/* Callgraph based interprocedural optimizations.
- Copyright (C) 2003, 2004, 2005, 2006, 2007, 2008
+ Copyright (C) 2003, 2004, 2005, 2006, 2007, 2008, 2009
Free Software Foundation, Inc.
Contributed by Jan Hubicka
This function is called once (source level) compilation unit is finalized
and it will no longer change.
- In the unit-at-a-time the call-graph construction and local function
+ In the the call-graph construction and local function
analysis takes place here. Bodies of unreachable functions are released
to conserve memory usage.
??? On the tree-ssa genericizing should take place here and we will avoid
need for these hooks (replacing them by genericizing hook)
- We implement two compilation modes.
-
- - unit-at-a-time: In this mode analyzing of all functions is deferred
+ Analyzing of all functions is deferred
to cgraph_finalize_compilation_unit and expansion into cgraph_optimize.
In cgraph_finalize_compilation_unit the reachable functions are
??? Reorganize code so variables are output very last and only if they
really has been referenced by produced code, so we catch more cases
- where reference has been optimized out.
-
- - non-unit-at-a-time
-
- All functions are variables are output as early as possible to conserve
- memory consumption. This may or may not result in less memory used but
- it is still needed for some legacy code that rely on particular ordering
- of things output from the compiler.
-
- Varpool data structures are not used and variables are output directly.
-
- Functions are output early using call of
- cgraph_assemble_pending_function from cgraph_finalize_function. The
- decision on whether function is needed is made more conservative so
- uninlininable static functions are needed too. During the call-graph
- construction the edge destinations are not marked as reachable and it
- is completely relied upn assemble_variable to mark them. */
+ where reference has been optimized out. */
#include "config.h"
#include "timevar.h"
#include "params.h"
#include "fibheap.h"
-#include "c-common.h"
#include "intl.h"
#include "function.h"
#include "ipa-prop.h"
-#include "tree-gimple.h"
+#include "gimple.h"
+#include "tree-iterator.h"
#include "tree-pass.h"
+#include "tree-dump.h"
#include "output.h"
+#include "coverage.h"
static void cgraph_expand_all_functions (void);
static void cgraph_mark_functions_to_output (void);
static void cgraph_expand_function (struct cgraph_node *);
static void cgraph_output_pending_asms (void);
+static void cgraph_analyze_function (struct cgraph_node *);
static FILE *cgraph_dump_file;
VEC_safe_push (tree, gc, static_dtors, fndecl);
DECL_STATIC_DESTRUCTOR (fndecl) = 0;
}
- DECL_INLINE (fndecl) = 1;
node = cgraph_node (fndecl);
node->local.disregard_inline_limits = 1;
cgraph_mark_reachable_node (node);
priority = p;
else if (p != priority)
break;
- append_to_statement_list (build_function_call_expr (fn, 0),
+ append_to_statement_list (build_function_call_expr (UNKNOWN_LOCATION,
+ fn, 0),
&body);
++i;
}
/* Determine if function DECL is needed. That is, visible to something
either outside this translation unit, something magic in the system
- configury, or (if not doing unit-at-a-time) to something we havn't
- seen yet. */
+ configury. */
-static bool
-decide_is_function_needed (struct cgraph_node *node, tree decl)
+bool
+cgraph_decide_is_function_needed (struct cgraph_node *node, tree decl)
{
- tree origin;
- if (MAIN_NAME_P (DECL_NAME (decl))
- && TREE_PUBLIC (decl))
- {
- node->local.externally_visible = true;
- return true;
- }
-
/* If the user told us it is used, then it must be so. */
if (node->local.externally_visible)
return true;
- if (!flag_unit_at_a_time && lookup_attribute ("used", DECL_ATTRIBUTES (decl)))
- return true;
-
/* ??? If the assembler name is set by hand, it is possible to assemble
the name later after finalizing the function and the fact is noticed
in assemble_name then. This is arguably a bug. */
|| (!optimize && !node->local.disregard_inline_limits
&& !DECL_DECLARED_INLINE_P (decl)
&& !node->origin))
- && !flag_whole_program)
+ && !flag_whole_program
+ && !flag_lto
+ && !flag_whopr)
&& !DECL_COMDAT (decl) && !DECL_EXTERNAL (decl))
return true;
if (DECL_STATIC_CONSTRUCTOR (decl) || DECL_STATIC_DESTRUCTOR (decl))
return true;
- if (flag_unit_at_a_time)
- return false;
-
- /* If not doing unit at a time, then we'll only defer this function
- if its marked for inlining. Otherwise we want to emit it now. */
-
- /* "extern inline" functions are never output locally. */
- if (DECL_EXTERNAL (decl))
- return false;
- /* Nested functions of extern inline function shall not be emit unless
- we inlined the origin. */
- for (origin = decl_function_context (decl); origin;
- origin = decl_function_context (origin))
- if (DECL_EXTERNAL (origin))
- return false;
- /* We want to emit COMDAT functions only when absolutely necessary. */
- if (DECL_COMDAT (decl))
- return false;
- if (!DECL_INLINE (decl)
- || (!node->local.disregard_inline_limits
- /* When declared inline, defer even the uninlinable functions.
- This allows them to be eliminated when unused. */
- && !DECL_DECLARED_INLINE_P (decl)
- && (!node->local.inlinable || !cgraph_default_inline_p (node, NULL))))
- return true;
-
return false;
}
it into reachable functions list. */
node->next_needed = NULL;
- node->needed = node->reachable = false;
cgraph_finalize_function (fndecl, false);
cgraph_mark_reachable_node (node);
output = true;
transformations that has been already performed on the whole
cgraph but not on this function. */
- tree_register_cfg_hooks ();
+ gimple_register_cfg_hooks ();
if (!node->analyzed)
cgraph_analyze_function (node);
push_cfun (DECL_STRUCT_FUNCTION (fndecl));
current_function_decl = fndecl;
- node->local.inlinable = tree_inlinable_function_p (fndecl);
- node->local.self_insns = estimate_num_insns (fndecl,
- &eni_inlining_weights);
- node->local.disregard_inline_limits
- |= DECL_DISREGARD_INLINE_LIMITS (fndecl);
- /* Inlining characteristics are maintained by the
- cgraph_mark_inline. */
- node->global.insns = node->local.self_insns;
- if (flag_really_no_inline && !node->local.disregard_inline_limits)
- node->local.inlinable = 0;
+ compute_inline_parameters (node);
if ((cgraph_state == CGRAPH_STATE_IPA_SSA
&& !gimple_in_ssa_p (DECL_STRUCT_FUNCTION (fndecl)))
/* When not optimizing, be sure we run early local passes anyway
to expand OMP. */
|| !optimize)
- execute_pass_list (pass_early_local_passes.sub);
+ execute_pass_list (pass_early_local_passes.pass.sub);
free_dominance_info (CDI_POST_DOMINATORS);
free_dominance_info (CDI_DOMINATORS);
pop_cfun ();
case CGRAPH_STATE_EXPANSION:
/* Functions created during expansion shall be compiled
directly. */
- node->output = 0;
+ node->process = 0;
cgraph_expand_function (node);
break;
gcc_unreachable ();
break;
}
+ cgraph_call_function_insertion_hooks (node);
}
return output;
}
-/* When not doing unit-at-a-time, output all functions enqueued.
