1 |
6 |
jlechner |
@c Copyright 1991, 1992, 1993, 1994, 1995, 1997, 1999, 2002
|
2 |
|
|
@c Free Software Foundation, Inc.
|
3 |
|
|
@c This is part of the GAS manual.
|
4 |
|
|
@c For copying conditions, see the file as.texinfo.
|
5 |
|
|
@ifset GENERIC
|
6 |
|
|
@page
|
7 |
|
|
@node Sparc-Dependent
|
8 |
|
|
@chapter SPARC Dependent Features
|
9 |
|
|
@end ifset
|
10 |
|
|
@ifclear GENERIC
|
11 |
|
|
@node Machine Dependencies
|
12 |
|
|
@chapter SPARC Dependent Features
|
13 |
|
|
@end ifclear
|
14 |
|
|
|
15 |
|
|
@cindex SPARC support
|
16 |
|
|
@menu
|
17 |
|
|
* Sparc-Opts:: Options
|
18 |
|
|
* Sparc-Aligned-Data:: Option to enforce aligned data
|
19 |
|
|
* Sparc-Syntax:: Syntax
|
20 |
|
|
* Sparc-Float:: Floating Point
|
21 |
|
|
* Sparc-Directives:: Sparc Machine Directives
|
22 |
|
|
@end menu
|
23 |
|
|
|
24 |
|
|
@node Sparc-Opts
|
25 |
|
|
@section Options
|
26 |
|
|
|
27 |
|
|
@cindex options for SPARC
|
28 |
|
|
@cindex SPARC options
|
29 |
|
|
@cindex architectures, SPARC
|
30 |
|
|
@cindex SPARC architectures
|
31 |
|
|
The SPARC chip family includes several successive versions, using the same
|
32 |
|
|
core instruction set, but including a few additional instructions at
|
33 |
|
|
each version. There are exceptions to this however. For details on what
|
34 |
|
|
instructions each variant supports, please see the chip's architecture
|
35 |
|
|
reference manual.
|
36 |
|
|
|
37 |
|
|
By default, @code{@value{AS}} assumes the core instruction set (SPARC
|
38 |
|
|
v6), but ``bumps'' the architecture level as needed: it switches to
|
39 |
|
|
successively higher architectures as it encounters instructions that
|
40 |
|
|
only exist in the higher levels.
|
41 |
|
|
|
42 |
|
|
If not configured for SPARC v9 (@code{sparc64-*-*}) GAS will not bump
|
43 |
|
|
past sparclite by default, an option must be passed to enable the
|
44 |
|
|
v9 instructions.
|
45 |
|
|
|
46 |
|
|
GAS treats sparclite as being compatible with v8, unless an architecture
|
47 |
|
|
is explicitly requested. SPARC v9 is always incompatible with sparclite.
|
48 |
|
|
|
49 |
|
|
@c The order here is the same as the order of enum sparc_opcode_arch_val
|
50 |
|
|
@c to give the user a sense of the order of the "bumping".
|
51 |
|
|
|
52 |
|
|
@table @code
|
53 |
|
|
@kindex -Av6
|
54 |
|
|
@kindex Av7
|
55 |
|
|
@kindex -Av8
|
56 |
|
|
@kindex -Asparclet
|
57 |
|
|
@kindex -Asparclite
|
58 |
|
|
@kindex -Av9
|
59 |
|
|
@kindex -Av9a
|
60 |
|
|
@item -Av6 | -Av7 | -Av8 | -Asparclet | -Asparclite
|
61 |
|
|
@itemx -Av8plus | -Av8plusa | -Av9 | -Av9a
|
62 |
|
|
Use one of the @samp{-A} options to select one of the SPARC
|
63 |
|
|
architectures explicitly. If you select an architecture explicitly,
|
64 |
|
|
@code{@value{AS}} reports a fatal error if it encounters an instruction
|
65 |
|
|
or feature requiring an incompatible or higher level.
|
66 |
|
|
|
67 |
|
|
@samp{-Av8plus} and @samp{-Av8plusa} select a 32 bit environment.
|
68 |
|
|
|
69 |
|
|
@samp{-Av9} and @samp{-Av9a} select a 64 bit environment and are not
|
70 |
|
|
available unless GAS is explicitly configured with 64 bit environment
|
71 |
|
|
support.
|
72 |
|
|
|
73 |
|
|
@samp{-Av8plusa} and @samp{-Av9a} enable the SPARC V9 instruction set with
|
74 |
|
|
UltraSPARC extensions.
|
75 |
|
|
|
76 |
|
|
@item -xarch=v8plus | -xarch=v8plusa
|
77 |
|
|
For compatibility with the SunOS v9 assembler. These options are
|
78 |
|
|
equivalent to -Av8plus and -Av8plusa, respectively.
|
79 |
|
|
|
80 |
|
|
@item -bump
|
81 |
|
|
Warn whenever it is necessary to switch to another level.
|
82 |
|
|
If an architecture level is explicitly requested, GAS will not issue
|
83 |
|
|
warnings until that level is reached, and will then bump the level
|
84 |
|
|
as required (except between incompatible levels).
|
85 |
|
|
|
86 |
|
|
@item -32 | -64
|
87 |
|
|
Select the word size, either 32 bits or 64 bits.
|
88 |
|
|
These options are only available with the ELF object file format,
|
89 |
|
|
and require that the necessary BFD support has been included.
|
90 |
|
|
@end table
|
91 |
|
|
|
92 |
|
|
@node Sparc-Aligned-Data
|
93 |
|
|
@section Enforcing aligned data
|
94 |
|
|
|
95 |
|
|
@cindex data alignment on SPARC
|
96 |
|
|
@cindex SPARC data alignment
|
97 |
|
|
SPARC GAS normally permits data to be misaligned. For example, it
|
98 |
|
|
permits the @code{.long} pseudo-op to be used on a byte boundary.
