blob: 8ae9899008843ae51b2699e41d52d3e2df95b648 [file] [log] [blame]
/* Convert function calls to rtl insns, for GNU C compiler.
Copyright (C) 1989-2017 Free Software Foundation, Inc.
This file is part of GCC.
GCC is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free
Software Foundation; either version 3, or (at your option) any later
version.
GCC is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
for more details.
You should have received a copy of the GNU General Public License
along with GCC; see the file COPYING3. If not see
<http://www.gnu.org/licenses/>. */
#include "config.h"
#include "system.h"
#include "coretypes.h"
#include "backend.h"
#include "target.h"
#include "rtl.h"
#include "tree.h"
#include "gimple.h"
#include "predict.h"
#include "memmodel.h"
#include "tm_p.h"
#include "stringpool.h"
#include "expmed.h"
#include "optabs.h"
#include "emit-rtl.h"
#include "cgraph.h"
#include "diagnostic-core.h"
#include "fold-const.h"
#include "stor-layout.h"
#include "varasm.h"
#include "internal-fn.h"
#include "dojump.h"
#include "explow.h"
#include "calls.h"
#include "expr.h"
#include "output.h"
#include "langhooks.h"
#include "except.h"
#include "dbgcnt.h"
#include "rtl-iter.h"
#include "tree-chkp.h"
#include "tree-vrp.h"
#include "tree-ssanames.h"
#include "rtl-chkp.h"
#include "intl.h"
#include "stringpool.h"
#include "attribs.h"
/* Like PREFERRED_STACK_BOUNDARY but in units of bytes, not bits. */
#define STACK_BYTES (PREFERRED_STACK_BOUNDARY / BITS_PER_UNIT)
/* Data structure and subroutines used within expand_call. */
struct arg_data
{
/* Tree node for this argument. */
tree tree_value;
/* Mode for value; TYPE_MODE unless promoted. */
machine_mode mode;
/* Current RTL value for argument, or 0 if it isn't precomputed. */
rtx value;
/* Initially-compute RTL value for argument; only for const functions. */
rtx initial_value;
/* Register to pass this argument in, 0 if passed on stack, or an
PARALLEL if the arg is to be copied into multiple non-contiguous
registers. */
rtx reg;
/* Register to pass this argument in when generating tail call sequence.
This is not the same register as for normal calls on machines with
register windows. */
rtx tail_call_reg;
/* If REG is a PARALLEL, this is a copy of VALUE pulled into the correct
form for emit_group_move. */
rtx parallel_value;
/* If value is passed in neither reg nor stack, this field holds a number
of a special slot to be used. */
rtx special_slot;
/* For pointer bounds hold an index of parm bounds are bound to. -1 if
there is no such pointer. */
int pointer_arg;
/* If pointer_arg refers a structure, then pointer_offset holds an offset
of a pointer in this structure. */
int pointer_offset;
/* If REG was promoted from the actual mode of the argument expression,
indicates whether the promotion is sign- or zero-extended. */
int unsignedp;
/* Number of bytes to put in registers. 0 means put the whole arg
in registers. Also 0 if not passed in registers. */
int partial;
/* Nonzero if argument must be passed on stack.
Note that some arguments may be passed on the stack
even though pass_on_stack is zero, just because FUNCTION_ARG says so.
pass_on_stack identifies arguments that *cannot* go in registers. */
int pass_on_stack;
/* Some fields packaged up for locate_and_pad_parm. */
struct locate_and_pad_arg_data locate;
/* Location on the stack at which parameter should be stored. The store
has already been done if STACK == VALUE. */
rtx stack;
/* Location on the stack of the start of this argument slot. This can
differ from STACK if this arg pads downward. This location is known
to be aligned to TARGET_FUNCTION_ARG_BOUNDARY. */
rtx stack_slot;
/* Place that this stack area has been saved, if needed. */
rtx save_area;
/* If an argument's alignment does not permit direct copying into registers,
copy in smaller-sized pieces into pseudos. These are stored in a
block pointed to by this field. The next field says how many
word-sized pseudos we made. */
rtx *aligned_regs;
int n_aligned_regs;
};
/* A vector of one char per byte of stack space. A byte if nonzero if
the corresponding stack location has been used.
This vector is used to prevent a function call within an argument from
clobbering any stack already set up. */
static char *stack_usage_map;
/* Size of STACK_USAGE_MAP. */
static int highest_outgoing_arg_in_use;
/* A bitmap of virtual-incoming stack space. Bit is set if the corresponding
stack location's tail call argument has been already stored into the stack.
This bitmap is used to prevent sibling call optimization if function tries
to use parent's incoming argument slots when they have been already
overwritten with tail call arguments. */
static sbitmap stored_args_map;
/* stack_arg_under_construction is nonzero when an argument may be
initialized with a constructor call (including a C function that
returns a BLKmode struct) and expand_call must take special action
to make sure the object being constructed does not overlap the
argument list for the constructor call. */
static int stack_arg_under_construction;
static void emit_call_1 (rtx, tree, tree, tree, HOST_WIDE_INT, HOST_WIDE_INT,
HOST_WIDE_INT, rtx, rtx, int, rtx, int,
cumulative_args_t);
static void precompute_register_parameters (int, struct arg_data *, int *);
static void store_bounds (struct arg_data *, struct arg_data *);
static int store_one_arg (struct arg_data *, rtx, int, int, int);
static void store_unaligned_arguments_into_pseudos (struct arg_data *, int);
static int finalize_must_preallocate (int, int, struct arg_data *,
struct args_size *);
static void precompute_arguments (int, struct arg_data *);
static int compute_argument_block_size (int, struct args_size *, tree, tree, int);
static void initialize_argument_information (int, struct arg_data *,
struct args_size *, int,
tree, tree,
tree, tree, cumulative_args_t, int,
rtx *, int *, int *, int *,
bool *, bool);
static void compute_argument_addresses (struct arg_data *, rtx, int);
static rtx rtx_for_function_call (tree, tree);
static void load_register_parameters (struct arg_data *, int, rtx *, int,
int, int *);
static int special_function_p (const_tree, int);
static int check_sibcall_argument_overlap_1 (rtx);
static int check_sibcall_argument_overlap (rtx_insn *, struct arg_data *, int);
static int combine_pending_stack_adjustment_and_call (int, struct args_size *,
unsigned int);
static tree split_complex_types (tree);
#ifdef REG_PARM_STACK_SPACE
static rtx save_fixed_argument_area (int, rtx, int *, int *);
static void restore_fixed_argument_area (rtx, rtx, int, int);
#endif
/* Force FUNEXP into a form suitable for the address of a CALL,
and return that as an rtx. Also load the static chain register
if FNDECL is a nested function.
CALL_FUSAGE points to a variable holding the prospective
CALL_INSN_FUNCTION_USAGE information. */
rtx
prepare_call_address (tree fndecl_or_type, rtx funexp, rtx static_chain_value,
rtx *call_fusage, int reg_parm_seen, int flags)
{
/* Make a valid memory address and copy constants through pseudo-regs,
but not for a constant address if -fno-function-cse. */
if (GET_CODE (funexp) != SYMBOL_REF)
{
/* If it's an indirect call by descriptor, generate code to perform
runtime identification of the pointer and load the descriptor. */
if ((flags & ECF_BY_DESCRIPTOR) && !flag_trampolines)
{
const int bit_val = targetm.calls.custom_function_descriptors;
rtx call_lab = gen_label_rtx ();
gcc_assert (fndecl_or_type && TYPE_P (fndecl_or_type));
fndecl_or_type
= build_decl (UNKNOWN_LOCATION, FUNCTION_DECL, NULL_TREE,
fndecl_or_type);
DECL_STATIC_CHAIN (fndecl_or_type) = 1;
rtx chain = targetm.calls.static_chain (fndecl_or_type, false);
if (GET_MODE (funexp) != Pmode)
funexp = convert_memory_address (Pmode, funexp);
/* Avoid long live ranges around function calls. */
funexp = copy_to_mode_reg (Pmode, funexp);
if (REG_P (chain))
emit_insn (gen_rtx_CLOBBER (VOIDmode, chain));
/* Emit the runtime identification pattern. */
rtx mask = gen_rtx_AND (Pmode, funexp, GEN_INT (bit_val));
emit_cmp_and_jump_insns (mask, const0_rtx, EQ, NULL_RTX, Pmode, 1,
call_lab);
/* Statically predict the branch to very likely taken. */
rtx_insn *insn = get_last_insn ();
if (JUMP_P (insn))
predict_insn_def (insn, PRED_BUILTIN_EXPECT, TAKEN);
/* Load the descriptor. */
rtx mem = gen_rtx_MEM (ptr_mode,
plus_constant (Pmode, funexp, - bit_val));
MEM_NOTRAP_P (mem) = 1;
mem = convert_memory_address (Pmode, mem);
emit_move_insn (chain, mem);
mem = gen_rtx_MEM (ptr_mode,
plus_constant (Pmode, funexp,
POINTER_SIZE / BITS_PER_UNIT
- bit_val));
MEM_NOTRAP_P (mem) = 1;
mem = convert_memory_address (Pmode, mem);
emit_move_insn (funexp, mem);
emit_label (call_lab);
if (REG_P (chain))
{
use_reg (call_fusage, chain);
STATIC_CHAIN_REG_P (chain) = 1;
}
/* Make sure we're not going to be overwritten below. */
gcc_assert (!static_chain_value);
}
/* If we are using registers for parameters, force the
function address into a register now. */
funexp = ((reg_parm_seen
&& targetm.small_register_classes_for_mode_p (FUNCTION_MODE))
? force_not_mem (memory_address (FUNCTION_MODE, funexp))
: memory_address (FUNCTION_MODE, funexp));
}
else
{
/* funexp could be a SYMBOL_REF represents a function pointer which is
of ptr_mode. In this case, it should be converted into address mode
to be a valid address for memory rtx pattern. See PR 64971. */
if (GET_MODE (funexp) != Pmode)
funexp = convert_memory_address (Pmode, funexp);
if (!(flags & ECF_SIBCALL))
{
if (!NO_FUNCTION_CSE && optimize && ! flag_no_function_cse)
funexp = force_reg (Pmode, funexp);
}
}
if (static_chain_value != 0
&& (TREE_CODE (fndecl_or_type) != FUNCTION_DECL
|| DECL_STATIC_CHAIN (fndecl_or_type)))
{
rtx chain;
chain = targetm.calls.static_chain (fndecl_or_type, false);
static_chain_value = convert_memory_address (Pmode, static_chain_value);
emit_move_insn (chain, static_chain_value);
if (REG_P (chain))
{
use_reg (call_fusage, chain);
STATIC_CHAIN_REG_P (chain) = 1;
}
}
return funexp;
}
/* Generate instructions to call function FUNEXP,
and optionally pop the results.
The CALL_INSN is the first insn generated.
FNDECL is the declaration node of the function. This is given to the
hook TARGET_RETURN_POPS_ARGS to determine whether this function pops
its own args.
FUNTYPE is the data type of the function. This is given to the hook
TARGET_RETURN_POPS_ARGS to determine whether this function pops its
own args. We used to allow an identifier for library functions, but
that doesn't work when the return type is an aggregate type and the
calling convention says that the pointer to this aggregate is to be
popped by the callee.
STACK_SIZE is the number of bytes of arguments on the stack,
ROUNDED_STACK_SIZE is that number rounded up to
PREFERRED_STACK_BOUNDARY; zero if the size is variable. This is
both to put into the call insn and to generate explicit popping
code if necessary.
STRUCT_VALUE_SIZE is the number of bytes wanted in a structure value.
It is zero if this call doesn't want a structure value.
NEXT_ARG_REG is the rtx that results from executing
targetm.calls.function_arg (&args_so_far, VOIDmode, void_type_node, true)
just after all the args have had their registers assigned.
This could be whatever you like, but normally it is the first
arg-register beyond those used for args in this call,
or 0 if all the arg-registers are used in this call.
It is passed on to `gen_call' so you can put this info in the call insn.
VALREG is a hard register in which a value is returned,
or 0 if the call does not return a value.
OLD_INHIBIT_DEFER_POP is the value that `inhibit_defer_pop' had before
the args to this call were processed.
We restore `inhibit_defer_pop' to that value.
