no type named ‘other’ in ‘struct A’ template > ^~~~
[Bug target/68014] ICE when using Flag Output Operands
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68014 Michał Mirosław changed: What|Removed |Added CC||mirq-gccboogs at rere dot qmqm.pl --- Comment #3 from Michał Mirosław --- Just accidentally tripped on this one using gcc 6.3.0.
[Bug c/77331] incorrect range location in -Wformat with a concatenated format literal
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=77331
Michał Mirosław changed:
What|Removed |Added
CC||mirq-gccboogs at rere dot
qmqm.pl
--- Comment #1 from Michał Mirosław ---
This looks similar (related to PR#52952?): when showing error for concatenated
format string, caret does not point to the triggering format.
void foo(const char *fmt, ...) __attribute__((format(printf, 1, 2)));
#define PREFIX "a:"
#define A(fmt, args...) foo(PREFIX fmt ".", ##args)
void bar()
{
A("%");
}
$ arm-linux-gnueabihf-gcc-7 -fsyntax-only -Wformat foo.c
foo.c: In function ‘bar’:
foo.c:3:16: warning: conversion lacks type at end of format [-Wformat=]
#define PREFIX "a:"
^
foo.c:4:29: note: in expansion of macro ‘PREFIX’
#define A(fmt, args...) foo(PREFIX fmt ".", ##args)
^~
foo.c:8:2: note: in expansion of macro ‘A’
A("%");
^
foo.c:4:41: note: format string is defined here
#define A(fmt, args...) foo(PREFIX fmt ".", ##args)
^
The same happens with gcc-6.3.0.
[Bug c/80585] -Wformat location
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80585 Michał Mirosław changed: What|Removed |Added Status|UNCONFIRMED |RESOLVED Resolution|--- |INVALID --- Comment #1 from Michał Mirosław --- Missclick, sorry for the noise.
[Bug c/80585] New: -Wformat location
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80585 Bug ID: 80585 Summary: -Wformat location Product: gcc Version: 6.3.0 Status: UNCONFIRMED Severity: normal Priority: P3 Component: c Assignee: unassigned at gcc dot gnu.org Reporter: mirq-gccboogs at rere dot qmqm.pl Target Milestone: ---
[Bug target/79355] New: poor code for AVX vector compare
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79355
Bug ID: 79355
Summary: poor code for AVX vector compare
Product: gcc
Version: 6.3.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: mirq-gccboogs at rere dot qmqm.pl
Target Milestone: ---
gcc-6.2 (and previous versions) generates a very inefficient code for AVX when
comparing 32-byte vectors:
$ cat a.c
#include
__v8su eq2(__v8su a, __v8su b)
{
return a == b;
}
$ gcc -S -Ofast -mavx a.c -o -
.file "a.c"
.text
.p2align 4,,15
.globl eq2
.type eq2, @function
eq2:
.LFB4856:
.cfi_startproc
vmovd %xmm0, %edx
vmovd %xmm1, %eax
leaq8(%rsp), %r10
.cfi_def_cfa 10, 0
vpextrd $1, %xmm0, %ecx
andq$-32, %rsp
cmpl%eax, %edx
[... extracting and comparing every element here ...]
