[PATCH] Fix handling of variable length fields in structures (PR c/82210)

[PATCH] Fix handling of variable length fields in structures (PR c/82210)

[Jakub Jelinek]

Hi!

When placing a variable length field into a structure, we need to update
rli->offset_align for the next field.  We do:
rli->offset_align = MIN (rli->offset_align, desired_align);
which updates it according to the start of that VLA field, the problem is
that if the field doesn't have a size that is a multiple of this alignment
rli->offset_align will not reflect properly the alignment of the end of that
field.  E.g. on the testcase, we have a VLA array aligned as a whole (the
field itself) to 16 bytes / 128 bits, so rli->offset_align remains 128.
The array has element size 2 bytes / 16 bits, times function argument,
so the end of the field is worst case aligned just to 16 bits; if we keep
rli->offset_align as 128 for the next field, then DECL_OFFSET_ALIGN is too
large. DECL_FIELD_OFFSET documented as:
/* In a FIELD_DECL, this is the field position, counting in bytes, of the
   DECL_OFFSET_ALIGN-bit-sized word containing the bit closest to the beginning
   of the structure.  */
and when gimplifying COMPONENT_REFs with that field we:
              tree offset = unshare_expr (component_ref_field_offset (t));
              tree field = TREE_OPERAND (t, 1);
              tree factor
                = size_int (DECL_OFFSET_ALIGN (field) / BITS_PER_UNIT);
   
              /* Divide the offset by its alignment.  */
              offset = size_binop_loc (loc, EXACT_DIV_EXPR, offset, factor);
and later on multiply it again by DECL_OFFSET_ALIGN.  The EXACT_DIV_EXPR
isn't exact.

Fixed by lowering the rli->offset_align if the size isn't a multiple of
the align.  We don't have a multiple_of_p variant that would compute
highest power of two number the expression is known to be a multiple of,
so I'm just checking the most common case, where the size is a multiple
of the starting alignment, and otherwise just compute it very
conservatively.  This will be lower than necessary say for
  __attribute__((aligned (16))) short field[2 * size];
- just 16 bits instead of 32.  In theory we could do a binary search
on power of two numbers in between that high initial rli->offset_align
for which the first multiple_of_p failed, and the conservative guess
we do to improve it.  If you think it is worth it, I can code it up.

Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?

2018-02-09  Jakub Jelinek  <jakub@redhat.com>

	PR c/82210
	* stor-layout.c (place_field): For variable length fields, adjust
	offset_align afterwards not just based on the field's alignment,
	but also on the size.

	* gcc.c-torture/execute/pr82210.c: New test.

--- gcc/stor-layout.c.jj	2018-01-16 16:07:57.000000000 +0100
+++ gcc/stor-layout.c	2018-02-08 13:48:32.380582662 +0100
@@ -1622,6 +1622,30 @@ place_field (record_layout_info rli, tre
 	= size_binop (PLUS_EXPR, rli->offset, DECL_SIZE_UNIT (field));
       rli->bitpos = bitsize_zero_node;
       rli->offset_align = MIN (rli->offset_align, desired_align);
+
+      if (!multiple_of_p (bitsizetype, DECL_SIZE (field),
+			  bitsize_int (rli->offset_align)))
+	{
+	  tree type = strip_array_types (TREE_TYPE (field));
+	  /* The above adjusts offset_align just based on the start of the
+	     field.  The field might not have a size that is a multiple of
+	     that offset_align though.  If the field is an array of fixed
+	     sized elements, assume there can be any multiple of those
+	     sizes.  If it is a variable length aggregate or array of
+	     variable length aggregates, assume worst that the end is
+	     just BITS_PER_UNIT aligned.  */
+	  if (TREE_CODE (TYPE_SIZE (type)) == INTEGER_CST)
+	    {
+	      if (TREE_INT_CST_LOW (TYPE_SIZE (type)))
+		{
+		  unsigned HOST_WIDE_INT sz
+		    = least_bit_hwi (TREE_INT_CST_LOW (TYPE_SIZE (type)));
+		  rli->offset_align = MIN (rli->offset_align, sz);
+		}
+	    }
+	  else
+	    rli->offset_align = MIN (rli->offset_align, BITS_PER_UNIT);
+	}
     }
   else if (targetm.ms_bitfield_layout_p (rli->t))
     {
--- gcc/testsuite/gcc.c-torture/execute/pr82210.c.jj	2018-02-08 13:59:37.247901958 +0100
+++ gcc/testsuite/gcc.c-torture/execute/pr82210.c	2018-02-08 13:59:14.185912469 +0100
@@ -0,0 +1,26 @@
+/* PR c/82210 */
+
+void
+foo (int size)
+{
+  int i;
+  struct S {
+    __attribute__((aligned (16))) struct T { short c; } a[size];
+    int b[size];
+  } s;
+
+  for (i = 0; i < size; i++)
+    s.a[i].c = 0x1234;
+  for (i = 0; i < size; i++)
+    s.b[i] = 0;
+  for (i = 0; i < size; i++)
+    if (s.a[i].c != 0x1234 || s.b[i] != 0)
+      __builtin_abort ();
+}
+
+int
+main ()
+{
+  foo (15);
+  return 0;
+}