|
99 |
|
|
However, the native SunOS assemblers issue an error when they see
|
100 |
|
|
misaligned data.
|
101 |
|
|
|
102 |
|
|
@kindex --enforce-aligned-data
|
103 |
|
|
You can use the @code{--enforce-aligned-data} option to make SPARC GAS
|
104 |
|
|
also issue an error about misaligned data, just as the SunOS
|
105 |
|
|
assemblers do.
|
106 |
|
|
|
107 |
|
|
The @code{--enforce-aligned-data} option is not the default because gcc
|
108 |
|
|
issues misaligned data pseudo-ops when it initializes certain packed
|
109 |
|
|
data structures (structures defined using the @code{packed} attribute).
|
110 |
|
|
You may have to assemble with GAS in order to initialize packed data
|
111 |
|
|
structures in your own code.
|
112 |
|
|
|
113 |
|
|
@cindex SPARC syntax
|
114 |
|
|
@cindex syntax, SPARC
|
115 |
|
|
@node Sparc-Syntax
|
116 |
|
|
@section Sparc Syntax
|
117 |
|
|
The assembler syntax closely follows The Sparc Architecture Manual,
|
118 |
|
|
versions 8 and 9, as well as most extensions defined by Sun
|
119 |
|
|
for their UltraSPARC and Niagara line of processors.
|
120 |
|
|
|
121 |
|
|
@menu
|
122 |
|
|
* Sparc-Chars:: Special Characters
|
123 |
|
|
* Sparc-Regs:: Register Names
|
124 |
|
|
* Sparc-Constants:: Constant Names
|
125 |
|
|
* Sparc-Relocs:: Relocations
|
126 |
|
|
* Sparc-Size-Translations:: Size Translations
|
127 |
|
|
@end menu
|
128 |
|
|
|
129 |
|
|
@node Sparc-Chars
|
130 |
|
|
@subsection Special Characters
|
131 |
|
|
|
132 |
|
|
@cindex line comment character, Sparc
|
133 |
|
|
@cindex Sparc line comment character
|
134 |
|
|
@samp{#} is the line comment character.
|
135 |
|
|
|
136 |
|
|
@cindex line separator, Sparc
|
137 |
|
|
@cindex statement separator, Sparc
|
138 |
|
|
@cindex Sparc line separator
|
139 |
|
|
@samp{;} can be used instead of a newline to separate statements.
|
140 |
|
|
|
141 |
|
|
@node Sparc-Regs
|
142 |
|
|
@subsection Register Names
|
143 |
|
|
@cindex Sparc registers
|
144 |
|
|
@cindex register names, Sparc
|
145 |
|
|
|
146 |
|
|
The Sparc integer register file is broken down into global,
|
147 |
|
|
outgoing, local, and incoming.
|
148 |
|
|
|
149 |
|
|
@itemize @bullet
|
150 |
|
|
@item
|
151 |
|
|
The 8 global registers are referred to as @samp{%g@var{n}}.
|
152 |
|
|
|
153 |
|
|
@item
|
154 |
|
|
The 8 outgoing registers are referred to as @samp{%o@var{n}}.
|
155 |
|
|
|
156 |
|
|
@item
|
157 |
|
|
The 8 local registers are referred to as @samp{%l@var{n}}.
|
158 |
|
|
|
159 |
|
|
@item
|
160 |
|
|
The 8 incoming registers are referred to as @samp{%i@var{n}}.
|
161 |
|
|
|
162 |
|
|
@item
|
163 |
|
|
The frame pointer register @samp{%i6} can be referenced using
|
164 |
|
|
the alias @samp{%fp}.
|
165 |
|
|
|
166 |
|
|
@item
|
167 |
|
|
The stack pointer register @samp{%o6} can be referenced using
|
168 |
|
|
the alias @samp{%sp}.
|
169 |
|
|
@end itemize
|
170 |
|
|
|
171 |
|
|
Floating point registers are simply referred to as @samp{%f@var{n}}.
|
172 |
|
|
When assembling for pre-V9, only 32 floating point registers
|
173 |
|
|
are available. For V9 and later there are 64, but there are
|
174 |
|
|
restrictions when referencing the upper 32 registers. They
|
175 |
|
|
can only be accessed as double or quad, and thus only even
|
176 |
|
|
or quad numbered accesses are allowed. For example, @samp{%f34}
|
177 |
|
|
is a legal floating point register, but @samp{%f35} is not.
|
178 |
|
|
|
179 |
|
|
Certain V9 instructions allow access to ancillary state registers.
|
180 |
|
|
Most simply they can be referred to as @samp{%asr@var{n}} where
|
181 |
|
|
@var{n} can be from 16 to 31. However, there are some aliases
|
182 |
|
|
defined to reference ASR registers defined for various UltraSPARC
|
183 |
|
|
processors:
|
184 |
|
|
|
185 |
|
|
@itemize @bullet
|
186 |
|
|
@item
|
187 |
|
|
The tick compare register is referred to as @samp{%tick_cmpr}.
|
188 |
|
|
|
189 |
|
|
@item
|
190 |
|
|
The system tick register is referred to as @samp{%stick}. An alias,
|
191 |
|
|
@samp{%sys_tick}, exists but is deprecated and should not be used
|
192 |
|
|
by new software.
|
193 |
|
|
|
194 |
|
|
@item
|
195 |
|
|
The system tick compare register is referred to as @samp{%stick_cmpr}.
