Re: Stack pointer alignment

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Facundo Ciccioli writes:
 > Hi. I'll save you all the details that I think are irrelevant and go
 > directly to my question.
 > I would like to know why the following code doesn't work:
 > 
 > int main() {
 >     asm("sub $2, %esp");
 >     printf("Hi, this is a test\n");
 >     asm("add $2, %esp");
 > 
 >     return 0;
 > }
 > 
 > By "doesn't work" I mean that the call to printf doesn't have any
 > effect. Of course if I replace the number 2 (that I substract and the
 > add to esp) by any multiple of 4, the program does what expected. I
 > wonder if any of you knows from where does the restriction that the
 > esp must by a multiple of 4 comes from. From printf? I know that it
 > doesn't come from the processor (x86).

Something in the library code, probably.  I thin the ABI specifies
that the stack poiunter will be aligned.  But you can't just randomly
mess with the stack pointer inside compiler-generated code and expect
reasonable results, and if you do trash a hard register, you must tell
the compiler you've done so:

     asm("sub $2, %%esp":::"esp");

Andrew.

[Index of Archives]     [Linux C Programming]     [Linux Kernel]     [eCos]     [Fedora Development]     [Fedora Announce]     [Autoconf]     [The DWARVES Debugging Tools]     [Yosemite Campsites]     [Yosemite News]     [Linux GCC]

  Powered by Linux