Skip to content
GitLab
Commit 32f2fa83 authored by Matthias Braun's avatar Matthias Braun
Browse files

remove old buggy escape_analysis 

You can always revive it from older revision, although I would strongly
suffest to start from scratch.
parent e135087d
Hide whitespace changes
Inline Side-by-side
include/libfirm/iroptimize.h
View file @ 32f2fa83
......@@ -59,41 +59,6 @@ FIRM_API void conv_opt(ir_graph *irg);
*/
typedef int (*check_alloc_entity_func)(ir_entity *ent);
/**
* Performs simple and fast escape analysis for one graph.
*
* @param irg the graph
* @param callback a callback function to check whether a
* given entity is a allocation call
*/
FIRM_API void escape_enalysis_irg(ir_graph *irg,
check_alloc_entity_func callback);
/**
* Performs simple and fast escape analysis for all graphs.
*
* This optimization implements a simple and fast but inexact
* escape analysis. Some addresses might be marked as 'escaped' even
* if they are not.
* The advantage is a low memory footprint and fast speed.
*
* @param run_scalar_replace if this flag in non-zero, scalar
* replacement optimization is run on graphs with removed
* allocation
* @param callback a callback function to check whether a
* given entity is a allocation call
*
* This optimization removes allocation which are not used (rare) and replace
* allocation that can be proved dead at the end of the graph which stack variables.
*
* The creation of stack variable allows scalar replacement to be run only
* on those graphs that have been changed.
*
* This is most effective on Java where no other stack variables exists.
*/
FIRM_API void escape_analysis(int run_scalar_replace,
check_alloc_entity_func callback);
/**
* Optimize function calls by handling const functions.
*
......
ir/Makefile.am
View file @ 32f2fa83
......@@ -191,7 +191,6 @@ libfirm_la_SOURCES = \
opt/convopt.c \
opt/critical_edges.c \
opt/dead_code_elimination.c \
opt/escape_ana.c \
opt/fp-vrp.c \
opt/funccall.c \
opt/garbage_collect.c \
......
ir/opt/escape_ana.c deleted 100644 → 0
View file @ e135087d
/*
* This file is part of libFirm.
* Copyright (C) 2012 University of Karlsruhe.
*/
/**
* @file
* @author Michael Beck
* @date 03.11.2005
* @brief A fast and simple Escape analysis.
*/
#include "iroptimize.h"
#include "cgana.h"
#include "irgraph_t.h"
#include "irnode_t.h"
#include "type_t.h"
#include "irgwalk.h"
#include "irouts.h"
#include "analyze_irg_args.h"
#include "irgmod.h"
#include "ircons.h"
#include "irprintf.h"
#include "debug.h"
#include "error.h"
#include "util.h"
/**
* walker environment
*/
typedef struct walk_env {
ir_node *found_allocs; /**< list of all found non-escaped allocs */
ir_node *dead_allocs; /**< list of all found dead alloc */
check_alloc_entity_func callback; /**< callback that checks a given entity for allocation */
unsigned nr_removed; /**< number of removed allocs (placed of frame) */
unsigned nr_changed; /**< number of changed allocs (allocated on stack now) */
unsigned nr_deads; /**< number of dead allocs */
/* these fields are only used in the global escape analysis */
ir_graph *irg; /**< the irg for this environment */
struct walk_env *next; /**< for linking environments */
} walk_env_t;
/** debug handle */
DEBUG_ONLY(static firm_dbg_module_t *dbgHandle;)
/**
* checks whether a Raise leaves a method
*/
static int is_method_leaving_raise(ir_node *raise)
{
int i;
ir_node *proj = NULL;
ir_node *n;
for (i = get_irn_n_outs(raise) - 1; i >= 0; --i) {
ir_node *succ = get_irn_out(raise, i);
/* there should be only one ProjX node */
if (get_Proj_proj(succ) == pn_Raise_X) {
proj = succ;
break;
}
}
if (! proj) {
/* Hmm: no ProjX from a Raise? This should be a verification
* error. For now we just assert and return.
*/
panic("No ProjX after Raise found");
}
if (get_irn_n_outs(proj) != 1) {
/* Hmm: more than one user of ProjX: This is a verification
* error.
