1 @c Copyright (C) 2016-2019 Free Software Foundation, Inc.
2 @c This is part of the GAS anual.
3 @c For copying conditions, see the file as.texinfo
9 @chapter RISC-V Dependent Features
12 @node Machine Dependencies
13 @chapter RISC-V Dependent Features
16 @cindex RISC-V support
18 * RISC-V-Options:: RISC-V Options
19 * RISC-V-Directives:: RISC-V Directives
20 * RISC-V-Formats:: RISC-V Instruction Formats
21 * RISC-V-ATTRIBUTE:: RISC-V Object Attribute
25 @section RISC-V Options
27 The following table lists all available RISC-V specific options.
32 @cindex @samp{-fpic} option, RISC-V
35 Generate position-independent code
37 @cindex @samp{-fno-pic} option, RISC-V
39 Don't generate position-independent code (default)
41 @cindex @samp{-march=ISA} option, RISC-V
43 Select the base isa, as specified by ISA. For example -march=rv32ima.
45 @cindex @samp{-mabi=ABI} option, RISC-V
47 Selects the ABI, which is either "ilp32" or "lp64", optionally followed
48 by "f", "d", or "q" to indicate single-precision, double-precision, or
49 quad-precision floating-point calling convention, or none to indicate
50 the soft-float calling convention. Also, "ilp32" can optionally be followed
51 by "e" to indicate the RVE ABI, which is always soft-float.
53 @cindex @samp{-mrelax} option, RISC-V
55 Take advantage of linker relaxations to reduce the number of instructions
56 required to materialize symbol addresses. (default)
58 @cindex @samp{-mno-relax} option, RISC-V
60 Don't do linker relaxations.
65 @node RISC-V-Directives
66 @section RISC-V Directives
67 @cindex machine directives, RISC-V
68 @cindex RISC-V machine directives
70 The following table lists all available RISC-V specific directives.
74 @cindex @code{align} directive
75 @item .align @var{size-log-2}
76 Align to the given boundary, with the size given as log2 the number of bytes to
79 @cindex Data directives
80 @item .half @var{value}
81 @itemx .word @var{value}
82 @itemx .dword @var{value}
83 Emits a half-word, word, or double-word value at the current position.
85 @cindex DTP-relative data directives
86 @item .dtprelword @var{value}
87 @itemx .dtpreldword @var{value}
88 Emits a DTP-relative word (or double-word) at the current position. This is
89 meant to be used by the compiler in shared libraries for DWARF debug info for
90 thread local variables.
94 Sets the current section to the BSS section.
96 @cindex LEB128 directives
97 @item .uleb128 @var{value}
98 @itemx .sleb128 @var{value}
99 Emits a signed or unsigned LEB128 value at the current position. This only
100 accepts constant expressions, because symbol addresses can change with
101 relaxation, and we don't support relocations to modify LEB128 values at link
104 @cindex Option directive
105 @cindex @code{option} directive
106 @item .option @var{argument}
107 Modifies RISC-V specific assembler options inline with the assembly code.
108 This is used when particular instruction sequences must be assembled with a
109 specific set of options. For example, since we relax addressing sequences to
110 shorter GP-relative sequences when possible the initial load of GP must not be
111 relaxed and should be emitted as something like
116 la gp, __global_pointer$
120 in order to produce after linker relaxation the expected
123 auipc gp, %pcrel_hi(__global_pointer$)
124 addi gp, gp, %pcrel_lo(__global_pointer$)
133 It's not expected that options are changed in this manner during regular use,
134 but there are a handful of esoteric cases like the one above where users need
135 to disable particular features of the assembler for particular code sequences.
136 The complete list of option arguments is shown below:
141 Pushes or pops the current option stack. These should be used whenever
142 changing an option in line with assembly code in order to ensure the user's
143 command-line options are respected for the bulk of the file being assembled.
147 Enables or disables the generation of compressed instructions. Instructions
148 are opportunistically compressed by the RISC-V assembler when possible, but
149 sometimes this behavior is not desirable.
153 Enables or disables position-independent code generation. Unless you really
154 know what you're doing, this should only be at the top of a file.
158 Enables or disables relaxation. The RISC-V assembler and linker
159 opportunistically relax some code sequences, but sometimes this behavior is not
163 @cindex INSN directives
164 @item .insn @var{value}
165 @itemx .insn @var{value}
166 This directive permits the numeric representation of an instructions
167 and makes the assembler insert the operands according to one of the
168 instruction formats for @samp{.insn} (@ref{RISC-V-Formats}).
169 For example, the instruction @samp{add a0, a1, a2} could be written as
170 @samp{.insn r 0x33, 0, 0, a0, a1, a2}.