- Return true when such a functions were found. */
-
-static bool
-cgraph_assemble_pending_functions (void)
-{
- bool output = false;
-
- if (flag_unit_at_a_time)
- return false;
-
- cgraph_output_pending_asms ();
-
- while (cgraph_nodes_queue)
- {
- struct cgraph_node *n = cgraph_nodes_queue;
-
- cgraph_nodes_queue = cgraph_nodes_queue->next_needed;
- n->next_needed = NULL;
- if (!n->global.inlined_to
- && !n->alias
- && !DECL_EXTERNAL (n->decl))
- {
- cgraph_expand_function (n);
- output = true;
- }
- output |= cgraph_process_new_functions ();
- }
-
- return output;
-}
-
-
/* As an GCC extension we allow redefinition of the function. The
semantics when both copies of bodies differ is not well defined.
We replace the old body with new body so in unit at a time mode
static void
cgraph_reset_node (struct cgraph_node *node)
{
- /* If node->output is set, then this is a unit-at-a-time compilation
- and we have already begun whole-unit analysis. This is *not*
- testing for whether we've already emitted the function. That
- case can be sort-of legitimately seen with real function
- redefinition errors. I would argue that the front end should
- never present us with such a case, but don't enforce that for now. */
- gcc_assert (!node->output);
+ /* If node->process is set, then we have already begun whole-unit analysis.
+ This is *not* testing for whether we've already emitted the function.
+ That case can be sort-of legitimately seen with real function redefinition
+ errors. I would argue that the front end should never present us with
+ such a case, but don't enforce that for now. */
+ gcc_assert (!node->process);
/* Reset our data structures so we can analyze the function again. */
memset (&node->local, 0, sizeof (node->local));
node->local.redefined_extern_inline = true;
node->local.finalized = false;
- if (!flag_unit_at_a_time)
- {
- struct cgraph_node *n, *next;
-
- for (n = cgraph_nodes; n; n = next)
- {
- next = n->next;
- if (n->global.inlined_to == node)
- cgraph_remove_node (n);
- }
- }
-
cgraph_node_remove_callees (node);
/* We may need to re-queue the node for assembling in case
{
if (node->lowered)
return;
+
+ if (node->nested)
+ lower_nested_functions (node->decl);
+ gcc_assert (!node->nested);
+
tree_lowering_passes (node->decl);
node->lowered = true;
}
node->pid = cgraph_max_pid ++;
notice_global_symbol (decl);
- node->decl = decl;
node->local.finalized = true;
node->lowered = DECL_STRUCT_FUNCTION (decl)->cfg != NULL;
+ node->finalized_by_frontend = true;
record_cdtor_fn (node->decl);
- if (node->nested)
- lower_nested_functions (decl);
- gcc_assert (!node->nested);
-
- /* If not unit at a time, then we need to create the call graph
- now, so that called functions can be queued and emitted now. */
- if (!flag_unit_at_a_time)
- cgraph_analyze_function (node);
- if (decide_is_function_needed (node, decl))
+ if (cgraph_decide_is_function_needed (node, decl))
cgraph_mark_needed_node (node);
/* Since we reclaim unreachable nodes at the end of every language
if ((TREE_PUBLIC (decl) && !DECL_COMDAT (decl) && !DECL_EXTERNAL (decl)))
cgraph_mark_reachable_node (node);
- /* If not unit at a time, go ahead and emit everything we've found
- to be reachable at this time. */
- if (!nested)
- {
- if (!cgraph_assemble_pending_functions ())
- ggc_collect ();
- }
-
/* If we've not yet emitted decl, tell the debug info about it. */
if (!TREE_ASM_WRITTEN (decl))
(*debug_hooks->deferred_inline_function) (decl);
/* Possibly warn about unused parameters. */
if (warn_unused_parameter)
do_warn_unused_parameter (decl);
+
+ if (!nested)
+ ggc_collect ();
+}
+
+/* C99 extern inline keywords allow changing of declaration after function
+ has been finalized. We need to re-decide if we want to mark the function as
+ needed then. */
+
+void
+cgraph_mark_if_needed (tree decl)
+{
+ struct cgraph_node *node = cgraph_node (decl);
+ if (node->local.finalized && cgraph_decide_is_function_needed (node, decl))
+ cgraph_mark_needed_node (node);
+}
+
+/* Return TRUE if NODE2 is equivalent to NODE or its clone. */
+static bool
+clone_of_p (struct cgraph_node *node, struct cgraph_node *node2)
+{
+ while (node != node2 && node2)
+ node2 = node2->clone_of;
+ return node2 != NULL;
}
/* Verify cgraph nodes of given cgraph node. */
verify_cgraph_node (struct cgraph_node *node)
{
struct cgraph_edge *e;
- struct cgraph_node *main_clone;
struct function *this_cfun = DECL_STRUCT_FUNCTION (node->decl);
struct function *saved_cfun = cfun;
basic_block this_block;
- block_stmt_iterator bsi;
+ gimple_stmt_iterator gsi;
bool error_found = false;
if (errorcount || sorrycount)
if (e->aux)
{
error ("aux field set for edge %s->%s",
- cgraph_node_name (e->caller), cgraph_node_name (e->callee));
+ identifier_to_locale (cgraph_node_name (e->caller)),
+ identifier_to_locale (cgraph_node_name (e->callee)));
error_found = true;
}
if (node->count < 0)
error ("Execution count is negative");
error_found = true;
}
+ if (node->global.inlined_to && node->local.externally_visible)
+ {
+ error ("Externally visible inline clone");
+ error_found = true;
+ }
+ if (node->global.inlined_to && node->address_taken)
+ {
+ error ("Inline clone with address taken");
+ error_found = true;
+ }
+ if (node->global.inlined_to && node->needed)
+ {
+ error ("Inline clone is needed");
+ error_found = true;
+ }
for (e = node->callers; e; e = e->next_caller)
{
if (e->count < 0)
error_found = true;
}
- for (main_clone = cgraph_node (node->decl); main_clone;
- main_clone = main_clone->next_clone)
- if (main_clone == node)
- break;
if (!cgraph_node (node->decl))
{
error ("node not found in cgraph_hash");
error_found = true;
}
- if (node->analyzed
- && DECL_SAVED_TREE (node->decl) && !TREE_ASM_WRITTEN (node->decl)
- && (!DECL_EXTERNAL (node->decl) || node->global.inlined_to))
+ if (node->clone_of)
+ {
+ struct cgraph_node *n;
+ for (n = node->clone_of->clones; n; n = n->next_sibling_clone)
+ if (n == node)
+ break;
+ if (!n)
+ {
+ error ("node has wrong clone_of");
+ error_found = true;
+ }
+ }
+ if (node->clones)
+ {
+ struct cgraph_node *n;
+ for (n = node->clones; n; n = n->next_sibling_clone)
+ if (n->clone_of != node)
+ break;
+ if (n)
+ {
+ error ("node has wrong clone list");
+ error_found = true;
+ }
+ }
+ if ((node->prev_sibling_clone || node->next_sibling_clone) && !node->clone_of)
+ {
+ error ("node is in clone list but it is not clone");
+ error_found = true;
+ }
+ if (!node->prev_sibling_clone && node->clone_of && node->clone_of->clones != node)
+ {
+ error ("node has wrong prev_clone pointer");
+ error_found = true;
+ }
+ if (node->prev_sibling_clone && node->prev_sibling_clone->next_sibling_clone != node)
+ {
+ error ("double linked list of clones corrupted");
+ error_found = true;
+ }
+
+ if (node->analyzed && gimple_has_body_p (node->decl)
