https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109486
Bug ID: 109486
Summary: Optimization regression with pseudodestructors
Product: gcc
Version: 11.2.1
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: rtl-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: pshevchuk at pshevchuk dot com
Target Milestone: ---
Created attachment 54840
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54840&action=edit
Source file, the same as in the text.
GCC 11 and higher seems to produce a loop when compiling the following code
with O3 (attached):
template <typename T>
class SillyVector final {
void clear() {
while (ptr_ > data_) {
(--ptr_)->~T();
}
}
private:
T *data_, *ptr_;
};
template class SillyVector<long>;
The (abridged) amd64 assembly generated by GCC11 is:
_ZN11SillyVectorIlE5clearEv:
.LFB1:
.cfi_startproc
movq 8(%rdi), %rax
movq (%rdi), %rdx
cmpq %rdx, %rax
jbe .L1
.p2align 4,,10
.p2align 3
.L3:
subq $8, %rax
movq %rax, 8(%rdi)
cmpq %rdx, %rax
ja .L3
.L1:
ret
GCC10 (and older versions of gcc in general) produce this:
_ZN11SillyVectorIlE5clearEv:
.LFB1:
.cfi_startproc
movq 8(%rdi), %rdx
movq (%rdi), %rcx
cmpq %rcx, %rdx
jbe .L1
leaq -1(%rdx), %rax
subq %rcx, %rax
notq %rax
andq $-8, %rax
addq %rdx, %rax
movq %rax, 8(%rdi)
.L1:
ret
GCC10 version seems superior for the most cases.
Compilers used are
GCC10:
Target: x86_64-redhat-linux
Configured with: ../configure --enable-bootstrap
--enable-languages=c,c++,fortran,lto --prefix=/opt/rh/devtoolset-10/root/usr
--mandir=/opt/rh/devtoolset-10/root/usr/share/man
--infodir=/opt/rh/devtoolset-10/root/usr/share/info
--with-bugurl=http://bugzilla.redhat.com/bugzilla --enable-shared
--enable-threads=posix --enable-checking=release --enable-multilib
--with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions
--enable-gnu-unique-object --enable-linker-build-id
--with-gcc-major-version-only --with-linker-hash-style=gnu
--with-default-libstdcxx-abi=gcc4-compatible --enable-plugin
--enable-initfini-array
--with-isl=/builddir/build/BUILD/gcc-10.2.1-20210130/obj-x86_64-redhat-linux/isl-install
--disable-libmpx --enable-gnu-indirect-function --with-tune=generic
--with-arch_32=x86-64 --build=x86_64-redhat-linux
Thread model: posix
Supported LTO compression algorithms: zlib
gcc version 10.2.1 20210130 (Red Hat 10.2.1-11) (GCC)
GCC11:
Configured with: ../configure --enable-bootstrap
--enable-languages=c,c++,fortran,lto --prefix=/opt/rh/devtoolset-11/root/usr
--mandir=/opt/rh/devtoolset-11/root/usr/share/man
--infodir=/opt/rh/devtoolset-11/root/usr/share/info
--with-bugurl=http://bugzilla.redhat.com/bugzilla --enable-shared
--enable-threads=posix --enable-checking=release --enable-multilib
--with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions
--enable-gnu-unique-object --enable-linker-build-id
--with-gcc-major-version-only --with-linker-hash-style=gnu
--with-default-libstdcxx-abi=gcc4-compatible --enable-plugin
--enable-initfini-array
--with-isl=/builddir/build/BUILD/gcc-11.2.1-20210728/obj-x86_64-redhat-linux/isl-install
--enable-gnu-indirect-function --with-tune=generic --with-arch_32=x86-64
--build=x86_64-redhat-linux
Thread model: posix
Supported LTO compression algorithms: zlib
gcc version 11.2.1 20210728 (Red Hat 11.2.1-1) (GCC)
Intermediate file is omitted as no preprocessor is necessary to reproduce.
arm64 versions seem to exhibit the same behavior: loop in gcc11, no loop in gcc
10
Call to pseudo destructor is necessary for the issue to arise, it doesn't
happen if such call is omitted
