You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

603 lines
33 KiB
HTML

<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html4/loose.dtd">
<html>
<!-- Copyright (C) 1988-2018 Free Software Foundation, Inc.
Permission is granted to copy, distribute and/or modify this document
under the terms of the GNU Free Documentation License, Version 1.3 or
any later version published by the Free Software Foundation; with the
Invariant Sections being "Funding Free Software", the Front-Cover
Texts being (a) (see below), and with the Back-Cover Texts being (b)
(see below). A copy of the license is included in the section entitled
"GNU Free Documentation License".
(a) The FSF's Front-Cover Text is:
A GNU Manual
(b) The FSF's Back-Cover Text is:
You have freedom to copy and modify this GNU Manual, like GNU
software. Copies published by the Free Software Foundation raise
funds for GNU development. -->
<!-- Created by GNU Texinfo 6.4, http://www.gnu.org/software/texinfo/ -->
<head>
<title>Register Arguments (GNU Compiler Collection (GCC) Internals)</title>
<meta name="description" content="Register Arguments (GNU Compiler Collection (GCC) Internals)">
<meta name="keywords" content="Register Arguments (GNU Compiler Collection (GCC) Internals)">
<meta name="resource-type" content="document">
<meta name="distribution" content="global">
<meta name="Generator" content="makeinfo">
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<link href="index.html#Top" rel="start" title="Top">
<link href="Option-Index.html#Option-Index" rel="index" title="Option Index">
<link href="index.html#SEC_Contents" rel="contents" title="Table of Contents">
<link href="Stack-and-Calling.html#Stack-and-Calling" rel="up" title="Stack and Calling">
<link href="Scalar-Return.html#Scalar-Return" rel="next" title="Scalar Return">
<link href="Stack-Arguments.html#Stack-Arguments" rel="prev" title="Stack Arguments">
<style type="text/css">
<!--
a.summary-letter {text-decoration: none}
blockquote.indentedblock {margin-right: 0em}
blockquote.smallindentedblock {margin-right: 0em; font-size: smaller}
blockquote.smallquotation {font-size: smaller}
div.display {margin-left: 3.2em}
div.example {margin-left: 3.2em}
div.lisp {margin-left: 3.2em}
div.smalldisplay {margin-left: 3.2em}
div.smallexample {margin-left: 3.2em}
div.smalllisp {margin-left: 3.2em}
kbd {font-style: oblique}
pre.display {font-family: inherit}
pre.format {font-family: inherit}
pre.menu-comment {font-family: serif}
pre.menu-preformatted {font-family: serif}
pre.smalldisplay {font-family: inherit; font-size: smaller}
pre.smallexample {font-size: smaller}
pre.smallformat {font-family: inherit; font-size: smaller}
pre.smalllisp {font-size: smaller}
span.nolinebreak {white-space: nowrap}
span.roman {font-family: initial; font-weight: normal}
span.sansserif {font-family: sans-serif; font-weight: normal}
ul.no-bullet {list-style: none}
-->
</style>
</head>
<body lang="en">
<a name="Register-Arguments"></a>
<div class="header">
<p>
Next: <a href="Scalar-Return.html#Scalar-Return" accesskey="n" rel="next">Scalar Return</a>, Previous: <a href="Stack-Arguments.html#Stack-Arguments" accesskey="p" rel="prev">Stack Arguments</a>, Up: <a href="Stack-and-Calling.html#Stack-and-Calling" accesskey="u" rel="up">Stack and Calling</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Option-Index.html#Option-Index" title="Index" rel="index">Index</a>]</p>
</div>
<hr>
<a name="Passing-Arguments-in-Registers"></a>
<h4 class="subsection">18.9.7 Passing Arguments in Registers</h4>
<a name="index-arguments-in-registers"></a>
<a name="index-registers-arguments"></a>
<p>This section describes the macros which let you control how various
types of arguments are passed in registers or how they are arranged in
the stack.
</p>
<dl>
<dt><a name="index-TARGET_005fFUNCTION_005fARG"></a>Target Hook: <em>rtx</em> <strong>TARGET_FUNCTION_ARG</strong> <em>(cumulative_args_t <var>ca</var>, machine_mode <var>mode</var>, const_tree <var>type</var>, bool <var>named</var>)</em></dt>
<dd><p>Return an RTX indicating whether a function argument is passed in a
register and if so, which register.
