This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
Re: Bug in if_convert
- To: Richard Earnshaw <rearnsha at arm dot com>
- Subject: Re: Bug in if_convert
- From: Richard Henderson <rth at cygnus dot com>
- Date: Thu, 4 May 2000 16:00:00 -0700
- Cc: gcc-bugs at gcc dot gnu dot org, gcc-patches at gcc dot gnu dot org
- References: <200005031809.TAA11206@cam-mail2.cambridge.arm.com>
On Wed, May 03, 2000 at 07:09:14PM +0100, Richard Earnshaw wrote:
> That is, we have a simple if-then-else sequence with both arms setting x;
> except that in this case the 'else' setter is a complex expression
> involving the condition code register (this was generated by a previous
> pass of if_convert, but cse has eliminated the comparison, since it was
> identical to the earlier one).
Fixed thus. Bootstrapped alpha-linux for sanity.
r~
* ifcvt.c (noce_process_if_block): Fail if A or B modified
between condition and jump.
Index: ifcvt.c
===================================================================
RCS file: /cvs/gcc/egcs/gcc/ifcvt.c,v
retrieving revision 1.8
diff -c -p -d -r1.8 ifcvt.c
*** ifcvt.c 2000/05/03 11:16:35 1.8
--- ifcvt.c 2000/05/04 22:53:08
*************** noce_process_if_block (test_bb, then_bb,
*** 1054,1064 ****
x = SET_DEST (set_a);
a = SET_SRC (set_a);
- /* X may not be mentioned between cond_earliest and the jump. */
- for (insn = jump; insn != if_info.cond_earliest; insn = PREV_INSN (insn))
- if (INSN_P (insn) && reg_mentioned_p (x, insn))
- return FALSE;
-
/* Look for the other potential set. Make sure we've got equivalent
destinations. */
/* ??? This is overconservative. Storing to two different mems is
--- 1054,1059 ----
*************** noce_process_if_block (test_bb, then_bb,
*** 1087,1092 ****
--- 1082,1098 ----
insn_b = set_b = NULL_RTX;
}
b = (set_b ? SET_SRC (set_b) : x);
+
+ /* X may not be mentioned in the range (cond_earliest, jump]. */
+ for (insn = jump; insn != if_info.cond_earliest; insn = PREV_INSN (insn))
+ if (INSN_P (insn) && reg_mentioned_p (x, insn))
+ return FALSE;
+
+ /* A and B may not be modified in the range [cond_earliest, jump). */
+ for (insn = if_info.cond_earliest; insn != jump; insn = NEXT_INSN (insn))
+ if (INSN_P (insn)
+ && (modified_in_p (a, insn) || modified_in_p (b, insn)))
+ return FALSE;
/* Only operate on register destinations, and even then avoid extending
the lifetime of hard registers on small register class machines. */