|
196 |
|
|
An alias, @samp{%sys_tick_cmpr}, exists but is deprecated and should
|
197 |
|
|
not be used by new software.
|
198 |
|
|
|
199 |
|
|
@item
|
200 |
|
|
The software interrupt register is referred to as @samp{%softint}.
|
201 |
|
|
|
202 |
|
|
@item
|
203 |
|
|
The set software interrupt register is referred to as @samp{%set_softint}.
|
204 |
|
|
The mnemonic @samp{%softint_set} is provided as an alias.
|
205 |
|
|
|
206 |
|
|
@item
|
207 |
|
|
The clear software interrupt register is referred to as
|
208 |
|
|
@samp{%clear_softint}. The mnemonic @samp{%softint_clear} is provided
|
209 |
|
|
as an alias.
|
210 |
|
|
|
211 |
|
|
@item
|
212 |
|
|
The performance instrumentation counters register is referred to as
|
213 |
|
|
@samp{%pic}.
|
214 |
|
|
|
215 |
|
|
@item
|
216 |
|
|
The performance control register is referred to as @samp{%pcr}.
|
217 |
|
|
|
218 |
|
|
@item
|
219 |
|
|
The graphics status register is referred to as @samp{%gsr}.
|
220 |
|
|
|
221 |
|
|
@item
|
222 |
|
|
The V9 dispatch control register is referred to as @samp{%dcr}.
|
223 |
|
|
@end itemize
|
224 |
|
|
|
225 |
|
|
Various V9 branch and conditional move instructions allow
|
226 |
|
|
specification of which set of integer condition codes to
|
227 |
|
|
test. These are referred to as @samp{%xcc} and @samp{%icc}.
|
228 |
|
|
|
229 |
|
|
In V9, there are 4 sets of floating point condition codes
|
230 |
|
|
which are referred to as @samp{%fcc@var{n}}.
|
231 |
|
|
|
232 |
|
|
Several special privileged and non-privileged registers
|
233 |
|
|
exist:
|
234 |
|
|
|
235 |
|
|
@itemize @bullet
|
236 |
|
|
@item
|
237 |
|
|
The V9 address space identifier register is referred to as @samp{%asi}.
|
238 |
|
|
|
239 |
|
|
@item
|
240 |
|
|
The V9 restorable windows register is referred to as @samp{%canrestore}.
|
241 |
|
|
|
242 |
|
|
@item
|
243 |
|
|
The V9 savable windows register is referred to as @samp{%cansave}.
|
244 |
|
|
|
245 |
|
|
@item
|
246 |
|
|
The V9 clean windows register is referred to as @samp{%cleanwin}.
|
247 |
|
|
|
248 |
|
|
@item
|
249 |
|
|
The V9 current window pointer register is referred to as @samp{%cwp}.
|
250 |
|
|
|
251 |
|
|
@item
|
252 |
|
|
The floating-point queue register is referred to as @samp{%fq}.
|
253 |
|
|
|
254 |
|
|
@item
|
255 |
|
|
The V8 co-processor queue register is referred to as @samp{%cq}.
|
256 |
|
|
|
257 |
|
|
@item
|
258 |
|
|
The floating point status register is referred to as @samp{%fsr}.
|
259 |
|
|
|
260 |
|
|
@item
|
261 |
|
|
The other windows register is referred to as @samp{%otherwin}.
|
262 |
|
|
|
263 |
|
|
@item
|
264 |
|
|
The V9 program counter register is referred to as @samp{%pc}.
|
265 |
|
|
|
266 |
|
|
@item
|
267 |
|
|
The V9 next program counter register is referred to as @samp{%npc}.
|
268 |
|
|
|
269 |
|
|
@item
|
270 |
|
|
The V9 processor interrupt level register is referred to as @samp{%pil}.
|
271 |
|
|
|
272 |
|
|
@item
|
273 |
|
|
The V9 processor state register is referred to as @samp{%pstate}.
|
274 |
|
|
|
275 |
|
|
@item
|
276 |
|
|
The trap base address register is referred to as @samp{%tba}.
|
277 |
|
|
|
278 |
|
|
@item
|
279 |
|
|
The V9 tick register is referred to as @samp{%tick}.
|
280 |
|
|
|
281 |
|
|
@item
|
282 |
|
|
The V9 trap level is referred to as @samp{%tl}.
|
283 |
|
|
|
284 |
|
|
@item
|
285 |
|
|
The V9 trap program counter is referred to as @samp{%tpc}.
|
286 |
|
|
|
287 |
|
|
@item
|
288 |
|
|
The V9 trap next program counter is referred to as @samp{%tnpc}.
|
289 |
|
|
|
290 |
|
|
@item
|
291 |
|
|
The V9 trap state is referred to as @samp{%tstate}.
|
292 |
|
|
|
293 |
|
|
@item
|
294 |
|
|
The V9 trap type is referred to as @samp{%tt}.
|
295 |
|
|
|
296 |
|
|
@item
|
297 |
|
|
The V9 condition codes is referred to as @samp{%ccr}.
|
298 |
|
|
|
299 |
|
|
@item
|
300 |
|
|
The V9 floating-point registers state is referred to as @samp{%fprs}.
|
301 |
|
|
|
302 |
|
|
@item
|
303 |
|
|
The V9 version register is referred to as @samp{%ver}.
|
304 |
|
|
|
305 |
|
|
@item
|
306 |
|
|
The V9 window state register is referred to as @samp{%wstate}.
|
307 |
|
|
|
308 |
|
|
@item
|
309 |
|
|
The Y register is referred to as @samp{%y}.
|
310 |
|
|
|
311 |
|
|
@item
|
312 |
|
|
The V8 window invalid mask register is referred to as @samp{%wim}.
