1
@c Copyright 2002, 2004 Free Software Foundation, Inc.
2
@c This is part of the GAS manual.
3
@c For copying conditions, see the file as.texinfo.
4
@c CRIS description contributed by Axis Communications.
8
@chapter CRIS Dependent Features
11
@node Machine Dependencies
12
@chapter CRIS Dependent Features
17
* CRIS-Opts:: Command-line Options
18
* CRIS-Expand:: Instruction expansion
19
* CRIS-Symbols:: Symbols
20
* CRIS-Syntax:: Syntax
24
@section Command-line Options
28
The CRIS version of @code{@value{AS}} has these
29
machine-dependent command-line options.
31
@cindex @option{--emulation=criself} command line option, CRIS
32
@cindex @option{--emulation=crisaout} command line option, CRIS
33
@cindex CRIS @option{--emulation=criself} command line option
34
@cindex CRIS @option{--emulation=crisaout} command line option
36
The format of the generated object files can be either ELF or
37
a.out, specified by the command-line options
38
@option{--emulation=crisaout} and @option{--emulation=criself}.
39
The default is ELF (criself), unless @code{@value{AS}} has been
40
configured specifically for a.out by using the configuration
41
name @code{cris-axis-aout}.
43
@cindex @option{--underscore} command line option, CRIS
44
@cindex @option{--no-underscore} command line option, CRIS
45
@cindex CRIS @option{--underscore} command line option
46
@cindex CRIS @option{--no-underscore} command line option
47
There are two different link-incompatible ELF object file
48
variants for CRIS, for use in environments where symbols are
49
expected to be prefixed by a leading @samp{_} character and for
50
environments without such a symbol prefix. The variant used for
51
GNU/Linux port has no symbol prefix. Which variant to produce
52
is specified by either of the options @option{--underscore} and
53
@option{--no-underscore}. The default is @option{--underscore}.
54
Since symbols in CRIS a.out objects are expected to have a
55
@samp{_} prefix, specifying @option{--no-underscore} when
56
generating a.out objects is an error. Besides the object format
57
difference, the effect of this option is to parse register names
58
differently (@pxref{crisnous}). The @option{--no-underscore}
59
option makes a @samp{$} register prefix mandatory.
61
@cindex @option{--pic} command line option, CRIS
62
@cindex CRIS @option{--pic} command line option
63
@cindex Position-independent code, CRIS
64
@cindex CRIS position-independent code
65
The option @option{--pic} must be passed to @code{@value{AS}} in
66
order to recognize the symbol syntax used for ELF (SVR4 PIC)
67
position-independent-code (@pxref{crispic}). This will also
68
affect expansion of instructions. The expansion with
69
@option{--pic} will use PC-relative rather than (slightly
70
faster) absolute addresses in those expansions. This option is only
71
valid when generating ELF format object files.
73
@cindex @option{--march=@var{architecture}} command line option, CRIS
74
@cindex CRIS @option{--march=@var{architecture}} command line option
75
@cindex Architecture variant option, CRIS
76
@cindex CRIS architecture variant option
77
The option @option{--march=@var{architecture}}
78
@anchor{march-option}specifies the recognized instruction set
79
and recognized register names. It also controls the
80
architecture type of the object file. Valid values for
81
@var{architecture} are:
85
All instructions and register names for any architecture variant
86
in the set v0@dots{}v10 are recognized. This is the
87
default if the target is configured as cris-*.
90
Only instructions and register names for CRIS v10 (as found in
91
ETRAX 100 LX) are recognized. This is the default if the target
92
is configured as crisv10-*.
95
Only instructions and register names for CRIS v32 (code name
96
Guinness) are recognized. This is the default if the target is
97
configured as crisv32-*. This value implies
98
@option{--no-mul-bug-abort}. (A subsequent
99
@option{--mul-bug-abort} will turn it back on.)
102
Only instructions with register names and addressing modes with
103
opcodes common to the v10 and v32 are recognized.
106
@cindex @option{-N} command line option, CRIS
107
@cindex CRIS @option{-N} command line option
108
When @option{-N} is specified, @code{@value{AS}} will emit a
109
warning when a 16-bit branch instruction is expanded into a
110
32-bit multiple-instruction construct (@pxref{CRIS-Expand}).
112
@cindex @option{--no-mul-bug-abort} command line option, CRIS
113
@cindex @option{--mul-bug-abort} command line option, CRIS
114
@cindex CRIS @option{--no-mul-bug-abort} command line option
115
@cindex CRIS @option{--mul-bug-abort} command line option
117
Some versions of the CRIS v10, for example in the Etrax 100 LX,
118
contain a bug that causes destabilizing memory accesses when a
119
multiply instruction is executed with certain values in the
120
first operand just before a cache-miss. When the
121
@option{--mul-bug-abort} command line option is active (the
122
default value), @code{@value{AS}} will refuse to assemble a file
123
containing a multiply instruction at a dangerous offset, one
124
that could be the last on a cache-line, or is in a section with
125
insufficient alignment. This placement checking does not catch
126
any case where the multiply instruction is dangerously placed
127
because it is located in a delay-slot. The
128
@option{--mul-bug-abort} command line option turns off the
132
@section Instruction expansion
134
@cindex instruction expansion, CRIS
135
@cindex CRIS instruction expansion
136
@code{@value{AS}} will silently choose an instruction that fits
137
the operand size for @samp{[register+constant]} operands. For
138
example, the offset @code{127} in @code{move.d [r3+127],r4} fits
139
in an instruction using a signed-byte offset. Similarly,
140
@code{move.d [r2+32767],r1} will generate an instruction using a
141
16-bit offset. For symbolic expressions and constants that do
142
not fit in 16 bits including the sign bit, a 32-bit offset is
145
For branches, @code{@value{AS}} will expand from a 16-bit branch
146
instruction into a sequence of instructions that can reach a
147
full 32-bit address. Since this does not correspond to a single
148
instruction, such expansions can optionally be warned about.
