« back to all changes in this revision
Viewing changes to debian/NEWS.gcc
- browse files at revision 165
- compare with another revision
- download diff
- download tarball
- view history from revision 165
- Committer: Package Import Robot
- Author(s): Matthias Klose
- Date: 2013-05-12 14:56:42 UTC
- mfrom: (1.20.14) (135.1.29 saucy-proposed)
- Revision ID: package-import@ubuntu.com-20130512145642-jkg0131tyv5io67o
Tags: 20130512-0ubuntu1
Snapshot, taken from the trunk (20130512)
- files added:
- debian/patches/aarch64-libjava.diff
- files removed:
- debian/g++-BV-spu.overrides
- files modified:
- debian/NEWS.gcc
added
removed
debian/NEWS.gcc
1
GCC 4.7 Release Series -- Changes, New Features, and Fixes
1
GCC 4.8 Release Series -- Changes, New Features, and Fixes
2
2
==========================================================
3
3
4
4
5
5
Caveats
6
6
=======
7
7
8
- The -fconserve-space flag has been deprecated. The flag had no
9
effect for most targets: only targets without a global .bss
10
section and without support for switchable sections. Furthermore,
11
the flag only had an effect for G++, where it could result in
12
wrong semantics (please refer to the GCC manual for further
13
details). The flag will be removed in GCC 4.8.
14
15
- Support for a number of older systems and recently unmaintained or
16
untested target ports of GCC has been declared obsolete in GCC 4.7.
17
Unless there is activity to revive them, the next release of GCC
18
will have their sources permanently removed.
19
20
- All GCC ports for the following processor architectures have been
21
declared obsolete:
22
23
- picoChip (picochip-*)
24
25
The following ports for individual systems on particular
26
architectures have been obsoleted:
27
28
- IRIX 6.5 (mips-sgi-irix6.5)
29
- MIPS OpenBSD (mips*-*-openbsd*)
30
- Solaris 8 (*-*-solaris2.8). Details can be found in the announcement.
31
- Tru64 UNIX V5.1 (alpha*-dec-osf5.1*)
32
33
- On ARM, when compiling for ARMv6 (but not ARMv6-M), ARMv7-A,
34
ARMv7-R, or ARMv7-M, the new option -munaligned-access is active by
35
default, which for some source codes generates code that accesses
36
memory on unaligned addresses. This will require the kernel of
37
those systems to enable such accesses (controlled by CP15 register
38
c1, refer to ARM documentation). Alternatively or for compatibility
39
with kernels where unaligned accesses are not supported, all code
40
has to be compiled with -mno-unaligned-access. Linux/ARM in
41
official releases has automatically and unconditionally supported
42
unaligned accesses as emitted by GCC due to this option being
43
active, since Linux version 2.6.28.
44
45
- Support on ARM for the legacy floating-point accelerator (FPA) and
46
the mixed-endian floating-point format that it used has been
47
obsoleted. The ports that still use this format have been obsoleted
48
as well. Many legacy ARM ports already provide an alternative that
49
uses the VFP floating-point format. The obsolete ports will be
50
deleted in the next release. The obsolete ports with alternatives
51
are:
52
53
- arm*-*-rtems (use arm*-*-rtemseabi)
54
- arm*-*-linux-gnu (use arm*-*-linux-gnueabi)
55
- arm*-*-elf (use arm*-*-eabi)
56
- arm*-*-uclinux* (use arm*-*-uclinux*eabi)
57
58
Note, however, that these alternatives are not binary compatible
59
with their legacy counterparts (although some can support running
60
legacy applications).
61
62
The obsolete ports that currently lack a modern alternative are:
63
64
- arm*-*-ecos-elf
65
- arm*-*-freebsd
66
- arm*-wince-pe*
67
68
New ports that support more recent versions of the architecture are
69
welcome.
70
71
- Support for the Maverick co-processor on ARM has been obsoleted. Code to
72
support it will be deleted in the next release.
73
74
- Support has been removed for Unix International threads on Solaris 2,
75
so the --enable-threads=solaris configure option and the -threads
76
compiler option don't work any longer.
77
78
- Support has been removed for the Solaris BSD Compatibility Package,
79
which lives in /usr/ucbinclude and /usr/ucblib. It has been removed
80
from Solaris 11, and was only intended as a migration aid from
81
SunOS 4 to SunOS 5. The -compat-bsd compiler option is not
82
recognized any longer.
83
84
- The AVR port's libgcc has been improved and its multilib structure
85
has been enhanced. As a result, all objects contributing to an
86
application must either be compiled with GCC versions up to 4.6.x
87
or with GCC versions 4.7.0 or later.
88
89
- The ARM port's -mwords-little-endian option has been deprecated. It will
90
be removed in a future release.
91
92
- Support has been removed for the NetWare x86 configuration obsoleted in
93
GCC 4.6.
94
95
- It is no longer possible to use the "l" constraint in MIPS16 asm statements.
96
97
- GCC versions 4.7.0 and 4.7.1 had changes to the C++ standard
98
library which affected the ABI in C++11 mode: a data member was
99
added to std::list changing its size and altering the definitions
100
of some member functions, and std::pair's move constructor was
101
non-trivial which altered the calling convention for functions with
102
std::pair arguments or return types. The ABI incompatibilities have
103
been fixed for GCC version 4.7.2 but as a result C++11 code
104
compiled with GCC 4.7.0 or 4.7.1 may be incompatible with C++11
105
code compiled with different GCC versions and with C++98/C++03 code
106
compiled with any version.
107
108
- On ARM, a bug has been fixed in GCC's implementation of the AAPCS
109
rules for the layout of vectors that could lead to wrong code being
110
generated. Vectors larger than 8 bytes in size are now by default
111
aligned to an 8-byte boundary. This is an ABI change: code that
112
makes explicit use of vector types may be incompatible with binary
113
objects built with older versions of GCC. Auto-vectorized code is
114
not affected by this change. (This change affects GCC versions
115
4.7.2 and later.)
116
117
- More information on porting to GCC 4.7 from previous versions of GCC can
118
be found in the porting_guide for this release.
119
120
121
General Optimizer Improvements
122
==============================
123
124
- Support for a new parameter --param case-values-threshold=n was
125
added to allow users to control the cutoff between doing switch
126
statements as a series of if statements and using a jump table.
8
GCC now uses C++ as its implementation language. This means that to build GCC
9
from sources, you will need a C++ compiler that understands C++ 2003. For more
10
details on the rationale and specific changes, please refer to the C++
11
conversion page.
12
13
To enable the Graphite framework for loop optimizations you now need CLooG
14
version 0.18.0 and ISL version 0.11.1. Both can be obtained from the GCC
15
infrastructure directory. The installation manual contains more information
16
about requirements to build GCC.
17
18
GCC now uses a more aggressive analysis to derive an upper bound for the
19
number of iterations of loops using constraints imposed by language standards.
20
This may cause non-conforming programs to no longer work as expected, such as
21
SPEC CPU 2006 464.h264ref and 416.gamess. A new option, -fno-aggressive-loop-
22
optimizations, was added to disable this aggressive analysis. In some loops
23
that have known constant number of iterations, but undefined behavior is known
24
to occur in the loop before reaching or during the last iteration, GCC will
25
warn about the undefined behavior in the loop instead of deriving lower upper
26
bound of the number of iterations for the loop. The warning can be disabled
27
with -Wno-aggressive-loop-optimizations.
28
29
On ARM, a bug has been fixed in GCC's implementation of the AAPCS rules for
30
the layout of vectors that could lead to wrong code being generated. Vectors
31
larger than 8 bytes in size are now by default aligned to an 8-byte boundary.