|
313 |
|
|
|
314 |
|
|
@item
|
315 |
|
|
The V8 processor state register is referred to as @samp{%psr}.
|
316 |
|
|
|
317 |
|
|
@item
|
318 |
|
|
The V9 global register level register is referred to as @samp{%gl}.
|
319 |
|
|
@end itemize
|
320 |
|
|
|
321 |
|
|
Several special register names exist for hypervisor mode code:
|
322 |
|
|
|
323 |
|
|
@itemize @bullet
|
324 |
|
|
@item
|
325 |
|
|
The hyperprivileged processor state register is referred to as
|
326 |
|
|
@samp{%hpstate}.
|
327 |
|
|
|
328 |
|
|
@item
|
329 |
|
|
The hyperprivileged trap state register is referred to as @samp{%htstate}.
|
330 |
|
|
|
331 |
|
|
@item
|
332 |
|
|
The hyperprivileged interrupt pending register is referred to as
|
333 |
|
|
@samp{%hintp}.
|
334 |
|
|
|
335 |
|
|
@item
|
336 |
|
|
The hyperprivileged trap base address register is referred to as
|
337 |
|
|
@samp{%htba}.
|
338 |
|
|
|
339 |
|
|
@item
|
340 |
|
|
The hyperprivileged implementation version register is referred
|
341 |
|
|
to as @samp{%hver}.
|
342 |
|
|
|
343 |
|
|
@item
|
344 |
|
|
The hyperprivileged system tick compare register is referred
|
345 |
|
|
to as @samp{%hstick_cmpr}. Note that there is no @samp{%hstick}
|
346 |
|
|
register, the normal @samp{%stick} is used.
|
347 |
|
|
@end itemize
|
348 |
|
|
|
349 |
|
|
@node Sparc-Constants
|
350 |
|
|
@subsection Constants
|
351 |
|
|
@cindex Sparc constants
|
352 |
|
|
@cindex constants, Sparc
|
353 |
|
|
|
354 |
|
|
Several Sparc instructions take an immediate operand field for
|
355 |
|
|
which mnemonic names exist. Two such examples are @samp{membar}
|
356 |
|
|
and @samp{prefetch}. Another example are the set of V9
|
357 |
|
|
memory access instruction that allow specification of an
|
358 |
|
|
address space identifier.
|
359 |
|
|
|
360 |
|
|
The @samp{membar} instruction specifies a memory barrier that is
|
361 |
|
|
the defined by the operand which is a bitmask. The supported
|
362 |
|
|
mask mnemonics are:
|
363 |
|
|
|
364 |
|
|
@itemize @bullet
|
365 |
|
|
@item
|
366 |
|
|
@samp{#Sync} requests that all operations (including nonmemory
|
367 |
|
|
reference operations) appearing prior to the @code{membar} must have
|
368 |
|
|
been performed and the effects of any exceptions become visible before
|
369 |
|
|
any instructions after the @code{membar} may be initiated. This
|
370 |
|
|
corresponds to @code{membar} cmask field bit 2.
|
371 |
|
|
|
372 |
|
|
@item
|
373 |
|
|
@samp{#MemIssue} requests that all memory reference operations
|
374 |
|
|
appearing prior to the @code{membar} must have been performed before
|
375 |
|
|
any memory operation after the @code{membar} may be initiated. This
|
376 |
|
|
corresponds to @code{membar} cmask field bit 1.
|
377 |
|
|
|
378 |
|
|
@item
|
379 |
|
|
@samp{#Lookaside} requests that a store appearing prior to the
|
380 |
|
|
@code{membar} must complete before any load following the
|
381 |
|
|
@code{membar} referencing the same address can be initiated. This
|
382 |
|
|
corresponds to @code{membar} cmask field bit 0.
|
383 |
|
|
|
384 |
|
|
@item
|
385 |
|
|
@samp{#StoreStore} defines that the effects of all stores appearing
|
386 |
|
|
prior to the @code{membar} instruction must be visible to all
|
387 |
|
|
processors before the effect of any stores following the
|
388 |
|
|
@code{membar}. Equivalent to the deprecated @code{stbar} instruction.
|
389 |
|
|
This corresponds to @code{membar} mmask field bit 3.
|
390 |
|
|
|
391 |
|
|
@item
|
392 |
|
|
@samp{#LoadStore} defines all loads appearing prior to the
|
393 |
|
|
@code{membar} instruction must have been performed before the effect
|
394 |
|
|
of any stores following the @code{membar} is visible to any other
|
395 |
|
|
processor. This corresponds to @code{membar} mmask field bit 2.
|
396 |
|
|
|
397 |
|
|
@item
|
398 |
|
|
@samp{#StoreLoad} defines that the effects of all stores appearing
|
399 |
|
|
prior to the @code{membar} instruction must be visible to all
|
400 |
|
|
processors before loads following the @code{membar} may be performed.
|
401 |
|
|
This corresponds to @code{membar} mmask field bit 1.
|
402 |
|
|
|
403 |
|
|
@item
|
404 |
|
|
@samp{#LoadLoad} defines that all loads appearing prior to the
|
405 |
|
|
@code{membar} instruction must have been performed before any loads
|
406 |
|
|
following the @code{membar} may be performed. This corresponds to
|
407 |
|
|
@code{membar} mmask field bit 0.