</p>
<p>The arguments are <var>ca</var>, which summarizes all the previous
arguments; <var>mode</var>, the machine mode of the argument; <var>type</var>,
the data type of the argument as a tree node or 0 if that is not known
(which happens for C support library functions); and <var>named</var>,
which is <code>true</code> for an ordinary argument and <code>false</code> for
nameless arguments that correspond to &lsquo;<samp>&hellip;</samp>&rsquo; in the called
function&rsquo;s prototype. <var>type</var> can be an incomplete type if a
syntax error has previously occurred.
</p>
<p>The return value is usually either a <code>reg</code> RTX for the hard
register in which to pass the argument, or zero to pass the argument
on the stack.
</p>
<p>The return value can be a <code>const_int</code> which means argument is
passed in a target specific slot with specified number. Target hooks
should be used to store or load argument in such case. See
<code>TARGET_STORE_BOUNDS_FOR_ARG</code> and <code>TARGET_LOAD_BOUNDS_FOR_ARG</code>
for more information.
</p>
<p>The value of the expression can also be a <code>parallel</code> RTX. This is
used when an argument is passed in multiple locations. The mode of the
<code>parallel</code> should be the mode of the entire argument. The
<code>parallel</code> holds any number of <code>expr_list</code> pairs; each one
describes where part of the argument is passed. In each
<code>expr_list</code> the first operand must be a <code>reg</code> RTX for the hard
register in which to pass this part of the argument, and the mode of the
register RTX indicates how large this part of the argument is. The
second operand of the <code>expr_list</code> is a <code>const_int</code> which gives
the offset in bytes into the entire argument of where this part starts.
As a special exception the first <code>expr_list</code> in the <code>parallel</code>
RTX may have a first operand of zero. This indicates that the entire
argument is also stored on the stack.
</p>
<p>The last time this hook is called, it is called with <code>MODE ==
VOIDmode</code>, and its result is passed to the <code>call</code> or <code>call_value</code>
pattern as operands 2 and 3 respectively.
</p>
<a name="index-stdarg_002eh-and-register-arguments"></a>
<p>The usual way to make the ISO library <samp>stdarg.h</samp> work on a
machine where some arguments are usually passed in registers, is to
cause nameless arguments to be passed on the stack instead. This is
done by making <code>TARGET_FUNCTION_ARG</code> return 0 whenever
<var>named</var> is <code>false</code>.
</p>
<a name="index-TARGET_005fMUST_005fPASS_005fIN_005fSTACK_002c-and-TARGET_005fFUNCTION_005fARG"></a>
<a name="index-REG_005fPARM_005fSTACK_005fSPACE_002c-and-TARGET_005fFUNCTION_005fARG"></a>
<p>You may use the hook <code>targetm.calls.must_pass_in_stack</code>
in the definition of this macro to determine if this argument is of a
type that must be passed in the stack. If <code>REG_PARM_STACK_SPACE</code>
is not defined and <code>TARGET_FUNCTION_ARG</code> returns nonzero for such an
argument, the compiler will abort. If <code>REG_PARM_STACK_SPACE</code> is
defined, the argument will be computed in the stack and then loaded into
a register.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fMUST_005fPASS_005fIN_005fSTACK"></a>Target Hook: <em>bool</em> <strong>TARGET_MUST_PASS_IN_STACK</strong> <em>(machine_mode <var>mode</var>, const_tree <var>type</var>)</em></dt>
<dd><p>This target hook should return <code>true</code> if we should not pass <var>type</var>
solely in registers. The file <samp>expr.h</samp> defines a
definition that is usually appropriate, refer to <samp>expr.h</samp> for additional
documentation.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fFUNCTION_005fINCOMING_005fARG"></a>Target Hook: <em>rtx</em> <strong>TARGET_FUNCTION_INCOMING_ARG</strong> <em>(cumulative_args_t <var>ca</var>, machine_mode <var>mode</var>, const_tree <var>type</var>, bool <var>named</var>)</em></dt>
<dd><p>Define this hook if the caller and callee on the target have different
views of where arguments are passed. Also define this hook if there are
functions that are never directly called, but are invoked by the hardware
and which have nonstandard calling conventions.
