1------------------------------------------------------------------------------
2--                                                                          --
3--                         GNAT COMPILER COMPONENTS                         --
4--                                                                          --
5--                             E X P _ P A K D                              --
6--                                                                          --
7--                                 S p e c                                  --
8--                                                                          --
9--          Copyright (C) 1992-2010, Free Software Foundation, Inc.         --
10--                                                                          --
11-- GNAT is free software;  you can  redistribute it  and/or modify it under --
12-- terms of the  GNU General Public License as published  by the Free Soft- --
13-- ware  Foundation;  either version 3,  or (at your option) any later ver- --
14-- sion.  GNAT is distributed in the hope that it will be useful, but WITH- --
15-- OUT ANY WARRANTY;  without even the  implied warranty of MERCHANTABILITY --
16-- or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License --
17-- for  more details.  You should have  received  a copy of the GNU General --
18-- Public License  distributed with GNAT; see file COPYING3.  If not, go to --
19-- http://www.gnu.org/licenses for a complete copy of the license.          --
20--                                                                          --
21-- GNAT was originally developed  by the GNAT team at  New York University. --
22-- Extensive contributions were provided by Ada Core Technologies Inc.      --
23--                                                                          --
24------------------------------------------------------------------------------
25
26--  Expand routines for manipulation of packed arrays
27
28with Types; use Types;
29
30package Exp_Pakd is
31
32   -------------------------------------
33   -- Implementation of Packed Arrays --
34   -------------------------------------
35
36   --  When a packed array (sub)type is frozen, we create a corresponding
37   --  type that will be used to hold the bits of the packed value, and
38   --  store the entity for this type in the Packed_Array_Type field of the
39   --  E_Array_Type or E_Array_Subtype entity for the packed array.
40
41   --  This packed array type has the name xxxPn, where xxx is the name
42   --  of the packed type, and n is the component size. The expanded
43   --  declaration declares a type that is one of the following:
44
45   --    For an unconstrained array with component size 1,2,4 or any other
46   --    odd component size. These are the cases in which we do not need
47   --    to align the underlying array.
48
49   --      type xxxPn is new Packed_Bytes1;
50
51   --    For an unconstrained array with component size that is divisible
52   --    by 2, but not divisible by 4 (other than 2 itself). These are the
53   --    cases in which we can generate better code if the underlying array
54   --    is 2-byte aligned (see System.Pack_14 in file s-pack14 for example).
55
56   --      type xxxPn is new Packed_Bytes2;
57
58   --    For an unconstrained array with component size that is divisible
59   --    by 4, other than powers of 2 (which either come under the 1,2,4
60   --    exception above, or are not packed at all). These are cases where
61   --    we can generate better code if the underlying array is 4-byte
62   --    aligned (see System.Pack_20 in file s-pack20 for example).
63
64   --      type xxxPn is new Packed_Bytes4;
65
66   --    For a constrained array with a static index type where the number
67   --    of bits does not exceed the size of Unsigned:
68
69   --      type xxxPn is new Unsigned range 0 .. 2 ** nbits - 1;
70
71   --    For a constrained array with a static index type where the number
72   --    of bits is greater than the size of Unsigned, but does not exceed
73   --    the size of Long_Long_Unsigned:
74
75   --       type xxxPn is new Long_Long_Unsigned range 0 .. 2 ** nbits - 1;
76
77   --    For all other constrained arrays, we use one of
78
79   --       type xxxPn is new Packed_Bytes1 (0 .. m);
80   --       type xxxPn is new Packed_Bytes2 (0 .. m);
81   --       type xxxPn is new Packed_Bytes4 (0 .. m);
82
83   --    where m is calculated (from the length of the original packed array)
84   --    to hold the required number of bits, and the choice of the particular
85   --    Packed_Bytes{1,2,4} type is made on the basis of alignment needs as
86   --    described above for the unconstrained case.
87
88   --  When a variable of packed array type is allocated, gigi will allocate
89   --  the amount of space indicated by the corresponding packed array type.
90   --  However, we do NOT attempt to rewrite the types of any references or
91   --  to retype the variable itself, since this would cause all kinds of
92   --  semantic problems in the front end (remember that expansion proceeds
93   --  at the same time as analysis).
94
95   --  For an indexed reference to a packed array, we simply convert the
96   --  reference to the appropriate equivalent reference to the object
97   --  of the packed array type (using unchecked conversion).