151
If the operand is found to fit the range, a @code{lapc} mnemonic
152
will translate to a @code{lapcq} instruction. Use @code{lapc.d}
153
to force the 32-bit @code{lapc} instruction.
155
Similarly, the @code{addo} mnemonic will translate to the
156
shortest fitting instruction of @code{addoq}, @code{addo.w} and
157
@code{addo.d}, when used with a operand that is a constant known
162
@cindex Symbols, built-in, CRIS
163
@cindex Symbols, CRIS, built-in
164
@cindex CRIS built-in symbols
165
@cindex Built-in symbols, CRIS
167
Some symbols are defined by the assembler. They're intended to
168
be used in conditional assembly, for example:
170
.if ..asm.arch.cris.v32
171
@var{code for CRIS v32}
172
.elseif ..asm.arch.cris.common_v10_v32
173
@var{code common to CRIS v32 and CRIS v10}
174
.elseif ..asm.arch.cris.v10 | ..asm.arch.cris.any_v0_v10
177
.error "Code needs to be added here."
181
These symbols are defined in the assembler, reflecting
182
command-line options, either when specified or the default.
183
They are always defined, to 0 or 1.
186
@item ..asm.arch.cris.any_v0_v10
187
This symbol is non-zero when @option{--march=v0_v10} is specified
190
@item ..asm.arch.cris.common_v10_v32
191
Set according to the option @option{--march=common_v10_v32}.
193
@item ..asm.arch.cris.v10
194
Reflects the option @option{--march=v10}.
196
@item ..asm.arch.cris.v32
197
Corresponds to @option{--march=v10}.
200
Speaking of symbols, when a symbol is used in code, it can have
201
a suffix modifying its value for use in position-independent
202
code. @xref{CRIS-Pic}.
207
There are different aspects of the CRIS assembly syntax.
210
* CRIS-Chars:: Special Characters
211
* CRIS-Pic:: Position-Independent Code Symbols
212
* CRIS-Regs:: Register Names
213
* CRIS-Pseudos:: Assembler Directives
217
@subsection Special Characters
218
@cindex line comment characters, CRIS
219
@cindex CRIS line comment characters
221
The character @samp{#} is a line comment character. It starts a
222
comment if and only if it is placed at the beginning of a line.
224
A @samp{;} character starts a comment anywhere on the line,
225
causing all characters up to the end of the line to be ignored.
227
A @samp{@@} character is handled as a line separator equivalent
228
to a logical new-line character (except in a comment), so
229
separate instructions can be specified on a single line.
232
@subsection Symbols in position-independent code
233
@cindex Symbols in position-independent code, CRIS
234
@cindex CRIS symbols in position-independent code
235
@cindex Position-independent code, symbols in, CRIS
237
When generating @anchor{crispic}position-independent code (SVR4
238
PIC) for use in cris-axis-linux-gnu or crisv32-axis-linux-gnu
239
shared libraries, symbol
240
suffixes are used to specify what kind of run-time symbol lookup
241
will be used, expressed in the object as different
242
@emph{relocation types}. Usually, all absolute symbol values
243
must be located in a table, the @emph{global offset table},
244
leaving the code position-independent; independent of values of
245
global symbols and independent of the address of the code. The
246
suffix modifies the value of the symbol, into for example an
247
index into the global offset table where the real symbol value
248
is entered, or a PC-relative value, or a value relative to the
249
start of the global offset table. All symbol suffixes start
250
with the character @samp{:} (omitted in the list below). Every
251
symbol use in code or a read-only section must therefore have a
252
PIC suffix to enable a useful shared library to be created.
253
Usually, these constructs must not be used with an additive
254
constant offset as is usually allowed, i.e.@: no 4 as in
255
@code{symbol + 4} is allowed. This restriction is checked at
256
link-time, not at assembly-time.
261
Attaching this suffix to a symbol in an instruction causes the
262
symbol to be entered into the global offset table. The value is
263
a 32-bit index for that symbol into the global offset table.
