This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Fix pr23135.c on the Blackfin
- From: Bernd Schmidt <bernds_cb1 at t-online dot de>
- To: GCC Patches <gcc-patches at gcc dot gnu dot org>
- Cc: Alexandre Oliva <aoliva at redhat dot com>
- Date: Wed, 11 Apr 2007 21:49:10 +0200
- Subject: Fix pr23135.c on the Blackfin
This patch fixes
FAIL: gcc.c-torture/execute/pr23135.c compilation, -O2 (internal
compiler error)
on the Blackfin.
We abort in delete_output_reload, in the loop that compares substed
against the contents of reg_equiv_alt_mem_list. Code in reload.c is
careful not to insert the normal MEM rtx for an equivalence into this
list, but only forms that differ. However, it uses a pointer equality
test, and parts of find_reloads_address can copy a MEM without modifying
it further, which causes it to be put into the list erroneously.
Fixed by just using rtx_equal_p in all instances. Bootstrapped &
regression tested on i686-linux; I've also verified that generated code
is identical on a large set of input files. Committed as 123724.
Bernd
--
This footer brought to you by insane German lawmakers.
Analog Devices GmbH Wilhelm-Wagenfeld-Str. 6 80807 Muenchen
Registergericht Muenchen HRB 40368
Geschaeftsfuehrer Thomas Wessel, Vincent Roche, Joseph E. McDonough
Index: ChangeLog
===================================================================
--- ChangeLog (revision 123723)
+++ ChangeLog (working copy)
@@ -1,3 +1,10 @@
+2007-04-11 Bernd Schmidt <bernd.schmidt@analog.com>
+
+ * reload.c (find_reloads_toplev, find_reloads_address,
+ find_reloads_address_1, find_reloads_subreg_address): Use rtx_equal_p,
+ not a pointer equality test, to decide if we need to call
+ push_reg_equiv_alt_mem.
+
2007-04-11 Sebastian Pop <sebastian.pop@inria.fr>
* tree-data-ref.c (affine_function_zero_p, constant_access_functions,
Index: reload.c
===================================================================
--- reload.c (revision 123723)
+++ reload.c (working copy)
@@ -4570,7 +4570,7 @@ find_reloads_toplev (rtx x, int opnum, e
x = mem;
i = find_reloads_address (GET_MODE (x), &x, XEXP (x, 0), &XEXP (x, 0),
opnum, type, ind_levels, insn);
- if (x != mem)
+ if (!rtx_equal_p (x, mem))
push_reg_equiv_alt_mem (regno, x);
if (address_reloaded)
*address_reloaded = i;
@@ -4785,7 +4785,7 @@ find_reloads_address (enum machine_mode
find_reloads_address (GET_MODE (tem), &tem, XEXP (tem, 0),
&XEXP (tem, 0), opnum,
ADDR_TYPE (type), ind_levels, insn);
- if (tem != orig)
+ if (!rtx_equal_p (tem, orig))
push_reg_equiv_alt_mem (regno, tem);
}
/* We can avoid a reload if the register's equivalent memory
@@ -5589,7 +5589,7 @@ find_reloads_address_1 (enum machine_mod
RELOAD_OTHER,
ind_levels, insn);
- if (tem != orig)
+ if (!rtx_equal_p (tem, orig))
push_reg_equiv_alt_mem (regno, tem);
/* Then reload the memory location into a base
@@ -5656,7 +5656,7 @@ find_reloads_address_1 (enum machine_mod
find_reloads_address (GET_MODE (tem), &tem, XEXP (tem, 0),
&XEXP (tem, 0), opnum, type,
ind_levels, insn);
- if (tem != orig)
+ if (!rtx_equal_p (tem, orig))
push_reg_equiv_alt_mem (regno, tem);
/* Put this inside a new increment-expression. */
x = gen_rtx_fmt_e (GET_CODE (x), GET_MODE (x), tem);
@@ -5811,7 +5811,7 @@ find_reloads_address_1 (enum machine_mod
find_reloads_address (GET_MODE (x), &x, XEXP (x, 0),
&XEXP (x, 0), opnum, ADDR_TYPE (type),
ind_levels, insn);
- if (x != tem)
+ if (!rtx_equal_p (x, tem))
push_reg_equiv_alt_mem (regno, x);
}
}
@@ -6037,7 +6037,7 @@ find_reloads_subreg_address (rtx x, int
&XEXP (tem, 0), opnum, type,
ind_levels, insn);
/* ??? Do we need to handle nonzero offsets somehow? */
- if (!offset && tem != orig)
+ if (!offset && !rtx_equal_p (tem, orig))
push_reg_equiv_alt_mem (regno, tem);
/* If this is not a toplevel operand, find_reloads doesn't see