+ && !TREE_ASM_WRITTEN (node->decl)
+ && (!DECL_EXTERNAL (node->decl) || node->global.inlined_to)
+ && !flag_wpa)
{
if (this_cfun->cfg)
{
/* Reach the trees by walking over the CFG, and note the
enclosing basic-blocks in the call edges. */
FOR_EACH_BB_FN (this_block, this_cfun)
- for (bsi = bsi_start (this_block); !bsi_end_p (bsi); bsi_next (&bsi))
+ for (gsi = gsi_start_bb (this_block);
+ !gsi_end_p (gsi);
+ gsi_next (&gsi))
{
- tree stmt = bsi_stmt (bsi);
- tree call = get_call_expr_in (stmt);
+ gimple stmt = gsi_stmt (gsi);
tree decl;
- if (call && (decl = get_callee_fndecl (call)))
+ if (is_gimple_call (stmt) && (decl = gimple_call_fndecl (stmt)))
{
struct cgraph_edge *e = cgraph_edge (node, stmt);
if (e)
if (e->aux)
{
error ("shared call_stmt:");
- debug_generic_stmt (stmt);
+ debug_gimple_stmt (stmt);
error_found = true;
}
- if (e->callee->decl != cgraph_node (decl)->decl
- && e->inline_failed)
+ if (!clone_of_p (cgraph_node (decl), e->callee)
+ && !e->callee->global.inlined_to)
{
error ("edge points to wrong declaration:");
debug_tree (e->callee->decl);
else
{
error ("missing callgraph edge for call stmt:");
- debug_generic_stmt (stmt);
+ debug_gimple_stmt (stmt);
error_found = true;
}
}
for (e = node->callees; e; e = e->next_callee)
{
- if (!e->aux)
+ if (!e->aux && !e->indirect_call)
{
error ("edge %s->%s has no corresponding call_stmt",
- cgraph_node_name (e->caller),
- cgraph_node_name (e->callee));
- debug_generic_stmt (e->call_stmt);
+ identifier_to_locale (cgraph_node_name (e->caller)),
+ identifier_to_locale (cgraph_node_name (e->callee)));
+ debug_gimple_stmt (e->call_stmt);
error_found = true;
}
e->aux = 0;
}
/* Analyze the function scheduled to be output. */
-void
+static void
cgraph_analyze_function (struct cgraph_node *node)
{
+ tree save = current_function_decl;
tree decl = node->decl;
current_function_decl = decl;
push_cfun (DECL_STRUCT_FUNCTION (decl));
+
+ /* Make sure to gimplify bodies only once. During analyzing a
+ function we lower it, which will require gimplified nested
+ functions, so we can end up here with an already gimplified
+ body. */
+ if (!gimple_body (decl))
+ gimplify_function_tree (decl);
+ dump_function (TDI_generic, decl);
+
cgraph_lower_function (node);
node->analyzed = true;
- if (!flag_unit_at_a_time)
- {
- bitmap_obstack_initialize (NULL);
- tree_register_cfg_hooks ();
- execute_pass_list (pass_early_local_passes.sub);
- free_dominance_info (CDI_POST_DOMINATORS);
- free_dominance_info (CDI_DOMINATORS);
- bitmap_obstack_release (NULL);
- }
-
pop_cfun ();
- current_function_decl = NULL;
+ current_function_decl = save;
}
/* Look for externally_visible and used attributes and mark cgraph nodes
if (lookup_attribute ("externally_visible", DECL_ATTRIBUTES (decl)))
{
if (! TREE_PUBLIC (node->decl))
- warning (OPT_Wattributes,
- "%J%<externally_visible%> attribute have effect only on public objects",
- node->decl);
- else
- {
- if (node->local.finalized)
- cgraph_mark_needed_node (node);
- node->local.externally_visible = true;
- }
+ warning_at (DECL_SOURCE_LOCATION (node->decl), OPT_Wattributes,
+ "%<externally_visible%>"
+ " attribute have effect only on public objects");
+ else if (node->local.finalized)
+ cgraph_mark_needed_node (node);
}
}
for (vnode = varpool_nodes; vnode != first_var; vnode = vnode->next)
if (lookup_attribute ("used", DECL_ATTRIBUTES (decl)))
{
mark_decl_referenced (decl);
+ vnode->force_output = true;
if (vnode->finalized)
varpool_mark_needed_node (vnode);
}
if (lookup_attribute ("externally_visible", DECL_ATTRIBUTES (decl)))
{
if (! TREE_PUBLIC (vnode->decl))
- warning (OPT_Wattributes,
- "%J%<externally_visible%> attribute have effect only on public objects",
- vnode->decl);
- else
- {
- if (vnode->finalized)
- varpool_mark_needed_node (vnode);
- vnode->externally_visible = true;
- }
+ warning_at (DECL_SOURCE_LOCATION (vnode->decl), OPT_Wattributes,
+ "%<externally_visible%>"
+ " attribute have effect only on public objects");
+ else if (vnode->finalized)
+ varpool_mark_needed_node (vnode);
}
}
}
{
fprintf (cgraph_dump_file, "Initial entry points:");
for (node = cgraph_nodes; node != first_analyzed; node = node->next)
- if (node->needed && DECL_SAVED_TREE (node->decl))
+ if (node->needed)
fprintf (cgraph_dump_file, " %s", cgraph_node_name (node));
fprintf (cgraph_dump_file, "\n");
}
/* ??? It is possible to create extern inline function and later using
weak alias attribute to kill its body. See
gcc.c-torture/compile/20011119-1.c */
- if (!DECL_SAVED_TREE (decl))
+ if (!DECL_STRUCT_FUNCTION (decl))
{
cgraph_reset_node (node);
continue;
}
- gcc_assert (!node->analyzed && node->reachable);
- gcc_assert (DECL_SAVED_TREE (decl));
-
- cgraph_analyze_function (node);
+ if (!node->analyzed)
+ cgraph_analyze_function (node);
for (edge = node->callees; edge; edge = edge->next_callee)
if (!edge->callee->reachable)
cgraph_mark_reachable_node (edge->callee);
+ /* If decl is a clone of an abstract function, mark that abstract
+ function so that we don't release its body. The DECL_INITIAL() of that
+ abstract function declaration will be later needed to output debug info. */
+ if (DECL_ABSTRACT_ORIGIN (decl))
+ {
+ struct cgraph_node *origin_node = cgraph_node (DECL_ABSTRACT_ORIGIN (decl));
+ origin_node->abstract_and_needed = true;
+ }
+
/* We finalize local static variables during constructing callgraph
edges. Process their attributes too. */
process_function_and_variable_attributes (first_processed,
{
fprintf (cgraph_dump_file, "Unit entry points:");
for (node = cgraph_nodes; node != first_analyzed; node = node->next)
- if (node->needed && DECL_SAVED_TREE (node->decl))
+ if (node->needed)
fprintf (cgraph_dump_file, " %s", cgraph_node_name (node));
fprintf (cgraph_dump_file, "\n\nInitial ");
dump_cgraph (cgraph_dump_file);
tree decl = node->decl;
next = node->next;
- if (node->local.finalized && !DECL_SAVED_TREE (decl))
+ if (node->local.finalized && !gimple_has_body_p (decl))
cgraph_reset_node (node);
- if (!node->reachable && DECL_SAVED_TREE (decl))
+ if (!node->reachable && gimple_has_body_p (decl))
{
if (cgraph_dump_file)
fprintf (cgraph_dump_file, " %s", cgraph_node_name (node));
}
else
node->next_needed = NULL;
- gcc_assert (!node->local.finalized || DECL_SAVED_TREE (decl));
+ gcc_assert (!node->local.finalized || gimple_has_body_p (decl));
gcc_assert (node->analyzed == node->local.finalized);
}
if (cgraph_dump_file)
ggc_collect ();
}
+
+/* Emit thunks for every node in the cgraph.
+ FIXME: We really ought to emit thunks only for functions that are needed. */
+
+static void
+cgraph_emit_thunks (void)
+{
+ struct cgraph_node *n;
+
+ for (n = cgraph_nodes; n; n = n->next)
+ {
+ /* Only emit thunks on functions defined in this TU.
+ Note that this may emit more thunks than strictly necessary.
