More info on gcc 3.3 ia64 bootstrap comparison failure

[H. J. Lu]

Jakub narrowed the problem down to sched-deps.c:

http://gcc.gnu.org/ml/gcc/2003-07/msg00780.html

cselib_subst_to_values is called 3 times in sched-deps.c. 2 of them
call cselib_lookup first. What is the purpose of calling cselib_lookup?
Is that to tell cselib that it is used? However, add_insn_mem_dependence
doesn't call cselib_lookup. I added

	cselib_lookup (XEXP (mem, 0), Pmode, 1);

to add_insn_mem_dependence. It doesn't solve the problem. On

(mem/s:DI (post_inc:DI (reg/f:DI 41 r44 [762])) [3 java_global_trees+0 S8 A64])

cselib_lookup (XEXP (mem, 0), Pmode, 1) returns NULL. However,
cselib_subst_to_values returns (value:DI) from new_cselib_val due to
POST_INC, which may lead to a dangling pointer. My qestions are

1. Should the POST_INC case be handled in add_insn_mem_dependence?
2. Is cselib_lookup needed in add_insn_mem_dependence?
3. If yes, what should add_insn_mem_dependence do when cselib_lookup
returns NULL?


H.J.