[Bug target/63890] [4.9/5 regression] Compiling trivial program with -O -p leads to misaligned stack
jakub at gcc dot gnu.org
gcc-bugzilla@gcc.gnu.org
Tue Feb 24 09:22:00 GMT 2015
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63890
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rth at gcc dot gnu.org
--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
I guess the question is, do we guarantee the standard stack alignment for the
_mcount call or not? In any case, this call is very special, has different
calling conventions than other functions, and sometimes (e.g. for fentry) is
performed before the prologue, so even less likely to have aligned stack.
So, do we want any stack guarantees for mcount and/or fentry on Linux?
Similarly other OSes. Aligning stack in the callers can be of course done, but
isn't without code size and performance implications.
More information about the Gcc-bugs
mailing list