+ During optimization some nodes may disappear. It would be
+ nice to only emit thunks only for the functions that will be
+ emitted, but we cannot know that until the inliner and other
+ IPA passes have run (see the sequencing of the call to
+ cgraph_mark_functions_to_output in cgraph_optimize). */
+ if (n->reachable
+ && !DECL_EXTERNAL (n->decl))
+ lang_hooks.callgraph.emit_associated_thunks (n->decl);
+ }
+}
+
+
/* Analyze the whole compilation unit once it is parsed completely. */
void
cgraph_finalize_compilation_unit (void)
{
- if (errorcount || sorrycount)
- return;
+ timevar_push (TV_CGRAPH);
- finish_aliases_1 ();
+ /* Do not skip analyzing the functions if there were errors, we
+ miss diagnostics for following functions otherwise. */
- if (!flag_unit_at_a_time)
- {
- cgraph_output_pending_asms ();
- cgraph_assemble_pending_functions ();
- varpool_output_debug_info ();
- return;
- }
+ /* Emit size functions we didn't inline. */
+ finalize_size_functions ();
+
+ /* Call functions declared with the "constructor" or "destructor"
+ attribute. */
+ cgraph_build_cdtor_fns ();
+
+ /* Mark alias targets necessary and emit diagnostics. */
+ finish_aliases_1 ();
if (!quiet_flag)
{
fflush (stderr);
}
- timevar_push (TV_CGRAPH);
+ /* Gimplify and lower all functions, compute reachability and
+ remove unreachable nodes. */
cgraph_analyze_functions ();
+
+ /* Emit thunks for reachable nodes, if needed. */
+ if (lang_hooks.callgraph.emit_associated_thunks)
+ cgraph_emit_thunks ();
+
+ /* Mark alias targets necessary and emit diagnostics. */
+ finish_aliases_1 ();
+
+ /* Gimplify and lower thunks. */
+ cgraph_analyze_functions ();
+
+ /* Finally drive the pass manager. */
+ cgraph_optimize ();
+
timevar_pop (TV_CGRAPH);
}
+
+
/* Figure out what functions we want to assemble. */
static void
tree decl = node->decl;
struct cgraph_edge *e;
- gcc_assert (!node->output);
+ gcc_assert (!node->process);
for (e = node->callers; e; e = e->next_caller)
if (e->inline_failed)
/* We need to output all local functions that are used and not
always inlined, as well as those that are reachable from
outside the current compilation unit. */
- if (DECL_SAVED_TREE (decl)
+ if (node->analyzed
&& !node->global.inlined_to
&& (node->needed
|| (e && node->reachable))
&& !TREE_ASM_WRITTEN (decl)
&& !DECL_EXTERNAL (decl))
- node->output = 1;
+ node->process = 1;
else
{
/* We should've reclaimed all functions that are not needed. */
#ifdef ENABLE_CHECKING
- if (!node->global.inlined_to && DECL_SAVED_TREE (decl)
+ if (!node->global.inlined_to
+ && gimple_has_body_p (decl)
&& !DECL_EXTERNAL (decl))
{
dump_cgraph_node (stderr, node);
internal_error ("failed to reclaim unneeded function");
}
#endif
- gcc_assert (node->global.inlined_to || !DECL_SAVED_TREE (decl)
+ gcc_assert (node->global.inlined_to
+ || !gimple_has_body_p (decl)
|| DECL_EXTERNAL (decl));
}
/* We ought to not compile any inline clones. */
gcc_assert (!node->global.inlined_to);
- if (flag_unit_at_a_time)
- announce_function (decl);
+ announce_function (decl);
+ node->process = 0;
gcc_assert (node->lowered);
/* Generate RTL for the body of DECL. */
- if (lang_hooks.callgraph.emit_associated_thunks)
- lang_hooks.callgraph.emit_associated_thunks (decl);
tree_rest_of_compilation (decl);
/* Make sure that BE didn't give up on compiling. */
- /* ??? Can happen with nested function of extern inline. */
- gcc_assert (TREE_ASM_WRITTEN (node->decl));
-
+ gcc_assert (TREE_ASM_WRITTEN (decl));
current_function_decl = NULL;
- if (!cgraph_preserve_function_body_p (node->decl))
- {
- cgraph_release_function_body (node);
- /* Eliminate all call edges. This is important so the call_expr no longer
- points to the dead function body. */
- cgraph_node_remove_callees (node);
- }
+ gcc_assert (!cgraph_preserve_function_body_p (decl));
+ cgraph_release_function_body (node);
+ /* Eliminate all call edges. This is important so the GIMPLE_CALL no longer
+ points to the dead function body. */