*/
panic("More than one user of ProjX");
}
n = get_irn_out(proj, 0);
assert(is_Block(n) && "Argh: user of ProjX is no block");
if (n == get_irg_end_block(get_irn_irg(n)))
return 1;
/* ok, we get here so the raise will not leave the function */
return 0;
}
/**
* returns an Alloc node if the node adr Select
* from one
*/
static ir_node *is_depend_alloc(ir_node *adr)
{
ir_node *alloc;
if (!is_Sel(adr))
return NULL;
/* should be a simple Sel */
if (get_Sel_n_indexs(adr) != 0)
return NULL;
alloc = skip_Proj(get_Sel_ptr(adr));
if (!is_Alloc(alloc))
return NULL;
/* hmm, we depend on this Alloc */
ir_printf("depend alloc %+F\n", alloc);
return NULL;
}
/**
* determine if a value calculated by n "escape", ie
* is stored somewhere we could not track
*/
static int can_escape(ir_node *n)
{
int i;
/* should always be pointer mode or we made some mistake */
assert(mode_is_reference(get_irn_mode(n)));
for (i = get_irn_n_outs(n) - 1; i >= 0; --i) {
ir_node *succ = get_irn_out(n, i);
switch (get_irn_opcode(succ)) {
case iro_Store:
if (get_Store_value(succ) == n) {
ir_node *adr = get_Store_ptr(succ);
/*
* if this Alloc depends on another one,
* we can enqueue it
*/
if (is_depend_alloc(adr))
break;
/*
* We are storing n. As long as we do not further
* evaluate things, the pointer 'escape' here
*/
return 1;
}
break;
case iro_Conv:
/*
* Should not happen, but if it does we leave the pointer
* path and do not track further
*/
return 1;
case iro_Call: { /* most complicated case */
ir_entity *callee = get_Call_callee(succ);
if (callee != NULL) {
/* we know the called entity */
for (size_t j = get_Call_n_params(succ); j > 0;) {
if (get_Call_param(succ, --j) == n) {
/* n is the j'th param of the call */
if (get_method_param_access(callee, j) & ptr_access_store)
/* n is store in ent */
return 1;
}
}
} else if (is_Sel(get_Call_ptr(succ))) {
size_t k;
/* go through all possible callees */
for (k = cg_get_call_n_callees(succ); k > 0;) {
size_t j;
ir_entity *ent = cg_get_call_callee(succ, --k);
if (is_unknown_entity(ent)) {
/* we don't know what will be called, a possible escape */
return 1;
}
for (j = get_Call_n_params(succ); j > 0;) {
if (get_Call_param(succ, --j) == n) {
/* n is the j'th param of the call */
if (get_method_param_access(ent, j) & ptr_access_store)
/* n is store in ent */
return 1;
}
}
}
} else /* we don't know want will called */
return 1;
break;
}
case iro_Return:
/* Bad: the allocate object is returned */
return 1;
case iro_Raise:
/* Hmm: if we do NOT leave the method, it's local */
if (is_method_leaving_raise(succ))
return 1;
break;
case iro_Tuple: {
ir_node *proj;
/* Bad: trace the tuple backwards */
int j = -1;
foreach_irn_in_r(succ, jj, pred) {
if (pred == n) {
j = jj;
break;
}
}
assert(j >= 0);
for (int k = get_irn_n_outs(succ); k >= 0; --k) {
proj = get_irn_out(succ, k);
if (get_Proj_proj(proj) == j) {
/* we found the right Proj */
succ = proj;
break;
}
}
/*
* If we haven't found the right Proj, succ is still
* the Tuple and the search will end here.
*/
break;
}
default:
break;
}
if (! mode_is_reference(get_irn_mode(succ)))
continue;
if (can_escape(succ))
return 1;
}
return 0;
}
/**
* walker: search for Alloc nodes and follow the usages
*/
static void find_allocations(ir_node *alloc, void *ctx)
{
/* TODO: check if we have a heap allocation */
(void)alloc;
(void)ctx;
return;
#if 0
walk_env_t *env = (walk_env_t*)ctx;
int i;
ir_node *adr;
adr = NULL;
for (i = get_irn_n_outs(alloc) - 1; i >= 0; --i) {
ir_node *proj = get_irn_out(alloc, i);
if (get_Proj_proj(proj) == pn_Alloc_res) {
adr = proj;
break;
}
}
if (! adr) {
/*
* bad: no-one wants the result, should NOT happen but
* if it does we could delete it.