</p>
<p>In this case <code>TARGET_FUNCTION_ARG</code> computes the register in
which the caller passes the value, and
<code>TARGET_FUNCTION_INCOMING_ARG</code> should be defined in a similar
fashion to tell the function being called where the arguments will
arrive.
</p>
<p><code>TARGET_FUNCTION_INCOMING_ARG</code> can also return arbitrary address
computation using hard register, which can be forced into a register,
so that it can be used to pass special arguments.
</p>
<p>If <code>TARGET_FUNCTION_INCOMING_ARG</code> is not defined,
<code>TARGET_FUNCTION_ARG</code> serves both purposes.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fUSE_005fPSEUDO_005fPIC_005fREG"></a>Target Hook: <em>bool</em> <strong>TARGET_USE_PSEUDO_PIC_REG</strong> <em>(void)</em></dt>
<dd><p>This hook should return 1 in case pseudo register should be created
for pic_offset_table_rtx during function expand.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fINIT_005fPIC_005fREG"></a>Target Hook: <em>void</em> <strong>TARGET_INIT_PIC_REG</strong> <em>(void)</em></dt>
<dd><p>Perform a target dependent initialization of pic_offset_table_rtx.
This hook is called at the start of register allocation.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fARG_005fPARTIAL_005fBYTES"></a>Target Hook: <em>int</em> <strong>TARGET_ARG_PARTIAL_BYTES</strong> <em>(cumulative_args_t <var>cum</var>, machine_mode <var>mode</var>, tree <var>type</var>, bool <var>named</var>)</em></dt>
<dd><p>This target hook returns the number of bytes at the beginning of an
argument that must be put in registers. The value must be zero for
arguments that are passed entirely in registers or that are entirely
pushed on the stack.
</p>
<p>On some machines, certain arguments must be passed partially in
registers and partially in memory. On these machines, typically the
first few words of arguments are passed in registers, and the rest
on the stack. If a multi-word argument (a <code>double</code> or a
structure) crosses that boundary, its first few words must be passed
in registers and the rest must be pushed. This macro tells the
compiler when this occurs, and how many bytes should go in registers.
</p>
<p><code>TARGET_FUNCTION_ARG</code> for these arguments should return the first
register to be used by the caller for this argument; likewise
<code>TARGET_FUNCTION_INCOMING_ARG</code>, for the called function.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fPASS_005fBY_005fREFERENCE"></a>Target Hook: <em>bool</em> <strong>TARGET_PASS_BY_REFERENCE</strong> <em>(cumulative_args_t <var>cum</var>, machine_mode <var>mode</var>, const_tree <var>type</var>, bool <var>named</var>)</em></dt>
<dd><p>This target hook should return <code>true</code> if an argument at the
position indicated by <var>cum</var> should be passed by reference. This
predicate is queried after target independent reasons for being
passed by reference, such as <code>TREE_ADDRESSABLE (type)</code>.
</p>
<p>If the hook returns true, a copy of that argument is made in memory and a
pointer to the argument is passed instead of the argument itself.
The pointer is passed in whatever way is appropriate for passing a pointer
to that type.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fCALLEE_005fCOPIES"></a>Target Hook: <em>bool</em> <strong>TARGET_CALLEE_COPIES</strong> <em>(cumulative_args_t <var>cum</var>, machine_mode <var>mode</var>, const_tree <var>type</var>, bool <var>named</var>)</em></dt>
<dd><p>The function argument described by the parameters to this hook is
known to be passed by reference. The hook should return true if the
function argument should be copied by the callee instead of copied
by the caller.
</p>
<p>For any argument for which the hook returns true, if it can be
determined that the argument is not modified, then a copy need
not be generated.
</p>
<p>The default version of this hook always returns false.
</p></dd></dl>
<dl>
<dt><a name="index-CUMULATIVE_005fARGS"></a>Macro: <strong>CUMULATIVE_ARGS</strong></dt>
<dd><p>A C type for declaring a variable that is used as the first argument
of <code>TARGET_FUNCTION_ARG</code> and other related values. For some
target machines, the type <code>int</code> suffices and can hold the number
of bytes of argument so far.
</p>
<p>There is no need to record in <code>CUMULATIVE_ARGS</code> anything about the
arguments that have been passed on the stack. The compiler has other
variables to keep track of that. For target machines on which all
arguments are passed on the stack, there is no need to store anything in
<code>CUMULATIVE_ARGS</code>; however, the data structure must exist and
should not be empty, so use <code>int</code>.