+ cgraph_node_remove_callees (node);
cgraph_function_flags_ready = true;
}
/* Return true when CALLER_DECL should be inlined into CALLEE_DECL. */
bool
-cgraph_inline_p (struct cgraph_edge *e, const char **reason)
+cgraph_inline_p (struct cgraph_edge *e, cgraph_inline_failed_t *reason)
{
*reason = e->inline_failed;
return !e->inline_failed;
{
struct cgraph_node *node;
struct cgraph_node **order = XCNEWVEC (struct cgraph_node *, cgraph_n_nodes);
- int order_pos = 0, new_order_pos = 0;
+ int order_pos, new_order_pos = 0;
int i;
order_pos = cgraph_postorder (order);
/* Garbage collector may remove inline clones we eliminate during
optimization. So we must be sure to not reference them. */
for (i = 0; i < order_pos; i++)
- if (order[i]->output)
+ if (order[i]->process)
order[new_order_pos++] = order[i];
for (i = new_order_pos - 1; i >= 0; i--)
{
node = order[i];
- if (node->output)
+ if (node->process)
{
gcc_assert (node->reachable);
- node->output = 0;
+ node->process = 0;
cgraph_expand_function (node);
}
}
/* This is used to sort the node types by the cgraph order number. */
+enum cgraph_order_sort_kind
+{
+ ORDER_UNDEFINED = 0,
+ ORDER_FUNCTION,
+ ORDER_VAR,
+ ORDER_ASM
+};
+
struct cgraph_order_sort
{
- enum { ORDER_UNDEFINED = 0, ORDER_FUNCTION, ORDER_VAR, ORDER_ASM } kind;
+ enum cgraph_order_sort_kind kind;
union
{
struct cgraph_node *f;
for (pf = cgraph_nodes; pf; pf = pf->next)
{
- if (pf->output)
+ if (pf->process)
{
i = pf->order;
gcc_assert (nodes[i].kind == ORDER_UNDEFINED);
nodes[i].u.a = pa;
}
+ /* In toplevel reorder mode we output all statics; mark them as needed. */
+ for (i = 0; i < max; ++i)
+ {
+ if (nodes[i].kind == ORDER_VAR)
+ {
+ varpool_mark_needed_node (nodes[i].u.v);
+ }
+ }
+ varpool_empty_needed_queue ();
+
for (i = 0; i < max; ++i)
{
switch (nodes[i].kind)
{
case ORDER_FUNCTION:
- nodes[i].u.f->output = 0;
+ nodes[i].u.f->process = 0;
cgraph_expand_function (nodes[i].u.f);
break;
cgraph_preserve_function_body_p (tree decl)
{
struct cgraph_node *node;
- if (!cgraph_global_info_ready)
- return (flag_really_no_inline
- ? DECL_DISREGARD_INLINE_LIMITS (decl)
- : DECL_INLINE (decl));
+
+ gcc_assert (cgraph_global_info_ready);
/* Look if there is any clone around. */
- for (node = cgraph_node (decl); node; node = node->next_clone)
- if (node->global.inlined_to)
- return true;
+ node = cgraph_node (decl);
+ if (node->clones)
+ return true;
return false;
}
{
set_cfun (NULL);
current_function_decl = NULL;
- tree_register_cfg_hooks ();
+ gimple_register_cfg_hooks ();
bitmap_obstack_initialize (NULL);
- execute_ipa_pass_list (all_ipa_passes);
+
+ if (!in_lto_p)
+ execute_ipa_pass_list (all_small_ipa_passes);
+
+ /* If pass_all_early_optimizations was not scheduled, the state of
+ the cgraph will not be properly updated. Update it now. */
+ if (cgraph_state < CGRAPH_STATE_IPA_SSA)
+ cgraph_state = CGRAPH_STATE_IPA_SSA;
+
+ if (!in_lto_p)
+ {
+ /* Generate coverage variables and constructors. */
+ coverage_finish ();
+
+ /* Process new functions added. */
+ set_cfun (NULL);
+ current_function_decl = NULL;
+ cgraph_process_new_functions ();
+
+ execute_ipa_summary_passes ((struct ipa_opt_pass_d *) all_regular_ipa_passes);
+ }
+ execute_ipa_summary_passes ((struct ipa_opt_pass_d *) all_lto_gen_passes);
+
+ if (!in_lto_p)
+ ipa_write_summaries ();
+
+ if (!flag_ltrans)
+ execute_ipa_pass_list (all_regular_ipa_passes);
+
bitmap_obstack_release (NULL);
}
+
/* Perform simple optimizations based on callgraph. */
void
verify_cgraph ();
#endif
- /* Call functions declared with the "constructor" or "destructor"
- attribute. */
- cgraph_build_cdtor_fns ();
- if (!flag_unit_at_a_time)
- {
- cgraph_assemble_pending_functions ();
- cgraph_process_new_functions ();
- cgraph_state = CGRAPH_STATE_FINISHED;
- cgraph_output_pending_asms ();
- varpool_assemble_pending_decls ();
- varpool_output_debug_info ();
- return;
- }
-
/* Frontend may output common variables after the unit has been finalized.