98
99   --  In some cases (for internally generated types, and for the subtypes
100   --  for record fields that depend on a discriminant), the corresponding
101   --  packed type cannot be easily generated in advance. In these cases,
102   --  we generate the required subtype on the fly at the reference point.
103
104   --  For the modular case, any unused bits are initialized to zero, and
105   --  all operations maintain these bits as zero (where necessary all
106   --  unchecked conversions from corresponding array values require
107   --  these bits to be clear, which is done automatically by gigi).
108
109   --  For the array cases, there can be unused bits in the last byte, and
110   --  these are neither initialized, nor treated specially in operations
111   --  (i.e. it is allowable for these bits to be clobbered, e.g. by not).
112
113   ---------------------------
114   -- Endian Considerations --
115   ---------------------------
116
117   --  The standard does not specify the way in which bits are numbered in
118   --  a packed array. There are two reasonable rules for deciding this:
119
120   --    Store the first bit at right end (low order) word. This means
121   --    that the scaled subscript can be used directly as a left shift
122   --    count (if we put bit 0 at the left end, then we need an extra
123   --    subtract to compute the shift count).
124
125   --    Layout the bits so that if the packed boolean array is overlaid on
126   --    a record, using unchecked conversion, then bit 0 of the array is
127   --    the same as the bit numbered bit 0 in a record representation
128   --    clause applying to the record. For example:
129
130   --       type Rec is record
131   --          C : Bits4;
132   --          D : Bits7;
133   --          E : Bits5;
134   --       end record;
135
136   --       for Rec use record
137   --          C at 0 range  0  .. 3;
138   --          D at 0 range  4 .. 10;
139   --          E at 0 range 11 .. 15;
140   --       end record;
141
142   --       type P16 is array (0 .. 15) of Boolean;
143   --       pragma Pack (P16);
144
145   --    Now if we use unchecked conversion to convert a value of the record
146   --    type to the packed array type, according to this second criterion,
147   --    we would expect field D to occupy bits 4..10 of the Boolean array.
148
149   --  Although not required, this correspondence seems a highly desirable
150   --  property, and is one that GNAT decides to guarantee. For a little
151   --  endian machine, we can also meet the first requirement, but for a
152   --  big endian machine, it will be necessary to store the first bit of
153   --  a Boolean array in the left end (most significant) bit of the word.
154   --  This may cost an extra instruction on some machines, but we consider
155   --  that a worthwhile price to pay for the consistency.
156
157   --  One more important point arises in the case where we have a constrained
158   --  subtype of an unconstrained array. Take the case of 20 bits. For the
159   --  unconstrained representation, we would use an array of bytes:
160
161   --     Little-endian case
162   --       8-7-6-5-4-3-2-1  16-15-14-13-12-11-10-9  x-x-x-x-20-19-18-17
163
164   --     Big-endian case
165   --       1-2-3-4-5-6-7-8  9-10-11-12-13-14-15-16  17-18-19-20-x-x-x-x
166
167   --   For the constrained case, we use a 20-bit modular value, but in
168   --   general this value may well be stored in 32 bits. Let's look at
169   --   what it looks like:
170
171   --     Little-endian case
172
173   --       x-x-x-x-x-x-x-x-x-x-x-x-20-19-18-17-...-10-9-8-7-6-5-4-3-2-1
174
175   --         which stored in memory looks like
176
177   --       8-7-...-2-1  16-15-...-10-9  x-x-x-x-20-19-18-17  x-x-x-x-x-x-x
178
179   --   An important rule is that the constrained and unconstrained cases
180   --   must have the same bit representation in memory, since we will often
181   --   convert from one to the other (e.g. when calling a procedure whose
182   --   formal is unconstrained). As we see, that criterion is met for the
183   --   little-endian case above. Now let's look at the big-endian case:
184
185   --     Big-endian case
186
187   --       x-x-x-x-x-x-x-x-x-x-x-x-1-2-3-4-5-6-7-8-9-10-...-17-18-19-20
188
189   --         which stored in memory looks like
190
191   --       x-x-x-x-x-x-x-x  x-x-x-x-1-2-3-4  5-6-...11-12  13-14-...-19-20
192
193   --   That won't do, the representation value in memory is NOT the same in
194   --   the constrained and unconstrained case. The solution is to store the
195   --   modular value left-justified:
196
197   --       1-2-3-4-5-6-7-8-9-10-...-17-18-19-20-x-x-x-x-x-x-x-x-x-x-x
198
199   --         which stored in memory looks like
200
201   --       1-2-...-7-8  9-10-...15-16  17-18-19-20-x-x-x-x  x-x-x-x-x-x-x-x
202
203   --   and now, we do indeed have the same representation for the memory
204   --   version in the constrained and unconstrained cases.