</p></dd></dl>
<dl>
<dt><a name="index-OVERRIDE_005fABI_005fFORMAT"></a>Macro: <strong>OVERRIDE_ABI_FORMAT</strong> <em>(<var>fndecl</var>)</em></dt>
<dd><p>If defined, this macro is called before generating any code for a
function, but after the <var>cfun</var> descriptor for the function has been
created. The back end may use this macro to update <var>cfun</var> to
reflect an ABI other than that which would normally be used by default.
If the compiler is generating code for a compiler-generated function,
<var>fndecl</var> may be <code>NULL</code>.
</p></dd></dl>
<dl>
<dt><a name="index-INIT_005fCUMULATIVE_005fARGS"></a>Macro: <strong>INIT_CUMULATIVE_ARGS</strong> <em>(<var>cum</var>, <var>fntype</var>, <var>libname</var>, <var>fndecl</var>, <var>n_named_args</var>)</em></dt>
<dd><p>A C statement (sans semicolon) for initializing the variable
<var>cum</var> for the state at the beginning of the argument list. The
variable has type <code>CUMULATIVE_ARGS</code>. The value of <var>fntype</var>
is the tree node for the data type of the function which will receive
the args, or 0 if the args are to a compiler support library function.
For direct calls that are not libcalls, <var>fndecl</var> contain the
declaration node of the function. <var>fndecl</var> is also set when
<code>INIT_CUMULATIVE_ARGS</code> is used to find arguments for the function
being compiled. <var>n_named_args</var> is set to the number of named
arguments, including a structure return address if it is passed as a
parameter, when making a call. When processing incoming arguments,
<var>n_named_args</var> is set to -1.
</p>
<p>When processing a call to a compiler support library function,
<var>libname</var> identifies which one. It is a <code>symbol_ref</code> rtx which
contains the name of the function, as a string. <var>libname</var> is 0 when
an ordinary C function call is being processed. Thus, each time this
macro is called, either <var>libname</var> or <var>fntype</var> is nonzero, but
never both of them at once.
</p></dd></dl>
<dl>
<dt><a name="index-INIT_005fCUMULATIVE_005fLIBCALL_005fARGS"></a>Macro: <strong>INIT_CUMULATIVE_LIBCALL_ARGS</strong> <em>(<var>cum</var>, <var>mode</var>, <var>libname</var>)</em></dt>
<dd><p>Like <code>INIT_CUMULATIVE_ARGS</code> but only used for outgoing libcalls,
it gets a <code>MODE</code> argument instead of <var>fntype</var>, that would be
<code>NULL</code>. <var>indirect</var> would always be zero, too. If this macro
is not defined, <code>INIT_CUMULATIVE_ARGS (cum, NULL_RTX, libname,
0)</code> is used instead.
</p></dd></dl>
<dl>
<dt><a name="index-INIT_005fCUMULATIVE_005fINCOMING_005fARGS"></a>Macro: <strong>INIT_CUMULATIVE_INCOMING_ARGS</strong> <em>(<var>cum</var>, <var>fntype</var>, <var>libname</var>)</em></dt>
<dd><p>Like <code>INIT_CUMULATIVE_ARGS</code> but overrides it for the purposes of
finding the arguments for the function being compiled. If this macro is
undefined, <code>INIT_CUMULATIVE_ARGS</code> is used instead.
</p>
<p>The value passed for <var>libname</var> is always 0, since library routines
with special calling conventions are never compiled with GCC. The
argument <var>libname</var> exists for symmetry with
<code>INIT_CUMULATIVE_ARGS</code>.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fFUNCTION_005fARG_005fADVANCE"></a>Target Hook: <em>void</em> <strong>TARGET_FUNCTION_ARG_ADVANCE</strong> <em>(cumulative_args_t <var>ca</var>, machine_mode <var>mode</var>, const_tree <var>type</var>, bool <var>named</var>)</em></dt>
<dd><p>This hook updates the summarizer variable pointed to by <var>ca</var> to
advance past an argument in the argument list. The values <var>mode</var>,
<var>type</var> and <var>named</var> describe that argument. Once this is done,
the variable <var>cum</var> is suitable for analyzing the <em>following</em>
argument with <code>TARGET_FUNCTION_ARG</code>, etc.