CALL_FUSAGE is either empty or an EXPR_LIST of USE expressions that
denote registers used by the called function. */
static void
emit_call_1 (rtx funexp, tree fntree ATTRIBUTE_UNUSED, tree fndecl ATTRIBUTE_UNUSED,
tree funtype ATTRIBUTE_UNUSED,
HOST_WIDE_INT stack_size ATTRIBUTE_UNUSED,
HOST_WIDE_INT rounded_stack_size,
HOST_WIDE_INT struct_value_size ATTRIBUTE_UNUSED,
rtx next_arg_reg ATTRIBUTE_UNUSED, rtx valreg,
int old_inhibit_defer_pop, rtx call_fusage, int ecf_flags,
cumulative_args_t args_so_far ATTRIBUTE_UNUSED)
{
rtx rounded_stack_size_rtx = GEN_INT (rounded_stack_size);
rtx call, funmem, pat;
int already_popped = 0;
HOST_WIDE_INT n_popped = 0;
/* Sibling call patterns never pop arguments (no sibcall(_value)_pop
patterns exist). Any popping that the callee does on return will
be from our caller's frame rather than ours. */
if (!(ecf_flags & ECF_SIBCALL))
{
n_popped += targetm.calls.return_pops_args (fndecl, funtype, stack_size);
#ifdef CALL_POPS_ARGS
n_popped += CALL_POPS_ARGS (*get_cumulative_args (args_so_far));
#endif
}
/* Ensure address is valid. SYMBOL_REF is already valid, so no need,
and we don't want to load it into a register as an optimization,
because prepare_call_address already did it if it should be done. */
if (GET_CODE (funexp) != SYMBOL_REF)
funexp = memory_address (FUNCTION_MODE, funexp);
funmem = gen_rtx_MEM (FUNCTION_MODE, funexp);
if (fndecl && TREE_CODE (fndecl) == FUNCTION_DECL)
{
tree t = fndecl;
/* Although a built-in FUNCTION_DECL and its non-__builtin
counterpart compare equal and get a shared mem_attrs, they
produce different dump output in compare-debug compilations,
if an entry gets garbage collected in one compilation, then
adds a different (but equivalent) entry, while the other
doesn't run the garbage collector at the same spot and then
shares the mem_attr with the equivalent entry. */
if (DECL_BUILT_IN_CLASS (t) == BUILT_IN_NORMAL)
{
tree t2 = builtin_decl_explicit (DECL_FUNCTION_CODE (t));
if (t2)
t = t2;
}
set_mem_expr (funmem, t);
}
else if (fntree)
set_mem_expr (funmem, build_simple_mem_ref (CALL_EXPR_FN (fntree)));
if (ecf_flags & ECF_SIBCALL)
{
if (valreg)
pat = targetm.gen_sibcall_value (valreg, funmem,
rounded_stack_size_rtx,
next_arg_reg, NULL_RTX);
else
pat = targetm.gen_sibcall (funmem, rounded_stack_size_rtx,
next_arg_reg, GEN_INT (struct_value_size));
}
/* If the target has "call" or "call_value" insns, then prefer them
if no arguments are actually popped. If the target does not have
"call" or "call_value" insns, then we must use the popping versions
even if the call has no arguments to pop. */
else if (n_popped > 0
|| !(valreg
? targetm.have_call_value ()
: targetm.have_call ()))
{
rtx n_pop = GEN_INT (n_popped);
/* If this subroutine pops its own args, record that in the call insn
if possible, for the sake of frame pointer elimination. */
if (valreg)
pat = targetm.gen_call_value_pop (valreg, funmem,
rounded_stack_size_rtx,
next_arg_reg, n_pop);
else
pat = targetm.gen_call_pop (funmem, rounded_stack_size_rtx,
next_arg_reg, n_pop);
already_popped = 1;
}
else
{
if (valreg)
pat = targetm.gen_call_value (valreg, funmem, rounded_stack_size_rtx,
next_arg_reg, NULL_RTX);
else
pat = targetm.gen_call (funmem, rounded_stack_size_rtx, next_arg_reg,
GEN_INT (struct_value_size));
}
emit_insn (pat);
/* Find the call we just emitted. */
rtx_call_insn *call_insn = last_call_insn ();
/* Some target create a fresh MEM instead of reusing the one provided
above. Set its MEM_EXPR. */
call = get_call_rtx_from (call_insn);
if (call
&& MEM_EXPR (XEXP (call, 0)) == NULL_TREE
&& MEM_EXPR (funmem) != NULL_TREE)
set_mem_expr (XEXP (call, 0), MEM_EXPR (funmem));
/* Mark instrumented calls. */
if (call && fntree)
CALL_EXPR_WITH_BOUNDS_P (call) = CALL_WITH_BOUNDS_P (fntree);
/* Put the register usage information there. */
add_function_usage_to (call_insn, call_fusage);
/* If this is a const call, then set the insn's unchanging bit. */
if (ecf_flags & ECF_CONST)
RTL_CONST_CALL_P (call_insn) = 1;
/* If this is a pure call, then set the insn's unchanging bit. */
if (ecf_flags & ECF_PURE)
RTL_PURE_CALL_P (call_insn) = 1;
/* If this is a const call, then set the insn's unchanging bit. */
if (ecf_flags & ECF_LOOPING_CONST_OR_PURE)
RTL_LOOPING_CONST_OR_PURE_CALL_P (call_insn) = 1;
/* Create a nothrow REG_EH_REGION note, if needed. */
make_reg_eh_region_note (call_insn, ecf_flags, 0);
if (ecf_flags & ECF_NORETURN)
add_reg_note (call_insn, REG_NORETURN, const0_rtx);
if (ecf_flags & ECF_RETURNS_TWICE)
{
add_reg_note (call_insn, REG_SETJMP, const0_rtx);
cfun->calls_setjmp = 1;
}
SIBLING_CALL_P (call_insn) = ((ecf_flags & ECF_SIBCALL) != 0);
/* Restore this now, so that we do defer pops for this call's args
if the context of the call as a whole permits. */
inhibit_defer_pop = old_inhibit_defer_pop;
if (n_popped > 0)
{
if (!already_popped)
CALL_INSN_FUNCTION_USAGE (call_insn)
= gen_rtx_EXPR_LIST (VOIDmode,
gen_rtx_CLOBBER (VOIDmode, stack_pointer_rtx),
CALL_INSN_FUNCTION_USAGE (call_insn));
rounded_stack_size -= n_popped;
rounded_stack_size_rtx = GEN_INT (rounded_stack_size);
stack_pointer_delta -= n_popped;
add_reg_note (call_insn, REG_ARGS_SIZE, GEN_INT (stack_pointer_delta));
/* If popup is needed, stack realign must use DRAP */
if (SUPPORTS_STACK_ALIGNMENT)
crtl->need_drap = true;
}
/* For noreturn calls when not accumulating outgoing args force
REG_ARGS_SIZE note to prevent crossjumping of calls with different
args sizes. */
else if (!ACCUMULATE_OUTGOING_ARGS && (ecf_flags & ECF_NORETURN) != 0)
add_reg_note (call_insn, REG_ARGS_SIZE, GEN_INT (stack_pointer_delta));
if (!ACCUMULATE_OUTGOING_ARGS)
{
/* If returning from the subroutine does not automatically pop the args,
we need an instruction to pop them sooner or later.
Perhaps do it now; perhaps just record how much space to pop later.
If returning from the subroutine does pop the args, indicate that the
stack pointer will be changed. */
if (rounded_stack_size != 0)
{
if (ecf_flags & ECF_NORETURN)
/* Just pretend we did the pop. */
stack_pointer_delta -= rounded_stack_size;
else if (flag_defer_pop && inhibit_defer_pop == 0
&& ! (ecf_flags & (ECF_CONST | ECF_PURE)))
pending_stack_adjust += rounded_stack_size;
else
adjust_stack (rounded_stack_size_rtx);
}
}
/* When we accumulate outgoing args, we must avoid any stack manipulations.
Restore the stack pointer to its original value now. Usually
ACCUMULATE_OUTGOING_ARGS targets don't get here, but there are exceptions.
On i386 ACCUMULATE_OUTGOING_ARGS can be enabled on demand, and
popping variants of functions exist as well.
??? We may optimize similar to defer_pop above, but it is
probably not worthwhile.
??? It will be worthwhile to enable combine_stack_adjustments even for
such machines. */
else if (n_popped)
anti_adjust_stack (GEN_INT (n_popped));
}
/* Determine if the function identified by FNDECL is one with
special properties we wish to know about. Modify FLAGS accordingly.
For example, if the function might return more than one time (setjmp), then
set ECF_RETURNS_TWICE.
Set ECF_MAY_BE_ALLOCA for any memory allocation function that might allocate
space from the stack such as alloca. */
static int
special_function_p (const_tree fndecl, int flags)
{
tree name_decl = DECL_NAME (fndecl);
/* For instrumentation clones we want to derive flags
from the original name. */
if (cgraph_node::get (fndecl)
&& cgraph_node::get (fndecl)->instrumentation_clone)
name_decl = DECL_NAME (cgraph_node::get (fndecl)->orig_decl);
if (fndecl && name_decl
&& IDENTIFIER_LENGTH (name_decl) <= 11
/* Exclude functions not at the file scope, or not `extern',
since they are not the magic functions we would otherwise
think they are.
FIXME: this should be handled with attributes, not with this
hacky imitation of DECL_ASSEMBLER_NAME. It's (also) wrong
because you can declare fork() inside a function if you
wish. */
&& (DECL_CONTEXT (fndecl) == NULL_TREE
|| TREE_CODE (DECL_CONTEXT (fndecl)) == TRANSLATION_UNIT_DECL)
&& TREE_PUBLIC (fndecl))
{
const char *name = IDENTIFIER_POINTER (name_decl);
const char *tname = name;
/* We assume that alloca will always be called by name. It
makes no sense to pass it as a pointer-to-function to
anything that does not understand its behavior. */
if (IDENTIFIER_LENGTH (name_decl) == 6
&& name[0] == 'a'
&& ! strcmp (name, "alloca"))
flags |= ECF_MAY_BE_ALLOCA;
/* Disregard prefix _ or __. */
if (name[0] == '_')
{
if (name[1] == '_')
tname += 2;
else
tname += 1;
}
/* ECF_RETURNS_TWICE is safe even for -ffreestanding. */
if (! strcmp (tname, "setjmp")
|| ! strcmp (tname, "sigsetjmp")
|| ! strcmp (name, "savectx")
|| ! strcmp (name, "vfork")
|| ! strcmp (name, "getcontext"))
flags |= ECF_RETURNS_TWICE;
}
if (DECL_BUILT_IN_CLASS (fndecl) == BUILT_IN_NORMAL
&& ALLOCA_FUNCTION_CODE_P (DECL_FUNCTION_CODE (fndecl)))
flags |= ECF_MAY_BE_ALLOCA;
return flags;
}
/* Similar to special_function_p; return a set of ERF_ flags for the
function FNDECL. */
static int
decl_return_flags (tree fndecl)
{
tree attr;
tree type = TREE_TYPE (fndecl);
if (!type)
return 0;
attr = lookup_attribute ("fn spec", TYPE_ATTRIBUTES (type));
if (!attr)
return 0;
attr = TREE_VALUE (TREE_VALUE (attr));
if (!attr || TREE_STRING_LENGTH (attr) < 1)
return 0;
switch (TREE_STRING_POINTER (attr)[0])
{
case '1':
case '2':
case '3':
case '4':
return ERF_RETURNS_ARG | (TREE_STRING_POINTER (attr)[0] - '1');
case 'm':
return ERF_NOALIAS;
case '.':
default:
return 0;
}
}
/* Return nonzero when FNDECL represents a call to setjmp. */
int
setjmp_call_p (const_tree fndecl)
{
if (DECL_IS_RETURNS_TWICE (fndecl))
return ECF_RETURNS_TWICE;
return special_function_p (fndecl, 0) & ECF_RETURNS_TWICE;
}
/* Return true if STMT may be an alloca call. */
bool
gimple_maybe_alloca_call_p (const gimple *stmt)
{
tree fndecl;
if (!is_gimple_call (stmt))
return false;
fndecl = gimple_call_fndecl (stmt);
if (fndecl && (special_function_p (fndecl, 0) & ECF_MAY_BE_ALLOCA))
return true;
return false;
}
/* Return true if STMT is a builtin alloca call. */
bool
gimple_alloca_call_p (const gimple *stmt)
{
tree fndecl;
if (!is_gimple_call (stmt))
return false;
fndecl = gimple_call_fndecl (stmt);
if (fndecl && DECL_BUILT_IN_CLASS (fndecl) == BUILT_IN_NORMAL)
switch (DECL_FUNCTION_CODE (fndecl))
{
CASE_BUILT_IN_ALLOCA:
return true;
default:
break;
}
return false;
}
/* Return true when exp contains a builtin alloca call. */
bool
alloca_call_p (const_tree exp)
{
tree fndecl;
if (TREE_CODE (exp) == CALL_EXPR
&& (fndecl = get_callee_fndecl (exp))
&& DECL_BUILT_IN_CLASS (fndecl) == BUILT_IN_NORMAL)
switch (DECL_FUNCTION_CODE (fndecl))
{
CASE_BUILT_IN_ALLOCA:
return true;
default:
break;
}
return false;
}
/* Return TRUE if FNDECL is either a TM builtin or a TM cloned
function. Return FALSE otherwise. */
static bool
is_tm_builtin (const_tree fndecl)
{
if (fndecl == NULL)
return false;
if (decl_is_tm_clone (fndecl))
return true;
if (DECL_BUILT_IN_CLASS (fndecl) == BUILT_IN_NORMAL)
{
switch (DECL_FUNCTION_CODE (fndecl))
