/* Thread edges through blocks and update the control flow and SSA graphs.
|
/* Thread edges through blocks and update the control flow and SSA graphs.
|
Copyright (C) 2004, 2005, 2006, 2007, 2008 Free Software Foundation,
|
Copyright (C) 2004, 2005, 2006, 2007, 2008 Free Software Foundation,
|
Inc.
|
Inc.
|
|
|
This file is part of GCC.
|
This file is part of GCC.
|
|
|
GCC is free software; you can redistribute it and/or modify
|
GCC is free software; you can redistribute it and/or modify
|
it under the terms of the GNU General Public License as published by
|
it under the terms of the GNU General Public License as published by
|
the Free Software Foundation; either version 3, or (at your option)
|
the Free Software Foundation; either version 3, or (at your option)
|
any later version.
|
any later version.
|
|
|
GCC is distributed in the hope that it will be useful,
|
GCC is distributed in the hope that it will be useful,
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
but WITHOUT ANY WARRANTY; without even the implied warranty of
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
GNU General Public License for more details.
|
GNU General Public License for more details.
|
|
|
You should have received a copy of the GNU General Public License
|
You should have received a copy of the GNU General Public License
|
along with GCC; see the file COPYING3. If not see
|
along with GCC; see the file COPYING3. If not see
|
<http://www.gnu.org/licenses/>. */
|
<http://www.gnu.org/licenses/>. */
|
|
|
#include "config.h"
|
#include "config.h"
|
#include "system.h"
|
#include "system.h"
|
#include "coretypes.h"
|
#include "coretypes.h"
|
#include "tm.h"
|
#include "tm.h"
|
#include "tree.h"
|
#include "tree.h"
|
#include "flags.h"
|
#include "flags.h"
|
#include "rtl.h"
|
#include "rtl.h"
|
#include "tm_p.h"
|
#include "tm_p.h"
|
#include "ggc.h"
|
#include "ggc.h"
|
#include "basic-block.h"
|
#include "basic-block.h"
|
#include "output.h"
|
#include "output.h"
|
#include "expr.h"
|
#include "expr.h"
|
#include "function.h"
|
#include "function.h"
|
#include "diagnostic.h"
|
#include "diagnostic.h"
|
#include "tree-flow.h"
|
#include "tree-flow.h"
|
#include "tree-dump.h"
|
#include "tree-dump.h"
|
#include "tree-pass.h"
|
#include "tree-pass.h"
|
#include "cfgloop.h"
|
#include "cfgloop.h"
|
|
|
/* Given a block B, update the CFG and SSA graph to reflect redirecting
|
/* Given a block B, update the CFG and SSA graph to reflect redirecting
|
one or more in-edges to B to instead reach the destination of an
|
one or more in-edges to B to instead reach the destination of an
|
out-edge from B while preserving any side effects in B.
|
out-edge from B while preserving any side effects in B.
|
|
|
i.e., given A->B and B->C, change A->B to be A->C yet still preserve the
|
i.e., given A->B and B->C, change A->B to be A->C yet still preserve the
|
side effects of executing B.
|
side effects of executing B.
|
|
|
1. Make a copy of B (including its outgoing edges and statements). Call
|
1. Make a copy of B (including its outgoing edges and statements). Call
|
the copy B'. Note B' has no incoming edges or PHIs at this time.
|
the copy B'. Note B' has no incoming edges or PHIs at this time.
|
|
|
2. Remove the control statement at the end of B' and all outgoing edges
|
2. Remove the control statement at the end of B' and all outgoing edges
|
except B'->C.
|
except B'->C.
|
|
|
3. Add a new argument to each PHI in C with the same value as the existing
|
3. Add a new argument to each PHI in C with the same value as the existing
|
argument associated with edge B->C. Associate the new PHI arguments
|
argument associated with edge B->C. Associate the new PHI arguments
|
with the edge B'->C.
|
with the edge B'->C.
|
|
|
4. For each PHI in B, find or create a PHI in B' with an identical
|
4. For each PHI in B, find or create a PHI in B' with an identical
|
PHI_RESULT. Add an argument to the PHI in B' which has the same
|
PHI_RESULT. Add an argument to the PHI in B' which has the same
|
value as the PHI in B associated with the edge A->B. Associate
|
value as the PHI in B associated with the edge A->B. Associate
|
the new argument in the PHI in B' with the edge A->B.