32
This is an ABI change: code that makes explicit use of vector types may be
33
incompatible with binary objects built with older versions of GCC. Auto-
34
vectorized code is not affected by this change.
35
36
On AVR, support has been removed for the command-line option -mshort-calls
37
deprecated in GCC 4.7.
38
39
On AVR, the configure option --with-avrlibc supported since GCC 4.7.2 is
40
turned on per default for all non-RTEMS configurations. This option arranges
41
for a better integration of AVR_Libc with avr-gcc. For technical details, see
42
PR54461. To turn off the option in non-RTEMS configurations, use --with-
43
avrlibc=no. If the compiler is configured for RTEMS, the option is always
44
turned off.
45
46
More information on porting to GCC 4.8 from previous versions of GCC can be
47
found in the porting guide (http://gcc.gnu.org/gcc-4.8/porting_to.html) for
48
this release.
49
50
General Optimizer Improvements (and Changes)
51
============================================
52
53
- DWARF4 is now the default when generating DWARF debug information. When -g
54
is used on a platform that uses DWARF debugging information, GCC will now
55
default to -gdwarf-4 -fno-debug-types-section.
56
GDB 7.5, Valgrind 3.8.0 and elfutils 0.154 debug information consumers
57
support DWARF4 by default. Before GCC 4.8 the default version used was
58
DWARF2. To make GCC 4.8 generate an older DWARF version use -g together with
59
-gdwarf-2 or -gdwarf-3. The default for Darwin and VxWorks is still
60
-gdwarf-2 -gstrict-dwarf.
61
62
- A new general optimization level, -Og, has been introduced. It addresses the
63
need for fast compilation and a superior debugging experience while
64
providing a reasonable level of runtime performance. Overall experience for
65
development should be better than the default optimization level -O0.
66
67
- A new option -ftree-partial-pre was added to control the partial redundancy
68
elimination (PRE) optimization. This option is enabled by default at the -O3
69
optimization level, and it makes PRE more aggressive.
70
71
- The option -fconserve-space has been removed; it was no longer useful on
72
most targets since GCC supports putting variables into BSS without making
73
them common.
74
75
- The struct reorg and matrix reorg optimizations (command-line-options
76
-fipa-struct-reorg and -fipa-matrix-reorg) have been -removed. They did not
77
always work correctly, nor did they work -with link-time optimization (LTO),
78
hence were only applicable to -programs consisting of a single translation
79
unit.
80
81
- Several scalability bottle-necks have been removed from GCC's optimization
82
passes. Compilation of extremely large functions, e.g. due to the use of the
83
flatten attribute in the "Eigen" C++ linear algebra templates library, is
84
significantly faster than previous releases of GCC.
127
85
128
86
- Link-time optimization (LTO) improvements:
129
87
130
- Improved scalability and reduced memory usage. Link time
131
optimization of Firefox now requires 3GB of RAM on a 64-bit
132
system, while over 8GB was needed previously. Linking time has
133
been improved, too. The serial stage of linking Firefox has been
134
sped up by about a factor of 10.
135
136
- Reduced size of object files and temporary storage used during linking.
137
138
- Streaming performance (both outbound and inbound) has been improved.
139
140
- ld -r is now supported with LTO.
141
142
- Several bug fixes, especially in symbol table handling and merging.
88
- LTO partitioning has been rewritten for better reliability and
89
maintanibility. Several important bugs leading to link failures have been
90
fixed.
143
91
144
92
- Interprocedural optimization improvements:
145
93
146
- Heuristics now take into account that after inlining code will be
147
optimized out because of known values (or properties) of function
148
parameters. For example:
149
150
void foo(int a)
151
{
152
if (a > 10)
153
... huge code ...
154
}
155
void bar (void)
156
{
157
foo (0);
158
}
159
160
The call of foo will be inlined into bar even when optimizing for
161
code size. Constructs based on __builtin_constant_p are now
162
understood by the inliner and code size estimates are evaluated a
163
lot more realistically.
164
165
- The representation of C++ virtual thunks and aliases (both
166
implicit and defined via the alias attribute) has been
167
re-engineered. Aliases no longer pose optimization barriers and
168
calls to an alias can be inlined and otherwise optimized.
169
170
- The inter-procedural constant propagation pass has been
171
rewritten. It now performs generic function specialization. For
172
example when compiling the following:
173
174
void foo(bool flag)
175
{
176
if (flag)
177
... do something ...
178
else
179
... do something else ...
180
}
181
void bar (void)
182
{
183
foo (false);
184
foo (true);
185
foo (false);
186
foo (true);
187
foo (false);
188
foo (true);
189
}
190
191
GCC will now produce two copies of foo. One with flag being true,
192
while other with flag being false. This leads to performance
193
improvements previously possible only by inlining all calls.
194
Cloning causes a lot less code size growth.
195
196
- A string length optimization pass has been added. It attempts to
197
track string lengths and optimize various standard C string
198
functions like strlen, strchr, strcpy, strcat, stpcpy and their
199
_FORTIFY_SOURCE counterparts into faster alternatives. This pass
200
is enabled by default at -O2 or above, unless optimizing for
201
size, and can be disabled by the - fno-optimize-strlen
202
option. The pass can e.g. optimize
203
204
char *bar (const char *a)
205
{
206
size_t l = strlen (a) + 2;
207
char *p = malloc (l); if (p == NULL) return p;
208
strcpy (p, a); strcat (p, "/"); return p;
209
}
210
211
into:
212
213
char *bar (const char *a)
214
{
215
size_t tmp = strlen (a);
216
char *p = malloc (tmp + 2); if (p == NULL) return p;
217
memcpy (p, a, tmp); memcpy (p + tmp, "/", 2); return p;
218
}
219
220
or for hosted compilations where stpcpy is available in the runtime
221
and headers provide its prototype, e.g.
222
223
void foo (char *a, const char *b, const char *c, const char *d)
224
{
225
strcpy (a, b); strcat (a, c); strcat (a, d);
226
}
227
228
can be optimized into:
229
230
void foo (char *a, const char *b, const char *c, const char *d)
231
{
232
strcpy (stpcpy (stpcpy (a, b), c), d);
233
}
94
- A new symbol table has been implemented. It builds on existing callgraph
95
and varpool modules and provide a new API. Unusual symbol visibilities and
96
aliases are handled more consistently leading to, for example, more
97
aggressive unreachable code removal with LTO.
98
99
- The inline heuristic can now bypass limits on the size of of inlined
100
functions when the inlining is particularly profitable. This happens, for
101
example, when loop bounds or array strides get propagated.
102
103
- Values passed through aggregates (either by value or reference) are now
104
propagated at the inter-procedural level leading to better inlining
105
decisions (for example in the case of Fortran array descriptors) and
106
devirtualization.
107
108
- AddressSanitizer, a fast memory error detector, has been added and can be
109
enabled via -fsanitize=address. Memory access instructions will be
110
instrumented to detect heap-, stack-, and global-buffer overflow as well as
111
use-after-free bugs. To get nicer stacktraces, use -fno-omit-frame-pointer.
112
The AddressSanitizer is available on IA-32/x86-64/x32/PowerPC/PowerPC64 GNU/
113
Linux and on x86-64 Darwin.
114
115
- ThreadSanitizer has been added and can be enabled via -fsanitize=thread.
116
Instructions will be instrumented to detect data races. The ThreadSanitizer
117
is available on x86-64 GNU/Linux.
234
118
235
119
236
120
New Languages and Language specific improvements
237
121
================================================
238
122
239
- Version 3.1 of the OpenMP_specification is now supported for the C, C++,
240
and Fortran compilers.
241
242
243
Ada
244
---
245
246
- The command-line option -feliminate-unused-debug-types has been re-
247
enabled by default, as it is for the other languages, leading to a
248
reduction in debug info size of 12.5% and more for relevant cases,
249
as well as to a small compilation speedup.