</p>
<p>This hook need not do anything if the argument in question was passed
on the stack. The compiler knows how to track the amount of stack space
used for arguments without any special help.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fFUNCTION_005fARG_005fOFFSET"></a>Target Hook: <em>HOST_WIDE_INT</em> <strong>TARGET_FUNCTION_ARG_OFFSET</strong> <em>(machine_mode <var>mode</var>, const_tree <var>type</var>)</em></dt>
<dd><p>This hook returns the number of bytes to add to the offset of an
argument of type <var>type</var> and mode <var>mode</var> when passed in memory.
This is needed for the SPU, which passes <code>char</code> and <code>short</code>
arguments in the preferred slot that is in the middle of the quad word
instead of starting at the top. The default implementation returns 0.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fFUNCTION_005fARG_005fPADDING"></a>Target Hook: <em>pad_direction</em> <strong>TARGET_FUNCTION_ARG_PADDING</strong> <em>(machine_mode <var>mode</var>, const_tree <var>type</var>)</em></dt>
<dd><p>This hook determines whether, and in which direction, to pad out
an argument of mode <var>mode</var> and type <var>type</var>. It returns
<code>PAD_UPWARD</code> to insert padding above the argument, <code>PAD_DOWNWARD</code>
to insert padding below the argument, or <code>PAD_NONE</code> to inhibit padding.
</p>
<p>The <em>amount</em> of padding is not controlled by this hook, but by
<code>TARGET_FUNCTION_ARG_ROUND_BOUNDARY</code>. It is always just enough
to reach the next multiple of that boundary.
</p>
<p>This hook has a default definition that is right for most systems.
For little-endian machines, the default is to pad upward. For
big-endian machines, the default is to pad downward for an argument of
constant size shorter than an <code>int</code>, and upward otherwise.
</p></dd></dl>
<dl>
<dt><a name="index-PAD_005fVARARGS_005fDOWN"></a>Macro: <strong>PAD_VARARGS_DOWN</strong></dt>
<dd><p>If defined, a C expression which determines whether the default
implementation of va_arg will attempt to pad down before reading the
next argument, if that argument is smaller than its aligned space as
controlled by <code>PARM_BOUNDARY</code>. If this macro is not defined, all such
arguments are padded down if <code>BYTES_BIG_ENDIAN</code> is true.
</p></dd></dl>
<dl>
<dt><a name="index-BLOCK_005fREG_005fPADDING"></a>Macro: <strong>BLOCK_REG_PADDING</strong> <em>(<var>mode</var>, <var>type</var>, <var>first</var>)</em></dt>
<dd><p>Specify padding for the last element of a block move between registers and
memory. <var>first</var> is nonzero if this is the only element. Defining this
macro allows better control of register function parameters on big-endian
machines, without using <code>PARALLEL</code> rtl. In particular,
<code>MUST_PASS_IN_STACK</code> need not test padding and mode of types in
registers, as there is no longer a &quot;wrong&quot; part of a register; For example,
a three byte aggregate may be passed in the high part of a register if so
required.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fFUNCTION_005fARG_005fBOUNDARY"></a>Target Hook: <em>unsigned int</em> <strong>TARGET_FUNCTION_ARG_BOUNDARY</strong> <em>(machine_mode <var>mode</var>, const_tree <var>type</var>)</em></dt>
<dd><p>This hook returns the alignment boundary, in bits, of an argument
with the specified mode and type. The default hook returns
<code>PARM_BOUNDARY</code> for all arguments.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fFUNCTION_005fARG_005fROUND_005fBOUNDARY"></a>Target Hook: <em>unsigned int</em> <strong>TARGET_FUNCTION_ARG_ROUND_BOUNDARY</strong> <em>(machine_mode <var>mode</var>, const_tree <var>type</var>)</em></dt>
<dd><p>Normally, the size of an argument is rounded up to <code>PARM_BOUNDARY</code>,
which is the default value for this hook. You can define this hook to
return a different value if an argument size must be rounded to a larger
value.