{
case BUILT_IN_TM_COMMIT:
case BUILT_IN_TM_COMMIT_EH:
case BUILT_IN_TM_ABORT:
case BUILT_IN_TM_IRREVOCABLE:
case BUILT_IN_TM_GETTMCLONE_IRR:
case BUILT_IN_TM_MEMCPY:
case BUILT_IN_TM_MEMMOVE:
case BUILT_IN_TM_MEMSET:
CASE_BUILT_IN_TM_STORE (1):
CASE_BUILT_IN_TM_STORE (2):
CASE_BUILT_IN_TM_STORE (4):
CASE_BUILT_IN_TM_STORE (8):
CASE_BUILT_IN_TM_STORE (FLOAT):
CASE_BUILT_IN_TM_STORE (DOUBLE):
CASE_BUILT_IN_TM_STORE (LDOUBLE):
CASE_BUILT_IN_TM_STORE (M64):
CASE_BUILT_IN_TM_STORE (M128):
CASE_BUILT_IN_TM_STORE (M256):
CASE_BUILT_IN_TM_LOAD (1):
CASE_BUILT_IN_TM_LOAD (2):
CASE_BUILT_IN_TM_LOAD (4):
CASE_BUILT_IN_TM_LOAD (8):
CASE_BUILT_IN_TM_LOAD (FLOAT):
CASE_BUILT_IN_TM_LOAD (DOUBLE):
CASE_BUILT_IN_TM_LOAD (LDOUBLE):
CASE_BUILT_IN_TM_LOAD (M64):
CASE_BUILT_IN_TM_LOAD (M128):
CASE_BUILT_IN_TM_LOAD (M256):
case BUILT_IN_TM_LOG:
case BUILT_IN_TM_LOG_1:
case BUILT_IN_TM_LOG_2:
case BUILT_IN_TM_LOG_4:
case BUILT_IN_TM_LOG_8:
case BUILT_IN_TM_LOG_FLOAT:
case BUILT_IN_TM_LOG_DOUBLE:
case BUILT_IN_TM_LOG_LDOUBLE:
case BUILT_IN_TM_LOG_M64:
case BUILT_IN_TM_LOG_M128:
case BUILT_IN_TM_LOG_M256:
return true;
default:
break;
}
}
return false;
}
/* Detect flags (function attributes) from the function decl or type node. */
int
flags_from_decl_or_type (const_tree exp)
{
int flags = 0;
if (DECL_P (exp))
{
/* The function exp may have the `malloc' attribute. */
if (DECL_IS_MALLOC (exp))
flags |= ECF_MALLOC;
/* The function exp may have the `returns_twice' attribute. */
if (DECL_IS_RETURNS_TWICE (exp))
flags |= ECF_RETURNS_TWICE;
/* Process the pure and const attributes. */
if (TREE_READONLY (exp))
flags |= ECF_CONST;
if (DECL_PURE_P (exp))
flags |= ECF_PURE;
if (DECL_LOOPING_CONST_OR_PURE_P (exp))
flags |= ECF_LOOPING_CONST_OR_PURE;
if (DECL_IS_NOVOPS (exp))
flags |= ECF_NOVOPS;
if (lookup_attribute ("leaf", DECL_ATTRIBUTES (exp)))
flags |= ECF_LEAF;
if (lookup_attribute ("cold", DECL_ATTRIBUTES (exp)))
flags |= ECF_COLD;
if (TREE_NOTHROW (exp))
flags |= ECF_NOTHROW;
if (flag_tm)
{
if (is_tm_builtin (exp))
flags |= ECF_TM_BUILTIN;
else if ((flags & (ECF_CONST|ECF_NOVOPS)) != 0
|| lookup_attribute ("transaction_pure",
TYPE_ATTRIBUTES (TREE_TYPE (exp))))
flags |= ECF_TM_PURE;
}
flags = special_function_p (exp, flags);
}
else if (TYPE_P (exp))
{
if (TYPE_READONLY (exp))
flags |= ECF_CONST;
if (flag_tm
&& ((flags & ECF_CONST) != 0
|| lookup_attribute ("transaction_pure", TYPE_ATTRIBUTES (exp))))
flags |= ECF_TM_PURE;
}
else
gcc_unreachable ();
if (TREE_THIS_VOLATILE (exp))
{
flags |= ECF_NORETURN;
if (flags & (ECF_CONST|ECF_PURE))
flags |= ECF_LOOPING_CONST_OR_PURE;
}
return flags;
}
/* Detect flags from a CALL_EXPR. */
int
call_expr_flags (const_tree t)
{
int flags;
tree decl = get_callee_fndecl (t);
if (decl)
flags = flags_from_decl_or_type (decl);
else if (CALL_EXPR_FN (t) == NULL_TREE)
flags = internal_fn_flags (CALL_EXPR_IFN (t));
else
{
tree type = TREE_TYPE (CALL_EXPR_FN (t));
if (type && TREE_CODE (type) == POINTER_TYPE)
flags = flags_from_decl_or_type (TREE_TYPE (type));
else
flags = 0;
if (CALL_EXPR_BY_DESCRIPTOR (t))
flags |= ECF_BY_DESCRIPTOR;
}
return flags;
}
/* Return true if TYPE should be passed by invisible reference. */
bool
pass_by_reference (CUMULATIVE_ARGS *ca, machine_mode mode,
tree type, bool named_arg)
{
if (type)
{
/* If this type contains non-trivial constructors, then it is
forbidden for the middle-end to create any new copies. */
if (TREE_ADDRESSABLE (type))
return true;
/* GCC post 3.4 passes *all* variable sized types by reference. */
if (!TYPE_SIZE (type) || TREE_CODE (TYPE_SIZE (type)) != INTEGER_CST)
return true;
/* If a record type should be passed the same as its first (and only)
member, use the type and mode of that member. */
if (TREE_CODE (type) == RECORD_TYPE && TYPE_TRANSPARENT_AGGR (type))
{
type = TREE_TYPE (first_field (type));
mode = TYPE_MODE (type);
}
}
return targetm.calls.pass_by_reference (pack_cumulative_args (ca), mode,
type, named_arg);
}
/* Return true if TYPE, which is passed by reference, should be callee
copied instead of caller copied. */
bool
reference_callee_copied (CUMULATIVE_ARGS *ca, machine_mode mode,
tree type, bool named_arg)
{
if (type && TREE_ADDRESSABLE (type))
return false;
return targetm.calls.callee_copies (pack_cumulative_args (ca), mode, type,
named_arg);
}
/* Precompute all register parameters as described by ARGS, storing values
into fields within the ARGS array.
NUM_ACTUALS indicates the total number elements in the ARGS array.
Set REG_PARM_SEEN if we encounter a register parameter. */
static void
precompute_register_parameters (int num_actuals, struct arg_data *args,
int *reg_parm_seen)
{
int i;
*reg_parm_seen = 0;
for (i = 0; i < num_actuals; i++)
if (args[i].reg != 0 && ! args[i].pass_on_stack)
{
*reg_parm_seen = 1;
if (args[i].value == 0)
{
push_temp_slots ();
args[i].value = expand_normal (args[i].tree_value);
preserve_temp_slots (args[i].value);
pop_temp_slots ();
}
/* If we are to promote the function arg to a wider mode,
do it now. */
if (args[i].mode != TYPE_MODE (TREE_TYPE (args[i].tree_value)))
args[i].value
= convert_modes (args[i].mode,
TYPE_MODE (TREE_TYPE (args[i].tree_value)),
args[i].value, args[i].unsignedp);
/* If the value is a non-legitimate constant, force it into a
pseudo now. TLS symbols sometimes need a call to resolve. */
if (CONSTANT_P (args[i].value)
&& !targetm.legitimate_constant_p (args[i].mode, args[i].value))
args[i].value = force_reg (args[i].mode, args[i].value);
/* If we're going to have to load the value by parts, pull the
parts into pseudos. The part extraction process can involve
non-trivial computation. */
if (GET_CODE (args[i].reg) == PARALLEL)
{
tree type = TREE_TYPE (args[i].tree_value);
args[i].parallel_value
= emit_group_load_into_temps (args[i].reg, args[i].value,
type, int_size_in_bytes (type));
}
/* If the value is expensive, and we are inside an appropriately
short loop, put the value into a pseudo and then put the pseudo
into the hard reg.
For small register classes, also do this if this call uses
register parameters. This is to avoid reload conflicts while
loading the parameters registers. */
else if ((! (REG_P (args[i].value)
|| (GET_CODE (args[i].value) == SUBREG
&& REG_P (SUBREG_REG (args[i].value)))))
&& args[i].mode != BLKmode
&& (set_src_cost (args[i].value, args[i].mode,
optimize_insn_for_speed_p ())
> COSTS_N_INSNS (1))
&& ((*reg_parm_seen
&& targetm.small_register_classes_for_mode_p (args[i].mode))
|| optimize))
args[i].value = copy_to_mode_reg (args[i].mode, args[i].value);
}
}
#ifdef REG_PARM_STACK_SPACE
/* The argument list is the property of the called routine and it
may clobber it. If the fixed area has been used for previous
parameters, we must save and restore it. */
static rtx
save_fixed_argument_area (int reg_parm_stack_space, rtx argblock, int *low_to_save, int *high_to_save)
{
int low;
int high;
/* Compute the boundary of the area that needs to be saved, if any. */
high = reg_parm_stack_space;
if (ARGS_GROW_DOWNWARD)
high += 1;
if (high > highest_outgoing_arg_in_use)
high = highest_outgoing_arg_in_use;
for (low = 0; low < high; low++)
if (stack_usage_map[low] != 0)
{
int num_to_save;
machine_mode save_mode;
int delta;
rtx addr;
rtx stack_area;
rtx save_area;
while (stack_usage_map[--high] == 0)
;
*low_to_save = low;
*high_to_save = high;
num_to_save = high - low + 1;
/* If we don't have the required alignment, must do this
in BLKmode. */
scalar_int_mode imode;
if (int_mode_for_size (num_to_save * BITS_PER_UNIT, 1).exists (&imode)
&& (low & (MIN (GET_MODE_SIZE (imode),
BIGGEST_ALIGNMENT / UNITS_PER_WORD) - 1)) == 0)
save_mode = imode;
else
save_mode = BLKmode;
if (ARGS_GROW_DOWNWARD)
delta = -high;
else
delta = low;
addr = plus_constant (Pmode, argblock, delta);
stack_area = gen_rtx_MEM (save_mode, memory_address (save_mode, addr));
set_mem_align (stack_area, PARM_BOUNDARY);
if (save_mode == BLKmode)
{
save_area = assign_stack_temp (BLKmode, num_to_save);
emit_block_move (validize_mem (save_area), stack_area,
GEN_INT (num_to_save), BLOCK_OP_CALL_PARM);
}
else
{
save_area = gen_reg_rtx (save_mode);
emit_move_insn (save_area, stack_area);
}
return save_area;
}
return NULL_RTX;
}
static void
restore_fixed_argument_area (rtx save_area, rtx argblock, int high_to_save, int low_to_save)
{
machine_mode save_mode = GET_MODE (save_area);
int delta;
rtx addr, stack_area;
if (ARGS_GROW_DOWNWARD)
delta = -high_to_save;
else
delta = low_to_save;
addr = plus_constant (Pmode, argblock, delta);
stack_area = gen_rtx_MEM (save_mode, memory_address (save_mode, addr));
set_mem_align (stack_area, PARM_BOUNDARY);
if (save_mode != BLKmode)
emit_move_insn (stack_area, save_area);
else
emit_block_move (stack_area, validize_mem (save_area),
GEN_INT (high_to_save - low_to_save + 1),
BLOCK_OP_CALL_PARM);
}
#endif /* REG_PARM_STACK_SPACE */
/* If any elements in ARGS refer to parameters that are to be passed in
registers, but not in memory, and whose alignment does not permit a
direct copy into registers. Copy the values into a group of pseudos
which we will later copy into the appropriate hard registers.
Pseudos for each unaligned argument will be stored into the array
args[argnum].aligned_regs. The caller is responsible for deallocating
the aligned_regs array if it is nonzero. */
static void
store_unaligned_arguments_into_pseudos (struct arg_data *args, int num_actuals)
{
int i, j;
for (i = 0; i < num_actuals; i++)
if (args[i].reg != 0 && ! args[i].pass_on_stack
&& GET_CODE (args[i].reg) != PARALLEL
&& args[i].mode == BLKmode
&& MEM_P (args[i].value)
&& (MEM_ALIGN (args[i].value)
< (unsigned int) MIN (BIGGEST_ALIGNMENT, BITS_PER_WORD)))
{
int bytes = int_size_in_bytes (TREE_TYPE (args[i].tree_value));
int endian_correction = 0;
if (args[i].partial)
{
gcc_assert (args[i].partial % UNITS_PER_WORD == 0);
args[i].n_aligned_regs = args[i].partial / UNITS_PER_WORD;
}
else
{
args[i].n_aligned_regs
= (bytes + UNITS_PER_WORD - 1) / UNITS_PER_WORD;
}
args[i].aligned_regs = XNEWVEC (rtx, args[i].n_aligned_regs);
/* Structures smaller than a word are normally aligned to the
least significant byte. On a BYTES_BIG_ENDIAN machine,
this means we must skip the empty high order bytes when
calculating the bit offset. */
if (bytes < UNITS_PER_WORD
#ifdef BLOCK_REG_PADDING
&& (BLOCK_REG_PADDING (args[i].mode,
TREE_TYPE (args[i].tree_value), 1)
== PAD_DOWNWARD)
#else
&& BYTES_BIG_ENDIAN
#endif
)
endian_correction = BITS_PER_WORD - bytes * BITS_PER_UNIT;
for (j = 0; j < args[i].n_aligned_regs; j++)
{
rtx reg = gen_reg_rtx (word_mode);
rtx word = operand_subword_force (args[i].value, j, BLKmode);
int bitsize = MIN (bytes * BITS_PER_UNIT, BITS_PER_WORD);
args[i].aligned_regs[j] = reg;
word = extract_bit_field (word, bitsize, 0, 1, NULL_RTX,
word_mode, word_mode, false, NULL);
/* There is no need to restrict this code to loading items
in TYPE_ALIGN sized hunks. The bitfield instructions can
load up entire word sized registers efficiently.
??? This may not be needed anymore.
