This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
must_alias question
- From: Jan Hubicka <hubicka at ucw dot cz>
- To: gcc at gcc dot gnu dot org, dnovillo at redhat dot com
- Date: Thu, 20 Nov 2003 11:04:31 +0100
- Subject: must_alias question
Hi,
I've localized the problem with tail call into combination of must_alias
and tail recursion ellimination pass, so I am trying to understand what
effect must_alias may have.
One thing that looks strange to me is clearing of
may_point_to_global_mem for variable whose address is taken. Isn't this
the other way around? (ie we can look for pointers that are never
assigned to global memory addresses or other pointers with this flag).
Next step is obviously to review all tail recursion decisions made by
compiler...
Honza