172 @cindex @code{.attribute} directive, RISC-V
173 @item .attribute @var{tag}, @var{value}
174 Set the object attribute @var{tag} to @var{value}.
176 The @var{tag} is either an attribute number, or one of the following:
177 @code{Tag_RISCV_arch}, @code{Tag_RISCV_stack_align},
178 @code{Tag_RISCV_unaligned_access}, @code{Tag_RISCV_priv_spec},
179 @code{Tag_RISCV_priv_spec_minor}, @code{Tag_RISCV_priv_spec_revision}.
184 @section Instruction Formats
185 @cindex instruction formats, risc-v
186 @cindex RISC-V instruction formats
188 The RISC-V Instruction Set Manual Volume I: User-Level ISA lists 12
189 instruction formats where some of the formats have multiple variants.
190 For the @samp{.insn} pseudo directive the assembler recognizes some
192 Typically, the most general variant of the instruction format is used
193 by the @samp{.insn} directive.
195 The following table lists the abbreviations used in the table of
199 @multitable @columnfractions .15 .40
200 @item opcode @tab Unsigned immediate or opcode name for 7-bits opcode.
201 @item opcode2 @tab Unsigned immediate or opcode name for 2-bits opcode.
202 @item func7 @tab Unsigned immediate for 7-bits function code.
203 @item func6 @tab Unsigned immediate for 6-bits function code.
204 @item func4 @tab Unsigned immediate for 4-bits function code.
205 @item func3 @tab Unsigned immediate for 3-bits function code.
206 @item func2 @tab Unsigned immediate for 2-bits function code.
207 @item rd @tab Destination register number for operand x, can be GPR or FPR.
208 @item rd' @tab Destination register number for operand x,
209 only accept s0-s1, a0-a5, fs0-fs1 and fa0-fa5.
210 @item rs1 @tab First source register number for operand x, can be GPR or FPR.
211 @item rs1' @tab First source register number for operand x,
212 only accept s0-s1, a0-a5, fs0-fs1 and fa0-fa5.
213 @item rs2 @tab Second source register number for operand x, can be GPR or FPR.
214 @item rs2' @tab Second source register number for operand x,
215 only accept s0-s1, a0-a5, fs0-fs1 and fa0-fa5.
216 @item simm12 @tab Sign-extended 12-bit immediate for operand x.
217 @item simm20 @tab Sign-extended 20-bit immediate for operand x.
218 @item simm6 @tab Sign-extended 6-bit immediate for operand x.
219 @item uimm8 @tab Unsigned 8-bit immediate for operand x.
220 @item symbol @tab Symbol or lable reference for operand x.
224 The following table lists all available opcode name:
230 Opcode space for compressed instructions.
233 Opcode space for load instructions.
236 Opcode space for floating-point load instructions.
239 Opcode space for store instructions.
242 Opcode space for floating-point store instructions.
245 Opcode space for auipc instruction.
248 Opcode space for lui instruction.
251 Opcode space for branch instructions.
254 Opcode space for jal instruction.
257 Opcode space for jalr instruction.
260 Opcode space for ALU instructions.
263 Opcode space for 32-bits ALU instructions.
266 Opcode space for ALU with immediate instructions.
269 Opcode space for 32-bits ALU with immediate instructions.
272 Opcode space for floating-point operation instructions.
275 Opcode space for madd instruction.
278 Opcode space for msub instruction.
281 Opcode space for nmadd instruction.
284 Opcode space for msub instruction.
287 Opcode space for atomic memory operation instructions.
290 Opcode space for misc instructions.
293 Opcode space for system instructions.
299 Opcode space for customize instructions.
303 An instruction is two or four bytes in length and must be aligned
304 on a 2 byte boundary. The first two bits of the instruction specify the
305 length of the instruction, 00, 01 and 10 indicates a two byte instruction,
306 11 indicates a four byte instruction.