We use to emit a clobber here but that doesn't let later
passes optimize the instructions we emit. By storing 0 into
the register later passes know the first AND to zero out the
bitfield being set in the register is unnecessary. The store
of 0 will be deleted as will at least the first AND. */
emit_move_insn (reg, const0_rtx);
bytes -= bitsize / BITS_PER_UNIT;
store_bit_field (reg, bitsize, endian_correction, 0, 0,
word_mode, word, false);
}
}
}
/* The limit set by -Walloc-larger-than=. */
static GTY(()) tree alloc_object_size_limit;
/* Initialize ALLOC_OBJECT_SIZE_LIMIT based on the -Walloc-size-larger-than=
setting if the option is specified, or to the maximum object size if it
is not. Return the initialized value. */
static tree
alloc_max_size (void)
{
if (!alloc_object_size_limit)
{
alloc_object_size_limit = TYPE_MAX_VALUE (ssizetype);
if (warn_alloc_size_limit)
{
char *end = NULL;
errno = 0;
unsigned HOST_WIDE_INT unit = 1;
unsigned HOST_WIDE_INT limit
= strtoull (warn_alloc_size_limit, &end, 10);
if (!errno)
{
if (end && *end)
{
/* Numeric option arguments are at most INT_MAX. Make it
possible to specify a larger value by accepting common
suffixes. */
if (!strcmp (end, "kB"))
unit = 1000;
else if (!strcasecmp (end, "KiB") || strcmp (end, "KB"))
unit = 1024;
else if (!strcmp (end, "MB"))
unit = HOST_WIDE_INT_UC (1000) * 1000;
else if (!strcasecmp (end, "MiB"))
unit = HOST_WIDE_INT_UC (1024) * 1024;
else if (!strcasecmp (end, "GB"))
unit = HOST_WIDE_INT_UC (1000) * 1000 * 1000;
else if (!strcasecmp (end, "GiB"))
unit = HOST_WIDE_INT_UC (1024) * 1024 * 1024;
else if (!strcasecmp (end, "TB"))
unit = HOST_WIDE_INT_UC (1000) * 1000 * 1000 * 1000;
else if (!strcasecmp (end, "TiB"))
unit = HOST_WIDE_INT_UC (1024) * 1024 * 1024 * 1024;
else if (!strcasecmp (end, "PB"))
unit = HOST_WIDE_INT_UC (1000) * 1000 * 1000 * 1000 * 1000;
else if (!strcasecmp (end, "PiB"))
unit = HOST_WIDE_INT_UC (1024) * 1024 * 1024 * 1024 * 1024;
else if (!strcasecmp (end, "EB"))
unit = HOST_WIDE_INT_UC (1000) * 1000 * 1000 * 1000 * 1000
* 1000;
else if (!strcasecmp (end, "EiB"))
unit = HOST_WIDE_INT_UC (1024) * 1024 * 1024 * 1024 * 1024
* 1024;
else
unit = 0;
}
if (unit)
{
widest_int w = wi::mul (limit, unit);
if (w < wi::to_widest (alloc_object_size_limit))
alloc_object_size_limit = wide_int_to_tree (ssizetype, w);
}
}
}
}
return alloc_object_size_limit;
}
/* Return true when EXP's range can be determined and set RANGE[] to it
after adjusting it if necessary to make EXP a valid size argument to
an allocation function declared with attribute alloc_size (whose
argument may be signed), or to a string manipulation function like
memset. */
bool
get_size_range (tree exp, tree range[2])
{
if (tree_fits_uhwi_p (exp))
{
/* EXP is a constant. */
range[0] = range[1] = exp;
return true;
}
wide_int min, max;
enum value_range_type range_type
= ((TREE_CODE (exp) == SSA_NAME && INTEGRAL_TYPE_P (TREE_TYPE (exp)))
? get_range_info (exp, &min, &max) : VR_VARYING);
if (range_type == VR_VARYING)
{
/* No range information available. */
range[0] = NULL_TREE;
range[1] = NULL_TREE;
return false;
}
tree exptype = TREE_TYPE (exp);
unsigned expprec = TYPE_PRECISION (exptype);
bool signed_p = !TYPE_UNSIGNED (exptype);
if (range_type == VR_ANTI_RANGE)
{
if (signed_p)
{
if (wi::les_p (max, 0))
{
/* EXP is not in a strictly negative range. That means
it must be in some (not necessarily strictly) positive
range which includes zero. Since in signed to unsigned
conversions negative values end up converted to large
positive values, and otherwise they are not valid sizes,
the resulting range is in both cases [0, TYPE_MAX]. */
min = wi::zero (expprec);
max = wi::to_wide (TYPE_MAX_VALUE (exptype));
}
else if (wi::les_p (min - 1, 0))
{
/* EXP is not in a negative-positive range. That means EXP
is either negative, or greater than max. Since negative
sizes are invalid make the range [MAX + 1, TYPE_MAX]. */
min = max + 1;
max = wi::to_wide (TYPE_MAX_VALUE (exptype));
}
else
{
max = min - 1;
min = wi::zero (expprec);
}
}
else if (wi::eq_p (0, min - 1))
{
/* EXP is unsigned and not in the range [1, MAX]. That means
it's either zero or greater than MAX. Even though 0 would
normally be detected by -Walloc-zero set the range to
[MAX, TYPE_MAX] so that when MAX is greater than the limit
the whole range is diagnosed. */
min = max + 1;
max = wi::to_wide (TYPE_MAX_VALUE (exptype));
}
else
{
max = min - 1;
min = wi::zero (expprec);
}
}
range[0] = wide_int_to_tree (exptype, min);
range[1] = wide_int_to_tree (exptype, max);
return true;
}
/* Diagnose a call EXP to function FN decorated with attribute alloc_size
whose argument numbers given by IDX with values given by ARGS exceed
the maximum object size or cause an unsigned oveflow (wrapping) when
multiplied. When ARGS[0] is null the function does nothing. ARGS[1]
may be null for functions like malloc, and non-null for those like
calloc that are decorated with a two-argument attribute alloc_size. */
void
maybe_warn_alloc_args_overflow (tree fn, tree exp, tree args[2], int idx[2])
{
/* The range each of the (up to) two arguments is known to be in. */
tree argrange[2][2] = { { NULL_TREE, NULL_TREE }, { NULL_TREE, NULL_TREE } };
/* Maximum object size set by -Walloc-size-larger-than= or SIZE_MAX / 2. */
tree maxobjsize = alloc_max_size ();
location_t loc = EXPR_LOCATION (exp);
bool warned = false;
/* Validate each argument individually. */
for (unsigned i = 0; i != 2 && args[i]; ++i)
{
if (TREE_CODE (args[i]) == INTEGER_CST)
{
argrange[i][0] = args[i];
argrange[i][1] = args[i];
if (tree_int_cst_lt (args[i], integer_zero_node))
{
warned = warning_at (loc, OPT_Walloc_size_larger_than_,
"%Kargument %i value %qE is negative",
exp, idx[i] + 1, args[i]);
}
else if (integer_zerop (args[i]))
{
/* Avoid issuing -Walloc-zero for allocation functions other
than __builtin_alloca that are declared with attribute
returns_nonnull because there's no portability risk. This
avoids warning for such calls to libiberty's xmalloc and
friends.
Also avoid issuing the warning for calls to function named
"alloca". */
if ((DECL_FUNCTION_CODE (fn) == BUILT_IN_ALLOCA
&& IDENTIFIER_LENGTH (DECL_NAME (fn)) != 6)
|| (DECL_FUNCTION_CODE (fn) != BUILT_IN_ALLOCA
&& !lookup_attribute ("returns_nonnull",
TYPE_ATTRIBUTES (TREE_TYPE (fn)))))
warned = warning_at (loc, OPT_Walloc_zero,
"%Kargument %i value is zero",
exp, idx[i] + 1);
}
else if (tree_int_cst_lt (maxobjsize, args[i]))
{
/* G++ emits calls to ::operator new[](SIZE_MAX) in C++98
mode and with -fno-exceptions as a way to indicate array
size overflow. There's no good way to detect C++98 here
so avoid diagnosing these calls for all C++ modes. */
if (i == 0
&& !args[1]
&& lang_GNU_CXX ()
&& DECL_IS_OPERATOR_NEW (fn)
&& integer_all_onesp (args[i]))
continue;
warned = warning_at (loc, OPT_Walloc_size_larger_than_,
"%Kargument %i value %qE exceeds "
"maximum object size %E",
exp, idx[i] + 1, args[i], maxobjsize);
}
}
else if (TREE_CODE (args[i]) == SSA_NAME
&& get_size_range (args[i], argrange[i]))
{
/* Verify that the argument's range is not negative (including
upper bound of zero). */
if (tree_int_cst_lt (argrange[i][0], integer_zero_node)
&& tree_int_cst_le (argrange[i][1], integer_zero_node))
{
warned = warning_at (loc, OPT_Walloc_size_larger_than_,
"%Kargument %i range [%E, %E] is negative",
exp, idx[i] + 1,
argrange[i][0], argrange[i][1]);
}
else if (tree_int_cst_lt (maxobjsize, argrange[i][0]))
{
warned = warning_at (loc, OPT_Walloc_size_larger_than_,
"%Kargument %i range [%E, %E] exceeds "
"maximum object size %E",
exp, idx[i] + 1,
argrange[i][0], argrange[i][1],
maxobjsize);
}
}
}
if (!argrange[0])
return;
/* For a two-argument alloc_size, validate the product of the two
arguments if both of their values or ranges are known. */
if (!warned && tree_fits_uhwi_p (argrange[0][0])
&& argrange[1][0] && tree_fits_uhwi_p (argrange[1][0])
&& !integer_onep (argrange[0][0])
&& !integer_onep (argrange[1][0]))
{
/* Check for overflow in the product of a function decorated with
attribute alloc_size (X, Y). */
unsigned szprec = TYPE_PRECISION (size_type_node);
wide_int x = wi::to_wide (argrange[0][0], szprec);
wide_int y = wi::to_wide (argrange[1][0], szprec);
bool vflow;
wide_int prod = wi::umul (x, y, &vflow);
if (vflow)
warned = warning_at (loc, OPT_Walloc_size_larger_than_,
"%Kproduct %<%E * %E%> of arguments %i and %i "
"exceeds %<SIZE_MAX%>",
exp, argrange[0][0], argrange[1][0],
idx[0] + 1, idx[1] + 1);
else if (wi::ltu_p (wi::to_wide (maxobjsize, szprec), prod))
warned = warning_at (loc, OPT_Walloc_size_larger_than_,
"%Kproduct %<%E * %E%> of arguments %i and %i "
"exceeds maximum object size %E",
exp, argrange[0][0], argrange[1][0],
idx[0] + 1, idx[1] + 1,
maxobjsize);
if (warned)
{
/* Print the full range of each of the two arguments to make
it clear when it is, in fact, in a range and not constant. */
if (argrange[0][0] != argrange [0][1])
inform (loc, "argument %i in the range [%E, %E]",
idx[0] + 1, argrange[0][0], argrange[0][1]);
if (argrange[1][0] != argrange [1][1])
inform (loc, "argument %i in the range [%E, %E]",
idx[1] + 1, argrange[1][0], argrange[1][1]);
}
}
if (warned)
{
location_t fnloc = DECL_SOURCE_LOCATION (fn);
if (DECL_IS_BUILTIN (fn))
inform (loc,
"in a call to built-in allocation function %qD", fn);
else
inform (fnloc,
"in a call to allocation function %qD declared here", fn);
}
}
/* If EXPR refers to a character array or pointer declared attribute
nonstring return a decl for that array or pointer and set *REF to
the referenced enclosing object or pointer. Otherwise returns
null. */
tree
get_attr_nonstring_decl (tree expr, tree *ref)
{
tree decl = expr;
if (TREE_CODE (decl) == SSA_NAME)
{
gimple *def = SSA_NAME_DEF_STMT (decl);
if (is_gimple_assign (def))
{
tree_code code = gimple_assign_rhs_code (def);
if (code == ADDR_EXPR
|| code == COMPONENT_REF
|| code == VAR_DECL)
decl = gimple_assign_rhs1 (def);
}
else if (tree var = SSA_NAME_VAR (decl))
decl = var;
}
if (TREE_CODE (decl) == ADDR_EXPR)
decl = TREE_OPERAND (decl, 0);
if (ref)
*ref = decl;
if (TREE_CODE (decl) == COMPONENT_REF)
decl = TREE_OPERAND (decl, 1);
if (DECL_P (decl)
&& lookup_attribute ("nonstring", DECL_ATTRIBUTES (decl)))
return decl;
return NULL_TREE;
}
/* Warn about passing a non-string array/pointer to a function that
expects a nul-terminated string argument. */
void
maybe_warn_nonstring_arg (tree fndecl, tree exp)
{
if (!fndecl || DECL_BUILT_IN_CLASS (fndecl) != BUILT_IN_NORMAL)
return;
bool with_bounds = CALL_WITH_BOUNDS_P (exp);
/* The bound argument to a bounded string function like strncpy. */
tree bound = NULL_TREE;
/* It's safe to call "bounded" string functions with a non-string
argument since the functions provide an explicit bound for this
purpose. */
switch (DECL_FUNCTION_CODE (fndecl))
{
case BUILT_IN_STPNCPY:
case BUILT_IN_STPNCPY_CHK:
case BUILT_IN_STRNCMP:
case BUILT_IN_STRNCASECMP:
case BUILT_IN_STRNCPY:
case BUILT_IN_STRNCPY_CHK:
bound = CALL_EXPR_ARG (exp, with_bounds ? 4 : 2);
break;
case BUILT_IN_STRNDUP:
bound = CALL_EXPR_ARG (exp, with_bounds ? 2 : 1);
break;
default:
break;
}
/* Determine the range of the bound argument (if specified). */
tree bndrng[2] = { NULL_TREE, NULL_TREE };
if (bound)
get_size_range (bound, bndrng);
/* Iterate over the built-in function's formal arguments and check
each const char* against the actual argument. If the actual
argument is declared attribute non-string issue a warning unless
the argument's maximum length is bounded. */
function_args_iterator it;
function_args_iter_init (&it, TREE_TYPE (fndecl));
for (unsigned argno = 0; ; ++argno, function_args_iter_next (&it))
{
tree argtype = function_args_iter_cond (&it);
if (!argtype)
break;
if (TREE_CODE (argtype) != POINTER_TYPE)
continue;
argtype = TREE_TYPE (argtype);
if (TREE_CODE (argtype) != INTEGER_TYPE
|| !TYPE_READONLY (argtype))
continue;
argtype = TYPE_MAIN_VARIANT (argtype);
if (argtype != char_type_node)
continue;
tree callarg = CALL_EXPR_ARG (exp, argno);
if (TREE_CODE (callarg) == ADDR_EXPR)
callarg = TREE_OPERAND (callarg, 0);
/* See if the destination is declared with attribute "nonstring". */
tree decl = get_attr_nonstring_decl (callarg);
if (!decl)
continue;
tree type = TREE_TYPE (decl);
offset_int wibnd = 0;
if (bndrng[0])
wibnd = wi::to_offset (bndrng[0]);
offset_int asize = wibnd;
if (TREE_CODE (type) == ARRAY_TYPE)
if (tree arrbnd = TYPE_DOMAIN (type))
{
if ((arrbnd = TYPE_MAX_VALUE (arrbnd)))
asize = wi::to_offset (arrbnd) + 1;
}
location_t loc = EXPR_LOCATION (exp);
bool warned = false;
if (wi::ltu_p (asize, wibnd))
warned = warning_at (loc, OPT_Wstringop_overflow_,
"%qD argument %i declared attribute %<nonstring%> "
"is smaller than the specified bound %E",
fndecl, argno + 1, bndrng[0]);
else if (!bound)
warned = warning_at (loc, OPT_Wstringop_overflow_,
"%qD argument %i declared attribute %<nonstring%>",
fndecl, argno + 1);
if (warned)
inform (DECL_SOURCE_LOCATION (decl),
"argument %qD declared here", decl);
}
}
/* Issue an error if CALL_EXPR was flagged as requiring
tall-call optimization. */
static void
maybe_complain_about_tail_call (tree call_expr, const char *reason)
{
gcc_assert (TREE_CODE (call_expr) == CALL_EXPR);
if (!CALL_EXPR_MUST_TAIL_CALL (call_expr))
return;
error_at (EXPR_LOCATION (call_expr), "cannot tail-call: %s", reason);
}
/* Fill in ARGS_SIZE and ARGS array based on the parameters found in
CALL_EXPR EXP.