250
251
123
252
124
C family
253
125
--------
254
126
255
- A new built-in, __builtin_assume_aligned, has been added, through which
256
the compiler can be hinted about pointer alignment and can use it to
257
improve generated code.
258
259
- A new -Wunused-local-typedefs warning was added for C, C++,
260
Objective-C and Objective-C++. This warning diagnoses typedefs
261
locally defined in a function, and otherwise not used.
262
263
- A new experimental -ftrack-macro-expansion option was added for C,
264
C++, Objective-C, Objective-C++ and Fortran. It allows the
265
compiler to emit diagnostic about the current macro expansion
266
stack when a compilation error occurs in a macro expansion.
267
268
- Experimental support for transactional memory has been added. It includes
269
support in the compiler, as well as a supporting runtime library called
270
libitm. To compile code with transactional memory constructs, use the -
271
fgnu-tm option.
272
273
Support is currently available for Alpha, ARM, PowerPC, SH, SPARC, and
274
32-bit/64-bit x86 platforms.
275
276
For more details on transactional memory see the_GCC_WiKi.
277
278
- Support for atomic operations specifying the C++11/C11 memory model has
279
been added. These new __atomic routines replace the existing __sync
280
built-in routines.
281
282
Atomic support is also available for memory blocks. Lock-free
283
instructions will be used if a memory block is the same size and
284
alignment as a supported integer type. Atomic operations which do not
285
have lock-free support are left as function calls. A set of library
286
functions is available on the GCC atomic wiki in the "External Atomics
287
Library" section.
288
289
For more details on the memory models and features, see the atomic_wiki.
290
291
- When a binary operation is performed on vector types and one of the
292
operands is a uniform vector, it is possible to replace the vector with
293
the generating element. For example:
294
295
typedef int v4si __attribute__ ((vector_size (16)));
296
v4si res, a = {1,2,3,4};
297
int x;
298
299
res = 2 + a; /* means {2,2,2,2} + a */
300
res = a - x; /* means a - {x,x,x,x} */
301
302
303
C
304
-
305
306
- There is support for some more features from the C11 revision of the ISO
307
C standard. GCC now accepts the options -std=c11 and -std=gnu11, in
308
addition to the previous -std=c1x and -std=gnu1x.
309
310
- Unicode strings (previously supported only with options such as
311
-std=gnu11, now supported with -std=c11), and the predefined macros
312
__STDC_UTF_16__ and __STDC_UTF_32__.
313
314
- Nonreturning functions (_Noreturn and <stdnoreturn.h>).
315
316
- Alignment support (_Alignas, _Alignof, max_align_t, <stdalign.h>).
317
318
- A built-in function __builtin_complex is provided to support C
319
library implementation of the CMPLX family of macros.
127
- Each diagnostic emitted now includes the original source line and a caret
128
'^' indicating the column. The option -fno-diagnostics-show-caret suppresses
129
this information.
130
131
- The option -ftrack-macro-expansion=2 is now enabled by default. This allows
132
the compiler to display the macro expansion stack in diagnostics. Combined
133
with the caret information, an example diagnostic showing these two features
134
is:
135
136
t.c:1:94: error: invalid operands to binary < (have ‘struct mystruct’ and ‘float’)
137
#define MYMAX(A,B) __extension__ ({ __typeof__(A) __a = (A);
138
__typeof__(B) __b = (B);
139
__a < __b ? __b : __a; })
140
^
141
t.c:7:7: note: in expansion of macro 'MYMAX'
142
X = MYMAX(P, F);
143
^
144
145
- A new -Wsizeof-pointer-memaccess warning has been added (also enabled by
146
-Wall) to warn about suspicious length parameters to certain string and
147
memory built-in functions if the argument uses sizeof. This warning warns
148
e.g. about memset (ptr, 0, sizeof (ptr)); if ptr is not an array, but a
149
pointer, and suggests a possible fix, or about
150
memcpy (&foo, ptr, sizeof (&foo));.
151
152
- The new option -Wpedantic is an alias for -pedantic, which is now
153
deprecated. The forms -Wno-pedantic, -Werror=pedantic, and -Wno-
154
error=pedantic work in the same way as for any other -W option. One caveat
155
is that -Werror=pedantic is not equivalent to -pedantic-errors, since the
156
latter makes into errors some warnings that are not controlled by
157
-Wpedantic, and the former only affects diagnostics that are disabled when
158
using -Wno-pedantic.
159
160
- The option -Wshadow no longer warns if a declaration shadows a function
161
declaration, unless the former declares a function or pointer to function,
162
because this is a_common_and_valid_case_in_real-world_code.
320
163
321
164
322
165
C++
323
166
---
324
167
325
- G++ now accepts the -std=c++11, -std=gnu++11, and -Wc++11-compat options,
326
which are equivalent to -std=c++0x, -std=gnu++0x, and -Wc++0x-compat,
327
respectively.
328
329
- G++ now implements C++11 extended friend syntax:
330
331
template<class W>
332
class Q
333
{
334
static const int I = 2;
335
public:
336
friend W;
337
};
338
339
struct B
340
{
341
int ar[Q<B>::I];
342
};
343
344
- Thanks to Ville Voutilainen, G++ now implements C++11 explicit override
345
control.
346
347
struct B {
348
virtual void f() const final;
349
virtual void f(int);
350
};
351
352
struct D : B {
353
void f() const; // error: D::f attempts to
354
override final B::f
355
void f(long) override; // error: doesn't override
356
anything
357
void f(int) override; // ok
358
};
359
360
struct E final { };
361
struct F: E { }; // error: deriving from final class
362
363
- G++ now implements C++11 non-static data member initializers.
364
365
struct A {
366
int i = 42;
367
} a; // initializes a.i to 42
368
369
- Thanks to Ed Smith-Rowland, G++ now implements C++11 user-defined
370
literals.
371
372
// Not actually a good approximation. :)
373
constexpr long double operator"" _degrees (long double d)
374
{ return d * 0.0175; }
375
long double pi = 180.0_degrees;
376
377
- G++ now implements C++11 alias-declarations.
378
379
template <class T> using Ptr = T*;
380
Ptr<int> ip; // decltype(ip) is int*
381
382
- Thanks to Ville Voutilainen and Pedro Lamarão, G++ now implements C++11
383
delegating constructors.
384
385
struct A {
386
A(int);
387
A(): A(42) { } // delegate to the A(int) constructor
388
};
389
390
- G++ now fully implements C++11 atomic classes rather than just integer
391
derived classes.
392
393
class POD {
394
int a;
395
int b;
396
};
397
std::atomic<POD> my_atomic_POD;
398
399
- G++ now sets the predefined macro __cplusplus to the correct value,
400
199711L for C++98/03, and 201103L for C++11.
401
402
- G++ now correctly implements the two-phase lookup rules such that an
403
unqualified name used in a template must have an appropriate declaration
404
found either in scope at the point of definition of the template or by
405
argument-dependent lookup at the point of instantiation. As a result,
406
code that relies on a second unqualified lookup at the point of
407
instantiation to find functions declared after the template or in
408
dependent bases will be rejected. The compiler will suggest ways to fix
409
affected code, and using the -fpermissive compiler flag will allow the
410
code to compile with a warning.
411
412
template <class T>
413
void f() { g(T()); } // error, g(int) not found by argument-
414
dependent lookup
415
void g(int) { } // fix by moving this declaration before the
416
declaration of f
417
418
template <class T>
419
struct A: T {
420
// error, B::g(B) not found by argument-dependent lookup
421
void f() { g(T()); } // fix by using this->g or A::g
422
};
423
424
struct B { void g(B); };
425
426
int main()
427
{
428
f<int>();
429
A<B>().f();
430
}
431
432
- G++ now properly re-uses stack space allocated for temporary objects when
433
their lifetime ends, which can significantly lower stack consumption for
434
some C++ functions. As a result of this, some code with undefined
435
behavior will now break:
436
437
const int &f(const int &i) { return i; }
438
....