264
The name of the corresponding relocation is
265
@samp{R_CRIS_32_GOT}. Example: @code{move.d
266
[$r0+extsym:GOT],$r9}
270
Same as for @samp{GOT}, but the value is a 16-bit index into the
271
global offset table. The corresponding relocation is
272
@samp{R_CRIS_16_GOT}. Example: @code{move.d
273
[$r0+asymbol:GOT16],$r10}
277
This suffix is used for function symbols. It causes a
278
@emph{procedure linkage table}, an array of code stubs, to be
279
created at the time the shared object is created or linked
280
against, together with a global offset table entry. The value
281
is a pc-relative offset to the corresponding stub code in the
282
procedure linkage table. This arrangement causes the run-time
283
symbol resolver to be called to look up and set the value of the
284
symbol the first time the function is called (at latest;
285
depending environment variables). It is only safe to leave the
286
symbol unresolved this way if all references are function calls.
287
The name of the relocation is @samp{R_CRIS_32_PLT_PCREL}.
288
Example: @code{add.d fnname:PLT,$pc}
292
Like PLT, but the value is relative to the beginning of the
293
global offset table. The relocation is
294
@samp{R_CRIS_32_PLT_GOTREL}. Example: @code{move.d
299
Similar to @samp{PLT}, but the value of the symbol is a 32-bit
300
index into the global offset table. This is somewhat of a mix
301
between the effect of the @samp{GOT} and the @samp{PLT} suffix;
302
the difference to @samp{GOT} is that there will be a procedure
303
linkage table entry created, and that the symbol is assumed to
304
be a function entry and will be resolved by the run-time
305
resolver as with @samp{PLT}. The relocation is
306
@samp{R_CRIS_32_GOTPLT}. Example: @code{jsr
311
A variant of @samp{GOTPLT} giving a 16-bit value. Its
312
relocation name is @samp{R_CRIS_16_GOTPLT}. Example: @code{jsr
313
[$r0+fnname:GOTPLT16]}
317
This suffix must only be attached to a local symbol, but may be
318
used in an expression adding an offset. The value is the
319
address of the symbol relative to the start of the global offset
320
table. The relocation name is @samp{R_CRIS_32_GOTREL}.
321
Example: @code{move.d [$r0+localsym:GOTOFF],r3}
325
@subsection Register names
326
@cindex register names, CRIS
327
@cindex CRIS register names
329
A @samp{$} character may always prefix a general or special
330
register name in an instruction operand but is mandatory when
331
the option @option{--no-underscore} is specified or when the
332
@code{.syntax register_prefix} directive is in effect
333
(@pxref{crisnous}). Register names are case-insensitive.
336
@subsection Assembler Directives
337
@cindex assembler directives, CRIS
338
@cindex pseudo-ops, CRIS
339
@cindex CRIS assembler directives
340
@cindex CRIS pseudo-ops
342
There are a few CRIS-specific pseudo-directives in addition to
343
the generic ones. @xref{Pseudo Ops}. Constants emitted by
344
pseudo-directives are in little-endian order for CRIS. There is
345
no support for floating-point-specific directives for CRIS.
348
@item .dword EXPRESSIONS
349
@cindex assembler directive .dword, CRIS
350
@cindex pseudo-op .dword, CRIS
351
@cindex CRIS assembler directive .dword
352
@cindex CRIS pseudo-op .dword
354
The @code{.dword} directive is a synonym for @code{.int},
355
expecting zero or more EXPRESSIONS, separated by commas. For
356
each expression, a 32-bit little-endian constant is emitted.
358
@item .syntax ARGUMENT
359
@cindex assembler directive .syntax, CRIS
360
@cindex pseudo-op .syntax, CRIS
361
@cindex CRIS assembler directive .syntax
362
@cindex CRIS pseudo-op .syntax
363
The @code{.syntax} directive takes as @var{ARGUMENT} one of the
364
following case-sensitive choices.
367
@item no_register_prefix
369
The @code{.syntax no_register_prefix} @anchor{crisnous}directive
370
makes a @samp{$} character prefix on all registers optional. It
371
overrides a previous setting, including the corresponding effect
372
of the option @option{--no-underscore}. If this directive is
373
used when ordinary symbols do not have a @samp{_} character
374
prefix, care must be taken to avoid ambiguities whether an
375
operand is a register or a symbol; using symbols with names the
376
same as general or special registers then invoke undefined
379
@item register_prefix
381
This directive makes a @samp{$} character prefix on all
382
registers mandatory. It overrides a previous setting, including
383
the corresponding effect of the option @option{--underscore}.
385
@item leading_underscore
387
This is an assertion directive, emitting an error if the
388
@option{--no-underscore} option is in effect.
390
@item no_leading_underscore
392
This is the opposite of the @code{.syntax leading_underscore}
393
directive and emits an error if the option @option{--underscore}
398
@cindex assembler directive .arch, CRIS
399
@cindex pseudo-op .arch, CRIS
400
@cindex CRIS assembler directive .arch
401
@cindex CRIS pseudo-op .arch
402
This is an assertion directive, giving an error if the specified
403
@var{ARGUMENT} is not the same as the specified or default value
404
for the @option{--march=@var{architecture}} option
405
(@pxref{march-option}).
407
@c If you compare with md_pseudo_table, you see that we don't
408
@c document ".file" and ".loc" here. This is because we're just
409
@c wrapping the corresponding ELF function and emitting an error for