145
|
1 /* Copyright (C) 2012-2020 Free Software Foundation, Inc.
|
111
|
2 Contributed by Richard Henderson <rth@redhat.com>.
|
|
3
|
|
4 This file is part of the GNU Atomic Library (libatomic).
|
|
5
|
|
6 Libatomic is free software; you can redistribute it and/or modify it
|
|
7 under the terms of the GNU General Public License as published by
|
|
8 the Free Software Foundation; either version 3 of the License, or
|
|
9 (at your option) any later version.
|
|
10
|
|
11 Libatomic is distributed in the hope that it will be useful, but WITHOUT ANY
|
|
12 WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
|
|
13 FOR A PARTICULAR PURPOSE. See the GNU General Public License for
|
|
14 more details.
|
|
15
|
|
16 Under Section 7 of GPL version 3, you are granted additional
|
|
17 permissions described in the GCC Runtime Library Exception, version
|
|
18 3.1, as published by the Free Software Foundation.
|
|
19
|
|
20 You should have received a copy of the GNU General Public License and
|
|
21 a copy of the GCC Runtime Library Exception along with this program;
|
|
22 see the files COPYING3 and COPYING.RUNTIME respectively. If not, see
|
|
23 <http://www.gnu.org/licenses/>. */
|
|
24
|
|
25 #include "libatomic_i.h"
|
|
26
|
|
27
|
|
28 /* If we natively support the load, and if we're unconcerned with extra
|
|
29 barriers (e.g. fully in-order cpu for which barriers are a nop), then
|
|
30 go ahead and expand the operation inline. */
|
|
31 #if !defined(WANT_SPECIALCASE_RELAXED) && !defined(__OPTIMIZE_SIZE__)
|
|
32 # define EXACT_INLINE(N, DEST, SRC, DONE) \
|
|
33 if (C2(HAVE_ATOMIC_LDST_,N)) \
|
|
34 { \
|
|
35 DEST = __atomic_load_n (SRC, __ATOMIC_SEQ_CST); \
|
|
36 DONE; \
|
|
37 }
|
|
38 #else
|
|
39 # define EXACT_INLINE(N, DEST, SRC, DONE)
|
|
40 #endif
|
|
41
|
|
42
|
|
43 #define EXACT_(N, DEST, SRC, DONE) \
|
|
44 do { \
|
|
45 EXACT_INLINE (N, DEST, SRC, DONE); \
|
|
46 DEST = C3(local_,load_,N) (SRC, smodel); \
|
|
47 DONE; \
|
|
48 } while (0)
|
|
49
|
|
50 #define EXACT(N) \
|
|
51 do { \
|
|
52 if (!C2(HAVE_INT,N)) break; \
|
|
53 if ((uintptr_t)mptr & (N - 1)) break; \
|
|
54 EXACT_(N, *PTR(N,rptr), PTR(N,mptr), return); \
|
|
55 } while (0)
|
|
56
|
|
57 #define LARGER(N) \
|
|
58 do { \
|
|
59 if (!C2(HAVE_INT,N)) break; \
|
|
60 r = (uintptr_t)mptr & (N - 1); \
|
|
61 a = (uintptr_t)mptr & -N; \
|
|
62 if (r + n <= N) \
|
|
63 EXACT_ (N, u.C2(i,N), PTR(N,a), goto Lfinish); \
|
|
64 } while (0)
|
|
65
|
|
66
|
|
67 void
|
|
68 libat_load (size_t n, void *mptr, void *rptr, int smodel)
|
|
69 {
|
|
70 union max_size_u u;
|
|
71 uintptr_t r, a;
|
|
72
|
|
73 switch (n)
|
|
74 {
|
|
75 case 0: return;
|
|
76 case 1: EXACT(1); goto L4;
|
|
77 case 2: EXACT(2); goto L4;
|
|
78 case 4: EXACT(4); goto L8;
|
|
79 case 8: EXACT(8); goto L16;
|
|
80 case 16: EXACT(16); break;
|
|
81
|
|
82 case 3: L4: LARGER(4); /* FALLTHRU */
|
|
83 case 5 ... 7: L8: LARGER(8); /* FALLTHRU */
|
|
84 case 9 ... 15: L16: LARGER(16); break;
|
|
85
|
|
86 Lfinish:
|
|
87 memcpy (rptr, u.b + r, n);
|
|
88 return;
|
|
89 }
|
|
90
|
|
91 pre_seq_barrier (smodel);
|
|
92 libat_lock_n (mptr, n);
|
|
93
|
|
94 memcpy (rptr, mptr, n);
|
|
95
|
|
96 libat_unlock_n (mptr, n);
|
|
97 post_seq_barrier (smodel);
|
|
98 }
|
|
99
|
|
100 EXPORT_ALIAS (load);
|