205
206   -----------------
207   -- Subprograms --
208   -----------------
209
210   procedure Create_Packed_Array_Type (Typ  : Entity_Id);
211   --  Typ is a array type or subtype to which pragma Pack applies. If the
212   --  Packed_Array_Type field of Typ is already set, then the call has no
213   --  effect, otherwise a suitable type or subtype is created and stored
214   --  in the Packed_Array_Type field of Typ. This created type is an Itype
215   --  so that Gigi will simply elaborate and freeze the type on first use
216   --  (which is typically the definition of the corresponding array type).
217   --
218   --  Note: although this routine is included in the expander package for
219   --  packed types, it is actually called unconditionally from Freeze,
220   --  whether or not expansion (and code generation) is enabled. We do this
221   --  since we want gigi to be able to properly compute type characteristics
222   --  (for the Data Decomposition Annex of ASIS, and possible other future
223   --  uses) even if code generation is not active. Strictly this means that
224   --  this procedure is not part of the expander, but it seems appropriate
225   --  to keep it together with the other expansion routines that have to do
226   --  with packed array types.
227
228   procedure Expand_Packed_Boolean_Operator (N : Node_Id);
229   --  N is an N_Op_And, N_Op_Or or N_Op_Xor node whose operand type is a
230   --  packed boolean array. This routine expands the appropriate operations
231   --  to carry out the logical operation on the packed arrays. It handles
232   --  both the modular and array representation cases.
233
234   procedure Expand_Packed_Element_Reference (N : Node_Id);
235   --  N is an N_Indexed_Component node whose prefix is a packed array. In
236   --  the bit packed case, this routine can only be used for the expression
237   --  evaluation case, not the assignment case, since the result is not a
238   --  variable. See Expand_Bit_Packed_Element_Set for how the assignment case
239   --  is handled in the bit packed case. For the enumeration case, the result
240   --  of this call is always a variable, so the call can be used for both the
241   --  expression evaluation and assignment cases.
242
243   procedure Expand_Bit_Packed_Element_Set (N : Node_Id);
244   --  N is an N_Assignment_Statement node whose name is an indexed
245   --  component of a bit-packed array. This procedure rewrites the entire
246   --  assignment statement with appropriate code to set the referenced
247   --  bits of the packed array type object. Note that this procedure is
248   --  used only for the bit-packed case, not for the enumeration case.
249
250   procedure Expand_Packed_Eq (N : Node_Id);
251   --  N is an N_Op_Eq node where the operands are packed arrays whose
252   --  representation is an array-of-bytes type (the case where a modular
253   --  type is used for the representation does not require any special
254   --  handling, because in the modular case, unused bits are zeroes.
255
256   procedure Expand_Packed_Not (N : Node_Id);
257   --  N is an N_Op_Not node where the operand is packed array of Boolean
258   --  in standard representation (i.e. component size is one bit). This
259   --  procedure expands the corresponding not operation. Note that the
260   --  non-standard representation case is handled by using a loop through
261   --  elements generated by the normal non-packed circuitry.
262
263   function Involves_Packed_Array_Reference (N : Node_Id) return Boolean;
264   --  N is the node for a name. This function returns true if the name
265   --  involves a packed array reference. A node involves a packed array
266   --  reference if it is itself an indexed component referring to a bit-
267   --  packed array, or it is a selected component whose prefix involves
268   --  a packed array reference.
269
270   procedure Expand_Packed_Address_Reference (N : Node_Id);
271   --  The node N is an attribute reference for the 'Address reference, where
272   --  the prefix involves a packed array reference. This routine expands the
273   --  necessary code for performing the address reference in this case.
274
275   procedure Expand_Packed_Bit_Reference (N : Node_Id);
276   --  The node N is an attribute reference for the 'Bit reference, where the
277   --  prefix involves a packed array reference. This routine expands the
278   --  necessary code for performing the bit reference in this case.
279
280end Exp_Pakd;
281