|
the new argument in the PHI in B' with the edge A->B.
|
|
|
5. Change the edge A->B to A->B'.
|
5. Change the edge A->B to A->B'.
|
|
|
5a. This automatically deletes any PHI arguments associated with the
|
5a. This automatically deletes any PHI arguments associated with the
|
edge A->B in B.
|
edge A->B in B.
|
|
|
5b. This automatically associates each new argument added in step 4
|
5b. This automatically associates each new argument added in step 4
|
with the edge A->B'.
|
with the edge A->B'.
|
|
|
6. Repeat for other incoming edges into B.
|
6. Repeat for other incoming edges into B.
|
|
|
7. Put the duplicated resources in B and all the B' blocks into SSA form.
|
7. Put the duplicated resources in B and all the B' blocks into SSA form.
|
|
|
Note that block duplication can be minimized by first collecting the
|
Note that block duplication can be minimized by first collecting the
|
set of unique destination blocks that the incoming edges should
|
set of unique destination blocks that the incoming edges should
|
be threaded to. Block duplication can be further minimized by using
|
be threaded to. Block duplication can be further minimized by using
|
B instead of creating B' for one destination if all edges into B are
|
B instead of creating B' for one destination if all edges into B are
|
going to be threaded to a successor of B.
|
going to be threaded to a successor of B.
|
|
|
We further reduce the number of edges and statements we create by
|
We further reduce the number of edges and statements we create by
|
not copying all the outgoing edges and the control statement in
|
not copying all the outgoing edges and the control statement in
|
step #1. We instead create a template block without the outgoing
|
step #1. We instead create a template block without the outgoing
|
edges and duplicate the template. */
|
edges and duplicate the template. */
|
|
|
|
|
/* Steps #5 and #6 of the above algorithm are best implemented by walking
|
/* Steps #5 and #6 of the above algorithm are best implemented by walking
|
all the incoming edges which thread to the same destination edge at
|
all the incoming edges which thread to the same destination edge at
|
the same time. That avoids lots of table lookups to get information
|
the same time. That avoids lots of table lookups to get information
|
for the destination edge.
|
for the destination edge.
|
|
|
To realize that implementation we create a list of incoming edges
|
To realize that implementation we create a list of incoming edges
|
which thread to the same outgoing edge. Thus to implement steps
|
which thread to the same outgoing edge. Thus to implement steps
|
#5 and #6 we traverse our hash table of outgoing edge information.
|
#5 and #6 we traverse our hash table of outgoing edge information.
|
For each entry we walk the list of incoming edges which thread to
|
For each entry we walk the list of incoming edges which thread to
|
the current outgoing edge. */
|
the current outgoing edge. */
|
|
|
struct el
|
struct el
|
{
|
{
|
edge e;
|
edge e;
|
struct el *next;
|
struct el *next;
|
};
|
};
|
|
|
/* Main data structure recording information regarding B's duplicate
|
/* Main data structure recording information regarding B's duplicate
|
blocks. */
|
blocks. */
|
|
|
/* We need to efficiently record the unique thread destinations of this
|
/* We need to efficiently record the unique thread destinations of this
|
block and specific information associated with those destinations. We
|
block and specific information associated with those destinations. We
|
may have many incoming edges threaded to the same outgoing edge. This
|
may have many incoming edges threaded to the same outgoing edge. This
|
can be naturally implemented with a hash table. */
|
can be naturally implemented with a hash table. */
|
|
|
struct redirection_data
|
struct redirection_data
|
{
|
{
|
/* A duplicate of B with the trailing control statement removed and which
|
/* A duplicate of B with the trailing control statement removed and which
|
targets a single successor of B. */
|
targets a single successor of B. */
|
basic_block dup_block;
|
basic_block dup_block;
|
|
|
/* An outgoing edge from B. DUP_BLOCK will have OUTGOING_EDGE->dest as
|
