Segfault with ESP outside the [stack] range in /proc/pid/maps?

Question

I am debugging a crash and i see following behaviour -

When i attach GDB to the process and do info registers, i see the following value for esp -

esp            0xfd2475d0       0xfd2475d0

Upon doing a disassembly of code where it is crashing, I see it's storing to memory pointed-to by the stack pointer -

81c886a:       c7 04 24 2c f9 8a 0c    movl   $0xc8af92c,(%esp)

And if i view maps file in /proc/<PID>/maps, I see stack address range as -

fff39000-fff59000 rwxp 7ffffffde000 00:00 0                              [stack]

Clearly, value of ESP 0xfd2475d0 in GDB is not in sync with the stack address in maps file.

Can this be a reason for crash. I think it should be as i am getting SIGSEGV. Also, how do I resolve this issue?

Answer 1

The most likely option is over allocation.

 void foo() {    
     double too_big[6000000];    // this would be located at 0xfd...... 
     int a;                      // this would be located at 0xfff3f000 ... 
 }