vpinsrd $1, %r11d, %xmm5, %xmm1
vpinsrd $1, %r9d, %xmm7, %xmm0
popq%r10
.cfi_def_cfa 10, 0
vpunpcklqdq %xmm3, %xmm0, %xmm0
vpunpcklqdq %xmm2, %xmm1, %xmm1
popq%rbp
leaq-8(%r10), %rsp
.cfi_def_cfa 7, 8
vinsertf128 $0x1, %xmm1, %ymm0, %ymm0
ret
When it could instead generate (i.e. split vector in half and combine
afterwards):
vextractf128$0x1, %ymm0, %xmm2
vextractf128$0x1, %ymm1, %xmm3
vpcmpeqd%xmm1, %xmm0, %xmm0
vpcmpeqd%xmm3, %xmm2, %xmm2
vinsertf128 $0x1, %xmm2, %ymm0, %ymm0
ret
$ gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/6/lto-wrapper
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Debian 6.3.0-5'
--with-bugurl=file:///usr/share/doc/gcc-6/README.Bugs
--enable-languages=c,ada,c++,java
,go,d,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-6
--program-prefix=x86_64-linux-gnu- --enable-shared --enable-linker-build-id
--libexecdir=/usr/lib --without-included-gettext --enable-threads=posix
--libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu
--enable-libstdcxx-debug --enable-libstdcxx-time=yes
--with-default-libstdcxx-abi=new --enable-gnu-unique-object
--disable-vtable-verify --enable-libmpx --enable-plugin --enable-default-pie
--with-system-zlib --disable-browser-plugin --enable-java-awt=gtk
--enable-gtk-cairo --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-6-amd64/jre
--enable-java-home --with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-6-amd64
--with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-6-amd64
--with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar
--with-target-system-zlib --enable-objc-gc=auto --enable-multiarch
--with-arch-32=i686 --with-abi=m64 --with-multilib-list=m32,m64,mx32
--enable-multilib --with-tune=generic --enable-checking=release
--build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
gcc version 6.3.0 20170124 (Debian 6.3.0-5)
[Bug middle-end/71927] stack alignment prologue not optimized out when no local variables remain
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71927 Micha³ Miros³aw changed: What|Removed |Added Attachment #38926|0 |1 is obsolete|| --- Comment #3 from Micha³ Miros³aw --- Created attachment 38927 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=38927=edit test case It seems that just the asm and array length > 2 are needed. $ gcc -O3 -S -o - c.c [...] x: .LFB0: .cfi_startproc ret .cfi_endproc [...] y: .LFB1: .cfi_startproc leaq8(%rsp), %r10 .cfi_def_cfa 10, 0 andq$-32, %rsp pushq -8(%r10) pushq %rbp .cfi_escape 0x10,0x6,0x2,0x76,0 movq%rsp, %rbp pushq %r10 .cfi_escape 0xf,0x3,0x76,0x78,0x6 popq%r10 .cfi_def_cfa 10, 0 popq%rbp leaq-8(%r10), %rsp .cfi_def_cfa 7, 8 ret .cfi_endproc
[Bug middle-end/71927] stack alignment prologue not optimized out when no local variables remain
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71927 --- Comment #2 from Micha³ Miros³aw --- gcc 4.9.3 generates the same code. $ gcc-4.9 -v Using built-in specs. COLLECT_GCC=gcc-4.9 COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/4.9/lto-wrapper Target: x86_64-linux-gnu Configured with: ../src/configure -v --with-pkgversion='Debian 4.9.3-14' --with-bugurl=file:///usr/share/doc/gcc-4.9/README.Bugs --enable-languages=c,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-4.9 --enable-shared --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.9 --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-gnu-unique-object --disable-vtable-verify --enable-plugin --with-system-zlib --disable-browser-plugin --enable-java-awt=gtk --enable-gtk-cairo --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64/jre --enable-java-home --with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64 --with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-4.9-amd64 --with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar --enable-objc-gc --enable-multiarch --with-arch-32=i586 --with-abi=m64 --with-multilib-list=m32,m64,mx32 --enable-multilib --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu Thread model: posix gcc version 4.9.3 (Debian 4.9.3-14)
[Bug middle-end/71927] stack alignment prologue not optimized out when no local variables remain
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71927 Micha³ Miros³aw changed: What|Removed |Added Attachment #38925|0 |1 is obsolete|| --- Comment #1 from Micha³ Miros³aw --- Created attachment 38926 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=38926=edit test case 2 Fails the same way when using non-SSE registers, but when requesting alignment of the local variable. $ gcc -O3 -S -o - b.c [...] y: .LFB1: .cfi_startproc leaq8(%rsp), %r10 .cfi_def_cfa 10, 0 andq$-32, %rsp pushq -8(%r10) pushq %rbp .cfi_escape 0x10,0x6,0x2,0x76,0 movq%rsp, %rbp pushq %r10 .cfi_escape 0xf,0x3,0x76,0x78,0x6 movl(%rdi), %eax movl%edx, (%rdi) movl4(%rdi), %edx movl%ecx, 4(%rdi) movl8(%rdi), %ecx movl%esi, 8(%rdi) popq%r10 .cfi_def_cfa 10, 0 popq%rbp leaq-8(%r10), %rsp .cfi_def_cfa 7, 8 ret .cfi_endproc
[Bug middle-end/71927] New: stack alignment prologue not optimized out when no local variables remain
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71927 Bug ID: 71927 Summary: stack alignment prologue not optimized out when no local variables remain Product: gcc Version: 5.4.0 Status: UNCONFIRMED Keywords: missed-optimization Severity: normal Priority: P3 Component: middle-end Assignee: unassigned at gcc dot gnu.org Reporter: mirq-gccboogs at rere dot qmqm.pl Target Milestone: --- Created attachment 38925 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=38925=edit test case gcc 5.4 optimizes out local array, but does not remove stack-alignment code. x() has 2-element __m256i array - get's optimized out correctly y() has 3-element __m256i array - stack alignment is left in place for the y() case, at -O2 the loop is not unrolled (unlike x()) $ gcc -mavx2 -O3 -S -o - a.c [...] x: .LFB4854: .cfi_startproc vmovdqa (%rdi), %ymm2 vpminud %ymm1, %ymm2, %ymm3 vpmaxud %ymm1, %ymm2, %ymm1 vmovdqa %ymm1, (%rdi) vmovdqa 32(%rdi), %ymm1 vpminud %ymm0, %ymm1, %ymm2 vpmaxud %ymm0, %ymm1, %ymm0 vmovdqa %ymm0, 32(%rdi) vzeroupper ret .cfi_endproc [...] y: .LFB4855: .cfi_startproc leaq8(%rsp), %r10 .cfi_def_cfa 10, 0 andq$-32, %rsp pushq -8(%r10) pushq %rbp .cfi_escape 0x10,0x6,0x2,0x76,0 movq%rsp, %rbp pushq %r10 .cfi_escape 0xf,0x3,0x76,0x78,0x6 vmovdqa (%rdi), %ymm0 vpminud %ymm0, %ymm3, %ymm4 vpmaxud %ymm0, %ymm3, %ymm0 vmovdqa %ymm0, (%rdi) vmovdqa 32(%rdi), %ymm0 vpminud %ymm0, %ymm2, %ymm3 vpmaxud %ymm0, %ymm2, %ymm0 vmovdqa %ymm0, 32(%rdi) vmovdqa 64(%rdi), %ymm0 vpminud %ymm0, %ymm1, %ymm2 vpmaxud %ymm0, %ymm1, %ymm0 vmovdqa %ymm0, 64(%rdi) vzeroupper popq%r10 .cfi_def_cfa 10, 0 popq%rbp leaq-8(%r10), %rsp .cfi_def_cfa 7, 8 ret .cfi_endproc $ gcc -mavx2 -O2 -S -o - a.c [... x() - same as -O3] y: .LFB4855: .cfi_startproc leaq8(%rsp), %r10 .cfi_def_cfa 10, 0 andq$-32, %rsp pushq -8(%r10) pushq %rbp .cfi_escape 0x10,0x6,0x2,0x76,0 movq%rsp, %rbp pushq %r10 .cfi_escape 