439
const int &x = f(1);
440
const int &y = f(2);
441
442
Here, x refers to the temporary allocated to hold the 1 argument, which
443
only lives until the end of the initialization; it immediately becomes a
444
dangling reference. So the next statement re-uses the stack slot to hold
445
the 2 argument, and users of x get that value instead.
446
Note that this should not cause any change of behavior for temporaries of
447
types with non-trivial destructors, as they are already destroyed at end
448
of full-expression; the change is that now the storage is released as
449
well.
450
451
- A new command-line option -Wdelete-non-virtual-dtor has been added to
452
warn when delete is used to destroy an instance of a class which has
453
virtual functions and non-virtual destructor. It is unsafe to delete an
454
instance of a derived class through a pointer to a base class if the base
455
class does not have a virtual destructor. This warning is enabled by -Wall.
456
457
- A new command-line option -Wzero-as-null-pointer-constant has been added
458
to warn when a literal '0' is used as null pointer constant. It can be
459
useful to facilitate the conversion to nullptr in C++11.
460
461
- As per C++98, access-declarations are now deprecated by G++. Using-
462
declarations are to be used instead. Furthermore, some efforts have been
463
made to improve the support of class scope using-declarations. In
464
particular, using-declarations referring to a dependent type now work as
465
expected (bug_c++/14258).
466
467
- The ELF symbol visibility of a template instantiation is now properly
468
constrained by the visibility of its template arguments (bug_c++/35688).
168
- G++ now implements the C++11 thread_local keyword; this differs from the GNU
169
__thread keyword primarily in that it allows dynamic initialization and
170
destruction semantics. Unfortunately, this support requires a run-time
171
penalty for references to non-function-local thread_local variables defined
172
in a different translation unit even if they don't need dynamic
173
initialization, so users may want to continue to use __thread for TLS
174
variables with static initialization semantics.
175
176
If the programmer can be sure that no use of the variable in a non-defining
177
TU needs to trigger dynamic initialization (either because the variable is
178
statically initialized, or a use of the variable in the defining TU will be
179
executed before any uses in another TU), they can avoid this overhead with
180
the -fno-extern-tls-init option.
181
182
OpenMP threadprivate variables now also support dynamic initialization and
183
destruction by the same mechanism.
184
185
- G++ now implements the C++11 attribute syntax, e.g.
186
187
[[noreturn]] void f();
188
189
and also the alignment specifier, e.g.
190
191
alignas(double) int i;
192
193
- G++ now implements C++11 inheriting constructors, e.g.
194
195
struct A { A(int); };
196
struct B: A { using A::A; }; // defines B::B(int)
197
B b(42); // OK
198
199
- As of GCC 4.8.1, G++ implements the change to decltype semantics from N3276.
200
201
struct A f();
202
decltype(f()) g(); // OK, return type of f() is not required
203
to be complete.
204
205
- G++ now supports a -std=c++1y option for experimentation with features
206
proposed for the next revision of the standard, expected around 2017.
207
Currently the only difference from -std=c++11 is support for return type
208
deduction in normal functions, as proposed in N3386.
209
210
- The G++ namespace association extension, __attribute ((strong)), has been
211
deprecated. Inline namespaces should be used instead.
212
213
- G++ now supports a -fext-numeric-literal option to control whether GNU
214
numeric literal suffixes are accepted as extensions or processed as C++11
215
user-defined numeric literal suffixes. The flag is on (use suffixes for GNU
216
literals) by default for -std=gnu++*, and -std=c++98. The flag is off (use
217
suffixes for user-defined literals) by default for -std=c++11 and later.
469
218
470
219
471
220
Runtime Library (libstdc++)
474
223
- Improved_experimental_support_for_the_new_ISO_C++_standard,_C++11,
475
224
including:
476
225
477
- using noexcept in most of the library;
478
479
- implementations of pointer_traits, allocator_traits and
480
scoped_allocator_adaptor;
481
482
- uses-allocator construction for tuple;
483
484
- vector meets the allocator-aware container requirements;
485
486
- replacing monotonic_clock with steady_clock;
487
488
- enabling the thread support library on most POSIX targets;
489
490
- many small improvements to conform to the FDIS.
491
492
- Added --enable-clocale=newlib configure option.
493
494
- Debug Mode iterators for unordered associative containers.
495
496
- Avoid polluting the global namespace and do not include <unistd.h>.
226
- forward_list meets the allocator-aware container requirements;
227
228
- this_thread::sleep_for(), this_thread::sleep_until() and this_thread::
229
yield() are defined without requiring the configure option
230
--enable-libstdcxx-time;
231
232
- Improvements to <random>:
233
234
- SSE optimized normal_distribution.
235
236
- Use of hardware RNG instruction for random_device on new x86 processors
237
(requires the assembler to support the instruction.)
238
239
and <ext/random>:
240
241
- New random number engine simd_fast_mersenne_twister_engine with an
242
optimized SSE implementation.
243
244
- New random number distributions beta_distribution, normal_mv_distribution,
245
rice_distribution, nakagami_distribution, pareto_distribution,
246
k_distribution, arcsine_distribution, hoyt_distribution.
247
248
- Added --disable-libstdcxx-verbose configure option to disable diagnostic
249
messages issued when a process terminates abnormally. This may be useful for
250
embedded systems to reduce the size of executables that link statically to
251
the library.
497
252
498
253
499
254
Fortran
500
255
-------
501
256
502
- The compile flag -fstack-arrays has been added, which causes all local
503
arrays to be put on stack memory. For some programs this will improve the
504
performance significantly. If your program uses very large local arrays,
505
it is possible that you will have to extend your runtime limits for stack
506
memory.
507
508
- The -Ofast flag now also implies -fno-protect-parens and -fstack-arrays.
509
510
- Front-end optimizations can now be selected by the -ffrontend-optimize
511
option and deselected by the -fno-frontend-optimize option.
512
513
- When front-end optimization removes a function call, -Wfunction-
514
elimination warns about that.
515
516
- When performing front-end-optimization, the -faggressive-function-
517
elimination option allows the removal of duplicate function calls even
518
for impure functions.
519
520
- The flag -Wreal-q-constant has been added, which warns if floating-point
521
literals have been specified using q (such as 1.0q0); the q marker is now
522
supported as a vendor extension to denote quad precision (REAL(16) or, if
523
not available, REAL(10)). Consider using a kind parameter (such as in
524
1.0_qp) instead, which can be obtained via SELECTED_REAL_KIND.
525
526
- The GFORTRAN_USE_STDERR environment variable has been removed. GNU
527
Fortran now always prints error messages to standard error. If you wish
528
to redirect standard error, please consult the manual for your OS, shell,
529
batch environment etc. as appropriate.
530
531
- The -fdump-core option and GFORTRAN_ERROR_DUMPCORE environment variable
532
have been removed. When encountering a serious error, gfortran will now
533
always abort the program. Whether a core dump is generated depends on the
534
user environment settings; see the ulimit -c setting for POSIX shells,
535
limit coredumpsize for C shells, and the WER_user-mode_dumps_settings on
536
Windows.
537
538
- The -fbacktrace option is now enabled by default. When encountering a
539
fatal error, gfortran will attempt to print a backtrace to standard error
540
before aborting. It can be disabled with -fno-backtrace. Note: On POSIX
541
targets with the addr2line utility from GNU binutils, GNU Fortran can
542
print a backtrace with function name, file name, line number information
543
in addition to the addresses; otherwise only the addresses are printed.