*/
set_irn_link(alloc, env->dead_allocs);
env->dead_allocs = alloc;
return;
}
if (! can_escape(adr)) {
set_irn_link(alloc, env->found_allocs);
env->found_allocs = alloc;
}
#endif
}
/**
* walker: search for allocation Call nodes and follow the usages
*/
static void find_allocation_calls(ir_node *call, void *ctx)
{
walk_env_t *env = (walk_env_t*)ctx;
int i;
ir_node *adr;
if (! is_Call(call))
return;
adr = get_Call_ptr(call);
if (!is_Address(adr))
return;
ir_entity *const ent = get_Address_entity(adr);
if (! env->callback(ent))
return;
adr = NULL;
for (i = get_irn_n_outs(call) - 1; i >= 0; --i) {
ir_node *res_proj = get_irn_out(call, i);
if (get_Proj_proj(res_proj) == pn_Call_T_result) {
for (i = get_irn_n_outs(res_proj) - 1; i >= 0; --i) {
ir_node *proj = get_irn_out(res_proj, i);
if (get_Proj_proj(proj) == 0) {
/* found first result */
adr = proj;
break;
}
}
break;
}
}
if (! adr) {
/*
* bad: no-one wants the result, should NOT happen but
* if it does we could delete it.
*/
set_irn_link(call, env->dead_allocs);
env->dead_allocs = call;
return;
}
if (! can_escape(adr)) {
set_irn_link(call, env->found_allocs);
env->found_allocs = call;
}
}
/**
* Do the necessary graph transformations to transform
* Alloc nodes.
*/
static void transform_allocs(ir_graph *irg, walk_env_t *env)
{
(void)irg;
(void)env;
/* TODO: fix */
#if 0
ir_node *alloc, *next, *mem, *sel, *size, *blk;
ir_type *ftp, *atp, *tp;
ir_entity *ent;
char name[128];
unsigned nr = 0;
dbg_info *dbg;
/* kill all dead allocs */
for (alloc = env->dead_allocs; alloc; alloc = next) {
next = (ir_node*)get_irn_link(alloc);
DBG((dbgHandle, LEVEL_1, "%+F allocation of %+F unused, deleted.\n", irg, alloc));
mem = get_Alloc_mem(alloc);
blk = get_nodes_block(alloc);
ir_node *const in[] = {
[pn_Alloc_M] = mem,
};
turn_into_tuple(alloc, ARRAY_SIZE(in), in);
++env->nr_deads;
}
/* convert all non-escaped heap allocs into frame variables */
ftp = get_irg_frame_type(irg);
for (alloc = env->found_allocs; alloc; alloc = next) {
next = (ir_node*)get_irn_link(alloc);
size = get_Alloc_size(alloc);
unsigned alignment = get_Alloc_alignment(alloc);
tp = NULL;
if (is_TypeConst(size) && get_TypeConst_kind(size) == typeconst_size) {
/* if the size is a type size and the types matched */
tp = atp;
} else if (is_Const(size)) {
ir_tarval *tv = get_Const_tarval(size);
if (tarval_is_long(tv) &&
get_type_state(atp) == layout_fixed &&
(unsigned)get_tarval_long(tv) == get_type_size_bytes(atp)) {
/* a already lowered type size */
tp = atp;
}
}
if (tp && !is_unknown_type(tp)) {
/* we could determine the type, so we could place it on the frame */
dbg = get_irn_dbg_info(alloc);
blk = get_nodes_block(alloc);
DBG((dbgHandle, LEVEL_DEFAULT, "%+F allocation of %+F type %+F placed on frame\n", irg, alloc, tp));
snprintf(name, sizeof(name), "%s_NE_%u", get_entity_name(get_irg_entity(irg)), nr++);
name[sizeof(name) - 1] = '\0';
ent = new_d_entity(ftp, new_id_from_str(name), get_Alloc_type(alloc), dbg);
sel = new_rd_simpleSel(dbg, get_nodes_block(alloc), get_irg_no_mem(irg), get_irg_frame(irg), ent);
mem = get_Alloc_mem(alloc);
ir_node *const in[] = {
[pn_Alloc_M] = mem,
[pn_Alloc_res] = sel,
};
turn_into_tuple(alloc, ARRAY_SIZE(in), in);
++env->nr_removed;
}
else {
/*
* We could not determine the type or it is variable size.