0xf,0x3,0x76,0x78,0x6 vmovdqa %ymm2, -112(%rbp) xorl%eax, %eax vmovdqa %ymm1, -80(%rbp) vmovdqa %ymm0, -48(%rbp) .L3: vmovdqa (%rdi,%rax), %ymm0 vmovdqa -112(%rbp,%rax), %ymm1 vpminud %ymm0, %ymm1, %ymm2 vpmaxud %ymm0, %ymm1, %ymm0 vmovdqa %ymm2, -112(%rbp,%rax) vmovdqa %ymm0, (%rdi,%rax) addq$32, %rax cmpq$96, %rax jne .L3 vmovdqa -48(%rbp), %ymm2 vmovdqa -80(%rbp), %ymm1 vmovdqa -112(%rbp), %ymm0 vzeroupper popq%r10 .cfi_def_cfa 10, 0 popq%rbp leaq-8(%r10), %rsp .cfi_def_cfa 7, 8 ret .cfi_endproc $ gcc -v Using built-in specs. COLLECT_GCC=gcc COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/5/lto-wrapper Target: x86_64-linux-gnu Configured with: ../src/configure -v --with-pkgversion='Debian 5.4.0-6' --with-bugurl=file:///usr/share/doc/gcc-5/README.Bugs --enable-languages=c,ada,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-5 --enable-shared --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --with-default-libstdcxx-abi=new --enable-gnu-unique-object --disable-vtable-verify --enable-libmpx --enable-plugin --with-system-zlib --disable-browser-plugin --enable-java-awt=gtk --enable-gtk-cairo --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-5-amd64/jre --enable-java-home --with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-5-amd64 --with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-5-amd64 --with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar --enable-objc-gc --enable-multiarch --with-arch-32=i686 --with-abi=m64 --with-multilib-list=m32,m64,mx32 --enable-multilib --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu Thread model: posix gcc version 5.4.0 20160609 (Debian 5.4.0-6)
[Bug c++/71871] ICE on mixing templates and vector extensions ternary operator
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71871 --- Comment #1 from Micha³ Miros³aw --- Created attachment 38894 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=38894=edit test case 2 Small change (second variable instead of constant) changes the ICE to: $ g++ -c b.cc b.cc: In function ‘void foo() [with unsigned int N = 0u]’: b.cc:7:2: internal compiler error: in make_decl_rtl, at varasm.c:1313 x = x ? x : y; ^ 0x94b229 make_decl_rtl(tree_node*) ../../src/gcc/varasm.c:1309 0xea0272 expand_expr_real_1(tree_node*, rtx_def*, machine_mode, expand_modifier, rtx_def**, bool) ../../src/gcc/expr.c:9581 0xea17b8 expand_expr_real(tree_node*, rtx_def*, machine_mode, expand_modifier, rtx_def**, bool) ../../src/gcc/expr.c:8018 0xea17b8 store_expr_with_bounds(tree_node*, rtx_def*, int, bool, tree_node*) ../../src/gcc/expr.c:5382 0xea0fbc expand_assignment(tree_node*, tree_node*, bool) ../../src/gcc/expr.c:5154 0xdfbf10 expand_gimple_stmt_1 ../../src/gcc/cfgexpand.c:3426 0xdfbf10 expand_gimple_stmt ../../src/gcc/cfgexpand.c:3522 0xdf4f18 expand_gimple_basic_block ../../src/gcc/cfgexpand.c:5534 0xdf4f18 execute ../../src/gcc/cfgexpand.c:6152 Please submit a full bug report, with preprocessed source if appropriate. Please include the complete backtrace with any bug report. See for instructions.
[Bug c++/71871] New: ICE on mixing templates and vector extensions ternary operator