257
- Compatibility notice:
258
259
- Module files: The version of module files (.mod) has been incremented.
260
Fortran MODULEs compiled by earlier GCC versions have to be recompiled,
261
when they are USEd by files compiled with GCC 4.8. GCC 4.8 is not able to
262
read .mod files created by earlier versions; attempting to do so gives an
263
error message.
264
Note: The ABI of the produced assembler data itself has not changed;
265
object files and libraries are fully compatible with older versions except
266
as noted below.
267
268
- ABI: Some internal names (used in the assembler/object file) have changed
269
for symbols declared in the specification part of a module. If an affected
270
module - or a file using it via use association - is recompiled, the
271
module and all files which directly use such symbols have to be recompiled
272
as well. This change only affects the following kind of module symbols:
273
274
- Procedure pointers. Note: C-interoperable function pointers (type
275
(c_funptr)) are not affected nor are procedure-pointer components.
276
- Deferred-length character strings.
277
278
- The BACKTRACE intrinsic subroutine has been added. It shows a backtrace at
279
an arbitrary place in user code; program execution continues normally
280
afterwards.
281
282
- The -Wc-binding-type warning option has been added (disabled by default). It
283
warns if the a variable might not be C interoperable; in particular, if the
284
variable has been declared using an intrinsic type with default kind instead
285
of using a kind parameter defined for C interoperability in the intrinsic
286
ISO_C_Binding module. Before, this warning was always printed. The
287
-Wc-binding-type option is enabled by -Wall.
288
289
- The -Wrealloc-lhs and -Wrealloc-lhs-all warning command-line options have
290
been added, which diagnose when code to is inserted for automatic
291
(re)allocation of a variable during assignment. This option can be used to
292
decide whether it is safe to use -fno-realloc-lhs. Additionally, it can be
293
used to find automatic (re)allocation in hot loops. (For arrays, replacing
294
var= by var(:)= disables the automatic reallocation.)
295
296
- The -Wcompare-reals command-line option has been added. When this is set,
297
warnings are issued when comparing REAL or COMPLEX types for equality and
298
inequality; consider replacing a == b by abs(a-b) < eps with a suitable eps.
299
-Wcompare-reals is enabled by -Wextra.
300
301
- The -Wtarget-lifetime command-line option has been added (enabled with
302
-Wall), which warns if the pointer in a pointer assignment might outlive its
303
target.
304
305
- Reading floating point numbers which use q for the exponential (such as
306
4.0q0) is now supported as vendor extension for better compatibility with
307
old data files. It is strongly recommended to use for I/O the equivalent but
308
standard conforming e (such as 4.0e0).
309
(For Fortran source code, consider replacing the q in floating-point
310
literals by a kind parameter (e.g. 4.0e0_qp with a suitable qp). Note that -
311
in Fortran source code - replacing q by a simple e is not equivalent.)
312
313
- The GFORTRAN_TMPDIR environment variable for specifying a non-default
314
directory for files opened with STATUS="SCRATCH", is not used anymore.
315
Instead gfortran checks the POSIX/GNU standard TMPDIR environment variable.
316
If TMPDIR is not defined, gfortran falls back to other methods to determine
317
the directory for temporary files as documented in the user_manual.
544
318
545
319
- Fortran_2003:
546
320
547
- Generic interface names which have the same name as derived types
548
are now supported, which allows to write constructor functions.
549
Note that Fortran does not support static constructor functions;
550
only default initialization or an explicit structure-constructor
551
initialization are available.
552
553
- Polymorphic (class) arrays are now supported.
554
555
- Fortran_2008:
556
557
- Support for the DO CONCURRENT construct has been added, which
558
allows the user to specify that individual loop iterations have no
559
interdependencies.
560
561
- Coarrays: Full single-image support except for polymorphic
562
coarrays. Additionally, preliminary support for multiple images via
563
an MPI-based coarray_communication_library has been added. Note:
564
The library version is not yet usable as remote coarray access is
565
not yet possible.
321
- Support for unlimited polymorphic variables (CLASS(*)) has been added.
322
Nonconstant character lengths are not yet supported.
566
323
567
324
- TS_29113:
568
325
569
- New flag -std=f2008ts permits programs that are expected to conform
570
to the Fortran 2008 standard and the draft Technical Specification
571
(TS) 29113 on Further Interoperability of Fortran with C.
572
573
- The OPTIONAL attribute is now allowed for dummy arguments of BIND
574
(C) procedures.
575
576
- The RANK intrinsic has been added.
577
578
- The implementation of the ASYNCHRONOUS attribute in GCC is
579
compatible with the candidate draft of TS 29113 (since GCC 4.6).
326
- Assumed types (TYPE(*)) are now supported.
327
328
- Experimental support for assumed-rank arrays (dimension(..)) has been
329
added. Note that currently gfortran's own array descriptor is used, which
330
is different from the one defined in TS29113, see gfortran's_header_file
331
or use the Chasm_Language_Interoperability_Tools.
580
332
581
333
582
334
Go
583
335
--
584
336
585
- GCC 4.7 implements the Go_1_language_standard. The library support in
586
4.7.0 is not quite complete, due to release timing. Release 4.7.1 is
587
expected to include complete support.
337
- GCC 4.8.0 implements a preliminary version of the upcoming Go 1.1 release.
338
The library support is not quite complete, due to release timing.
588
339
589
- Go has been tested on GNU/Linux and Solaris platforms. It may work on
590
other platforms as well.
340
- Go has been tested on GNU/Linux and Solaris platforms for various processors
341
including x86, x86_64, PowerPC, SPARC, and Alpha. It may work on other
342
platforms as well.
591
343
592
344
593
345
New Targets and Target Specific Improvements
594
346
============================================
595
347
348
349
AArch64
350
-------
351
352
- A new port has been added to support AArch64, the new 64-bit architecture
353
from ARM. Note that this is a separate port from the existing 32-bit ARM
354
port.
355
- The port provides initial support for the Cortex-A53 and the Cortex-A57
356
processors with the command line options -mcpu=cortex-a53 and
357
-mcpu=cortex-a57.
358
359
596
360
ARM
597
361
---
598
362
599
- GCC now supports the Cortex-A7 processor implementing the v7-a version of
600
the architecture using the option -mcpu=cortex-a7.
601
602
- The default vector size in auto-vectorization for NEON is now 128 bits.
603
If vectorization fails thusly, the vectorizer tries again with 64-bit
604
vectors.
605
606
- A new option -mvectorize-with-neon-double was added to allow users to
607
change the vector size to 64 bits.
363
- Initial support has been added for the AArch32 extensions defined in the
364
ARMv8 architecture.
365
366
- Code generation improvements for the Cortex-A7 and Cortex-A15 CPUs.
367
368
- A new option, -mcpu=marvell-pj4, has been added to generate code for the
369
Marvell PJ4 processor.
370
371
- The compiler can now automatically generate the VFMA, VFMS, REVSH and REV16
372
instructions.
373
374
- A new vectorizer cost model for Advanced SIMD configurations to improve the
375
auto-vectorization strategies used.
376
377
- The scheduler now takes into account the number of live registers to reduce
378
the amount of spilling that can occur. This should improve code performance
379
in large functions. The limit can be removed by using the option -fno-sched-
380
pressure.
381
382
- Improvements have been made to the Marvell iWMMX code generation and support
383
for the iWMMX2 SIMD unit has been added. The option -mcpu=iwmmxt2 can be
384
used to enable code generation for the latter.
385
386
- A number of code generation improvements for Thumb2 to reduce code size when
387
compiling for the M-profile processors.
388
389
- The RTEMS (arm-rtems) port has been updated to use the EABI.
