https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87952
Giuseppe D'Angelo <dangelog at gmail dot com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |dangelog at gmail dot com
--- Comment #2 from Giuseppe D'Angelo <dangelog at gmail dot com> ---
Still valid as 10.2.
There's just been some discussion related to this missed optimization on the
std-proposals mailing list. It's also worth noticing that neither GCC nor Clang
remove the branch from some slightly different code:
int& test(std::variant<int, bool> &v)
{
return *std::get_if<int>(&v);
}
GCC 10.2 generates:
test(std::variant<int, double>&):
xorl %eax, %eax
cmpb $0, 8(%rdi)
cmove %rdi, %rax
ret
Clang 11 generates:
test(std::__1::variant<int, double>&): # @test(std::__1::variant<int, double>&)
xorl %eax, %eax
cmpl $0, 8(%rdi)
cmoveq %rdi, %rax
retq
If one pushes things a little bit, both compilers remove the branch:
int& test(std::variant<int, double>& v)
{
auto result = std::get_if<int>(&v);
if (!result) __builtin_unreachable();
return *result;
}
GCC:
test(std::variant<int, double>&):
movq %rdi, %rax
ret
Clang:
test(std::__1::variant<int, double>&): # @test(std::__1::variant<int, double>&)
movq %rdi, %rax
retq
Could you please elaborate on what you mean by "might end up as a security
issue"?