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71871 Bug ID: 71871 Summary: ICE on mixing templates and vector extensions ternary operator Product: gcc Version: 5.4.0 Status: UNCONFIRMED Severity: normal Priority: P3 Component: c++ Assignee: unassigned at gcc dot gnu.org Reporter: mirq-gccboogs at rere dot qmqm.pl Target Milestone: --- Created attachment 38893 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=38893=edit test case Attached causes an ICE. $ g++ -c a.cc a.cc: In function ‘void foo() [with unsigned int N = 0u]’: a.cc:7:2: internal compiler error: in gimplify_expr, at gimplify.c:8858 x = x ? x : -1; ^ 0x74c6d9 gimplify_expr(tree_node**, gimple_statement_base**, gimple_statement_base**, bool (*)(tree_node*), int) ../../src/gcc/gimplify.c:8858 0xefc41c internal_get_tmp_var ../../src/gcc/gimplify.c:555 0x12b62d1 get_initialized_tmp_var(tree_node*, gimple_statement_base**, gimple_statement_base**) ../../src/gcc/gimplify.c:598 0x12b62d1 gimplify_save_expr ../../src/gcc/gimplify.c:4887 0xef8d54 gimplify_expr(tree_node**, gimple_statement_base**, gimple_statement_base**, bool (*)(tree_node*), int) ../../src/gcc/gimplify.c:8412 0xef80cf gimplify_expr(tree_node**, gimple_statement_base**, gimple_statement_base**, bool (*)(tree_node*), int) ../../src/gcc/gimplify.c:8228 0x7516c5 gimplify_init_constructor ../../src/gcc/gimplify.c:4103 0xefbeb0 gimplify_modify_expr_rhs ../../src/gcc/gimplify.c:4274 0xefb4a9 gimplify_modify_expr ../../src/gcc/gimplify.c:4603 0xef778f gimplify_expr(tree_node**, gimple_statement_base**, gimple_statement_base**, bool (*)(tree_node*), int) ../../src/gcc/gimplify.c:8126 0xefc82b gimplify_stmt(tree_node**, gimple_statement_base**) ../../src/gcc/gimplify.c:5545 0xefc82b gimplify_and_add(tree_node*, gimple_statement_base**) ../../src/gcc/gimplify.c:423 0xefc82b internal_get_tmp_var ../../src/gcc/gimplify.c:568 0xef7c8c get_formal_tmp_var(tree_node*, gimple_statement_base**) ../../src/gcc/gimplify.c:589 0xef7c8c gimplify_expr(tree_node**, gimple_statement_base**, gimple_statement_base**, bool (*)(tree_node*), int) ../../src/gcc/gimplify.c:9050 0x74b6d6 gimplify_expr(tree_node**, gimple_statement_base**, gimple_statement_base**, bool (*)(tree_node*), int) ../../src/gcc/gimplify.c:8844 0xefb54b gimplify_modify_expr ../../src/gcc/gimplify.c:4644 0xef778f gimplify_expr(tree_node**, gimple_statement_base**, gimple_statement_base**, bool (*)(tree_node*), int) ../../src/gcc/gimplify.c:8126 0xef716e gimplify_stmt(tree_node**, gimple_statement_base**) ../../src/gcc/gimplify.c:5545 0xef8570 gimplify_cleanup_point_expr ../../src/gcc/gimplify.c:5321 Please submit a full bug report, with preprocessed source if appropriate. Please include the complete backtrace with any bug report. See for instructions. $ g++ -v Using built-in specs. COLLECT_GCC=g++ COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/5/lto-wrapper Target: x86_64-linux-gnu Configured with: ../src/configure -v --with-pkgversion='Debian 5.4.0-6' --with-bugurl=file:///usr/share/doc/gcc-5/README.Bugs --enable-languages=c,ada,c++,jau Thread model: posix gcc version 5.4.0 20160609 (Debian 5.4.0-6) also fails on: $ g++ -v Using built-in specs. COLLECT_GCC=/usr/bin/g++-4.9.real COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/4.9/lto-wrapper Target: x86_64-linux-gnu Configured with: ../src/configure -v --with-pkgversion='Debian 4.9.2-10' --with-bugurl=file:///usr/share/doc/gcc-4.9/README.Bugs --enable-languages=c,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-4.9 --enable-shared --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.9 --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-gnu-unique-object --disable-vtable-verify --enable-plugin --with-system-zlib --disable-browser-plugin --enable-java-awt=gtk --enable-gtk-cairo --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64/jre --enable-java-home --with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64 --with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-4.9-amd64 --with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar --enable-objc-gc --enable-multiarch --with-arch-32=i586 --with-abi=m64 --with-multilib-list=m32,m64,mx32 --enable-multilib --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu Thread model: posix gcc version 4.9.2 (Debian 4.9.2-10)