NUM_ACTUALS is the total number of parameters.
N_NAMED_ARGS is the total number of named arguments.
STRUCT_VALUE_ADDR_VALUE is the implicit argument for a struct return
value, or null.
FNDECL is the tree code for the target of this call (if known)
ARGS_SO_FAR holds state needed by the target to know where to place
the next argument.
REG_PARM_STACK_SPACE is the number of bytes of stack space reserved
for arguments which are passed in registers.
OLD_STACK_LEVEL is a pointer to an rtx which olds the old stack level
and may be modified by this routine.
OLD_PENDING_ADJ, MUST_PREALLOCATE and FLAGS are pointers to integer
flags which may be modified by this routine.
MAY_TAILCALL is cleared if we encounter an invisible pass-by-reference
that requires allocation of stack space.
CALL_FROM_THUNK_P is true if this call is the jump from a thunk to
the thunked-to function. */
static void
initialize_argument_information (int num_actuals ATTRIBUTE_UNUSED,
struct arg_data *args,
struct args_size *args_size,
int n_named_args ATTRIBUTE_UNUSED,
tree exp, tree struct_value_addr_value,
tree fndecl, tree fntype,
cumulative_args_t args_so_far,
int reg_parm_stack_space,
rtx *old_stack_level, int *old_pending_adj,
int *must_preallocate, int *ecf_flags,
bool *may_tailcall, bool call_from_thunk_p)
{
CUMULATIVE_ARGS *args_so_far_pnt = get_cumulative_args (args_so_far);
location_t loc = EXPR_LOCATION (exp);
/* Count arg position in order args appear. */
int argpos;
int i;
args_size->constant = 0;
args_size->var = 0;
bitmap_obstack_initialize (NULL);
/* In this loop, we consider args in the order they are written.
We fill up ARGS from the back. */
i = num_actuals - 1;
{
int j = i, ptr_arg = -1;
call_expr_arg_iterator iter;
tree arg;
bitmap slots = NULL;
if (struct_value_addr_value)
{
args[j].tree_value = struct_value_addr_value;
j--;
/* If we pass structure address then we need to
create bounds for it. Since created bounds is
a call statement, we expand it right here to avoid
fixing all other places where it may be expanded. */
if (CALL_WITH_BOUNDS_P (exp))
{
args[j].value = gen_reg_rtx (targetm.chkp_bound_mode ());
args[j].tree_value
= chkp_make_bounds_for_struct_addr (struct_value_addr_value);
expand_expr_real (args[j].tree_value, args[j].value, VOIDmode,
EXPAND_NORMAL, 0, false);
args[j].pointer_arg = j + 1;
j--;
}
}
argpos = 0;
FOR_EACH_CALL_EXPR_ARG (arg, iter, exp)
{
tree argtype = TREE_TYPE (arg);
/* Remember last param with pointer and associate it
with following pointer bounds. */
if (CALL_WITH_BOUNDS_P (exp)
&& chkp_type_has_pointer (argtype))
{
if (slots)
BITMAP_FREE (slots);
ptr_arg = j;
if (!BOUNDED_TYPE_P (argtype))
{
slots = BITMAP_ALLOC (NULL);
chkp_find_bound_slots (argtype, slots);
}
}
else if (CALL_WITH_BOUNDS_P (exp)
&& pass_by_reference (NULL, TYPE_MODE (argtype), argtype,
argpos < n_named_args))
{
if (slots)
BITMAP_FREE (slots);
ptr_arg = j;
}
else if (POINTER_BOUNDS_TYPE_P (argtype))
{
/* We expect bounds in instrumented calls only.
Otherwise it is a sign we lost flag due to some optimization
and may emit call args incorrectly. */
gcc_assert (CALL_WITH_BOUNDS_P (exp));
/* For structures look for the next available pointer. */
if (ptr_arg != -1 && slots)
{
unsigned bnd_no = bitmap_first_set_bit (slots);
args[j].pointer_offset =
bnd_no * POINTER_SIZE / BITS_PER_UNIT;
bitmap_clear_bit (slots, bnd_no);
/* Check we have no more pointers in the structure. */
if (bitmap_empty_p (slots))
BITMAP_FREE (slots);
}
args[j].pointer_arg = ptr_arg;
/* Check we covered all pointers in the previous
non bounds arg. */
if (!slots)
ptr_arg = -1;
}
else
ptr_arg = -1;
if (targetm.calls.split_complex_arg
&& argtype
&& TREE_CODE (argtype) == COMPLEX_TYPE
&& targetm.calls.split_complex_arg (argtype))
{
tree subtype = TREE_TYPE (argtype);
args[j].tree_value = build1 (REALPART_EXPR, subtype, arg);
j--;
args[j].tree_value = build1 (IMAGPART_EXPR, subtype, arg);
}
else
args[j].tree_value = arg;
j--;
argpos++;
}
if (slots)
BITMAP_FREE (slots);
}
bitmap_obstack_release (NULL);
/* Extract attribute alloc_size and if set, store the indices of
the corresponding arguments in ALLOC_IDX, and then the actual
argument(s) at those indices in ALLOC_ARGS. */
int alloc_idx[2] = { -1, -1 };
if (tree alloc_size
= (fndecl ? lookup_attribute ("alloc_size",
TYPE_ATTRIBUTES (TREE_TYPE (fndecl)))
: NULL_TREE))
{
tree args = TREE_VALUE (alloc_size);
alloc_idx[0] = TREE_INT_CST_LOW (TREE_VALUE (args)) - 1;
if (TREE_CHAIN (args))
alloc_idx[1] = TREE_INT_CST_LOW (TREE_VALUE (TREE_CHAIN (args))) - 1;
}
/* Array for up to the two attribute alloc_size arguments. */
tree alloc_args[] = { NULL_TREE, NULL_TREE };
/* I counts args in order (to be) pushed; ARGPOS counts in order written. */
for (argpos = 0; argpos < num_actuals; i--, argpos++)
{
tree type = TREE_TYPE (args[i].tree_value);
int unsignedp;
machine_mode mode;
/* Replace erroneous argument with constant zero. */
if (type == error_mark_node || !COMPLETE_TYPE_P (type))
args[i].tree_value = integer_zero_node, type = integer_type_node;
/* If TYPE is a transparent union or record, pass things the way
we would pass the first field of the union or record. We have
already verified that the modes are the same. */
if ((TREE_CODE (type) == UNION_TYPE || TREE_CODE (type) == RECORD_TYPE)
&& TYPE_TRANSPARENT_AGGR (type))
type = TREE_TYPE (first_field (type));
/* Decide where to pass this arg.
args[i].reg is nonzero if all or part is passed in registers.
args[i].partial is nonzero if part but not all is passed in registers,
and the exact value says how many bytes are passed in registers.
args[i].pass_on_stack is nonzero if the argument must at least be
computed on the stack. It may then be loaded back into registers
if args[i].reg is nonzero.
These decisions are driven by the FUNCTION_... macros and must agree
with those made by function.c. */
/* See if this argument should be passed by invisible reference. */
if (pass_by_reference (args_so_far_pnt, TYPE_MODE (type),
type, argpos < n_named_args))
{
bool callee_copies;
tree base = NULL_TREE;
callee_copies
= reference_callee_copied (args_so_far_pnt, TYPE_MODE (type),
type, argpos < n_named_args);
/* If we're compiling a thunk, pass through invisible references
instead of making a copy. */
if (call_from_thunk_p
|| (callee_copies
&& !TREE_ADDRESSABLE (type)
&& (base = get_base_address (args[i].tree_value))
&& TREE_CODE (base) != SSA_NAME
&& (!DECL_P (base) || MEM_P (DECL_RTL (base)))))
{
/* We may have turned the parameter value into an SSA name.
Go back to the original parameter so we can take the
address. */
if (TREE_CODE (args[i].tree_value) == SSA_NAME)
{
gcc_assert (SSA_NAME_IS_DEFAULT_DEF (args[i].tree_value));
args[i].tree_value = SSA_NAME_VAR (args[i].tree_value);
gcc_assert (TREE_CODE (args[i].tree_value) == PARM_DECL);
}
/* Argument setup code may have copied the value to register. We
revert that optimization now because the tail call code must
use the original location. */
if (TREE_CODE (args[i].tree_value) == PARM_DECL
&& !MEM_P (DECL_RTL (args[i].tree_value))
&& DECL_INCOMING_RTL (args[i].tree_value)
&& MEM_P (DECL_INCOMING_RTL (args[i].tree_value)))
set_decl_rtl (args[i].tree_value,
DECL_INCOMING_RTL (args[i].tree_value));
mark_addressable (args[i].tree_value);
/* We can't use sibcalls if a callee-copied argument is
stored in the current function's frame. */
if (!call_from_thunk_p && DECL_P (base) && !TREE_STATIC (base))
{
*may_tailcall = false;
maybe_complain_about_tail_call (exp,
"a callee-copied argument is"
" stored in the current "
" function's frame");
}
args[i].tree_value = build_fold_addr_expr_loc (loc,
args[i].tree_value);
type = TREE_TYPE (args[i].tree_value);
if (*ecf_flags & ECF_CONST)
*ecf_flags &= ~(ECF_CONST | ECF_LOOPING_CONST_OR_PURE);
}
else
{
/* We make a copy of the object and pass the address to the
function being called. */
rtx copy;
if (!COMPLETE_TYPE_P (type)
|| TREE_CODE (TYPE_SIZE_UNIT (type)) != INTEGER_CST
|| (flag_stack_check == GENERIC_STACK_CHECK
&& compare_tree_int (TYPE_SIZE_UNIT (type),
STACK_CHECK_MAX_VAR_SIZE) > 0))
{
/* This is a variable-sized object. Make space on the stack
for it. */
rtx size_rtx = expr_size (args[i].tree_value);
if (*old_stack_level == 0)
{
emit_stack_save (SAVE_BLOCK, old_stack_level);
*old_pending_adj = pending_stack_adjust;
pending_stack_adjust = 0;
}
/* We can pass TRUE as the 4th argument because we just
saved the stack pointer and will restore it right after
the call. */
copy = allocate_dynamic_stack_space (size_rtx,
TYPE_ALIGN (type),
TYPE_ALIGN (type),
max_int_size_in_bytes
(type),
true);
copy = gen_rtx_MEM (BLKmode, copy);
set_mem_attributes (copy, type, 1);
}
else
copy = assign_temp (type, 1, 0);
store_expr (args[i].tree_value, copy, 0, false, false);
/* Just change the const function to pure and then let
the next test clear the pure based on
callee_copies. */
if (*ecf_flags & ECF_CONST)
{
*ecf_flags &= ~ECF_CONST;
*ecf_flags |= ECF_PURE;
}
if (!callee_copies && *ecf_flags & ECF_PURE)
*ecf_flags &= ~(ECF_PURE | ECF_LOOPING_CONST_OR_PURE);
args[i].tree_value
= build_fold_addr_expr_loc (loc, make_tree (type, copy));
type = TREE_TYPE (args[i].tree_value);
*may_tailcall = false;
maybe_complain_about_tail_call (exp,
"argument must be passed"
" by copying");
}
}
unsignedp = TYPE_UNSIGNED (type);
mode = promote_function_mode (type, TYPE_MODE (type), &unsignedp,
fndecl ? TREE_TYPE (fndecl) : fntype, 0);
args[i].unsignedp = unsignedp;
args[i].mode = mode;
targetm.calls.warn_parameter_passing_abi (args_so_far, type);
args[i].reg = targetm.calls.function_arg (args_so_far, mode, type,
argpos < n_named_args);
if (args[i].reg && CONST_INT_P (args[i].reg))
{
args[i].special_slot = args[i].reg;
args[i].reg = NULL;
}
/* If this is a sibling call and the machine has register windows, the
register window has to be unwinded before calling the routine, so
arguments have to go into the incoming registers. */
if (targetm.calls.function_incoming_arg != targetm.calls.function_arg)
args[i].tail_call_reg
= targetm.calls.function_incoming_arg (args_so_far, mode, type,
argpos < n_named_args);
else
args[i].tail_call_reg = args[i].reg;
if (args[i].reg)
args[i].partial
= targetm.calls.arg_partial_bytes (args_so_far, mode, type,
argpos < n_named_args);
args[i].pass_on_stack = targetm.calls.must_pass_in_stack (mode, type);
/* If FUNCTION_ARG returned a (parallel [(expr_list (nil) ...) ...]),
it means that we are to pass this arg in the register(s) designated
by the PARALLEL, but also to pass it in the stack. */
if (args[i].reg && GET_CODE (args[i].reg) == PARALLEL
&& XEXP (XVECEXP (args[i].reg, 0, 0), 0) == 0)
args[i].pass_on_stack = 1;
/* If this is an addressable type, we must preallocate the stack
since we must evaluate the object into its final location.