It is safe to deal with them here as they are always zero initialized. */
varpool_analyze_pending_decls ();
- cgraph_analyze_functions ();
timevar_push (TV_CGRAPHOPT);
if (pre_ipa_mem_report)
if (!quiet_flag)
fprintf (stderr, "Performing interprocedural optimizations\n");
cgraph_state = CGRAPH_STATE_IPA;
-
+
/* Don't run the IPA passes if there was any error or sorry messages. */
if (errorcount == 0 && sorrycount == 0)
ipa_passes ();
+ /* Do nothing else if any IPA pass found errors. */
+ if (errorcount || sorrycount)
+ {
+ timevar_pop (TV_CGRAPHOPT);
+ return;
+ }
+
/* This pass remove bodies of extern inline functions we never inlined.
Do this later so other IPA passes see what is really going on. */
cgraph_remove_unreachable_nodes (false, dump_file);
timevar_pop (TV_CGRAPHOPT);
/* Output everything. */
+ (*debug_hooks->assembly_start) ();
if (!quiet_flag)
fprintf (stderr, "Assembling functions:\n");
#ifdef ENABLE_CHECKING
verify_cgraph ();
#endif
+ cgraph_materialize_all_clones ();
cgraph_mark_functions_to_output ();
cgraph_state = CGRAPH_STATE_EXPANSION;
varpool_remove_unreferenced_decls ();
varpool_assemble_pending_decls ();
- varpool_output_debug_info ();
}
cgraph_process_new_functions ();
cgraph_state = CGRAPH_STATE_FINISHED;
verify_cgraph ();
/* Double check that all inline clones are gone and that all
function bodies have been released from memory. */
- if (flag_unit_at_a_time
- && !(sorrycount || errorcount))
+ if (!(sorrycount || errorcount))
{
struct cgraph_node *node;
bool error_found = false;
for (node = cgraph_nodes; node; node = node->next)
if (node->analyzed
&& (node->global.inlined_to
- || DECL_SAVED_TREE (node->decl)))
+ || gimple_has_body_p (node->decl)))
{
error_found = true;
dump_cgraph_node (stderr, node);
}
if (error_found)
- internal_error ("nodes with no released memory found");
+ internal_error ("nodes with unreleased memory found");
}
#endif
}
+
+
/* Generate and emit a static constructor or destructor. WHICH must
be one of 'I' (for a constructor) or 'D' (for a destructor). BODY
is a STATEMENT_LIST containing GENERIC statements. PRIORITY is the
- initialization priority fot this constructor or destructor. */
+ initialization priority for this constructor or destructor. */
void
cgraph_build_static_cdtor (char which, tree body, int priority)
sprintf (which_buf, "%c_%.5d_%d", which, priority, counter++);
name = get_file_function_name (which_buf);
- decl = build_decl (FUNCTION_DECL, name,
+ decl = build_decl (input_location, FUNCTION_DECL, name,
build_function_type (void_type_node, void_list_node));
current_function_decl = decl;
- resdecl = build_decl (RESULT_DECL, NULL_TREE, void_type_node);
+ resdecl = build_decl (input_location,
+ RESULT_DECL, NULL_TREE, void_type_node);
DECL_ARTIFICIAL (resdecl) = 1;
DECL_RESULT (decl) = resdecl;
+ DECL_CONTEXT (resdecl) = decl;
allocate_struct_function (decl, false);
struct cgraph_edge *e;
gcc_assert (new_version);
+
+ /* Update the call expr on the edges to call the new version. */
for (e = new_version->callers; e; e = e->next_caller)
- /* Update the call expr on the edges
- to call the new version. */
- TREE_OPERAND (CALL_EXPR_FN (get_call_expr_in (e->call_stmt)), 0) = new_version->decl;
+ {
+ struct function *inner_function = DECL_STRUCT_FUNCTION (e->caller->decl);
+ gimple_call_set_fndecl (e->call_stmt, new_version->decl);
+ maybe_clean_eh_stmt_fn (inner_function, e->call_stmt);
+ }
}
also cloned. */
for (e = old_version->callees;e; e=e->next_callee)
{
- new_e = cgraph_clone_edge (e, new_version, e->call_stmt, 0, e->frequency,
+ new_e = cgraph_clone_edge (e, new_version, e->call_stmt,
+ e->lto_stmt_uid, 0, e->frequency,
e->loop_nest, true);
new_e->count = e->count;
}
TREE_MAP is a mapping of tree nodes we want to replace with
new ones (according to results of prior analysis).
OLD_VERSION_NODE is the node that is versioned.
- It returns the new version's cgraph node. */
+ It returns the new version's cgraph node.