390
391
- Code generation support for the old FPA and Maverick floating-point
392
architectures has been removed. Ports that previously relied on these
393
features have also been removed. This includes the targets:
394
395
- arm*-*-linux-gnu (use arm*-*-linux-gnueabi)
396
- arm*-*-elf (use arm*-*-eabi)
397
- arm*-*-uclinux* (use arm*-*-uclinux*eabi)
398
- arm*-*-ecos-elf (no alternative)
399
- arm*-*-freebsd (no alternative)
400
- arm*-wince-pe* (no alternative).
608
401
609
402
610
403
AVR
611
404
---
612
405
613
- GCC now supports the XMEGA architecture. This requires GNU binutils 2.22
614
or later.
615
616
- Support for the named_address_spaces __flash, __flash1, …, __flash5 and
617
__memx has been added. These address spaces locate read-only data in
618
flash memory and allow reading from flash memory by means of ordinary C
619
code, i.e. without the need of (inline) assembler code:
620
621
const __flash int values[] = { 42, 31 };
622
623
int add_values (const __flash int *p, int i)
624
{
625
return values[i] + *p;
626
}
627
628
- Support has been added for a new AVR-specific configure option
629
--with-avrlibc=yes in order to arrange for better integration of
630
AVR-Libc. This configure option is supported in avr-gcc 4.7.2 and
631
newer and will only take effect in non-RTEMS configurations. If
632
avr-gcc is configured for RTEMS, the option will be ignored which is
633
the same as specifying --with-avrlibc=no. See PR54461 for more
634
technical details.
635
636
- Support for AVR-specific built-in_functions has been added.
637
638
- Support has been added for the signed and unsigned 24-bit scalar integer
639
types __int24 and __uint24.
640
641
- New command-line options -maccumulate-args, -mbranch-cost=cost and
642
-mstrict-X were added to allow better fine-tuning of code optimization.
643
644
- Many optimizations to:
645
646
- 64-bit integer arithmetic
647
648
- Widening multiplication
649
650
- Integer division by a constant
651
652
- Avoid constant reloading in multi-byte instructions.
653
654
- Micro-optimizations for special instruction sequences.
655
656
- Generic built-in functions like __builtin_ffs*, __builtin_clz*, etc.
657
658
- If-else decision trees generated by switch instructions
659
660
- Merging of data located in flash memory
661
662
- New libgcc variants for devices with 8-bit wide stack pointer
663
664
- Better documentation:
665
666
- Handling of EIND and indirect jumps on devices with more than 128
667
KiB of program memory.
668
669
- Handling of the RAMPD, RAMPX, RAMPY and RAMPZ special function
670
registers.
671
672
- Function attributes OS_main and OS_task.
673
- AVR-specific built-in macros.
674
675
676
C6X
677
---
678
679
- Support has been added for the Texas Instruments C6X family of processors.
680
681
682
CR16
683
----
684
685
- Support has been added for National Semiconductor's CR16 architecture.
686
687
688
Epiphany
689
--------
690
691
- Support has been added for Adapteva's Epiphany architecture.
406
- Support for the "Embedded C" fixed-point has been added. For details, see
407
the GCC_wiki and the user_manual. The support is not complete.
408
- A new print modifier %r for register operands in inline assembler is
409
supported. It will print the raw register number without the register prefix
410
'r':
411
412
/* Return the most significant byte of 'val', a 64-bit value. */
413
414
unsigned char msb (long long val)
415
{
416
unsigned char c;
417
__asm__ ("mov %0, %r1+7" : "=r" (c) : "r" (val));
418
return c;
419
}
420
421
The inline assembler in this example will generate code like
422
423
mov r24, 8+7
424
425
provided c is allocated to R24 and val is allocated to R8...R15. This works
426
because the GNU assembler accepts plain register numbers without register
427
prefix.
428
429
- Static initializers with 3-byte symbols are supported now:
430
431
extern const __memx char foo;
432
const __memx void *pfoo = &foo;
433
434
This requires at least Binutils 2.23.
692
435
693
436
694
437
IA-32/x86-64
695
438
------------
696
439
697
- Support for Intel AVX2 intrinsics, built-in functions and code generation
698
is available via -mavx2.
699
700
- Support for Intel BMI2 intrinsics, built-in functions and code generation
701
is available via -mbmi2.
702
703
- Implementation and automatic generation of __builtin_clz* using the lzcnt
704
instruction is available via -mlzcnt.
705
706
- Support for Intel FMA3 intrinsics and code generation is available via
707
-mfma.
708
709
- A new -mfsgsbase command-line option is available that makes GCC generate
710
new segment register read/write instructions through dedicated built-ins.
711
712
- Support for the new Intel rdrnd instruction is available via -mrdrnd.
713
714
- Two additional AVX vector conversion instructions are available via
715
-mf16c.
716
717
- Support for new Intel processor codename IvyBridge with RDRND, FSGSBASE
718
and F16C is available through -march=core-avx-i.
719
720
- Support for the new Intel processor codename Haswell with AVX2, FMA, BMI,
721
BMI2, LZCNT is available through -march=core-avx2.
722
723
- Support for new AMD family 15h processors (Piledriver core) is now
724
available through -march=bdver2 and -mtune=bdver2 options.
725
726
- Support for the_x32_psABI is now available through the -mx32 option.
727
728
- Windows mingw targets are using the -mms-bitfields option by default.
729
730
- Windows x86 targets are using the __thiscall calling convention for C++
731
class-member functions.
732
733
- Support for the configure option --with-threads=posix for Windows mingw
734
targets.
440
- Allow -mpreferred-stack-boundary=3 for the x86-64 architecture with SSE
441
extensions disabled. Since the x86-64 ABI requires 16 byte stack alignment,
442
this is ABI incompatible and intended to be used in controlled environments
443
where stack space is an important limitation. This option will lead to wrong
444
code when functions compiled with 16 byte stack alignment (such as functions
445
from a standard library) are called with misaligned stack. In this case, SSE
446
instructions may lead to misaligned memory access traps. In addition,
447
variable arguments will be handled incorrectly for 16 byte aligned objects
448
(including x87 long double and __int128), leading to wrong results. You must
449
build all modules with -mpreferred-stack-boundary=3, including any
450
libraries. This includes the system libraries and startup modules.
451
452
- Support for the new Intel processor codename Broadwell with RDSEED, ADCX,
453
ADOX, PREFETCHW is available through -madx, -mprfchw, -mrdseed command-line
454
options.
455
456
- Support for the Intel RTM and HLE intrinsics, built-in functions and code
457
generation is available via -mrtm and -mhle.
458
459
- Support for the Intel FXSR, XSAVE and XSAVEOPT instruction sets. Intrinsics
460
and built-in functions are available via -mfxsr, -mxsave and -mxsaveopt
461
respectively.
462
463
- New -maddress-mode=[short|long] options for x32. -maddress-mode=short
464
overrides default 64-bit addresses to 32-bit by emitting the 0x67 address-
465
size override prefix. This is the default address mode for x32.
466
467
- New built-in functions to detect run-time CPU type and ISA:
468
469
- A built-in function __builtin_cpu_is has been added to detect if the run-
470
time CPU is of a particular type. It returns a positive integer on a match
471
and zero otherwise. It accepts one string literal argument, the CPU name.
472
For example, __builtin_cpu_is("westmere") returns a positive integer if
473
the run-time CPU is an Intel Core i7 Westmere processor. Please refer to
474
the user_manual for the list of valid CPU names recognized.
