1@c Copyright (C) 2000-2016 Free Software Foundation, Inc.
2@c This is part of the GAS manual.
3@c For copying conditions, see the file as.texinfo.
4@ifset GENERIC
5@page
6@node i860-Dependent
7@chapter Intel i860 Dependent Features
8@end ifset
9@ifclear GENERIC
10@node Machine Dependencies
11@chapter Intel i860 Dependent Features
12@end ifclear
13
14@ignore
15@c FIXME: This is basically a stub for i860. There is tons more information
16that I will add later (jle@cygnus.com).
17@end ignore
18
19@cindex i860 support
20@menu
21* Notes-i860::                  i860 Notes
22* Options-i860::                i860 Command-line Options
23* Directives-i860::             i860 Machine Directives
24* Opcodes for i860::            i860 Opcodes
25* Syntax of i860::              i860 Syntax
26@end menu
27
28@node Notes-i860
29@section i860 Notes
30This is a fairly complete i860 assembler which is compatible with the
31UNIX System V/860 Release 4 assembler. However, it does not currently
32support SVR4 PIC (i.e., @code{@@GOT, @@GOTOFF, @@PLT}).
33
34Like the SVR4/860 assembler, the output object format is ELF32. Currently,
35this is the only supported object format. If there is sufficient interest,
36other formats such as COFF may be implemented.
37
38Both the Intel and AT&T/SVR4 syntaxes are supported, with the latter
39being the default.  One difference is that AT&T syntax requires the '%'
40prefix on register names while Intel syntax does not.  Another difference
41is in the specification of relocatable expressions.  The Intel syntax
42is @code{ha%expression} whereas the SVR4 syntax is @code{[expression]@@ha}
43(and similarly for the "l" and "h" selectors).
44@node Options-i860
45@section i860 Command-line Options
46@subsection SVR4 compatibility options
47@table @code
48@item -V
49Print assembler version.
50@item -Qy
51Ignored.
52@item -Qn
53Ignored.
54@end table
55@subsection Other options
56@table @code
57@item -EL
58Select little endian output (this is the default).
59@item -EB
60Select big endian output. Note that the i860 always reads instructions
61as little endian data, so this option only effects data and not
62instructions.
63@item -mwarn-expand
64Emit a warning message if any pseudo-instruction expansions occurred.
65For example, a @code{or} instruction with an immediate larger than 16-bits
66will be expanded into two instructions. This is a very undesirable feature to
67rely on, so this flag can help detect any code where it happens. One
68use of it, for instance, has been to find and eliminate any place
69where @code{gcc} may emit these pseudo-instructions.
70@item -mxp
71Enable support for the i860XP instructions and control registers.  By default,
72this option is disabled so that only the base instruction set (i.e., i860XR)
73is supported.
74@item -mintel-syntax
75The i860 assembler defaults to AT&T/SVR4 syntax.  This option enables the
76Intel syntax.
77@end table
78
79@node Directives-i860
80@section i860 Machine Directives
81
82@cindex machine directives, i860
83@cindex i860 machine directives
84
85@table @code
86@cindex @code{dual} directive, i860
87@item .dual
88Enter dual instruction mode. While this directive is supported, the
89preferred way to use dual instruction mode is to explicitly code
90the dual bit with the @code{d.} prefix.
91@end table
92
93@table @code
94@cindex @code{enddual} directive, i860
95@item .enddual
96Exit dual instruction mode. While this directive is supported, the
97preferred way to use dual instruction mode is to explicitly code
98the dual bit with the @code{d.} prefix.
99@end table
100
101@table @code
102@cindex @code{atmp} directive, i860
103@item .atmp
104Change the temporary register used when expanding pseudo operations. The
105default register is @code{r31}.
106@end table
107
108The @code{.dual}, @code{.enddual}, and @code{.atmp} directives are available only in the Intel syntax mode.
109
110Both syntaxes allow for the standard @code{.align} directive.  However,
111the Intel syntax additionally allows keywords for the alignment
112parameter: "@code{.align type}", where `type' is one of @code{.short}, @code{.long},
113@code{.quad}, @code{.single}, @code{.double} representing alignments of 2, 4,
11416, 4, and 8, respectively.
115
116@node Opcodes for i860
117@section i860 Opcodes
118
119@cindex opcodes, i860
120@cindex i860 opcodes
121All of the Intel i860XR and i860XP machine instructions are supported. Please see
122either @emph{i860 Microprocessor Programmer's Reference Manual} or @emph{i860 Microprocessor Architecture} for more information.
123@subsection Other instruction support (pseudo-instructions)
124For compatibility with some other i860 assemblers, a number of
125pseudo-instructions are supported. While these are supported, they are
126a very undesirable feature that should be avoided -- in particular, when
127they result in an expansion to multiple actual i860 instructions. Below
128are the pseudo-instructions that result in expansions.
129@itemize @bullet
130@item Load large immediate into general register:
131
132The pseudo-instruction @code{mov imm,%rn} (where the immediate does
133not fit within a signed 16-bit field) will be expanded into:
134@smallexample
135orh large_imm@@h,%r0,%rn
136or large_imm@@l,%rn,%rn
137@end smallexample
138@item Load/store with relocatable address expression:
139
140For example, the pseudo-instruction @code{ld.b addr_exp(%rx),%rn}
141will be expanded into:
142@smallexample
143orh addr_exp@@ha,%rx,%r31
144ld.l addr_exp@@l(%r31),%rn
145@end smallexample
146
147The analogous expansions apply to @code{ld.x, st.x, fld.x, pfld.x, fst.x}, and @code{pst.x} as well.
148@item Signed large immediate with add/subtract:
149
150If any of the arithmetic operations @code{adds, addu, subs, subu} are used
151with an immediate larger than 16-bits (signed), then they will be expanded.
152For instance, the pseudo-instruction @code{adds large_imm,%rx,%rn} expands to:
153@smallexample
154orh large_imm@@h,%r0,%r31
155or large_imm@@l,%r31,%r31
156adds %r31,%rx,%rn
157@end smallexample
158@item Unsigned large immediate with logical operations:
159
160Logical operations (@code{or, andnot, or, xor}) also result in expansions.
161The pseudo-instruction @code{or large_imm,%rx,%rn} results in:
162@smallexample
163orh large_imm@@h,%rx,%r31
164or large_imm@@l,%r31,%rn
165@end smallexample
166
167Similarly for the others, except for @code{and} which expands to:
168@smallexample
169andnot (-1 - large_imm)@@h,%rx,%r31
170andnot (-1 - large_imm)@@l,%r31,%rn
171@end smallexample
172@end itemize
173
174@node Syntax of i860
175@section i860 Syntax
176@menu
177* i860-Chars::                Special Characters
178@end menu
179
180@node i860-Chars
181@subsection Special Characters
182
183@cindex line comment character, i860
184@cindex i860 line comment character
185The presence of a @samp{#} appearing anywhere on a line indicates the
186start of a comment that extends to the end of that line.
187
188If a @samp{#} appears as the first character of a line then the whole
189line is treated as a comment, but in this case the line can also be a
190logical line number directive (@pxref{Comments}) or a preprocessor
191control command (@pxref{Preprocessing}).
192
193@cindex line separator, i860
194@cindex statement separator, i860
195@cindex i860 line separator
196The @samp{;} character can be used to separate statements on the same
197line.
198