If this is to be passed in both registers and the stack, it is simpler
to preallocate. */
if (TREE_ADDRESSABLE (type)
|| (args[i].pass_on_stack && args[i].reg != 0))
*must_preallocate = 1;
/* No stack allocation and padding for bounds. */
if (POINTER_BOUNDS_P (args[i].tree_value))
;
/* Compute the stack-size of this argument. */
else if (args[i].reg == 0 || args[i].partial != 0
|| reg_parm_stack_space > 0
|| args[i].pass_on_stack)
locate_and_pad_parm (mode, type,
#ifdef STACK_PARMS_IN_REG_PARM_AREA
1,
#else
args[i].reg != 0,
#endif
reg_parm_stack_space,
args[i].pass_on_stack ? 0 : args[i].partial,
fndecl, args_size, &args[i].locate);
#ifdef BLOCK_REG_PADDING
else
/* The argument is passed entirely in registers. See at which
end it should be padded. */
args[i].locate.where_pad =
BLOCK_REG_PADDING (mode, type,
int_size_in_bytes (type) <= UNITS_PER_WORD);
#endif
/* Update ARGS_SIZE, the total stack space for args so far. */
args_size->constant += args[i].locate.size.constant;
if (args[i].locate.size.var)
ADD_PARM_SIZE (*args_size, args[i].locate.size.var);
/* Increment ARGS_SO_FAR, which has info about which arg-registers
have been used, etc. */
targetm.calls.function_arg_advance (args_so_far, TYPE_MODE (type),
type, argpos < n_named_args);
/* Store argument values for functions decorated with attribute
alloc_size. */
if (argpos == alloc_idx[0])
alloc_args[0] = args[i].tree_value;
else if (argpos == alloc_idx[1])
alloc_args[1] = args[i].tree_value;
}
if (alloc_args[0])
{
/* Check the arguments of functions decorated with attribute
alloc_size. */
maybe_warn_alloc_args_overflow (fndecl, exp, alloc_args, alloc_idx);
}
/* Detect passing non-string arguments to functions expecting
nul-terminated strings. */
maybe_warn_nonstring_arg (fndecl, exp);
}
/* Update ARGS_SIZE to contain the total size for the argument block.
Return the original constant component of the argument block's size.
REG_PARM_STACK_SPACE holds the number of bytes of stack space reserved
for arguments passed in registers. */
static int
compute_argument_block_size (int reg_parm_stack_space,
struct args_size *args_size,
tree fndecl ATTRIBUTE_UNUSED,
tree fntype ATTRIBUTE_UNUSED,
int preferred_stack_boundary ATTRIBUTE_UNUSED)
{
int unadjusted_args_size = args_size->constant;
/* For accumulate outgoing args mode we don't need to align, since the frame
will be already aligned. Align to STACK_BOUNDARY in order to prevent
backends from generating misaligned frame sizes. */
if (ACCUMULATE_OUTGOING_ARGS && preferred_stack_boundary > STACK_BOUNDARY)
preferred_stack_boundary = STACK_BOUNDARY;
/* Compute the actual size of the argument block required. The variable
and constant sizes must be combined, the size may have to be rounded,
and there may be a minimum required size. */
if (args_size->var)
{
args_size->var = ARGS_SIZE_TREE (*args_size);
args_size->constant = 0;
preferred_stack_boundary /= BITS_PER_UNIT;
if (preferred_stack_boundary > 1)
{
/* We don't handle this case yet. To handle it correctly we have
to add the delta, round and subtract the delta.
Currently no machine description requires this support. */
gcc_assert (!(stack_pointer_delta & (preferred_stack_boundary - 1)));
args_size->var = round_up (args_size->var, preferred_stack_boundary);
}
if (reg_parm_stack_space > 0)
{
args_size->var
= size_binop (MAX_EXPR, args_size->var,
ssize_int (reg_parm_stack_space));
/* The area corresponding to register parameters is not to count in
the size of the block we need. So make the adjustment. */
if (! OUTGOING_REG_PARM_STACK_SPACE ((!fndecl ? fntype : TREE_TYPE (fndecl))))
args_size->var
= size_binop (MINUS_EXPR, args_size->var,
ssize_int (reg_parm_stack_space));
}
}
else
{
preferred_stack_boundary /= BITS_PER_UNIT;
if (preferred_stack_boundary < 1)
preferred_stack_boundary = 1;
args_size->constant = (((args_size->constant
+ stack_pointer_delta
+ preferred_stack_boundary - 1)
/ preferred_stack_boundary
* preferred_stack_boundary)
- stack_pointer_delta);
args_size->constant = MAX (args_size->constant,
reg_parm_stack_space);
if (! OUTGOING_REG_PARM_STACK_SPACE ((!fndecl ? fntype : TREE_TYPE (fndecl))))
args_size->constant -= reg_parm_stack_space;
}
return unadjusted_args_size;
}
/* Precompute parameters as needed for a function call.
FLAGS is mask of ECF_* constants.
NUM_ACTUALS is the number of arguments.
ARGS is an array containing information for each argument; this
routine fills in the INITIAL_VALUE and VALUE fields for each
precomputed argument. */
static void
precompute_arguments (int num_actuals, struct arg_data *args)
{
int i;
/* If this is a libcall, then precompute all arguments so that we do not
get extraneous instructions emitted as part of the libcall sequence. */
/* If we preallocated the stack space, and some arguments must be passed
on the stack, then we must precompute any parameter which contains a
function call which will store arguments on the stack.
Otherwise, evaluating the parameter may clobber previous parameters
which have already been stored into the stack. (we have code to avoid
such case by saving the outgoing stack arguments, but it results in
worse code) */
if (!ACCUMULATE_OUTGOING_ARGS)
return;
for (i = 0; i < num_actuals; i++)
{
tree type;
machine_mode mode;
if (TREE_CODE (args[i].tree_value) != CALL_EXPR)
continue;
/* If this is an addressable type, we cannot pre-evaluate it. */
type = TREE_TYPE (args[i].tree_value);
gcc_assert (!TREE_ADDRESSABLE (type));
args[i].initial_value = args[i].value
= expand_normal (args[i].tree_value);
mode = TYPE_MODE (type);
if (mode != args[i].mode)
{
int unsignedp = args[i].unsignedp;
args[i].value
= convert_modes (args[i].mode, mode,
args[i].value, args[i].unsignedp);
/* CSE will replace this only if it contains args[i].value
pseudo, so convert it down to the declared mode using
a SUBREG. */
if (REG_P (args[i].value)
&& GET_MODE_CLASS (args[i].mode) == MODE_INT
&& promote_mode (type, mode, &unsignedp) != args[i].mode)
{
args[i].initial_value
= gen_lowpart_SUBREG (mode, args[i].value);
SUBREG_PROMOTED_VAR_P (args[i].initial_value) = 1;
SUBREG_PROMOTED_SET (args[i].initial_value, args[i].unsignedp);
}
}
}
}
/* Given the current state of MUST_PREALLOCATE and information about
arguments to a function call in NUM_ACTUALS, ARGS and ARGS_SIZE,
compute and return the final value for MUST_PREALLOCATE. */
static int
finalize_must_preallocate (int must_preallocate, int num_actuals,
struct arg_data *args, struct args_size *args_size)
{
/* See if we have or want to preallocate stack space.
If we would have to push a partially-in-regs parm
before other stack parms, preallocate stack space instead.
If the size of some parm is not a multiple of the required stack
alignment, we must preallocate.
If the total size of arguments that would otherwise create a copy in
a temporary (such as a CALL) is more than half the total argument list
size, preallocation is faster.
Another reason to preallocate is if we have a machine (like the m88k)
where stack alignment is required to be maintained between every
pair of insns, not just when the call is made. However, we assume here
that such machines either do not have push insns (and hence preallocation
would occur anyway) or the problem is taken care of with
PUSH_ROUNDING. */
if (! must_preallocate)
{
int partial_seen = 0;
int copy_to_evaluate_size = 0;
int i;
for (i = 0; i < num_actuals && ! must_preallocate; i++)
{
if (args[i].partial > 0 && ! args[i].pass_on_stack)
partial_seen = 1;
else if (partial_seen && args[i].reg == 0)
must_preallocate = 1;
/* We preallocate in case there are bounds passed
in the bounds table to have precomputed address
for bounds association. */
else if (POINTER_BOUNDS_P (args[i].tree_value)
&& !args[i].reg)
must_preallocate = 1;
if (TYPE_MODE (TREE_TYPE (args[i].tree_value)) == BLKmode
&& (TREE_CODE (args[i].tree_value) == CALL_EXPR
|| TREE_CODE (args[i].tree_value) == TARGET_EXPR
|| TREE_CODE (args[i].tree_value) == COND_EXPR
|| TREE_ADDRESSABLE (TREE_TYPE (args[i].tree_value))))
copy_to_evaluate_size
+= int_size_in_bytes (TREE_TYPE (args[i].tree_value));
}
if (copy_to_evaluate_size * 2 >= args_size->constant
&& args_size->constant > 0)
must_preallocate = 1;
}
return must_preallocate;
}
/* If we preallocated stack space, compute the address of each argument
and store it into the ARGS array.
We need not ensure it is a valid memory address here; it will be
validized when it is used.
ARGBLOCK is an rtx for the address of the outgoing arguments. */
static void
compute_argument_addresses (struct arg_data *args, rtx argblock, int num_actuals)
{
if (argblock)
{
rtx arg_reg = argblock;
int i, arg_offset = 0;
if (GET_CODE (argblock) == PLUS)
arg_reg = XEXP (argblock, 0), arg_offset = INTVAL (XEXP (argblock, 1));
for (i = 0; i < num_actuals; i++)
{
rtx offset = ARGS_SIZE_RTX (args[i].locate.offset);
rtx slot_offset = ARGS_SIZE_RTX (args[i].locate.slot_offset);
rtx addr;
unsigned int align, boundary;
unsigned int units_on_stack = 0;
machine_mode partial_mode = VOIDmode;
/* Skip this parm if it will not be passed on the stack. */
if (! args[i].pass_on_stack
&& args[i].reg != 0
&& args[i].partial == 0)
continue;
/* Pointer Bounds are never passed on the stack. */
if (POINTER_BOUNDS_P (args[i].tree_value))
continue;
addr = simplify_gen_binary (PLUS, Pmode, arg_reg, offset);
addr = plus_constant (Pmode, addr, arg_offset);
if (args[i].partial != 0)
{
/* Only part of the parameter is being passed on the stack.