|
408 |
|
|
|
409 |
|
|
@end itemize
|
410 |
|
|
|
411 |
|
|
These values can be ored together, for example:
|
412 |
|
|
|
413 |
|
|
@example
|
414 |
|
|
membar #Sync
|
415 |
|
|
membar #StoreLoad | #LoadLoad
|
416 |
|
|
membar #StoreLoad | #StoreStore
|
417 |
|
|
@end example
|
418 |
|
|
|
419 |
|
|
The @code{prefetch} and @code{prefetcha} instructions take a prefetch
|
420 |
|
|
function code. The following prefetch function code constant
|
421 |
|
|
mnemonics are available:
|
422 |
|
|
|
423 |
|
|
@itemize @bullet
|
424 |
|
|
@item
|
425 |
|
|
@samp{#n_reads} requests a prefetch for several reads, and corresponds
|
426 |
|
|
to a prefetch function code of 0.
|
427 |
|
|
|
428 |
|
|
@samp{#one_read} requests a prefetch for one read, and corresponds
|
429 |
|
|
to a prefetch function code of 1.
|
430 |
|
|
|
431 |
|
|
@samp{#n_writes} requests a prefetch for several writes (and possibly
|
432 |
|
|
reads), and corresponds to a prefetch function code of 2.
|
433 |
|
|
|
434 |
|
|
@samp{#one_write} requests a prefetch for one write, and corresponds
|
435 |
|
|
to a prefetch function code of 3.
|
436 |
|
|
|
437 |
|
|
@samp{#page} requests a prefetch page, and corresponds to a prefetch
|
438 |
|
|
function code of 4.
|
439 |
|
|
|
440 |
|
|
@samp{#invalidate} requests a prefetch invalidate, and corresponds to
|
441 |
|
|
a prefetch function code of 16.
|
442 |
|
|
|
443 |
|
|
@samp{#unified} requests a prefetch to the nearest unified cache, and
|
444 |
|
|
corresponds to a prefetch function code of 17.
|
445 |
|
|
|
446 |
|
|
@samp{#n_reads_strong} requests a strong prefetch for several reads,
|
447 |
|
|
and corresponds to a prefetch function code of 20.
|
448 |
|
|
|
449 |
|
|
@samp{#one_read_strong} requests a strong prefetch for one read,
|
450 |
|
|
and corresponds to a prefetch function code of 21.
|
451 |
|
|
|
452 |
|
|
@samp{#n_writes_strong} requests a strong prefetch for several writes,
|
453 |
|
|
and corresponds to a prefetch function code of 22.
|
454 |
|
|
|
455 |
|
|
@samp{#one_write_strong} requests a strong prefetch for one write,
|
456 |
|
|
and corresponds to a prefetch function code of 23.
|
457 |
|
|
|
458 |
|
|
Onle one prefetch code may be specified. Here are some examples:
|
459 |
|
|
|
460 |
|
|
@example
|
461 |
|
|
prefetch [%l0 + %l2], #one_read
|
462 |
|
|
prefetch [%g2 + 8], #n_writes
|
463 |
|
|
prefetcha [%g1] 0x8, #unified
|
464 |
|
|
prefetcha [%o0 + 0x10] %asi, #n_reads
|
465 |
|
|
@end example
|
466 |
|
|
|
467 |
|
|
The actual behavior of a given prefetch function code is processor
|
468 |
|
|
specific. If a processor does not implement a given prefetch
|
469 |
|
|
function code, it will treat the prefetch instruction as a nop.
|
470 |
|
|
|
471 |
|
|
For instructions that accept an immediate address space identifier,
|
472 |
|
|
@code{@value{AS}} provides many mnemonics corresponding to
|
473 |
|
|
V9 defined as well as UltraSPARC and Niagara extended values.
|
474 |
|
|
For example, @samp{#ASI_P} and @samp{#ASI_BLK_INIT_QUAD_LDD_AIUS}.
|
475 |
|
|
See the V9 and processor specific manuals for details.
|
476 |
|
|
|
477 |
|
|
@end itemize
|
478 |
|
|
|
479 |
|
|
@node Sparc-Relocs
|
480 |
|
|
@subsection Relocations
|
481 |
|
|
@cindex Sparc relocations
|
482 |
|
|
@cindex relocations, Sparc
|
483 |
|
|
|
484 |
|
|
ELF relocations are available as defined in the 32-bit and 64-bit
|
485 |
|
|
Sparc ELF specifications.
|
486 |
|
|
|
487 |
|
|
@code{R_SPARC_HI22} is obtained using @samp{%hi} and @code{R_SPARC_LO10}
|
488 |
|
|
is obtained using @samp{%lo}. Likewise @code{R_SPARC_HIX22} is
|
489 |
|
|
obtained from @samp{%hix} and @code{R_SPARC_LOX10} is obtained
|
490 |
|
|
using @samp{%lox}. For example:
|
491 |
|
|
|
492 |
|
|
@example
|
493 |
|
|
sethi %hi(symbol), %g1
|
494 |
|
|
or %g1, %lo(symbol), %g1
|
495 |
|
|
|
496 |
|
|
sethi %hix(symbol), %g1
|
497 |
|
|
xor %g1, %lox(symbol), %g1
|
498 |
|
|
@end example
|
499 |
|
|
|
500 |
|
|
These ``high'' mnemonics extract bits 31:10 of their operand,
|
501 |
|
|
and the ``low'' mnemonics extract bits 9:0 of their operand.
|
502 |
|
|
|
503 |
|
|
V9 code model relocations can be requested as follows:
|
504 |
|
|
|
505 |
|
|
@itemize @bullet
|
506 |
|
|
@item
|
507 |
|
|
@code{R_SPARC_HH22} is requested using @samp{%hh}. It can
|
508 |
|
|
also be generated using @samp{%uhi}.
|
509 |
|
|
@item
|
510 |
|
|
@code{R_SPARC_HM10} is requested using @samp{%hm}. It can
|
511 |
|
|
also be generated using @samp{%ulo}.