+ ARGS_TO_SKIP lists arguments to be omitted from functions
+ */
struct cgraph_node *
cgraph_function_versioning (struct cgraph_node *old_version_node,
VEC(cgraph_edge_p,heap) *redirect_callers,
- varray_type tree_map)
+ VEC (ipa_replace_map_p,gc)* tree_map,
+ bitmap args_to_skip)
{
tree old_decl = old_version_node->decl;
struct cgraph_node *new_version_node = NULL;
/* Make a new FUNCTION_DECL tree node for the
new version. */
- new_decl = copy_node (old_decl);
+ if (!args_to_skip)
+ new_decl = copy_node (old_decl);
+ else
+ new_decl = build_function_decl_skip_args (old_decl, args_to_skip);
/* Create the new version's call-graph node.
and update the edges of the new node. */
redirect_callers);
/* Copy the OLD_VERSION_NODE function tree to the new version. */
- tree_function_versioning (old_decl, new_decl, tree_map, false);
- /* Update the call_expr on the edges to call the new version node. */
- update_call_expr (new_version_node);
+ tree_function_versioning (old_decl, new_decl, tree_map, false, args_to_skip);
/* Update the new version's properties.
- Make The new version visible only within this translation unit.
+ Make The new version visible only within this translation unit. Make sure
+ that is not weak also.
??? We cannot use COMDAT linkage because there is no
ABI support for this. */
DECL_EXTERNAL (new_version_node->decl) = 0;
- DECL_ONE_ONLY (new_version_node->decl) = 0;
+ DECL_COMDAT_GROUP (new_version_node->decl) = NULL_TREE;
TREE_PUBLIC (new_version_node->decl) = 0;
DECL_COMDAT (new_version_node->decl) = 0;
+ DECL_WEAK (new_version_node->decl) = 0;
+ DECL_VIRTUAL_P (new_version_node->decl) = 0;
new_version_node->local.externally_visible = 0;
new_version_node->local.local = 1;
new_version_node->lowered = true;
+
+ /* Update the call_expr on the edges to call the new version node. */
+ update_call_expr (new_version_node);
+
+ cgraph_call_function_insertion_hooks (new_version_node);
return new_version_node;
}
struct cgraph_node *
save_inline_function_body (struct cgraph_node *node)
{
- struct cgraph_node *first_clone;
+ struct cgraph_node *first_clone, *n;
gcc_assert (node == cgraph_node (node->decl));
cgraph_lower_function (node);
- /* In non-unit-at-a-time we construct full fledged clone we never output to
- assembly file. This clone is pointed out by inline_decl of original function
- and inlining infrastructure knows how to deal with this. */
- if (!flag_unit_at_a_time)
- {
- struct cgraph_edge *e;
-
- first_clone = cgraph_clone_node (node, node->count, 0, CGRAPH_FREQ_BASE,
- false);
- first_clone->needed = 0;
- first_clone->reachable = 1;
- /* Recursively clone all bodies. */
- for (e = first_clone->callees; e; e = e->next_callee)
- if (!e->inline_failed)
- cgraph_clone_inlined_nodes (e, true, false);
- }
- else
- first_clone = node->next_clone;
+ first_clone = node->clones;
first_clone->decl = copy_node (node->decl);
- node->next_clone = NULL;
- if (!flag_unit_at_a_time)
- node->inline_decl = first_clone->decl;
- first_clone->prev_clone = NULL;
cgraph_insert_node_to_hashtable (first_clone);
gcc_assert (first_clone == cgraph_node (first_clone->decl));
+ if (first_clone->next_sibling_clone)
+ {
+ for (n = first_clone->next_sibling_clone; n->next_sibling_clone; n = n->next_sibling_clone)
+ n->clone_of = first_clone;
+ n->clone_of = first_clone;
+ n->next_sibling_clone = first_clone->clones;
+ if (first_clone->clones)
+ first_clone->clones->prev_sibling_clone = n;
+ first_clone->clones = first_clone->next_sibling_clone;
+ first_clone->next_sibling_clone->prev_sibling_clone = NULL;
+ first_clone->next_sibling_clone = NULL;
+ gcc_assert (!first_clone->prev_sibling_clone);
+ }
+ first_clone->clone_of = NULL;
+ node->clones = NULL;
+
+ if (first_clone->clones)
+ for (n = first_clone->clones; n != first_clone;)
+ {
+ gcc_assert (n->decl == node->decl);
+ n->decl = first_clone->decl;
+ if (n->clones)
+ n = n->clones;
+ else if (n->next_sibling_clone)
+ n = n->next_sibling_clone;
+ else
+ {
+ while (n != first_clone && !n->next_sibling_clone)
+ n = n->clone_of;
+ if (n != first_clone)
+ n = n->next_sibling_clone;
+ }
+ }
/* Copy the OLD_VERSION_NODE function tree to the new version. */
- tree_function_versioning (node->decl, first_clone->decl, NULL, true);
+ tree_function_versioning (node->decl, first_clone->decl, NULL, true, NULL);
DECL_EXTERNAL (first_clone->decl) = 0;
- DECL_ONE_ONLY (first_clone->decl) = 0;
+ DECL_COMDAT_GROUP (first_clone->decl) = NULL_TREE;
TREE_PUBLIC (first_clone->decl) = 0;
DECL_COMDAT (first_clone->decl) = 0;
+ VEC_free (ipa_opt_pass, heap,
+ DECL_STRUCT_FUNCTION (first_clone->decl)->ipa_transforms_to_apply);
+ DECL_STRUCT_FUNCTION (first_clone->decl)->ipa_transforms_to_apply = NULL;
- for (node = first_clone->next_clone; node; node = node->next_clone)
- node->decl = first_clone->decl;
#ifdef ENABLE_CHECKING
verify_cgraph_node (first_clone);
#endif
return first_clone;
}
+/* Given virtual clone, turn it into actual clone. */
+static void
+cgraph_materialize_clone (struct cgraph_node *node)
+{
+ bitmap_obstack_initialize (NULL);
+ /* Copy the OLD_VERSION_NODE function tree to the new version. */
+ tree_function_versioning (node->clone_of->decl, node->decl,
+ node->clone.tree_map, true,
+ node->clone.args_to_skip);
+ if (cgraph_dump_file)
+ {
+ dump_function_to_file (node->clone_of->decl, cgraph_dump_file, dump_flags);
+ dump_function_to_file (node->decl, cgraph_dump_file, dump_flags);
+ }
+
+ /* Function is no longer clone. */
+ if (node->next_sibling_clone)
+ node->next_sibling_clone->prev_sibling_clone = node->prev_sibling_clone;
+ if (node->prev_sibling_clone)
+ node->prev_sibling_clone->next_sibling_clone = node->next_sibling_clone;
+ else
+ node->clone_of->clones = node->next_sibling_clone;
+ node->next_sibling_clone = NULL;
+ node->prev_sibling_clone = NULL;
+ node->clone_of = NULL;
+ bitmap_obstack_release (NULL);
+}
+
+/* Once all functions from compilation unit are in memory, produce all clones
+ and update all calls.