Generate a simple memory reference of the correct size. */
units_on_stack = args[i].locate.size.constant;
unsigned int bits_on_stack = units_on_stack * BITS_PER_UNIT;
partial_mode = int_mode_for_size (bits_on_stack, 1).else_blk ();
args[i].stack = gen_rtx_MEM (partial_mode, addr);
set_mem_size (args[i].stack, units_on_stack);
}
else
{
args[i].stack = gen_rtx_MEM (args[i].mode, addr);
set_mem_attributes (args[i].stack,
TREE_TYPE (args[i].tree_value), 1);
}
align = BITS_PER_UNIT;
boundary = args[i].locate.boundary;
if (args[i].locate.where_pad != PAD_DOWNWARD)
align = boundary;
else if (CONST_INT_P (offset))
{
align = INTVAL (offset) * BITS_PER_UNIT | boundary;
align = least_bit_hwi (align);
}
set_mem_align (args[i].stack, align);
addr = simplify_gen_binary (PLUS, Pmode, arg_reg, slot_offset);
addr = plus_constant (Pmode, addr, arg_offset);
if (args[i].partial != 0)
{
/* Only part of the parameter is being passed on the stack.
Generate a simple memory reference of the correct size.
*/
args[i].stack_slot = gen_rtx_MEM (partial_mode, addr);
set_mem_size (args[i].stack_slot, units_on_stack);
}
else
{
args[i].stack_slot = gen_rtx_MEM (args[i].mode, addr);
set_mem_attributes (args[i].stack_slot,
TREE_TYPE (args[i].tree_value), 1);
}
set_mem_align (args[i].stack_slot, args[i].locate.boundary);
/* Function incoming arguments may overlap with sibling call
outgoing arguments and we cannot allow reordering of reads
from function arguments with stores to outgoing arguments
of sibling calls. */
set_mem_alias_set (args[i].stack, 0);
set_mem_alias_set (args[i].stack_slot, 0);
}
}
}
/* Given a FNDECL and EXP, return an rtx suitable for use as a target address
in a call instruction.
FNDECL is the tree node for the target function. For an indirect call
FNDECL will be NULL_TREE.
ADDR is the operand 0 of CALL_EXPR for this call. */
static rtx
rtx_for_function_call (tree fndecl, tree addr)
{
rtx funexp;
/* Get the function to call, in the form of RTL. */
if (fndecl)
{
if (!TREE_USED (fndecl) && fndecl != current_function_decl)
TREE_USED (fndecl) = 1;
/* Get a SYMBOL_REF rtx for the function address. */
funexp = XEXP (DECL_RTL (fndecl), 0);
}
else
/* Generate an rtx (probably a pseudo-register) for the address. */
{
push_temp_slots ();
funexp = expand_normal (addr);
pop_temp_slots (); /* FUNEXP can't be BLKmode. */
}
return funexp;
}
/* Internal state for internal_arg_pointer_based_exp and its helpers. */
static struct
{
/* Last insn that has been scanned by internal_arg_pointer_based_exp_scan,
or NULL_RTX if none has been scanned yet. */
rtx_insn *scan_start;
/* Vector indexed by REGNO - FIRST_PSEUDO_REGISTER, recording if a pseudo is
based on crtl->args.internal_arg_pointer. The element is NULL_RTX if the
pseudo isn't based on it, a CONST_INT offset if the pseudo is based on it
with fixed offset, or PC if this is with variable or unknown offset. */
vec<rtx> cache;
} internal_arg_pointer_exp_state;
static rtx internal_arg_pointer_based_exp (const_rtx, bool);
/* Helper function for internal_arg_pointer_based_exp. Scan insns in
the tail call sequence, starting with first insn that hasn't been
scanned yet, and note for each pseudo on the LHS whether it is based
on crtl->args.internal_arg_pointer or not, and what offset from that
that pointer it has. */
static void
internal_arg_pointer_based_exp_scan (void)
{
rtx_insn *insn, *scan_start = internal_arg_pointer_exp_state.scan_start;
if (scan_start == NULL_RTX)
insn = get_insns ();
else
insn = NEXT_INSN (scan_start);
while (insn)
{
rtx set = single_set (insn);
if (set && REG_P (SET_DEST (set)) && !HARD_REGISTER_P (SET_DEST (set)))
{
rtx val = NULL_RTX;
unsigned int idx = REGNO (SET_DEST (set)) - FIRST_PSEUDO_REGISTER;
/* Punt on pseudos set multiple times. */
if (idx < internal_arg_pointer_exp_state.cache.length ()
&& (internal_arg_pointer_exp_state.cache[idx]
!= NULL_RTX))
val = pc_rtx;
else
val = internal_arg_pointer_based_exp (SET_SRC (set), false);
if (val != NULL_RTX)
{
if (idx >= internal_arg_pointer_exp_state.cache.length ())
internal_arg_pointer_exp_state.cache
.safe_grow_cleared (idx + 1);
internal_arg_pointer_exp_state.cache[idx] = val;
}
}
if (NEXT_INSN (insn) == NULL_RTX)
scan_start = insn;
insn = NEXT_INSN (insn);
}
internal_arg_pointer_exp_state.scan_start = scan_start;
}
/* Compute whether RTL is based on crtl->args.internal_arg_pointer. Return
NULL_RTX if RTL isn't based on it, a CONST_INT offset if RTL is based on
it with fixed offset, or PC if this is with variable or unknown offset.
TOPLEVEL is true if the function is invoked at the topmost level. */
static rtx
internal_arg_pointer_based_exp (const_rtx rtl, bool toplevel)
{
if (CONSTANT_P (rtl))
return NULL_RTX;
if (rtl == crtl->args.internal_arg_pointer)
return const0_rtx;
if (REG_P (rtl) && HARD_REGISTER_P (rtl))
return NULL_RTX;
if (GET_CODE (rtl) == PLUS && CONST_INT_P (XEXP (rtl, 1)))
{
rtx val = internal_arg_pointer_based_exp (XEXP (rtl, 0), toplevel);
if (val == NULL_RTX || val == pc_rtx)
return val;
return plus_constant (Pmode, val, INTVAL (XEXP (rtl, 1)));
}
/* When called at the topmost level, scan pseudo assignments in between the
last scanned instruction in the tail call sequence and the latest insn
in that sequence. */
if (toplevel)
internal_arg_pointer_based_exp_scan ();
if (REG_P (rtl))
{
unsigned int idx = REGNO (rtl) - FIRST_PSEUDO_REGISTER;
if (idx < internal_arg_pointer_exp_state.cache.length ())
return internal_arg_pointer_exp_state.cache[idx];
return NULL_RTX;
}
subrtx_iterator::array_type array;
FOR_EACH_SUBRTX (iter, array, rtl, NONCONST)
{
const_rtx x = *iter;
if (REG_P (x) && internal_arg_pointer_based_exp (x, false) != NULL_RTX)
return pc_rtx;
if (MEM_P (x))
iter.skip_subrtxes ();
}
return NULL_RTX;
}
/* Return true if and only if SIZE storage units (usually bytes)
starting from address ADDR overlap with already clobbered argument
area. This function is used to determine if we should give up a
sibcall. */
static bool
mem_overlaps_already_clobbered_arg_p (rtx addr, unsigned HOST_WIDE_INT size)
{
HOST_WIDE_INT i;
rtx val;
if (bitmap_empty_p (stored_args_map))
return false;
val = internal_arg_pointer_based_exp (addr, true);
if (val == NULL_RTX)
return false;
else if (val == pc_rtx)
return true;
else
i = INTVAL (val);
if (STACK_GROWS_DOWNWARD)
i -= crtl->args.pretend_args_size;
else
i += crtl->args.pretend_args_size;
if (ARGS_GROW_DOWNWARD)
i = -i - size;
if (size > 0)
{
unsigned HOST_WIDE_INT k;
for (k = 0; k < size; k++)
if (i + k < SBITMAP_SIZE (stored_args_map)
&& bitmap_bit_p (stored_args_map, i + k))
return true;
}
return false;
}
/* Do the register loads required for any wholly-register parms or any
parms which are passed both on the stack and in a register. Their
expressions were already evaluated.
Mark all register-parms as living through the call, putting these USE
insns in the CALL_INSN_FUNCTION_USAGE field.
When IS_SIBCALL, perform the check_sibcall_argument_overlap
checking, setting *SIBCALL_FAILURE if appropriate. */
static void
load_register_parameters (struct arg_data *args, int num_actuals,
rtx *call_fusage, int flags, int is_sibcall,
int *sibcall_failure)
{
int i, j;
for (i = 0; i < num_actuals; i++)
{
rtx reg = ((flags & ECF_SIBCALL)
? args[i].tail_call_reg : args[i].reg);
if (reg)
{
int partial = args[i].partial;
int nregs;
int size = 0;
rtx_insn *before_arg = get_last_insn ();
/* Set non-negative if we must move a word at a time, even if
just one word (e.g, partial == 4 && mode == DFmode). Set
to -1 if we just use a normal move insn. This value can be
zero if the argument is a zero size structure. */
nregs = -1;
if (GET_CODE (reg) == PARALLEL)
;
else if (partial)
{
gcc_assert (partial % UNITS_PER_WORD == 0);
nregs = partial / UNITS_PER_WORD;
}
else if (TYPE_MODE (TREE_TYPE (args[i].tree_value)) == BLKmode)
{
size = int_size_in_bytes (TREE_TYPE (args[i].tree_value));
nregs = (size + (UNITS_PER_WORD - 1)) / UNITS_PER_WORD;
}
else
size = GET_MODE_SIZE (args[i].mode);
/* Handle calls that pass values in multiple non-contiguous
locations. The Irix 6 ABI has examples of this. */
if (GET_CODE (reg) == PARALLEL)
emit_group_move (reg, args[i].parallel_value);
/* If simple case, just do move. If normal partial, store_one_arg
has already loaded the register for us. In all other cases,
load the register(s) from memory. */
else if (nregs == -1)
{
emit_move_insn (reg, args[i].value);
#ifdef BLOCK_REG_PADDING
/* Handle case where we have a value that needs shifting
up to the msb. eg. a QImode value and we're padding
upward on a BYTES_BIG_ENDIAN machine. */
if (size < UNITS_PER_WORD
&& (args[i].locate.where_pad
== (BYTES_BIG_ENDIAN ? PAD_UPWARD : PAD_DOWNWARD)))
{
rtx x;
int shift = (UNITS_PER_WORD - size) * BITS_PER_UNIT;
/* Assigning REG here rather than a temp makes CALL_FUSAGE
report the whole reg as used. Strictly speaking, the
call only uses SIZE bytes at the msb end, but it doesn't
seem worth generating rtl to say that. */
reg = gen_rtx_REG (word_mode, REGNO (reg));
x = expand_shift (LSHIFT_EXPR, word_mode, reg, shift, reg, 1);
if (x != reg)
emit_move_insn (reg, x);
}
#endif
}
/* If we have pre-computed the values to put in the registers in
the case of non-aligned structures, copy them in now. */
else if (args[i].n_aligned_regs != 0)
for (j = 0; j < args[i].n_aligned_regs; j++)
emit_move_insn (gen_rtx_REG (word_mode, REGNO (reg) + j),
args[i].aligned_regs[j]);
else if (partial == 0 || args[i].pass_on_stack)
{
rtx mem = validize_mem (copy_rtx (args[i].value));
/* Check for overlap with already clobbered argument area,
providing that this has non-zero size. */
if (is_sibcall
&& size != 0
&& (mem_overlaps_already_clobbered_arg_p
(XEXP (args[i].value, 0), size)))
*sibcall_failure = 1;
if (size % UNITS_PER_WORD == 0
|| MEM_ALIGN (mem) % BITS_PER_WORD == 0)
move_block_to_reg (REGNO (reg), mem, nregs, args[i].mode);
else
{
if (nregs > 1)
move_block_to_reg (REGNO (reg), mem, nregs - 1,
args[i].mode);
rtx dest = gen_rtx_REG (word_mode, REGNO (reg) + nregs - 1);
unsigned int bitoff = (nregs - 1) * BITS_PER_WORD;
unsigned int bitsize = size * BITS_PER_UNIT - bitoff;
rtx x = extract_bit_field (mem, bitsize, bitoff, 1, dest,
word_mode, word_mode, false,
NULL);
if (BYTES_BIG_ENDIAN)
x = expand_shift (LSHIFT_EXPR, word_mode, x,
BITS_PER_WORD - bitsize, dest, 1);
if (x != dest)
emit_move_insn (dest, x);
}
/* Handle a BLKmode that needs shifting. */
if (nregs == 1 && size < UNITS_PER_WORD
#ifdef BLOCK_REG_PADDING
&& args[i].locate.where_pad == PAD_DOWNWARD
#else
&& BYTES_BIG_ENDIAN
#endif
)
{
rtx dest = gen_rtx_REG (word_mode, REGNO (reg));
int shift = (UNITS_PER_WORD - size) * BITS_PER_UNIT;
enum tree_code dir = (BYTES_BIG_ENDIAN
? RSHIFT_EXPR : LSHIFT_EXPR);
rtx x;
x = expand_shift (dir, word_mode, dest, shift, dest, 1);
if (x != dest)
emit_move_insn (dest, x);
}
}
/* When a parameter is a block, and perhaps in other cases, it is
possible that it did a load from an argument slot that was
already clobbered. */
if (is_sibcall
&& check_sibcall_argument_overlap (before_arg, &args[i], 0))
*sibcall_failure = 1;
/* Handle calls that pass values in multiple non-contiguous
locations. The Irix 6 ABI has examples of this. */
if (GET_CODE (reg) == PARALLEL)
use_group_regs (call_fusage, reg);
else if (nregs == -1)
use_reg_mode (call_fusage, reg,
TYPE_MODE (TREE_TYPE (args[i].tree_value)));
else if (nregs > 0)
use_regs (call_fusage, REGNO (reg), nregs);
}
}
}
/* We need to pop PENDING_STACK_ADJUST bytes. But, if the arguments
wouldn't fill up an even multiple of PREFERRED_UNIT_STACK_BOUNDARY
bytes, then we would need to push some additional bytes to pad the
arguments. So, we compute an adjust to the stack pointer for an
amount that will leave the stack under-aligned by UNADJUSTED_ARGS_SIZE
bytes. Then, when the arguments are pushed the stack will be perfectly
aligned. ARGS_SIZE->CONSTANT is set to the number of bytes that should
be popped after the call. Returns the adjustment. */
static int
combine_pending_stack_adjustment_and_call (int unadjusted_args_size,
struct args_size *args_size,
unsigned int preferred_unit_stack_boundary)
{
/* The number of bytes to pop so that the stack will be
under-aligned by UNADJUSTED_ARGS_SIZE bytes. */
HOST_WIDE_INT adjustment;
/* The alignment of the stack after the arguments are pushed, if we
just pushed the arguments without adjust the stack here. */
unsigned HOST_WIDE_INT unadjusted_alignment;
unadjusted_alignment
= ((stack_pointer_delta + unadjusted_args_size)
% preferred_unit_stack_boundary);
/* We want to get rid of as many of the PENDING_STACK_ADJUST bytes
as possible -- leaving just enough left to cancel out the
UNADJUSTED_ALIGNMENT. In other words, we want to ensure that the
PENDING_STACK_ADJUST is non-negative, and congruent to
-UNADJUSTED_ALIGNMENT modulo the PREFERRED_UNIT_STACK_BOUNDARY. */
/* Begin by trying to pop all the bytes. */
unadjusted_alignment
= (unadjusted_alignment
- (pending_stack_adjust % preferred_unit_stack_boundary));
adjustment = pending_stack_adjust;
/* Push enough additional bytes that the stack will be aligned
after the arguments are pushed. */
if (preferred_unit_stack_boundary > 1 && unadjusted_alignment)
adjustment -= preferred_unit_stack_boundary - unadjusted_alignment;
/* Now, sets ARGS_SIZE->CONSTANT so that we pop the right number of
bytes after the call. The right number is the entire
PENDING_STACK_ADJUST less our ADJUSTMENT plus the amount required
by the arguments in the first place. */
args_size->constant
= pending_stack_adjust - adjustment + unadjusted_args_size;
return adjustment;
}
/* Scan X expression if it does not dereference any argument slots
we already clobbered by tail call arguments (as noted in stored_args_map
bitmap).
