COMMON vs bss/sbss sections

[Ian Lance Taylor]

   Date: Thu, 06 Apr 2000 12:29:38 -0700
   From: Peter Popov <ppopov@redcreek.com>

   What's the difference between the COMMON and bss/sbss sections? Both
   refer to uninitialized data which must be zeroed out, so why have a
   COMMON section -- shouldn't those variables be part of the bss section? 
   This is a powerpc-eabi, linux hosted cross compiler I'm using
   (gcc-2.95.2, binutils 2.9.1).  

The difference is that the COMMON section is a dummy section used to
hold common variables in input files, whereas the bss/sbss sections
hold uninitialized variables.

It is legal to have an definition for a common variable in another
object file (see the documentation for the --warn-common option in the
GNU linker).  If you have a definition for a bss/sbss variable in
another object file, the linker will report a multiple definition
error.

If there is no definition for a common variable, then it should
normally be put in the bss/sbss section of the output file.  You can
not have a common variable in an output file.

Ian

------
Want more information?  See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/
Want to unsubscribe? Send a note to crossgcc-unsubscribe@sourceware.cygnus.com