+ We might also do this on demand if we don't want to bring all functions to
+ memory prior compilation, but current WHOPR implementation does that and it is
+ is bit easier to keep everything right in this order. */
+void
+cgraph_materialize_all_clones (void)
+{
+ struct cgraph_node *node;
+ bool stabilized = false;
+
+ if (cgraph_dump_file)
+ fprintf (cgraph_dump_file, "Materializing clones\n");
+#ifdef ENABLE_CHECKING
+ verify_cgraph ();
+#endif
+
+ /* We can also do topological order, but number of iterations should be
+ bounded by number of IPA passes since single IPA pass is probably not
+ going to create clones of clones it created itself. */
+ while (!stabilized)
+ {
+ stabilized = true;
+ for (node = cgraph_nodes; node; node = node->next)
+ {
+ if (node->clone_of && node->decl != node->clone_of->decl
+ && !gimple_has_body_p (node->decl))
+ {
+ if (gimple_has_body_p (node->clone_of->decl))
+ {
+ if (cgraph_dump_file)
+ {
+ fprintf (cgraph_dump_file, "clonning %s to %s\n",
+ cgraph_node_name (node->clone_of),
+ cgraph_node_name (node));
+ if (node->clone.tree_map)
+ {
+ unsigned int i;
+ fprintf (cgraph_dump_file, " replace map: ");
+ for (i = 0; i < VEC_length (ipa_replace_map_p,
+ node->clone.tree_map);
+ i++)
+ {
+ struct ipa_replace_map *replace_info;
+ replace_info = VEC_index (ipa_replace_map_p,
+ node->clone.tree_map,
+ i);
+ print_generic_expr (cgraph_dump_file, replace_info->old_tree, 0);
+ fprintf (cgraph_dump_file, " -> ");
+ print_generic_expr (cgraph_dump_file, replace_info->new_tree, 0);
+ fprintf (cgraph_dump_file, "%s%s;",
+ replace_info->replace_p ? "(replace)":"",
+ replace_info->ref_p ? "(ref)":"");
+ }
+ fprintf (cgraph_dump_file, "\n");
+ }
+ if (node->clone.args_to_skip)
+ {
+ fprintf (cgraph_dump_file, " args_to_skip: ");
+ dump_bitmap (cgraph_dump_file, node->clone.args_to_skip);
+ }
+ if (node->clone.args_to_skip)
+ {
+ fprintf (cgraph_dump_file, " combined_args_to_skip:");
+ dump_bitmap (cgraph_dump_file, node->clone.combined_args_to_skip);
+ }
+ }
+ cgraph_materialize_clone (node);
+ }
+ else
+ stabilized = false;
+ }
+ }
+ }
+ if (cgraph_dump_file)
+ fprintf (cgraph_dump_file, "Updating call sites\n");
+ for (node = cgraph_nodes; node; node = node->next)
+ if (node->analyzed && gimple_has_body_p (node->decl)
+ && (!node->clone_of || node->clone_of->decl != node->decl))
+ {
+ struct cgraph_edge *e;
+
+ current_function_decl = node->decl;
+ push_cfun (DECL_STRUCT_FUNCTION (node->decl));
+ for (e = node->callees; e; e = e->next_callee)
+ {
+ tree decl = gimple_call_fndecl (e->call_stmt);
+ /* When function gets inlined, indirect inlining might've invented
+ new edge for orginally indirect stmt. Since we are not
+ preserving clones in the original form, we must not update here
+ since other inline clones don't need to contain call to the same
+ call. Inliner will do the substitution for us later. */
+ if (decl && decl != e->callee->decl)
+ {
+ gimple new_stmt;
+ gimple_stmt_iterator gsi;
+
+ if (cgraph_dump_file)
+ {
+ fprintf (cgraph_dump_file, "updating call of %s in %s:",
+ cgraph_node_name (node),
+ cgraph_node_name (e->callee));
+ print_gimple_stmt (cgraph_dump_file, e->call_stmt, 0, dump_flags);
+ }
+
+ if (e->callee->clone.combined_args_to_skip)
+ new_stmt = gimple_call_copy_skip_args (e->call_stmt,
+ e->callee->clone.combined_args_to_skip);
+ else
+ new_stmt = e->call_stmt;
+ if (gimple_vdef (new_stmt)
+ && TREE_CODE (gimple_vdef (new_stmt)) == SSA_NAME)
+ SSA_NAME_DEF_STMT (gimple_vdef (new_stmt)) = new_stmt;
+ gimple_call_set_fndecl (new_stmt, e->callee->decl);
+
+ gsi = gsi_for_stmt (e->call_stmt);
+ gsi_replace (&gsi, new_stmt, true);
+
+ /* Update EH information too, just in case. */
+ maybe_clean_or_replace_eh_stmt (e->call_stmt, new_stmt);
+
+ cgraph_set_call_stmt_including_clones (node, e->call_stmt, new_stmt);
+
+ if (cgraph_dump_file)
+ {
+ fprintf (cgraph_dump_file, " updated to:");
+ print_gimple_stmt (cgraph_dump_file, e->call_stmt, 0, dump_flags);
+ }
+ }
+ }
+ pop_cfun ();
+ current_function_decl = NULL;
+#ifdef ENABLE_CHECKING
+ verify_cgraph_node (node);
+#endif
+ }
+#ifdef ENABLE_CHECKING
+ verify_cgraph ();
+#endif
+ cgraph_remove_unreachable_nodes (false, cgraph_dump_file);
+}
+
#include "gt-cgraphunit.h"
OSDN Git Service