308 The following table lists the RISC-V instruction formats that are available
309 with the @samp{.insn} pseudo directive:
312 @item R type: .insn r opcode, func3, func7, rd, rs1, rs2
314 +-------+-----+-----+-------+----+-------------+
315 | func7 | rs2 | rs1 | func3 | rd | opcode |
316 +-------+-----+-----+-------+----+-------------+
320 @item R type with 4 register operands: .insn r opcode, func3, func2, rd, rs1, rs2, rs3
322 +-----+-------+-----+-----+-------+----+-------------+
323 | rs3 | func2 | rs2 | rs1 | func3 | rd | opcode |
324 +-----+-------+-----+-----+-------+----+-------------+
325 31 27 25 20 15 12 7 0
328 @item I type: .insn i opcode, func3, rd, rs1, simm12
330 +-------------+-----+-------+----+-------------+
331 | simm12 | rs1 | func3 | rd | opcode |
332 +-------------+-----+-------+----+-------------+
336 @item S type: .insn s opcode, func3, rd, rs1, simm12
338 +--------------+-----+-----+-------+-------------+-------------+
339 | simm12[11:5] | rs2 | rs1 | func3 | simm12[4:0] | opcode |
340 +--------------+-----+-----+-------+-------------+-------------+
344 @item SB type: .insn sb opcode, func3, rd, rs1, symbol
345 @itemx SB type: .insn sb opcode, func3, rd, simm12(rs1)
347 +--------------+-----+-----+-------+-------------+-------------+
348 | simm21[11:5] | rs2 | rs1 | func3 | simm12[4:0] | opcode |
349 +--------------+-----+-----+-------+-------------+-------------+
353 @item U type: .insn u opcode, rd, simm20
355 +---------------------------+----+-------------+
356 | simm20 | rd | opcode |
357 +---------------------------+----+-------------+
361 @item UJ type: .insn uj opcode, rd, symbol
363 +------------+--------------+------------+---------------+----+-------------+
364 | simm20[20] | simm20[10:1] | simm20[11] | simm20[19:12] | rd | opcode |
365 +------------+--------------+------------+---------------+----+-------------+
369 @item CR type: .insn cr opcode2, func4, rd, rs2
371 +---------+--------+-----+---------+
372 | func4 | rd/rs1 | rs2 | opcode2 |
373 +---------+--------+-----+---------+
377 @item CI type: .insn ci opcode2, func3, rd, simm6
379 +---------+-----+--------+-----+---------+
380 | func3 | imm | rd/rs1 | imm | opcode2 |
381 +---------+-----+--------+-----+---------+
385 @item CIW type: .insn ciw opcode2, func3, rd, uimm8
387 +---------+--------------+-----+---------+
388 | func3 | imm | rd' | opcode2 |
389 +---------+--------------+-----+---------+
393 @item CA type: .insn ca opcode2, func6, func2, rd, rs2
395 +---------+----------+-------+------+--------+
396 | func6 | rd'/rs1' | func2 | rs2' | opcode |
397 +---------+----------+-------+------+--------+
401 @item CB type: .insn cb opcode2, func3, rs1, symbol
403 +---------+--------+------+--------+---------+
404 | func3 | offset | rs1' | offset | opcode2 |
405 +---------+--------+------+--------+---------+
409 @item CJ type: .insn cj opcode2, symbol
411 +---------+--------------------+---------+
412 | func3 | jump target | opcode2 |
413 +---------+--------------------+---------+
420 For the complete list of all instruction format variants see
421 The RISC-V Instruction Set Manual Volume I: User-Level ISA.
423 @node RISC-V-ATTRIBUTE
424 @section RISC-V Object Attribute
425 @cindex Object Attribute, RISC-V
427 RISC-V attributes have a string value if the tag number is odd and an integer
428 value if the tag number is even.
431 @item Tag_RISCV_stack_align (4)
432 Tag_RISCV_strict_align records the N-byte stack alignment for this object. The
433 default value is 16 for RV32I or RV64I, and 4 for RV32E.
435 The smallest value will be used if object files with different
436 Tag_RISCV_stack_align values are merged.
438 @item Tag_RISCV_arch (5)
439 Tag_RISCV_arch contains a string for the target architecture taken from the
440 option @option{-march}. Different architectures will be integrated into a
441 superset when object files are merged.
443 Note that the version information of the target architecture must be presented
444 explicitly in the attribute and abbreviations must be expanded. The version
445 information, if not given by @option{-march}, must be in accordance with the
446 default specified by the tool. For example, the architecture @code{RV32I} has
447 to be recorded in the attribute as @code{RV32I2P0} in which @code{2P0} stands
448 for the default version of its base ISA. On the other hand, the architecture
449 @code{RV32G} has to be presented as @code{RV32I2P0_M2P0_A2P0_F2P0_D2P0} in
450 which the abbreviation @code{G} is expanded to the @code{IMAFD} combination
451 with default versions of the standard extensions.
453 @item Tag_RISCV_unaligned_access (6)
454 Tag_RISCV_unaligned_access is 0 for files that do not allow any unaligned
455 memory accesses, and 1 for files that do allow unaligned memory accesses.
457 @item Tag_RISCV_priv_spec (8)
458 @item Tag_RISCV_priv_spec_minor (10)
459 @item Tag_RISCV_priv_spec_revision (12)
460 Tag_RISCV_priv_spec contains the major/minor/revision version information of
461 the privileged specification. It will report errors if object files of
462 different privileged specification versions are merged.