This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
[RFC PATCH 2/3] Misaligned MEM_REFs within handled_components on LHS of assignments
- From: Martin Jambor <mjambor at suse dot cz>
- To: GCC Patches <gcc-patches at gcc dot gnu dot org>
- Cc: Richard Guenther <rguenther at suse dot de>
- Date: Tue, 28 Feb 2012 02:30:38 +0100
- Subject: [RFC PATCH 2/3] Misaligned MEM_REFs within handled_components on LHS of assignments
- References: <20120228013036.494202900@virgil.suse.cz>
Hi,
the second patch in the series handles MEM_REFs on LHS which are parts
of a handled_component, usually a COMPONENT_REF. The failing testcase
which requires it is not actually the one in the patch but it is
gcc.c-torture/execute/mayalias-3.c which fails at -O1 without this
change but with the third patch applied.
The mechanism is the same as in the previous patch, it also
piggy-backs a store_bit_field call onto generation of movmisalign
operation, if that is not available but the access would be
SLOW_UNALIGNED_ACCESS. I also needed to add
!mem_ref_refers_to_non_mem_p condition for the same reasons.
Additionally, complex numbers and accesses to their components are
already handled by the existing code (the new testcase passes on all
platforms I tried and so I added it to make sure it continues to) and
actually messing with it further leads to failures (of
g++.dg/torture/pr34099.C at -O1). Therefore I added a condition to
punt on complex modes.
Thanks,
Martin
2012-02-28 Martin Jambor <mjambor@suse.cz>
* expr.c (expand_assignment): Handle misaligned scalar writes to
memory through MEM_REFs within handled_components by calling
store_bit_field.
* testsuite/gcc.dg/misaligned-expand-3.c: New test.
Index: src/gcc/expr.c
===================================================================
--- src.orig/gcc/expr.c
+++ src/gcc/expr.c
@@ -4575,7 +4575,7 @@ expand_assignment (tree to, tree from, b
rtx result;
enum machine_mode mode;
unsigned int align;
- enum insn_code icode;
+ enum insn_code icode = (enum insn_code) 0;
/* Don't crash if the lhs of the assignment was erroneous. */
if (TREE_CODE (to) == ERROR_MARK)
@@ -4689,10 +4689,13 @@ expand_assignment (tree to, tree from, b
mode = TYPE_MODE (TREE_TYPE (tem));
if (TREE_CODE (tem) == MEM_REF
&& mode != BLKmode
+ && !COMPLEX_MODE_P (mode)
+ && !mem_ref_refers_to_non_mem_p (tem)
&& ((align = get_object_or_type_alignment (tem))
< GET_MODE_ALIGNMENT (mode))
- && ((icode = optab_handler (movmisalign_optab, mode))
- != CODE_FOR_nothing))
+ && (((icode = optab_handler (movmisalign_optab, mode))
+ != CODE_FOR_nothing)
+ || SLOW_UNALIGNED_ACCESS (mode, align)))
{
misalignp = true;
to_rtx = gen_reg_rtx (mode);
@@ -4871,13 +4874,18 @@ expand_assignment (tree to, tree from, b
if (TREE_THIS_VOLATILE (tem))
MEM_VOLATILE_P (mem) = 1;
- create_fixed_operand (&ops[0], mem);
- create_input_operand (&ops[1], to_rtx, mode);
- /* The movmisalign<mode> pattern cannot fail, else the assignment
- would silently be omitted. */
- expand_insn (icode, 2, ops);
+ if (icode != CODE_FOR_nothing)
+ {
+ create_fixed_operand (&ops[0], mem);
+ create_input_operand (&ops[1], to_rtx, mode);
+ /* The movmisalign<mode> pattern cannot fail, else the assignment
+ would silently be omitted. */
+ expand_insn (icode, 2, ops);
+ }
+ else
+ store_bit_field (mem, GET_MODE_BITSIZE (mode),
+ 0, 0, 0, mode, to_rtx);
}
-
if (result)
preserve_temp_slots (result);
free_temp_slots ();
Index: src/gcc/testsuite/gcc.dg/misaligned-expand-3.c
===================================================================
--- /dev/null
+++ src/gcc/testsuite/gcc.dg/misaligned-expand-3.c
@@ -0,0 +1,43 @@
+/* Test that expand can generate correct stores to misaligned data of complex
+ type even on strict alignment platforms. */
+
+/* { dg-do run } */
+/* { dg-options "-O0" } */
+
+extern void abort ();
+
+typedef _Complex float mycmplx __attribute__((aligned(1)));
+
+void
+foo (mycmplx *p, float r, float i)
+{
+ __real__ *p = r;
+ __imag__ *p = i;
+}
+
+#define cvr 3.2f
+#define cvi 2.5f
+#define NUM 8
+
+struct blah
+{
+ char c;
+ mycmplx x[NUM];
+} __attribute__((packed));
+
+struct blah g;
+
+int
+main (int argc, char **argv)
+{
+ int k;
+
+ for (k = 0; k < NUM; k++)
+ {
+ foo (&g.x[k], cvr, cvi);
+ if (__real__ g.x[k] != cvr
+ || __imag__ g.x[k] != cvi)
+ abort ();
+ }
+ return 0;
+}