[Bug target/71467] extra store for -mavx -mno-avx
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71467 --- Comment #1 from Micha³ Miros³aw --- It gets even worse with AVX also disabled. $ gcc -O3 -mno-avx -mno-avx2 -S -o - a.c
[Bug target/71467] New: extra store for -mavx -mno-avx
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71467
Bug ID: 71467
Summary: extra store for -mavx -mno-avx
Product: gcc
Version: 5.3.1
Status: UNCONFIRMED
Keywords: missed-optimization, ssemmx
Severity: minor
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: mirq-gccboogs at rere dot qmqm.pl
Target Milestone: ---
Host: x86_64-linux-gnu
Target: x86_64-linux-gnu
Created attachment 38664
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=38664=edit
test case
Code:
typedef int __v8si __attribute__ ((__vector_size__ (32)));
void vz(__v8si *v)
{
*v <<= 1;
}
Compiles to:
vz:
.LFB0:
.cfi_startproc
leaq8(%rsp), %r10
.cfi_def_cfa 10, 0
andq$-32, %rsp
pushq -8(%r10)
pushq %rbp
.cfi_escape 0x10,0x6,0x2,0x76,0
movq%rsp, %rbp
pushq %r10
.cfi_escape 0xf,0x3,0x76,0x78,0x6
vmovdqa (%rdi), %xmm2
vmovdqa 16(%rdi), %xmm4
vpslld $1, %xmm2, %xmm1
vpslld $1, %xmm4, %xmm3
vmovaps %xmm1, -48(%rbp)
vmovaps %xmm3, -32(%rbp)
vmovdqa -48(%rbp), %ymm5
vmovdqa %ymm5, (%rdi)
vzeroupper
popq%r10
.cfi_def_cfa 10, 0
popq%rbp
leaq-8(%r10), %rsp
.cfi_def_cfa 7, 8
ret
.cfi_endproc
.LFE0:
.size vz, .-vz
This should be no more than:
vmovdqa (%rdi), %xmm2
vmovdqa 16(%rdi), %xmm4
vpslld $1, %xmm2, %xmm1
vpslld $1, %xmm4, %xmm3
vmovaps %xmm1, (%rdi)
vmovaps %xmm3, 16(%rdi)
ret
Compiled with:
$ gcc -O3 -mavx -mno-avx2 -S -o - a.c
$ gcc -v
Using built-in specs.
COLLECT_GCC=gcc
COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/5/lto-wrapper
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Debian 5.3.1-21'
--with-bugurl=file:///usr/share/doc/gcc-5/README.Bugs
--enable-languages=c,ada,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr
--program-suffix=-5 --enable-shared --enable-linker-build-id
--libexecdir=/usr/lib --without-included-gettext --enable-threads=posix
--libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu
--enable-libstdcxx-debug --enable-libstdcxx-time=yes
--with-default-libstdcxx-abi=new --enable-gnu-unique-object
--disable-vtable-verify --enable-libmpx --enable-plugin --with-system-zlib
--disable-browser-plugin --enable-java-awt=gtk --enable-gtk-cairo
--with-java-home=/usr/lib/jvm/java-1.5.0-gcj-5-amd64/jre --enable-java-home
--with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-5-amd64
--with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-5-amd64
--with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar
--enable-objc-gc --enable-multiarch --with-arch-32=i686 --with-abi=m64
--with-multilib-list=m32,m64,mx32 --enable-multilib --with-tune=generic
--enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu
--target=x86_64-linux-gnu
Thread model: posix
gcc version 5.3.1 20160528 (Debian 5.3.1-21)
Same with:
Using built-in specs.