* At least, we could place it on the stack
*/
DBG((dbgHandle, LEVEL_DEFAULT, "%+F allocation of %+F type %+F placed on stack\n", irg, alloc));
set_Alloc_where(alloc, stack_alloc);
++env->nr_changed;
}
}
/* if allocs were removed somehow */
if (env->nr_removed && env->nr_deads) {
confirm_irg_properties(irg, IR_GRAPH_PROPERTIES_NONE);
}
#endif
}
/**
* Do the necessary graph transformations to transform
* Call nodes.
*/
static void transform_alloc_calls(ir_graph *irg, walk_env_t *env)
{
ir_node *call, *next, *mem, *blk;
/* kill all dead allocs */
for (call = env->dead_allocs; call; call = next) {
next = (ir_node*)get_irn_link(call);
DBG((dbgHandle, LEVEL_1, "%+F allocation of %+F unused, deleted.\n", irg, call));
mem = get_Call_mem(call);
blk = get_nodes_block(call);
ir_node *const in[] = {
[pn_Call_M] = mem,
[pn_Call_T_result] = new_r_Bad(irg, mode_T),
[pn_Call_X_regular] = new_r_Jmp(blk),
[pn_Call_X_except] = new_r_Bad(irg, mode_X),
};
turn_into_tuple(call, ARRAY_SIZE(in), in);
++env->nr_deads;
}
/* convert all non-escaped heap allocs into frame variables */
for (call = env->found_allocs; call; call = next) {
next = (ir_node*)get_irn_link(call);
}
confirm_irg_properties(irg, IR_GRAPH_PROPERTIES_NONE);
}
/* Do simple and fast escape analysis for one graph. */
void escape_enalysis_irg(ir_graph *irg, check_alloc_entity_func callback)
{
walk_env_t env;
if (get_irg_callee_info_state(irg) != irg_callee_info_consistent) {
/* no way yet to calculate this for one irg */
assert(! "need callee info");
return;
}
if (irg_has_properties(irg, IR_GRAPH_PROPERTY_CONSISTENT_OUTS))
compute_irg_outs(irg);
env.found_allocs = NULL;
env.dead_allocs = NULL;
env.callback = callback;
env.nr_removed = 0;
env.nr_changed = 0;
env.nr_deads = 0;
if (callback) {
/* search for Calls */
irg_walk_graph(irg, NULL, find_allocation_calls, &env);
transform_alloc_calls(irg, &env);
} else {
/* search for Alloc nodes */
irg_walk_graph(irg, NULL, find_allocations, &env);
transform_allocs(irg, &env);
}
}
/* Do simple and fast escape analysis for all graphs. */
void escape_analysis(int run_scalar_replace, check_alloc_entity_func callback)
{
size_t i, n;
struct obstack obst;
walk_env_t *env, *elist;
(void) run_scalar_replace;
if (get_irp_callee_info_state() != irg_callee_info_consistent) {
assert(! "need callee info");
return;
}
FIRM_DBG_REGISTER(dbgHandle, "firm.opt.escape_ana");
/*
* We treat memory for speed: we first collect all info in a
* list of environments, than do the transformation.
* Doing it this way, no analysis info gets invalid while we run
* over graphs
*/
obstack_init(&obst);
elist = NULL;
env = OALLOC(&obst, walk_env_t);
env->found_allocs = NULL;
env->dead_allocs = NULL;
env->callback = callback;
for (i = 0, n = get_irp_n_irgs(); i < n; ++i) {
ir_graph *irg = get_irp_irg(i);
assure_irg_properties(irg, IR_GRAPH_PROPERTY_CONSISTENT_OUTS);
if (callback) {
/* search for Calls */
irg_walk_graph(irg, NULL, find_allocation_calls, env);
} else {
/* search for Alloc nodes */
irg_walk_graph(irg, NULL, find_allocations, env);
}
if (env->found_allocs || env->dead_allocs) {
env->nr_removed = 0;
env->nr_deads = 0;
env->irg = irg;
env->next = elist;
elist = env;
env = OALLOC(&obst, walk_env_t);
env->found_allocs = NULL;
env->dead_allocs = NULL;
env->callback = callback;
}
}
if (callback) {
for (env = elist; env; env = env->next) {
transform_alloc_calls(env->irg, env);
}
} else {
for (env = elist; env; env = env->next) {
transform_allocs(env->irg, env);
}
}
obstack_free(&obst, NULL);
}
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!