Return nonzero if X expression dereferences such argument slots,
zero otherwise. */
static int
check_sibcall_argument_overlap_1 (rtx x)
{
RTX_CODE code;
int i, j;
const char *fmt;
if (x == NULL_RTX)
return 0;
code = GET_CODE (x);
/* We need not check the operands of the CALL expression itself. */
if (code == CALL)
return 0;
if (code == MEM)
return mem_overlaps_already_clobbered_arg_p (XEXP (x, 0),
GET_MODE_SIZE (GET_MODE (x)));
/* Scan all subexpressions. */
fmt = GET_RTX_FORMAT (code);
for (i = 0; i < GET_RTX_LENGTH (code); i++, fmt++)
{
if (*fmt == 'e')
{
if (check_sibcall_argument_overlap_1 (XEXP (x, i)))
return 1;
}
else if (*fmt == 'E')
{
for (j = 0; j < XVECLEN (x, i); j++)
if (check_sibcall_argument_overlap_1 (XVECEXP (x, i, j)))
return 1;
}
}
return 0;
}
/* Scan sequence after INSN if it does not dereference any argument slots
we already clobbered by tail call arguments (as noted in stored_args_map
bitmap). If MARK_STORED_ARGS_MAP, add stack slots for ARG to
stored_args_map bitmap afterwards (when ARG is a register MARK_STORED_ARGS_MAP
should be 0). Return nonzero if sequence after INSN dereferences such argument
slots, zero otherwise. */
static int
check_sibcall_argument_overlap (rtx_insn *insn, struct arg_data *arg,
int mark_stored_args_map)
{
int low, high;
if (insn == NULL_RTX)
insn = get_insns ();
else
insn = NEXT_INSN (insn);
for (; insn; insn = NEXT_INSN (insn))
if (INSN_P (insn)
&& check_sibcall_argument_overlap_1 (PATTERN (insn)))
break;
if (mark_stored_args_map)
{
if (ARGS_GROW_DOWNWARD)
low = -arg->locate.slot_offset.constant - arg->locate.size.constant;
else
low = arg->locate.slot_offset.constant;
for (high = low + arg->locate.size.constant; low < high; low++)
bitmap_set_bit (stored_args_map, low);
}
return insn != NULL_RTX;
}
/* Given that a function returns a value of mode MODE at the most
significant end of hard register VALUE, shift VALUE left or right
as specified by LEFT_P. Return true if some action was needed. */
bool
shift_return_value (machine_mode mode, bool left_p, rtx value)
{
HOST_WIDE_INT shift;
gcc_assert (REG_P (value) && HARD_REGISTER_P (value));
shift = GET_MODE_BITSIZE (GET_MODE (value)) - GET_MODE_BITSIZE (mode);
if (shift == 0)
return false;
/* Use ashr rather than lshr for right shifts. This is for the benefit
of the MIPS port, which requires SImode values to be sign-extended
when stored in 64-bit registers. */
if (!force_expand_binop (GET_MODE (value), left_p ? ashl_optab : ashr_optab,
value, GEN_INT (shift), value, 1, OPTAB_WIDEN))
gcc_unreachable ();
return true;
}
/* If X is a likely-spilled register value, copy it to a pseudo
register and return that register. Return X otherwise. */
static rtx
avoid_likely_spilled_reg (rtx x)
{
rtx new_rtx;
if (REG_P (x)
&& HARD_REGISTER_P (x)
&& targetm.class_likely_spilled_p (REGNO_REG_CLASS (REGNO (x))))
{
/* Make sure that we generate a REG rather than a CONCAT.
Moves into CONCATs can need nontrivial instructions,
and the whole point of this function is to avoid
using the hard register directly in such a situation. */
generating_concat_p = 0;
new_rtx = gen_reg_rtx (GET_MODE (x));
generating_concat_p = 1;
emit_move_insn (new_rtx, x);
return new_rtx;
}
return x;
}
/* Helper function for expand_call.
Return false is EXP is not implementable as a sibling call. */
static bool
can_implement_as_sibling_call_p (tree exp,
rtx structure_value_addr,
tree funtype,
int reg_parm_stack_space ATTRIBUTE_UNUSED,
tree fndecl,
int flags,
tree addr,
const args_size &args_size)
{
if (!targetm.have_sibcall_epilogue ())
{
maybe_complain_about_tail_call
(exp,
"machine description does not have"
" a sibcall_epilogue instruction pattern");
return false;
}
/* Doing sibling call optimization needs some work, since
structure_value_addr can be allocated on the stack.
It does not seem worth the effort since few optimizable
sibling calls will return a structure. */
if (structure_value_addr != NULL_RTX)
{
maybe_complain_about_tail_call (exp, "callee returns a structure");
return false;
}
#ifdef REG_PARM_STACK_SPACE
/* If outgoing reg parm stack space changes, we can not do sibcall. */
if (OUTGOING_REG_PARM_STACK_SPACE (funtype)
!= OUTGOING_REG_PARM_STACK_SPACE (TREE_TYPE (current_function_decl))
|| (reg_parm_stack_space != REG_PARM_STACK_SPACE (current_function_decl)))
{
maybe_complain_about_tail_call (exp,
"inconsistent size of stack space"
" allocated for arguments which are"
" passed in registers");
return false;
}
#endif
/* Check whether the target is able to optimize the call
into a sibcall. */
if (!targetm.function_ok_for_sibcall (fndecl, exp))
{
maybe_complain_about_tail_call (exp,
"target is not able to optimize the"
" call into a sibling call");
return false;
}
/* Functions that do not return exactly once may not be sibcall
optimized. */
if (flags & ECF_RETURNS_TWICE)
{
maybe_complain_about_tail_call (exp, "callee returns twice");
return false;
}
if (flags & ECF_NORETURN)
{
maybe_complain_about_tail_call (exp, "callee does not return");
return false;
}
if (TYPE_VOLATILE (TREE_TYPE (TREE_TYPE (addr))))
{
maybe_complain_about_tail_call (exp, "volatile function type");
return false;
}
/* If the called function is nested in the current one, it might access
some of the caller's arguments, but could clobber them beforehand if
the argument areas are shared. */
if (fndecl && decl_function_context (fndecl) == current_function_decl)
{
maybe_complain_about_tail_call (exp, "nested function");
return false;
}
/* If this function requires more stack slots than the current
function, we cannot change it into a sibling call.
crtl->args.pretend_args_size is not part of the
stack allocated by our caller. */
if (args_size.constant > (crtl->args.size - crtl->args.pretend_args_size))
{
maybe_complain_about_tail_call (exp,
"callee required more stack slots"
" than the caller");
return false;
}
/* If the callee pops its own arguments, then it must pop exactly
the same number of arguments as the current function. */
if (targetm.calls.return_pops_args (fndecl, funtype, args_size.constant)
!= targetm.calls.return_pops_args (current_function_decl,
TREE_TYPE (current_function_decl),
crtl->args.size))
{
maybe_complain_about_tail_call (exp,
"inconsistent number of"
" popped arguments");
return false;
}
if (!lang_hooks.decls.ok_for_sibcall (fndecl))
{
maybe_complain_about_tail_call (exp, "frontend does not support"
" sibling call");
return false;
}
/* All checks passed. */
return true;
}
/* Generate all the code for a CALL_EXPR exp
and return an rtx for its value.
Store the value in TARGET (specified as an rtx) if convenient.
If the value is stored in TARGET then TARGET is returned.
If IGNORE is nonzero, then we ignore the value of the function call. */
rtx
expand_call (tree exp, rtx target, int ignore)
{
/* Nonzero if we are currently expanding a call. */
static int currently_expanding_call = 0;
/* RTX for the function to be called. */
rtx funexp;
/* Sequence of insns to perform a normal "call". */
rtx_insn *normal_call_insns = NULL;
/* Sequence of insns to perform a tail "call". */
rtx_insn *tail_call_insns = NULL;
/* Data type of the function. */
tree funtype;
tree type_arg_types;
tree rettype;
/* Declaration of the function being called,
or 0 if the function is computed (not known by name). */
tree fndecl = 0;
/* The type of the function being called. */
tree fntype;
bool try_tail_call = CALL_EXPR_TAILCALL (exp);
bool must_tail_call = CALL_EXPR_MUST_TAIL_CALL (exp);
int pass;
/* Register in which non-BLKmode value will be returned,
or 0 if no value or if value is BLKmode. */
rtx valreg;
/* Register(s) in which bounds are returned. */
rtx valbnd = NULL;
/* Address where we should return a BLKmode value;
0 if value not BLKmode. */
rtx structure_value_addr = 0;
/* Nonzero if that address is being passed by treating it as
an extra, implicit first parameter. Otherwise,
it is passed by being copied directly into struct_value_rtx. */
int structure_value_addr_parm = 0;
/* Holds the value of implicit argument for the struct value. */
tree structure_value_addr_value = NULL_TREE;
/* Size of aggregate value wanted, or zero if none wanted
or if we are using the non-reentrant PCC calling convention
or expecting the value in registers. */
HOST_WIDE_INT struct_value_size = 0;
/* Nonzero if called function returns an aggregate in memory PCC style,
by returning the address of where to find it. */
int pcc_struct_value = 0;
rtx struct_value = 0;
/* Number of actual parameters in this call, including struct value addr. */
int num_actuals;
/* Number of named args. Args after this are anonymous ones
and they must all go on the stack. */
int n_named_args;
/* Number of complex actual arguments that need to be split. */
int num_complex_actuals = 0;
/* Vector of information about each argument.
Arguments are numbered in the order they will be pushed,
not the order they are written. */
struct arg_data *args;
/* Total size in bytes of all the stack-parms scanned so far. */
struct args_size args_size;
struct args_size adjusted_args_size;
/* Size of arguments before any adjustments (such as rounding). */
int unadjusted_args_size;
/* Data on reg parms scanned so far. */
CUMULATIVE_ARGS args_so_far_v;
cumulative_args_t args_so_far;
/* Nonzero if a reg parm has been scanned. */
int reg_parm_seen;
/* Nonzero if this is an indirect function call. */
/* Nonzero if we must avoid push-insns in the args for this call.
If stack space is allocated for register parameters, but not by the
caller, then it is preallocated in the fixed part of the stack frame.
So the entire argument block must then be preallocated (i.e., we
ignore PUSH_ROUNDING in that case). */
int must_preallocate = !PUSH_ARGS;
/* Size of the stack reserved for parameter registers. */
int reg_parm_stack_space = 0;
/* Address of space preallocated for stack parms
(on machines that lack push insns), or 0 if space not preallocated. */
rtx argblock = 0;
/* Mask of ECF_ and ERF_ flags. */
int flags = 0;
int return_flags = 0;
#ifdef REG_PARM_STACK_SPACE
/* Define the boundary of the register parm stack space that needs to be
saved, if any. */
int low_to_save, high_to_save;
rtx save_area = 0; /* Place that it is saved */
#endif
int initial_highest_arg_in_use = highest_outgoing_arg_in_use;