This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
[PATCH] Fix powerpc bootstrap when using --enable-checking=rtl
- From: Peter Bergner <bergner at vnet dot ibm dot com>
- To: gcc-patches at gcc dot gnu dot org
- Cc: Andreas Jaeger <aj at suse dot de>, Andrew Pinski <pinskia at physics dot uc dot edu>, David Edelsohn <dje at watson dot ibm dot com>
- Date: Fri, 10 Nov 2006 23:52:11 -0600
- Subject: [PATCH] Fix powerpc bootstrap when using --enable-checking=rtl
This patch fixes the bootstrap problem on powerpc that Andreas ran into
using --enable-checking=rtl. The problem was that adjacent_mem_locations
wasn't setup to handle store instructions embedded inside a parallel.
However, store with update insns contain stores embedded in parallels,
so the solution here is to reach inside the parallel for the actual
store.
The passed bootstrap (with and without --enable-checking=rtl) and
regtesting on powerpc64-linux. Is this ok for mainline? This problem
also exists on 4.2, should we fix it there too?
I'll note that I'm heading on vacation for a week tomorrow (Saturday),
so if this patch is ok, either someone can check it in for me, or I'll
check it in when I get back.
Peter
2006-11-10 Peter Bergner <bergner@vnet.ibm.com>
* config/rs6000/rs6000.c (get_store_dest): New.
(adjacent_mem_locations): Use get_store_dest() to get
the rtl of the store destination.
Index: config/rs6000/rs6000.c
===================================================================
--- config/rs6000/rs6000.c (revision 118684)
+++ config/rs6000/rs6000.c (working copy)
@@ -674,6 +674,7 @@
static bool is_cracked_insn (rtx);
static bool is_branch_slot_insn (rtx);
static bool is_load_insn (rtx);
+static rtx get_store_dest (rtx pat);
static bool is_store_insn (rtx);
static bool set_to_load_agen (rtx,rtx);
static bool adjacent_mem_locations (rtx,rtx);
@@ -17004,9 +17005,9 @@
adjacent_mem_locations (rtx insn1, rtx insn2)
{
- rtx a = SET_DEST (PATTERN (insn1));
- rtx b = SET_DEST (PATTERN (insn2));
-
+ rtx a = get_store_dest (PATTERN (insn1));
+ rtx b = get_store_dest (PATTERN (insn2));
+
if ((GET_CODE (XEXP (a, 0)) == REG
|| (GET_CODE (XEXP (a, 0)) == PLUS
&& GET_CODE (XEXP (XEXP (a, 0), 1)) == CONST_INT))
@@ -17260,6 +17261,28 @@
return is_store_insn1 (PATTERN (insn));
}
+/* Return the dest of a store insn. */
+
+static rtx
+get_store_dest (rtx pat)
+{
+ gcc_assert (is_store_insn1 (pat));
+
+ if (GET_CODE (pat) == SET)
+ return SET_DEST (pat);
+ else if (GET_CODE (pat) == PARALLEL)
+ {
+ int i;
+
+ for (i = 0; i < XVECLEN (pat, 0); i++)
+ if (is_store_insn1 (XVECEXP (pat, 0, i)))
+ return XVECEXP (pat, 0, i);
+ }
+ /* We shouldn't get here, because we should have either a simple
+ store insn or a store with update which are covered above. */
+ gcc_assert (0);
+}
+
/* Returns whether the dependence between INSN and NEXT is considered
costly by the given target. */