COLLECT_GCC=gcc-4.9
COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/4.9/lto-wrapper
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Debian 4.9.3-14'
--with-bugurl=file:///usr/share/doc/gcc-4.9/README.Bugs
--enable-languages=c,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr
--program-suffix=-4.9 --enable-shared --enable-linker-build-id
--libexecdir=/usr/lib --without-included-gettext --enable-threads=posix
--with-gxx-include-dir=/usr/include/c++/4.9 --libdir=/usr/lib --enable-nls
--with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug
--enable-libstdcxx-time=yes --enable-gnu-unique-object --disable-vtable-verify
--enable-plugin --with-system-zlib --disable-browser-plugin
--enable-java-awt=gtk --enable-gtk-cairo
--with-java-home=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64/jre --enable-java-home
--with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64
--with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-4.9-amd64
--with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar
--enable-objc-gc --enable-multiarch --with-arch-32=i586 --with-abi=m64
--with-multilib-list=m32,m64,mx32 --enable-multilib --with-tune=generic
--enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu
--target=x86_64-linux-gnu
Thread model: posix
gcc version 4.9.3 (Debian 4.9.3-14)
[Bug inline-asm/71086] New: miscompiled inline-asm + alternative constraints + __builtin_expect + always_inline
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71086 Bug ID: 71086 Summary: miscompiled inline-asm + alternative constraints + __builtin_expect + always_inline Product: gcc Version: 4.9.3 Status: UNCONFIRMED Severity: normal Priority: P3 Component: inline-asm Assignee: unassigned at gcc dot gnu.org Reporter: mirq-gccboogs at rere dot qmqm.pl Target Milestone: --- Created attachment 38474 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=38474=edit test case source The attached code get's miscompiled when -DNO_EXPECT is not passed. With __builtin_expect used, the operand to inline asm in inc_overflow() get's replaced with uninitialized stack entry. This happens only with exact combination of: -O1, inline-asm has multiple alternative constraints, __builtin_expect is used in inc_overflow(), there are other not-optimized-out variables on stack. gcc versions 4.9.2 (-10 Debian) and 4.9.3 (-13ubuntu2) miscompile, gcc 5.3.1 generates correct code. --- good: g++ -O1 -S -DNO_EXCEPT bug_inline.cc -o - +++ bad: g++ -O1 -S bug_inline.cc -o - @@ -5,55 +5,49 @@ _Z6test_smj: .LFB2: .cfi_startproc - pushq %rbp - .cfi_def_cfa_offset 16 - .cfi_offset 6, -16 pushq %rbx - .cfi_def_cfa_offset 24 - .cfi_offset 3, -24 - subq$40, %rsp + .cfi_def_cfa_offset 16 + .cfi_offset 3, -16 + subq$48, %rsp .cfi_def_cfa_offset 64 - movq%rdi, %rbp + movq%rdi, 8(%rsp) movq%fs:40, %rax - movq%rax, 24(%rsp) + movq%rax, 40(%rsp) xorl%eax, %eax cmpl$10, %esi jbe .L2 movl%esi, %ebx - leaq8(%rsp), %rdi + leaq24(%rsp), %rdi call_Z3xdoRm -.L5: - movq%rbp, %rdi +.L4: #APP # 18 "bug_inline.cpp" 1 - add %rbx, %rdi + add %rbx, 32(%rsp)<--- 32(%rsp) is not initialized, but should be with 'v' (saved at 8(%rsp), or in %rbp in good code) setc %sil # 0 "" 2 #NO_APP testb %sil, %sil - movq$-1, %rax - cmovne %rax, %rdi - movq%rdi, 16(%rsp) - leaq16(%rsp), %rdi + je .L3 + movq$-1, 32(%rsp) +.L3: + leaq32(%rsp), %rdi call_Z3xdoRm - movq8(%rsp), %rax - movq24(%rsp), %rdx + movq24(%rsp), %rax + movq40(%rsp), %rdx xorq%fs:40, %rdx - je .L6 - jmp .L7 + je .L5 + jmp .L6 .L2: - leaq8(%rsp), %rdi + leaq24(%rsp), %rdi call_Z3xdoRm movl$4294967295, %ebx - jmp .L5 -.L7: - call__stack_chk_fail + jmp .L4 .L6: - addq$40, %rsp - .cfi_def_cfa_offset 24 - popq%rbx + call__stack_chk_fail +.L5: + addq$48, %rsp .cfi_def_cfa_offset 16 - popq%rbp + popq%rbx .cfi_def_cfa_offset 8 ret .cfi_endproc # Debian 4.9.2 $ gcc -v Using built-in specs. COLLECT_GCC=/usr/bin/gcc-4.9.real COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/4.9/lto-wrapper Target: x86_64-linux-gnu Configured with: ../src/configure -v --with-pkgversion='Debian 4.9.2-10' --with-bugurl=file:///usr/share/doc/gcc-4.9/README.Bugs --enable-languages=c,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-4.9 --enable-shared --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.9 --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-gnu-unique-object --disable-vtable-verify --enable-plugin --with-system-zlib --disable-browser-plugin --enable-java-awt=gtk --enable-gtk-cairo --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64/jre --enable-java-home --with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64 --with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-4.9-amd64 --with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar --enable-objc-gc --enable-multiarch --with-arch-32=i586 --with-abi=m64 --with-multilib-list=m32,m64,mx32 --enable-multilib --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu Thread model: posix gcc version 4.9.2 (Debian 4.9.2-10) # Ubuntu 4.9.3 $ g++-4.9 -v Using built-in specs. COLLECT_GCC=g++-4.9 COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/4.9/lto-wrapper Target: x86_64-linux-gnu Configured with: ../src/configure -v --with-pkgversion='Ubuntu 4.9.3-13ubuntu2' --with-bugurl=file:///usr/share/doc/gcc-4.9/README.Bugs --enable-languages=c,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-4.9 --enable-shared --enable-linker-build-id --lib
[Bug target/68228] __builtin_ia32_pbroadcastd256 generates wrong insn at >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68228 --- Comment #4 from Micha³ Miros³aw --- Created attachment 36673 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36673=edit gcc 4.8 assembler output for -O1 gcc-4.8 also generates correct VPBROADCASTD, though with VMOVD before it.
[Bug target/68228] New: __builtin_ia32_pbroadcastd256 generates wrong insn at >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68228 Bug ID: 68228 Summary: __builtin_ia32_pbroadcastd256 generates wrong insn at >= -O1 Product: gcc Version: 4.9.2 Status: UNCONFIRMED Severity: normal Priority: P3 Component: target Assignee: unassigned at gcc dot gnu.org Reporter: mirq-gccboogs at rere dot qmqm.pl Target Milestone: --- Created attachment 36656 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36656=edit test source __builtin_ia32_pbroadcastd256() generates VBROADCASTSS instead of VPBROADCASTD with optimization turned on. At -O0 there's a lot of useless moving data to/from stack, but in the end the correct VPBROADCASTD is issued. command line: gcc -march=core-avx2 -O1 -S test.c $ gcc -v Using built-in specs. COLLECT_GCC=gcc COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-linux-gnu/4.9/lto-wrapper Target: x86_64-linux-gnu Configured with: ../src/configure -v --with-pkgversion='Debian 4.9.2-10' --with-bugurl=file:///usr/share/doc/gcc-4.9/README.Bugs --enable-languages=c,c++,java,go,d,fortran,objc,obj-c++ --prefix=/usr --program-suffix=-4.9 --enable-shared --enable-linker-build-id --libexecdir=/usr/lib --without-included-gettext --enable-threads=posix --with-gxx-include-dir=/usr/include/c++/4.9 --libdir=/usr/lib --enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug --enable-libstdcxx-time=yes --enable-gnu-unique-object --disable-vtable-verify --enable-plugin --with-system-zlib --disable-browser-plugin --enable-java-awt=gtk --enable-gtk-cairo --with-java-home=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64/jre --enable-java-home --with-jvm-root-dir=/usr/lib/jvm/java-1.5.0-gcj-4.9-amd64 --with-jvm-jar-dir=/usr/lib/jvm-exports/java-1.5.0-gcj-4.9-amd64 --with-arch-directory=amd64 --with-ecj-jar=/usr/share/java/eclipse-ecj.jar --enable-objc-gc --enable-multiarch --with-arch-32=i586 --with-abi=m64 --with-multilib-list=m32,m64,mx32 --enable-multilib --with-tune=generic --enable-checking=release --build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu Thread model: posix gcc version 4.9.2 (Debian 4.9.2-10)
[Bug target/68228] __builtin_ia32_pbroadcastd256 generates wrong insn at >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68228 --- Comment #1 from Micha³ Miros³aw --- Created attachment 36657 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36657=edit gcc assembler output for -O1
[Bug target/68228] __builtin_ia32_pbroadcastd256 generates wrong insn at >= -O1
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68228 --- Comment #2 from Micha³ Miros³aw --- Created attachment 36658 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36658=edit gcc assembler output for -O0