475
476
- A built-in function __builtin_cpu_supports has been added to detect if the
477
run-time CPU supports a particular ISA feature. It returns a positive
478
integer on a match and zero otherwise. It accepts one string literal
479
argument, the ISA feature. For example, __builtin_cpu_supports("ssse3")
480
returns a positive integer if the run-time CPU supports SSSE3
481
instructions. Please refer to the user_manual for the list of valid ISA
482
names recognized.
483
484
Caveat: If these built-in functions are called before any static
485
constructors are invoked, like during IFUNC initialization, then the CPU
486
detection initialization must be explicitly run using this newly provided
487
built-in function, __builtin_cpu_init. The initialization needs to be done
488
only once. For example, this is how the invocation would look like inside an
489
IFUNC initializer:
490
491
static void (*some_ifunc_resolver(void))(void)
492
{
493
__builtin_cpu_init();
494
if (__builtin_cpu_is("amdfam10h") ...
495
if (__builtin_cpu_supports("popcnt") ...
496
}
497
498
- Function Multiversioning Support with G++:
499
It is now possible to create multiple function versions each targeting a
500
specific processor and/or ISA. Function versions have the same signature but
501
different target attributes. For example, here is a program with function
502
versions:
503
504
__attribute__ ((target ("default")))
505
int foo(void)
506
{
507
return 1;
508
}
509
510
__attribute__ ((target ("sse4.2")))
511
int foo(void)
512
{
513
return 2;
514
}
515
516
int main (void)
517
{
518
int (*p) = &foo;
519
assert ((*p)() == foo());
520
return 0;
521
}
522
523
Please refer to this wiki for more information.
524
525
- The x86 backend has been improved to allow option -fschedule-insns to work
526
reliably. This option can be used to schedule instructions better and leads
527
to improved performace in certain cases.
528
529
- Windows MinGW-w64 targets (*-w64-mingw*) require at least r5437 from the
530
Mingw-w64 trunk.
531
532
- Support for new AMD family 15h processors (Steamroller core) is now
533
available through the -march=bdver3 and -mtune=bdver3 options.
534
535
- Support for new AMD family 16h processors (Jaguar core) is now available
536
through the -march=btver2 and -mtune=btver2 options.
537
538
539
FRV
540
---
541
542
- This target now supports the -fstack-usage command-line option.
735
543
736
544
737
545
MIPS
738
546
----
739
547
740
- GCC now supports thread-local storage (TLS) for MIPS16. This requires GNU
741
binutils 2.22 or later.
742
743
- GCC can now generate code specifically for the Cavium Octeon+ and Octeon2
744
processors. The associated command-line options are -march=octeon+ and
745
-march=octeon2 respectively. Both options require GNU binutils 2.22 or
746
later.
747
748
- GCC can now work around certain 24k errata, under the control of the
749
command-line option -mfix-24k. These workarounds require GNU binutils
750
2.20 or later.
751
752
- 32-bit MIPS GNU/Linux targets such as mips-linux-gnu can now build n32
753
and n64 multilibs. The result is effectively a 64-bit GNU/Linux toolchain
754
that generates 32-bit code by default. Use the configure-time option
755
--enable-targets=all to select these extra multilibs.
756
757
- Passing -fno-delayed-branch now also stops the assembler from
758
automatically filling delay slots.
759
760
761
PowerPC/PowerPC64
762
-----------------
763
764
- Vectors of type vector long long or vector long are passed and returned
765
using the same method as other vectors with the VSX instruction set.
766
Previously the GCC compiler did not adhere to the ABI for 128-bit vectors
767
with 64-bit integer base types (PR 48857). This will also be fixed in the
768
GCC 4.6.1 and 4.5.4 releases.
769
770
- A new option -mno-pointers-to-nested-functions was added to allow AIX 32-
771
bit/64-bit and GNU/Linux 64-bit PowerPC users to specify that the
772
compiler should not load up the chain register (r11) before calling a
773
function through a pointer. If you use this option, you cannot call
774
nested functions through a pointer, or call other languages that might
775
use the static chain.
776
777
- A new option msave-toc-indirect was added to allow AIX 32-bit/64-bit and
778
GNU/Linux 64-bit PowerPC users control whether we save the TOC in the
779
prologue for indirect calls or generate the save inline. This can speed
780
up some programs that call through a function pointer a lot, but it can
781
slow down other functions that only call through a function pointer in
782
exceptional cases.
783
784
- The PowerPC port will now enable machine-specific built-in functions when
785
the user switches the target machine using the #pragma GCC target or
786
__attribute__ ((__target__ ("target"))) code sequences. In addition, the
787
target macros are updated. However, due to the way the -save-temps switch
788
is implemented, you won't see the effect of these additional macros being
789
defined in preprocessor output.
548
- GCC can now generate code specifically for the R4700, Broadcom XLP and MIPS
549
34kn processors. The associated -march options are -march=r4700, -march=xlp
550
and -march=34kn respectively.
551
552
- GCC now generates better DSP code for MIPS 74k cores thanks to further
553
scheduling optimizations.
554
555
- The MIPS port now supports the -fstack-check option.
556
557
- GCC now passes the -mmcu and -mno-mcu options to the assembler.
558
559
- Previous versions of GCC would silently accept -fpic and -fPIC for
560
-mno-abicalls targets like mips*-elf. This combination was not intended or
561
supported, and did not generate position-independent code. GCC 4.8 now
562
reports an error when this combination is used.
563
564
565
PowerPC / PowerPC64 / RS6000
566
----------------------------
567
568
- SVR4 configurations (GNU/Linux, FreeBSD, NetBSD) no longer save, restore or
569
update the VRSAVE register by default. The respective operating systems
570
manage the VRSAVE register directly.
571
572
- Large TOC support has been added for AIX through the command line option
573
-mcmodel=large.
574
575
- Native Thread-Local Storage support has been added for AIX.
576
577
- VMX (Altivec) and VSX instruction sets now are enabled implicitly when
578
targetting processors that support those hardware features on AIX 6.1 and
579
above.
580
581
582
RX
583
--
584
585
- This target will now issue a warning message whenever multiple fast
586
interrupt handlers are found in the same cpmpilation unit. This feature can
587
be turned off by the new -mno-warn-multiple-fast-interrupts command-line
588
option.
589
590
591
S/390, System z
592
---------------
593
594
- Support for the IBM zEnterprise zEC12 processor has been added. When using
595
the -march=zEC12 option, the compiler will generate code making use of the
596
following new instructions:
597
598
- load and trap instructions
599
- 2 new compare and trap instructions
600
- rotate and insert selected bits - without CC clobber
601
602
The -mtune=zEC12 option enables zEC12 specific instruction scheduling
603
without making use of new instructions.
604
605
- Register pressure sensitive instruction scheduling is enabled by default.
606
607
- The ifunc function attribute is enabled by default.
608
609
- memcpy and memcmp invokations on big memory chunks or with run time lengths
610
are not generated inline anymore when tuning for z10 or higher. The purpose
611
is to make use of the IFUNC optimized versions in Glibc.
790
612
791
613
792
614
SH
793
615
--
794
616
795
- A new option -msoft-atomic has been added. When it is specified, GCC will
796
generate GNU/Linux-compatible gUSA atomic sequences for the new __atomic
797
routines.
798
799
- Since it is neither supported by GAS nor officially documented, code
800
generation for little endian SH2A has been disabled. Specifying -ml with
801
-m2a* will now result in a compiler error.
802
803
- The defunct -mbranch-cost option has been fixed.
804
805
- Some improvements to the generated code of:
806
807
- Utilization of the tst #imm,R0 instruction.
808
809
- Dynamic shift instructions on SH2A.
810
811
- Integer absolute value calculations.
617
- The default alignment settings have been reduced to be less aggressive. This
618
results in more compact code for optimization levels other than -Os.
