1 @c Copyright 2005, 2006, 2009, 2010
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.
9 @node Blackfin-Dependent
10 @chapter Blackfin Dependent Features
14 @node Machine Dependencies
15 @chapter Blackfin Dependent Features
18 @cindex Blackfin support
20 * Blackfin Options:: Blackfin Options
21 * Blackfin Syntax:: Blackfin Syntax
22 * Blackfin Directives:: Blackfin Directives
25 @node Blackfin Options
27 @cindex Blackfin options (none)
28 @cindex options for Blackfin (none)
33 @cindex @code{-mcpu=} command line option, Blackfin
34 @item -mcpu=@var{processor}@r{[}-@var{sirevision}@r{]}
35 This option specifies the target processor. The optional @var{sirevision}
36 is not used in assembler. It's here such that GCC can easily pass down its
37 @code{-mcpu=} option. The assembler will issue an
38 error message if an attempt is made to assemble an instruction which
39 will not execute on the target processor. The following processor names are
57 @code{bf535} (not implemented yet),
76 @cindex @code{-mfdpic} command line option, Blackfin
78 Assemble for the FDPIC ABI.
80 @cindex @code{-mno-fdpic} command line option, Blackfin
81 @cindex @code{-mnopic} command line option, Blackfin
90 @cindex Blackfin syntax
91 @cindex syntax, Blackfin
94 @item Special Characters
95 Assembler input is free format and may appear anywhere on the line.
96 One instruction may extend across multiple lines or more than one
97 instruction may appear on the same line. White space (space, tab,
98 comments or newline) may appear anywhere between tokens. A token must
99 not have embedded spaces. Tokens include numbers, register names,
100 keywords, user identifiers, and also some multicharacter special
101 symbols like "+=", "/*" or "||".
103 @item Instruction Delimiting
104 A semicolon must terminate every instruction. Sometimes a complete
105 instruction will consist of more than one operation. There are two
106 cases where this occurs. The first is when two general operations
107 are combined. Normally a comma separates the different parts, as in
110 a0= r3.h * r2.l, a1 = r3.l * r2.h ;
113 The second case occurs when a general instruction is combined with one
114 or two memory references for joint issue. The latter portions are
115 set off by a "||" token.
118 a0 = r3.h * r2.l || r1 = [p3++] || r4 = [i2++];
123 The assembler treats register names and instruction keywords in a case
124 insensitive manner. User identifiers are case sensitive. Thus, R3.l,
125 R3.L, r3.l and r3.L are all equivalent input to the assembler.
127 Register names are reserved and may not be used as program identifiers.
129 Some operations (such as "Move Register") require a register pair.
130 Register pairs are always data registers and are denoted using a colon,
131 eg., R3:2. The larger number must be written firsts. Note that the
132 hardware only supports odd-even pairs, eg., R7:6, R5:4, R3:2, and R1:0.
134 Some instructions (such as --SP (Push Multiple)) require a group of
135 adjacent registers. Adjacent registers are denoted in the syntax by
136 the range enclosed in parentheses and separated by a colon, eg., (R7:3).
137 Again, the larger number appears first.
139 Portions of a particular register may be individually specified. This
140 is written with a dot (".") following the register name and then a
141 letter denoting the desired portion. For 32-bit registers, ".H"
142 denotes the most significant ("High") portion. ".L" denotes the
143 least-significant portion. The subdivisions of the 40-bit registers
147 The set of 40-bit registers A1 and A0 that normally contain data that
148 is being manipulated. Each accumulator can be accessed in four ways.
151 @item one 40-bit register
152 The register will be referred to as A1 or A0.
153 @item one 32-bit register
154 The registers are designated as A1.W or A0.W.
155 @item two 16-bit registers
156 The registers are designated as A1.H, A1.L, A0.H or A0.L.
157 @item one 8-bit register
158 The registers are designated as A1.X or A0.X for the bits that
159 extend beyond bit 31.
163 The set of 32-bit registers (R0, R1, R2, R3, R4, R5, R6 and R7) that
164 normally contain data for manipulation. These are abbreviated as
165 D-register or Dreg. Data registers can be accessed as 32-bit registers
166 or as two independent 16-bit registers. The least significant 16 bits
167 of each register is called the "low" half and is designated with ".L"
168 following the register name. The most significant 16 bits are called
169 the "high" half and is designated with ".H" following the name.
172 R7.L, r2.h, r4.L, R0.H
175 @item Pointer Registers
176 The set of 32-bit registers (P0, P1, P2, P3, P4, P5, SP and FP) that
177 normally contain byte addresses of data structures. These are
178 abbreviated as P-register or Preg.
184 @item Stack Pointer SP
185 The stack pointer contains the 32-bit address of the last occupied
186 byte location in the stack. The stack grows by decrementing the
189 @item Frame Pointer FP
190 The frame pointer contains the 32-bit address of the previous frame
191 pointer in the stack. It is located at the top of a frame.
194 LT0 and LT1. These registers contain the 32-bit address of the top of
195 a zero overhead loop.
198 LC0 and LC1. These registers contain the 32-bit counter of the zero
199 overhead loop executions.
202 LB0 and LB1. These registers contain the 32-bit address of the bottom
203 of a zero overhead loop.
205 @item Index Registers
206 The set of 32-bit registers (I0, I1, I2, I3) that normally contain byte
207 addresses of data structures. Abbreviated I-register or Ireg.
209 @item Modify Registers
210 The set of 32-bit registers (M0, M1, M2, M3) that normally contain
211 offset values that are added and subracted to one of the index
212 registers. Abbreviated as Mreg.
214 @item Length Registers
215 The set of 32-bit registers (L0, L1, L2, L3) that normally contain the
216 length in bytes of the circular buffer. Abbreviated as Lreg. Clear
217 the Lreg to disable circular addressing for the corresponding Ireg.
220 The set of 32-bit registers (B0, B1, B2, B3) that normally contain the
221 base address in bytes of the circular buffer. Abbreviated as Breg.
224 The Blackfin family has no hardware floating point but the .float
225 directive generates ieee floating point numbers for use with software
226 floating point libraries.
228 @item Blackfin Opcodes
229 For detailed information on the Blackfin machine instruction set, see
230 the Blackfin(r) Processor Instruction Set Reference.
234 @node Blackfin Directives
236 @cindex Blackfin directives
237 @cindex directives, Blackfin
239 The following directives are provided for compatibility with the VDSP assembler.
243 Initializes a four byte data object.
245 Initializes a two byte data object.
253 Define and initialize a 32 bit data object.