</p></dd></dl>
<dl>
<dt><a name="index-FUNCTION_005fARG_005fREGNO_005fP"></a>Macro: <strong>FUNCTION_ARG_REGNO_P</strong> <em>(<var>regno</var>)</em></dt>
<dd><p>A C expression that is nonzero if <var>regno</var> is the number of a hard
register in which function arguments are sometimes passed. This does
<em>not</em> include implicit arguments such as the static chain and
the structure-value address. On many machines, no registers can be
used for this purpose since all function arguments are pushed on the
stack.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fSPLIT_005fCOMPLEX_005fARG"></a>Target Hook: <em>bool</em> <strong>TARGET_SPLIT_COMPLEX_ARG</strong> <em>(const_tree <var>type</var>)</em></dt>
<dd><p>This hook should return true if parameter of type <var>type</var> are passed
as two scalar parameters. By default, GCC will attempt to pack complex
arguments into the target&rsquo;s word size. Some ABIs require complex arguments
to be split and treated as their individual components. For example, on
AIX64, complex floats should be passed in a pair of floating point
registers, even though a complex float would fit in one 64-bit floating
point register.
</p>
<p>The default value of this hook is <code>NULL</code>, which is treated as always
false.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fBUILD_005fBUILTIN_005fVA_005fLIST"></a>Target Hook: <em>tree</em> <strong>TARGET_BUILD_BUILTIN_VA_LIST</strong> <em>(void)</em></dt>
<dd><p>This hook returns a type node for <code>va_list</code> for the target.
The default version of the hook returns <code>void*</code>.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fENUM_005fVA_005fLIST_005fP"></a>Target Hook: <em>int</em> <strong>TARGET_ENUM_VA_LIST_P</strong> <em>(int <var>idx</var>, const char **<var>pname</var>, tree *<var>ptree</var>)</em></dt>
<dd><p>This target hook is used in function <code>c_common_nodes_and_builtins</code>
to iterate through the target specific builtin types for va_list. The
variable <var>idx</var> is used as iterator. <var>pname</var> has to be a pointer
to a <code>const char *</code> and <var>ptree</var> a pointer to a <code>tree</code> typed
variable.
The arguments <var>pname</var> and <var>ptree</var> are used to store the result of
this macro and are set to the name of the va_list builtin type and its
internal type.
If the return value of this macro is zero, then there is no more element.
Otherwise the <var>IDX</var> should be increased for the next call of this
macro to iterate through all types.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fFN_005fABI_005fVA_005fLIST"></a>Target Hook: <em>tree</em> <strong>TARGET_FN_ABI_VA_LIST</strong> <em>(tree <var>fndecl</var>)</em></dt>
<dd><p>This hook returns the va_list type of the calling convention specified by
<var>fndecl</var>.
The default version of this hook returns <code>va_list_type_node</code>.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fCANONICAL_005fVA_005fLIST_005fTYPE"></a>Target Hook: <em>tree</em> <strong>TARGET_CANONICAL_VA_LIST_TYPE</strong> <em>(tree <var>type</var>)</em></dt>
<dd><p>This hook returns the va_list type of the calling convention specified by the
type of <var>type</var>. If <var>type</var> is not a valid va_list type, it returns
<code>NULL_TREE</code>.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fGIMPLIFY_005fVA_005fARG_005fEXPR"></a>Target Hook: <em>tree</em> <strong>TARGET_GIMPLIFY_VA_ARG_EXPR</strong> <em>(tree <var>valist</var>, tree <var>type</var>, gimple_seq *<var>pre_p</var>, gimple_seq *<var>post_p</var>)</em></dt>
<dd><p>This hook performs target-specific gimplification of
<code>VA_ARG_EXPR</code>. The first two parameters correspond to the
arguments to <code>va_arg</code>; the latter two are as in
<code>gimplify.c:gimplify_expr</code>.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fVALID_005fPOINTER_005fMODE"></a>Target Hook: <em>bool</em> <strong>TARGET_VALID_POINTER_MODE</strong> <em>(scalar_int_mode <var>mode</var>)</em></dt>
<dd><p>Define this to return nonzero if the port can handle pointers
with machine mode <var>mode</var>. The default version of this
hook returns true for both <code>ptr_mode</code> and <code>Pmode</code>.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fREF_005fMAY_005fALIAS_005fERRNO"></a>Target Hook: <em>bool</em> <strong>TARGET_REF_MAY_ALIAS_ERRNO</strong> <em>(struct ao_ref *<var>ref</var>)</em></dt>
<dd><p>Define this to return nonzero if the memory reference <var>ref</var> may alias with the system C library errno location. The default version of this hook assumes the system C library errno location is either a declaration of type int or accessed by dereferencing a pointer to int.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fSCALAR_005fMODE_005fSUPPORTED_005fP"></a>Target Hook: <em>bool</em> <strong>TARGET_SCALAR_MODE_SUPPORTED_P</strong> <em>(scalar_mode <var>mode</var>)</em></dt>
<dd><p>Define this to return nonzero if the port is prepared to handle
insns involving scalar mode <var>mode</var>. For a scalar mode to be
considered supported, all the basic arithmetic and comparisons
must work.
