Adding code size bloats to virtual destructor

I'm cheating with C ++ on a cortex m3 arm with the latest available arm-none-eabi-gcc 6.3. I made a dummy class and created a global object of this class:

class B
{};

class A : B
{
public:
    A()
    {
        RCC->APB2ENR |= RCC_APB2ENR_IOPCEN;
        GPIOC->CRH = 1;
        GPIOC->ODR |= 1<<8;    
    }

    virtual void foo()
    {}

    virtual ~A()
    {}
};

A a;

int main(void) { while(1); }

      

        
        
        
      

    

This code is 1620 bytes. If I remove the virtual dtor it compiles to 1304 bytes. This is a tangible difference.

I went and looked into the assembly and the .map file and saw that many different execution functions are linked to my binary if I use a virtual dtor. There's malloc and free and _static_initialization_and_destruction, etc.

Strange thing: I don't understand what they are called? Main is called like this:

 8000260:   d3f9        bcc.n   8000256 <FillZerobss>
 8000262:   f000 f89f   bl  80003a4 <SystemInit>
 8000266:   f000 f957   bl  8000518 <__libc_init_array>
 800026a:   f000 f865   bl  8000338 <main>
 800026e:   4770        bx  lr

      

        
        
        
      

    

After returning from main (which never happens by the way) the execution just sits there on bx lr, it keeps jumping to the same address.

So, I don't see a way to deinitialize a static object. Why isn't it optimized as unreachable code?

I compile like this: arm-none-eabi-g++ -c -fmessage-length=0 -mcpu=cortex-m3 -mthumb -fdata-sections -ffunction-sections -fno-rtti -fno-exceptions -fno-threadsafe-statics

and a link like thisarm-none-eabi-g++ -mcpu=cortex-m3 -mthumb --specs=nosys.specs --specs=nano.specs -Wl,--gc-sections -T "${ProjDirPath}/src/Startup/STM32F100XB_FLASH.ld" -ffreestanding

I tried adding -fno-use-cxa-atexit or making it a dummy one __cxa_atexit

- and it makes the binary a little smaller (which is even weirder than no effect at all).

Is there a way to completely disable the destruction of static objects?

UPDATE:

• adding -O to the code done is shorthand, but there is still a difference with / without virtual dtor (1244 vs 1040 bytes), so the question remains.
• here's the compile output with -S with dtor , without dtor