	Jakub

Re: [PATCH] Fix handling of variable length fields in structures (PR c/82210)

[Jeff Law]

On 02/08/2018 11:22 PM, Jakub Jelinek wrote:
> Hi!
> 
> When placing a variable length field into a structure, we need to update
> rli->offset_align for the next field.  We do:
> rli->offset_align = MIN (rli->offset_align, desired_align);
> which updates it according to the start of that VLA field, the problem is
> that if the field doesn't have a size that is a multiple of this alignment
> rli->offset_align will not reflect properly the alignment of the end of that
> field.  E.g. on the testcase, we have a VLA array aligned as a whole (the
> field itself) to 16 bytes / 128 bits, so rli->offset_align remains 128.
> The array has element size 2 bytes / 16 bits, times function argument,
> so the end of the field is worst case aligned just to 16 bits; if we keep
> rli->offset_align as 128 for the next field, then DECL_OFFSET_ALIGN is too
> large. DECL_FIELD_OFFSET documented as:
> /* In a FIELD_DECL, this is the field position, counting in bytes, of the
>    DECL_OFFSET_ALIGN-bit-sized word containing the bit closest to the beginning
>    of the structure.  */
> and when gimplifying COMPONENT_REFs with that field we:
>               tree offset = unshare_expr (component_ref_field_offset (t));
>               tree field = TREE_OPERAND (t, 1);
>               tree factor
>                 = size_int (DECL_OFFSET_ALIGN (field) / BITS_PER_UNIT);
>    
>               /* Divide the offset by its alignment.  */
>               offset = size_binop_loc (loc, EXACT_DIV_EXPR, offset, factor);
> and later on multiply it again by DECL_OFFSET_ALIGN.  The EXACT_DIV_EXPR
> isn't exact.
> 
> Fixed by lowering the rli->offset_align if the size isn't a multiple of
> the align.  We don't have a multiple_of_p variant that would compute
> highest power of two number the expression is known to be a multiple of,
> so I'm just checking the most common case, where the size is a multiple
> of the starting alignment, and otherwise just compute it very
> conservatively.  This will be lower than necessary say for
>   __attribute__((aligned (16))) short field[2 * size];
> - just 16 bits instead of 32.  In theory we could do a binary search
> on power of two numbers in between that high initial rli->offset_align
> for which the first multiple_of_p failed, and the conservative guess
> we do to improve it.  If you think it is worth it, I can code it up.
> 
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
> 
> 2018-02-09  Jakub Jelinek  <jakub@redhat.com>
> 
> 	PR c/82210
> 	* stor-layout.c (place_field): For variable length fields, adjust
> 	offset_align afterwards not just based on the field's alignment,
> 	but also on the size.
> 
> 	* gcc.c-torture/execute/pr82210.c: New test.
OK.
jeff