|
512 |
|
|
@item
|
513 |
|
|
@code{R_SPARC_LM22} is requested using @samp{%lm}.
|
514 |
|
|
|
515 |
|
|
@item
|
516 |
|
|
@code{R_SPARC_H44} is requested using @samp{%h44}.
|
517 |
|
|
@item
|
518 |
|
|
@code{R_SPARC_M44} is requested using @samp{%m44}.
|
519 |
|
|
@item
|
520 |
|
|
@code{R_SPARC_L44} is requested using @samp{%l44}.
|
521 |
|
|
@end itemize
|
522 |
|
|
|
523 |
|
|
The PC relative relocation @code{R_SPARC_PC22} can be obtained by
|
524 |
|
|
enclosing an operand inside of @samp{%pc22}. Likewise, the
|
525 |
|
|
@code{R_SPARC_PC10} relocation can be obtained using @samp{%pc10}.
|
526 |
|
|
These are mostly used when assembling PIC code. For example, the
|
527 |
|
|
standard PIC sequence on Sparc to get the base of the global offset
|
528 |
|
|
table, PC relative, into a register, can be performed as:
|
529 |
|
|
|
530 |
|
|
@example
|
531 |
|
|
sethi %pc22(_GLOBAL_OFFSET_TABLE_-4), %l7
|
532 |
|
|
add %l7, %pc10(_GLOBAL_OFFSET_TABLE_+4), %l7
|
533 |
|
|
@end example
|
534 |
|
|
|
535 |
|
|
Several relocations exist to allow the link editor to potentially
|
536 |
|
|
optimize GOT data references. The @code{R_SPARC_GOTDATA_OP_HIX22}
|
537 |
|
|
relocation can obtained by enclosing an operand inside of
|
538 |
|
|
@samp{%gdop_hix22}. The @code{R_SPARC_GOTDATA_OP_LOX10}
|
539 |
|
|
relocation can obtained by enclosing an operand inside of
|
540 |
|
|
@samp{%gdop_lox10}. Likewise, @code{R_SPARC_GOTDATA_OP} can be
|
541 |
|
|
obtained by enclosing an operand inside of @samp{%gdop}.
|
542 |
|
|
For example, assuming the GOT base is in register @code{%l7}:
|
543 |
|
|
|
544 |
|
|
@example
|
545 |
|
|
sethi %gdop_hix22(symbol), %l1
|
546 |
|
|
xor %l1, %gdop_lox10(symbol), %l1
|
547 |
|
|
ld [%l7 + %l1], %l2, %gdop(symbol)
|
548 |
|
|
@end example
|
549 |
|
|
|
550 |
|
|
There are many relocations that can be requested for access to
|
551 |
|
|
thread local storage variables. All of the Sparc TLS mnemonics
|
552 |
|
|
are supported:
|
553 |
|
|
|
554 |
|
|
@itemize @bullet
|
555 |
|
|
@item
|
556 |
|
|
@code{R_SPARC_TLS_GD_HI22} is requested using @samp{%tgd_hi22}.
|
557 |
|
|
@item
|
558 |
|
|
@code{R_SPARC_TLS_GD_LO10} is requested using @samp{%tgd_lo10}.
|
559 |
|
|
@item
|
560 |
|
|
@code{R_SPARC_TLS_GD_ADD} is requested using @samp{%tgd_add}.
|
561 |
|
|
@item
|
562 |
|
|
@code{R_SPARC_TLS_GD_CALL} is requested using @samp{%tgd_call}.
|
563 |
|
|
|
564 |
|
|
@item
|
565 |
|
|
@code{R_SPARC_TLS_LDM_HI22} is requested using @samp{%tldm_hi22}.
|
566 |
|
|
@item
|
567 |
|
|
@code{R_SPARC_TLS_LDM_LO10} is requested using @samp{%tldm_lo10}.
|
568 |
|
|
@item
|
569 |
|
|
@code{R_SPARC_TLS_LDM_ADD} is requested using @samp{%tldm_add}.
|
570 |
|
|
@item
|
571 |
|
|
@code{R_SPARC_TLS_LDM_CALL} is requested using @samp{%tldm_call}.
|
572 |
|
|
|
573 |
|
|
@item
|
574 |
|
|
@code{R_SPARC_TLS_LDO_HIX22} is requested using @samp{%tldo_hix22}.
|
575 |
|
|
@item
|
576 |
|
|
@code{R_SPARC_TLS_LDO_LOX10} is requested using @samp{%tldo_lox10}.
|
577 |
|
|
@item
|
578 |
|
|
@code{R_SPARC_TLS_LDO_ADD} is requested using @samp{%tldo_add}.
|
579 |
|
|
|
580 |
|
|
@item
|
581 |
|
|
@code{R_SPARC_TLS_IE_HI22} is requested using @samp{%tie_hi22}.
|
582 |
|
|
@item
|
583 |
|
|
@code{R_SPARC_TLS_IE_LO10} is requested using @samp{%tie_lo10}.
|
584 |
|
|
@item
|
585 |
|
|
@code{R_SPARC_TLS_IE_LD} is requested using @samp{%tie_ld}.
|
586 |
|
|
@item
|
587 |
|
|
@code{R_SPARC_TLS_IE_LDX} is requested using @samp{%tie_ldx}.
|
588 |
|
|
@item
|
589 |
|
|
@code{R_SPARC_TLS_IE_ADD} is requested using @samp{%tie_add}.
|
590 |
|
|
|
591 |
|
|
@item
|
592 |
|
|
@code{R_SPARC_TLS_LE_HIX22} is requested using @samp{%tle_hix22}.