</p>
<p>The default version of this hook returns true for any mode
required to handle the basic C types (as defined by the port).
Included here are the double-word arithmetic supported by the
code in <samp>optabs.c</samp>.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fVECTOR_005fMODE_005fSUPPORTED_005fP"></a>Target Hook: <em>bool</em> <strong>TARGET_VECTOR_MODE_SUPPORTED_P</strong> <em>(machine_mode <var>mode</var>)</em></dt>
<dd><p>Define this to return nonzero if the port is prepared to handle
insns involving vector mode <var>mode</var>. At the very least, it
must have move patterns for this mode.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fARRAY_005fMODE"></a>Target Hook: <em>opt_machine_mode</em> <strong>TARGET_ARRAY_MODE</strong> <em>(machine_mode <var>mode</var>, unsigned HOST_WIDE_INT <var>nelems</var>)</em></dt>
<dd><p>Return the mode that GCC should use for an array that has
<var>nelems</var> elements, with each element having mode <var>mode</var>.
Return no mode if the target has no special requirements. In the
latter case, GCC looks for an integer mode of the appropriate size
if available and uses BLKmode otherwise. Usually the search for the
integer mode is limited to <code>MAX_FIXED_MODE_SIZE</code>, but the
<code>TARGET_ARRAY_MODE_SUPPORTED_P</code> hook allows a larger mode to be
used in specific cases.
</p>
<p>The main use of this hook is to specify that an array of vectors should
also have a vector mode. The default implementation returns no mode.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fARRAY_005fMODE_005fSUPPORTED_005fP"></a>Target Hook: <em>bool</em> <strong>TARGET_ARRAY_MODE_SUPPORTED_P</strong> <em>(machine_mode <var>mode</var>, unsigned HOST_WIDE_INT <var>nelems</var>)</em></dt>
<dd><p>Return true if GCC should try to use a scalar mode to store an array
of <var>nelems</var> elements, given that each element has mode <var>mode</var>.
Returning true here overrides the usual <code>MAX_FIXED_MODE</code> limit
and allows GCC to use any defined integer mode.
</p>
<p>One use of this hook is to support vector load and store operations
that operate on several homogeneous vectors. For example, ARM NEON
has operations like:
</p>
<div class="smallexample">
<pre class="smallexample">int8x8x3_t vld3_s8 (const int8_t *)
</pre></div>
<p>where the return type is defined as:
</p>
<div class="smallexample">
<pre class="smallexample">typedef struct int8x8x3_t
{
int8x8_t val[3];
} int8x8x3_t;
</pre></div>
<p>If this hook allows <code>val</code> to have a scalar mode, then
<code>int8x8x3_t</code> can have the same mode. GCC can then store
<code>int8x8x3_t</code>s in registers rather than forcing them onto the stack.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fLIBGCC_005fFLOATING_005fMODE_005fSUPPORTED_005fP"></a>Target Hook: <em>bool</em> <strong>TARGET_LIBGCC_FLOATING_MODE_SUPPORTED_P</strong> <em>(scalar_float_mode <var>mode</var>)</em></dt>
<dd><p>Define this to return nonzero if libgcc provides support for the
floating-point mode <var>mode</var>, which is known to pass
<code>TARGET_SCALAR_MODE_SUPPORTED_P</code>. The default version of this
hook returns true for all of <code>SFmode</code>, <code>DFmode</code>,
<code>XFmode</code> and <code>TFmode</code>, if such modes exist.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fFLOATN_005fMODE"></a>Target Hook: <em>opt_scalar_float_mode</em> <strong>TARGET_FLOATN_MODE</strong> <em>(int <var>n</var>, bool <var>extended</var>)</em></dt>
<dd><p>Define this to return the machine mode to use for the type
<code>_Float<var>n</var></code>, if <var>extended</var> is false, or the type
<code>_Float<var>n</var>x</code>, if <var>extended</var> is true. If such a type is not
supported, return <code>opt_scalar_float_mode ()</code>. The default version of
this hook returns <code>SFmode</code> for <code>_Float32</code>, <code>DFmode</code> for
<code>_Float64</code> and <code>_Float32x</code> and <code>TFmode</code> for