619
620
- Improved support for the __atomic built-in functions:
621
622
- A new option -matomic-model=model selects the model for the generated
623
atomic sequences. The following models are supported:
624
625
soft-gusa
626
Software gUSA sequences (SH3* and SH4* only). On SH4A targets this
627
will now also partially utilize the movco.l and movli.l
628
instructions. This is the default when the target is sh3*-*-linux*
629
or sh4*-*-linux*.
630
hard-llcs
631
Hardware movco.l / movli.l sequences (SH4A only).
632
soft-tcb
633
Software thread control block sequences.
634
soft-imask
635
Software interrupt flipping sequences (privileged mode only). This
636
is the default when the target is sh1*-*-linux* or sh2*-*-linux*.
637
none
638
Generates function calls to the respective __atomic built-in
639
functions. This is the default for SH64 targets or when the target
640
is not sh*-*-linux*.
641
642
- The option -msoft-atomic has been deprecated. It is now an alias for
643
-matomic-model=soft-gusa.
644
645
- A new option -mtas makes the compiler generate the tas.b instruction for
646
the __atomic_test_and_set built-in function regardless of the selected
647
atomic model.
648
649
- The __sync functions in libgcc now reflect the selected atomic model when
650
building the toolchain.
651
652
- Added support for the mov.b and mov.w instructions with displacement
653
addressing.
654
655
- Added support for the SH2A instructions movu.b and movu.w.
656
657
- Various improvements to code generated for integer arithmetic.
658
659
- Improvements to conditional branches and code that involves the T bit. A new
660
option -mzdcbranch tells the compiler to favor zero-displacement branches.
661
This is enabled by default for SH4* targets.
662
663
- The pref instruction will now be emitted by the __builtin_prefetch built-in
664
function for SH3* targets.
665
666
- The fmac instruction will now be emitted by the fmaf standard function and
667
the __builtin_fmaf built-in function.
668
669
- The -mfused-madd option has been deprecated in favor of the machine-
670
independent -ffp-contract option. Notice that the fmac instruction will now
671
be generated by default for expressions like a * b + c. This is due to the
672
compiler default setting -ffp-contract=fast.
673
674
- Added new options -mfsrra and -mfsca to allow the compiler using the fsrra
675
and fsca instructions on targets other than SH4A (where they are already
676
enabled by default).
677
678
- Added support for the __builtin_bswap32 built-in function. It is now
679
expanded as a sequence of swap.b and swap.w instructions instead of a
680
library function call.
681
682
- The behavior of the -mieee option has been fixed and the negative form
683
-mno-ieee has been added to control the IEEE conformance of floating point
684
comparisons. By default -mieee is now enabled and the option -ffinite-math-
685
only implicitly sets -mno-ieee.
686
687
- Added support for the built-in functions __builtin_thread_pointer and
688
__builtin_set_thread_pointer. This assumes that GBR is used to hold the
689
thread pointer of the current thread. Memory loads and stores relative to
690
the address returned by __builtin_thread_pointer will now also utilize GBR
691
based displacement address modes.
812
692
813
693
814
694
SPARC
815
695
-----
816
696
817
- The option -mflat has been reinstated. When it is specified, the compiler
818
will generate code for a single register window model. This is
819
essentially a new implementation and the corresponding debugger support
820
has been added to GDB 7.4.
821
822
- Support for the options -mtune=native and -mcpu=native has been added on
823
selected native platforms (GNU/Linux and Solaris).
824
825
- Support for the SPARC T3 (Niagara 3) processor has been added.
826
827
- VIS:
828
829
- An intrinsics header visintrin.h has been added.
830
- Builtin intrinsics for the VIS 1.0 edge handling and pixel compare
831
instructions have been added.
832
- The little-endian version of alignaddr is now supported.
833
- When possible, VIS builtins are marked const, which should increase
834
the compiler's ability to optimize VIS operations.
835
- The compiler now properly tracks the %gsr register and how it
836
behaves as an input for various VIS instructions.
837
- Akin to fzero, the compiler can now generate fone instructions in
838
order to set all of the bits of a floating-point register to 1.
839
- The documentation for the VIS intrinsics in the GCC manual has been
840
brought up to date and many inaccuracies were fixed.
841
- Intrinsics for the VIS 2.0 bmask, bshuffle, and non-condition-code
842
setting edge instructions have been added. Their availability is
843
controlled by the new -mvis2 and -mno-vis2 options. They are
844
enabled by default on UltraSPARC-III and later CPUs.
845
846
- Support for UltraSPARC Fused Multiply-Add floating-point extensions has
847
been added. These instructions are enabled by default on SPARC T3
848
(Niagara 3) and later CPUs.
849
850
851
TILE-Gx/TILEPro
852
---------------
853
854
- Support has been added for the Tilera TILE-Gx and TILEPro families of
855
processors.
856
857
858
Other significant improvements
859
==============================
860
861
- A new option (-grecord-gcc-switches) was added that appends compiler
862
command-line options that might affect code generation to the
863
DW_AT_producer attribute string in the DWARF debugging information.
864
865
- GCC now supports various new GNU extensions to the DWARF debugging
866
information format, like entry_value and call_site information, typed
867
DWARF_stack or a_more_compact_macro_representation. Support for these
868
extensions has been added to GDB 7.4. They can be disabled through the
869
-gstrict-dwarf command-line option.
870
871
872
GCC 4.7.1
873
=========
874
875
This is the list of problem reports (PRs) from GCC's bug tracking
876
system that are known to be fixed in the 4.7.1 release. This list
877
might not be complete (that is, it is possible that some PRs that have
878
been fixed are not listed here).
879
880
http://gcc.gnu.org/bugzilla/buglist.cgi?bug_status=RESOLVED&resolution=FIXED&target_milestone=4.7.1
881
882
The Go frontend in the 4.7.1 release fully supports the Go 1 language
883
standard (http://weekly.golang.org/doc/go1.html).
884
885
886
GCC 4.7.2
887
=========
888
889
This is the list of problem reports (PRs) from GCC's bug tracking
890
system that are known to be fixed in the 4.7.2 release. This list
891
might not be complete (that is, it is possible that some PRs that have
892
been fixed are not listed here).
893
894
http://gcc.gnu.org/bugzilla/buglist.cgi?bug_status=RESOLVED&resolution=FIXED&target_milestone=4.7.2
697
- Added optimized instruction scheduling for Niagara4.
698
699
700
TILE-Gx
701
-------
702
703
- Added support for the -mcmodel=MODEL command-line option. The models
704
supported are small and large.
705
706
707
V850
708
----
709
710
- This target now supports the E3V5 architecture via the use of the new
711
-mv850e3v5 command-line option. It also has experimental support for the e3v5
712
LOOP instruction which can be enabled via the new -mloop command-line
713
option.
714
715
716
XStormy16
717
---------
718
719
- This target now supports the -fstack-usage command-line option.
720
721
722
Operating Systems
723
=================
724
725
726
Windows (Cygwin)
727
----------------
728
729
- Executables are now linked against shared libgcc by default. The previous
730
default was to link statically, which can still be done by explicitly
731
specifying -static or -static-libgcc on the command line. However it is
732
strongly advised against, as it will cause problems for any application that
733
makes use of DLLs compiled by GCC. It should be alright for a monolithic
734
stand-alone application that only links against the Windows OS DLLs, but
735
offers little or no benefit.
895
736
896
737
897
738
For questions related to the use of GCC, please consult these web
905
746
Verbatim copying and distribution of this entire article is
906
747
permitted in any medium, provided this notice is preserved.
907
748
908
These pages are maintained_by_the_GCC_team.
749
These pages are maintained by the GCC team.
909
750
910
Last modified 2012-03-26.
751
Last modified 2013-03-23.
Loggerhead is a web-based interface for Breezy
Version: 2.0.1