|
593 |
|
|
@item
|
594 |
|
|
@code{R_SPARC_TLS_LE_LOX10} is requested using @samp{%tle_lox10}.
|
595 |
|
|
@end itemize
|
596 |
|
|
|
597 |
|
|
Here are some example TLS model sequences.
|
598 |
|
|
|
599 |
|
|
First, General Dynamic:
|
600 |
|
|
|
601 |
|
|
@example
|
602 |
|
|
sethi %tgd_hi22(symbol), %l1
|
603 |
|
|
add %l1, %tgd_lo10(symbol), %l1
|
604 |
|
|
add %l7, %l1, %o0, %tgd_add(symbol)
|
605 |
|
|
call __tls_get_addr, %tgd_call(symbol)
|
606 |
|
|
nop
|
607 |
|
|
@end example
|
608 |
|
|
|
609 |
|
|
Local Dynamic:
|
610 |
|
|
|
611 |
|
|
@example
|
612 |
|
|
sethi %tldm_hi22(symbol), %l1
|
613 |
|
|
add %l1, %tldm_lo10(symbol), %l1
|
614 |
|
|
add %l7, %l1, %o0, %tldm_add(symbol)
|
615 |
|
|
call __tls_get_addr, %tldm_call(symbol)
|
616 |
|
|
nop
|
617 |
|
|
|
618 |
|
|
sethi %tldo_hix22(symbol), %l1
|
619 |
|
|
xor %l1, %tldo_lox10(symbol), %l1
|
620 |
|
|
add %o0, %l1, %l1, %tldo_add(symbol)
|
621 |
|
|
@end example
|
622 |
|
|
|
623 |
|
|
Initial Exec:
|
624 |
|
|
|
625 |
|
|
@example
|
626 |
|
|
sethi %tie_hi22(symbol), %l1
|
627 |
|
|
add %l1, %tie_lo10(symbol), %l1
|
628 |
|
|
ld [%l7 + %l1], %o0, %tie_ld(symbol)
|
629 |
|
|
add %g7, %o0, %o0, %tie_add(symbol)
|
630 |
|
|
|
631 |
|
|
sethi %tie_hi22(symbol), %l1
|
632 |
|
|
add %l1, %tie_lo10(symbol), %l1
|
633 |
|
|
ldx [%l7 + %l1], %o0, %tie_ldx(symbol)
|
634 |
|
|
add %g7, %o0, %o0, %tie_add(symbol)
|
635 |
|
|
@end example
|
636 |
|
|
|
637 |
|
|
And finally, Local Exec:
|
638 |
|
|
|
639 |
|
|
@example
|
640 |
|
|
sethi %tle_hix22(symbol), %l1
|
641 |
|
|
add %l1, %tle_lox10(symbol), %l1
|
642 |
|
|
add %g7, %l1, %l1
|
643 |
|
|
@end example
|
644 |
|
|
|
645 |
|
|
When assembling for 64-bit, and a secondary constant addend is
|
646 |
|
|
specified in an address expression that would normally generate
|
647 |
|
|
an @code{R_SPARC_LO10} relocation, the assembler will emit an
|
648 |
|
|
@code{R_SPARC_OLO10} instead.
|
649 |
|
|
|
650 |
|
|
@node Sparc-Size-Translations
|
651 |
|
|
@subsection Size Translations
|
652 |
|
|
@cindex Sparc size translations
|
653 |
|
|
@cindex size, translations, Sparc
|
654 |
|
|
|
655 |
|
|
Often it is desirable to write code in an operand size agnostic
|
656 |
|
|
manner. @code{@value{AS}} provides support for this via
|
657 |
|
|
operand size opcode translations. Translations are supported
|
658 |
|
|
for loads, stores, shifts, compare-and-swap atomics, and the
|
659 |
|
|
@samp{clr} synthetic instruction.
|
660 |
|
|
|
661 |
|
|
If generating 32-bit code, @code{@value{AS}} will generate the
|
662 |
|
|
32-bit opcode. Whereas if 64-bit code is being generated,
|
663 |
|
|
the 64-bit opcode will be emitted. For example @code{ldn}
|
664 |
|
|
will be transformed into @code{ld} for 32-bit code and
|
665 |
|
|
@code{ldx} for 64-bit code.
|
666 |
|
|
|
667 |
|
|
Here is an example meant to demonstrate all the supported
|
668 |
|
|
opcode translations:
|
669 |
|
|
|
670 |
|
|
@example
|
671 |
|
|
ldn [%o0], %o1
|
672 |
|
|
ldna [%o0] %asi, %o2
|
673 |
|
|
stn %o1, [%o0]
|
674 |
|
|
stna %o2, [%o0] %asi
|
675 |
|
|
slln %o3, 3, %o3
|
676 |
|
|
srln %o4, 8, %o4
|
677 |
|
|
sran %o5, 12, %o5
|
678 |
|
|
casn [%o0], %o1, %o2
|
679 |
|
|
casna [%o0] %asi, %o1, %o2
|
680 |
|
|
clrn %g1
|
681 |
|
|
@end example
|
682 |
|
|
|
683 |
|
|
In 32-bit mode @code{@value{AS}} will emit:
|
684 |
|
|
|
685 |
|
|
@example
|
686 |
|
|
ld [%o0], %o1
|
687 |
|
|
lda [%o0] %asi, %o2
|
688 |
|
|
st %o1, [%o0]
|
689 |
|
|
sta %o2, [%o0] %asi
|
690 |
|
|
sll %o3, 3, %o3
|
691 |
|
|
srl %o4, 8, %o4
|
692 |
|
|
sra %o5, 12, %o5
|
693 |
|
|
cas [%o0], %o1, %o2
|
694 |
|
|
casa [%o0] %asi, %o1, %o2
|
695 |
|
|
clr %g1
|
696 |
|
|
@end example
|
697 |
|
|
|
698 |
|
|
And in 64-bit mode @code{@value{AS}} will emit:
|
699 |
|
|
|
700 |
|
|
@example
|
701 |
|
|
ldx [%o0], %o1
|
702 |
|
|
ldxa [%o0] %asi, %o2
|
703 |
|
|
stx %o1, [%o0]
|
704 |
|
|
stxa %o2, [%o0] %asi
|
705 |
|
|
sllx %o3, 3, %o3
|
706 |
|
|
srlx %o4, 8, %o4
|
707 |
|
|
srax %o5, 12, %o5
|
708 |
|
|
casx [%o0], %o1, %o2
|
709 |
|
|
casxa [%o0] %asi, %o1, %o2
|
710 |
|
|
clrx %g1
|
711 |
|
|
@end example
|
712 |
|
|
|
713 |
|
|
Finally, the @samp{.nword} translating directive is supported
|
714 |
|
|
as well. It is documented in the section on Sparc machine
|
715 |
|
|
directives.