<code>_Float128</code>, if those modes exist and satisfy the requirements for
those types and pass <code>TARGET_SCALAR_MODE_SUPPORTED_P</code> and
<code>TARGET_LIBGCC_FLOATING_MODE_SUPPORTED_P</code>; for <code>_Float64x</code>, it
returns the first of <code>XFmode</code> and <code>TFmode</code> that exists and
satisfies the same requirements; for other types, it returns
<code>opt_scalar_float_mode ()</code>. The hook is only called for values
of <var>n</var> and <var>extended</var> that are valid according to
ISO/IEC TS 18661-3:2015; that is, <var>n</var> is one of 32, 64, 128, or,
if <var>extended</var> is false, 16 or greater than 128 and a multiple of 32.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fFLOATN_005fBUILTIN_005fP"></a>Target Hook: <em>bool</em> <strong>TARGET_FLOATN_BUILTIN_P</strong> <em>(int <var>func</var>)</em></dt>
<dd><p>Define this to return true if the <code>_Float<var>n</var></code> and
<code>_Float<var>n</var>x</code> built-in functions should implicitly enable the
built-in function without the <code>__builtin_</code> prefix in addition to the
normal built-in function with the <code>__builtin_</code> prefix. The default is
to only enable built-in functions without the <code>__builtin_</code> prefix for
the GNU C langauge. In strict ANSI/ISO mode, the built-in function without
the <code>__builtin_</code> prefix is not enabled. The argument <code>FUNC</code> is the
<code>enum built_in_function</code> id of the function to be enabled.
</p></dd></dl>
<dl>
<dt><a name="index-TARGET_005fSMALL_005fREGISTER_005fCLASSES_005fFOR_005fMODE_005fP"></a>Target Hook: <em>bool</em> <strong>TARGET_SMALL_REGISTER_CLASSES_FOR_MODE_P</strong> <em>(machine_mode <var>mode</var>)</em></dt>
<dd><p>Define this to return nonzero for machine modes for which the port has
small register classes. If this target hook returns nonzero for a given
<var>mode</var>, the compiler will try to minimize the lifetime of registers
in <var>mode</var>. The hook may be called with <code>VOIDmode</code> as argument.
In this case, the hook is expected to return nonzero if it returns nonzero
for any mode.
</p>
<p>On some machines, it is risky to let hard registers live across arbitrary
insns. Typically, these machines have instructions that require values
to be in specific registers (like an accumulator), and reload will fail
if the required hard register is used for another purpose across such an
insn.
</p>
<p>Passes before reload do not know which hard registers will be used
in an instruction, but the machine modes of the registers set or used in
the instruction are already known. And for some machines, register
classes are small for, say, integer registers but not for floating point
registers. For example, the AMD x86-64 architecture requires specific
registers for the legacy x86 integer instructions, but there are many
SSE registers for floating point operations. On such targets, a good
strategy may be to return nonzero from this hook for <code>INTEGRAL_MODE_P</code>
machine modes but zero for the SSE register classes.
</p>
<p>The default version of this hook returns false for any mode. It is always
safe to redefine this hook to return with a nonzero value. But if you
unnecessarily define it, you will reduce the amount of optimizations
that can be performed in some cases. If you do not define this hook
to return a nonzero value when it is required, the compiler will run out
of spill registers and print a fatal error message.
</p></dd></dl>
<hr>
<div class="header">
<p>
Next: <a href="Scalar-Return.html#Scalar-Return" accesskey="n" rel="next">Scalar Return</a>, Previous: <a href="Stack-Arguments.html#Stack-Arguments" accesskey="p" rel="prev">Stack Arguments</a>, Up: <a href="Stack-and-Calling.html#Stack-and-Calling" accesskey="u" rel="up">Stack and Calling</a> &nbsp; [<a href="index.html#SEC_Contents" title="Table of contents" rel="contents">Contents</a>][<a href="Option-Index.html#Option-Index" title="Index" rel="index">Index</a>]</p>
</div>
</body>
</html>