|
716 |
|
|
|
717 |
|
|
@node Sparc-Float
|
718 |
|
|
@section Floating Point
|
719 |
|
|
|
720 |
|
|
@cindex floating point, SPARC (@sc{ieee})
|
721 |
|
|
@cindex SPARC floating point (@sc{ieee})
|
722 |
|
|
The Sparc uses @sc{ieee} floating-point numbers.
|
723 |
|
|
|
724 |
|
|
@node Sparc-Directives
|
725 |
|
|
@section Sparc Machine Directives
|
726 |
|
|
|
727 |
|
|
@cindex SPARC machine directives
|
728 |
|
|
@cindex machine directives, SPARC
|
729 |
|
|
The Sparc version of @code{@value{AS}} supports the following additional
|
730 |
|
|
machine directives:
|
731 |
|
|
|
732 |
|
|
@table @code
|
733 |
|
|
@cindex @code{align} directive, SPARC
|
734 |
|
|
@item .align
|
735 |
|
|
This must be followed by the desired alignment in bytes.
|
736 |
|
|
|
737 |
|
|
@cindex @code{common} directive, SPARC
|
738 |
|
|
@item .common
|
739 |
|
|
This must be followed by a symbol name, a positive number, and
|
740 |
|
|
@code{"bss"}. This behaves somewhat like @code{.comm}, but the
|
741 |
|
|
syntax is different.
|
742 |
|
|
|
743 |
|
|
@cindex @code{half} directive, SPARC
|
744 |
|
|
@item .half
|
745 |
|
|
This is functionally identical to @code{.short}.
|
746 |
|
|
|
747 |
|
|
@cindex @code{nword} directive, SPARC
|
748 |
|
|
@item .nword
|
749 |
|
|
On the Sparc, the @code{.nword} directive produces native word sized value,
|
750 |
|
|
ie. if assembling with -32 it is equivalent to @code{.word}, if assembling
|
751 |
|
|
with -64 it is equivalent to @code{.xword}.
|
752 |
|
|
|
753 |
|
|
@cindex @code{proc} directive, SPARC
|
754 |
|
|
@item .proc
|
755 |
|
|
This directive is ignored. Any text following it on the same
|
756 |
|
|
line is also ignored.
|
757 |
|
|
|
758 |
|
|
@cindex @code{register} directive, SPARC
|
759 |
|
|
@item .register
|
760 |
|
|
This directive declares use of a global application or system register.
|
761 |
|
|
It must be followed by a register name %g2, %g3, %g6 or %g7, comma and
|
762 |
|
|
the symbol name for that register. If symbol name is @code{#scratch},
|
763 |
|
|
it is a scratch register, if it is @code{#ignore}, it just suppresses any
|
764 |
|
|
errors about using undeclared global register, but does not emit any
|
765 |
|
|
information about it into the object file. This can be useful e.g. if you
|
766 |
|
|
save the register before use and restore it after.
|
767 |
|
|
|
768 |
|
|
@cindex @code{reserve} directive, SPARC
|
769 |
|
|
@item .reserve
|
770 |
|
|
This must be followed by a symbol name, a positive number, and
|
771 |
|
|
@code{"bss"}. This behaves somewhat like @code{.lcomm}, but the
|
772 |
|
|
syntax is different.
|
773 |
|
|
|
774 |
|
|
@cindex @code{seg} directive, SPARC
|
775 |
|
|
@item .seg
|
776 |
|
|
This must be followed by @code{"text"}, @code{"data"}, or
|
777 |
|
|
@code{"data1"}. It behaves like @code{.text}, @code{.data}, or
|
778 |
|
|
@code{.data 1}.
|
779 |
|
|
|
780 |
|
|
@cindex @code{skip} directive, SPARC
|
781 |
|
|
@item .skip
|
782 |
|
|
This is functionally identical to the @code{.space} directive.
|
783 |
|
|
|
784 |
|
|
@cindex @code{word} directive, SPARC
|
785 |
|
|
@item .word
|
786 |
|
|
On the Sparc, the @code{.word} directive produces 32 bit values,
|
787 |
|
|
instead of the 16 bit values it produces on many other machines.
|
788 |
|
|
|
789 |
|
|
@cindex @code{xword} directive, SPARC
|
790 |
|
|
@item .xword
|
791 |
|
|
On the Sparc V9 processor, the @code{.xword} directive produces
|
792 |
|
|
64 bit values.
|
793 |
|
|
@end table
|