/* An outgoing edge from B. DUP_BLOCK will have OUTGOING_EDGE->dest as
|
its single successor. */
|
its single successor. */
|
edge outgoing_edge;
|
edge outgoing_edge;
|
|
|
/* A list of incoming edges which we want to thread to
|
/* A list of incoming edges which we want to thread to
|
OUTGOING_EDGE->dest. */
|
OUTGOING_EDGE->dest. */
|
struct el *incoming_edges;
|
struct el *incoming_edges;
|
|
|
/* Flag indicating whether or not we should create a duplicate block
|
/* Flag indicating whether or not we should create a duplicate block
|
for this thread destination. This is only true if we are threading
|
for this thread destination. This is only true if we are threading
|
all incoming edges and thus are using BB itself as a duplicate block. */
|
all incoming edges and thus are using BB itself as a duplicate block. */
|
bool do_not_duplicate;
|
bool do_not_duplicate;
|
};
|
};
|
|
|
/* Main data structure to hold information for duplicates of BB. */
|
/* Main data structure to hold information for duplicates of BB. */
|
static htab_t redirection_data;
|
static htab_t redirection_data;
|
|
|
/* Data structure of information to pass to hash table traversal routines. */
|
/* Data structure of information to pass to hash table traversal routines. */
|
struct local_info
|
struct local_info
|
{
|
{
|
/* The current block we are working on. */
|
/* The current block we are working on. */
|
basic_block bb;
|
basic_block bb;
|
|
|
/* A template copy of BB with no outgoing edges or control statement that
|
/* A template copy of BB with no outgoing edges or control statement that
|
we use for creating copies. */
|
we use for creating copies. */
|
basic_block template_block;
|
basic_block template_block;
|
|
|
/* TRUE if we thread one or more jumps, FALSE otherwise. */
|
/* TRUE if we thread one or more jumps, FALSE otherwise. */
|
bool jumps_threaded;
|
bool jumps_threaded;
|
};
|
};
|
|
|
/* Passes which use the jump threading code register jump threading
|
/* Passes which use the jump threading code register jump threading
|
opportunities as they are discovered. We keep the registered
|
opportunities as they are discovered. We keep the registered
|
jump threading opportunities in this vector as edge pairs
|
jump threading opportunities in this vector as edge pairs
|
(original_edge, target_edge). */
|
(original_edge, target_edge). */
|
static VEC(edge,heap) *threaded_edges;
|
static VEC(edge,heap) *threaded_edges;
|
|
|
|
|
/* Jump threading statistics. */
|
/* Jump threading statistics. */
|
|
|
struct thread_stats_d
|
struct thread_stats_d
|
{
|
{
|
unsigned long num_threaded_edges;
|
unsigned long num_threaded_edges;
|
};
|
};
|
|
|
struct thread_stats_d thread_stats;
|
struct thread_stats_d thread_stats;
|
|
|
|
|
/* Remove the last statement in block BB if it is a control statement
|
/* Remove the last statement in block BB if it is a control statement
|
Also remove all outgoing edges except the edge which reaches DEST_BB.
|
Also remove all outgoing edges except the edge which reaches DEST_BB.
|
If DEST_BB is NULL, then remove all outgoing edges. */
|
If DEST_BB is NULL, then remove all outgoing edges. */
|
|
|
static void
|
static void
|
remove_ctrl_stmt_and_useless_edges (basic_block bb, basic_block dest_bb)
|
remove_ctrl_stmt_and_useless_edges (basic_block bb, basic_block dest_bb)
|
{
|
{
|
gimple_stmt_iterator gsi;
|
gimple_stmt_iterator gsi;
|
edge e;
|
edge e;
|
edge_iterator ei;
|
edge_iterator ei;
|
|
|
gsi = gsi_last_bb (bb);
|
gsi = gsi_last_bb (bb);
|
|
|
/* If the duplicate ends with a control statement, then remove it.
|
/* If the duplicate ends with a control statement, then remove it.
|
|
|
Note that if we are duplicating the template block rather than the
|
Note that if we are duplicating the template block rather than the
|
original basic block, then the duplicate might not have any real
|
original basic block, then the duplicate might not have any real
|
statements in it. */
|
statements in it. */
|
if (!gsi_end_p (gsi)
|
if (!gsi_end_p (gsi)
|
&& gsi_stmt (gsi)
|
&& gsi_stmt (gsi)
|
&& (gimple_code (gsi_stmt (gsi)) == GIMPLE_COND
|
&& (gimple_code (gsi_stmt (gsi)) == GIMPLE_COND
|
|| gimple_code (gsi_stmt (gsi)) == GIMPLE_GOTO
|
|| gimple_code (gsi_stmt (gsi)) == GIMPLE_GOTO
|
|| gimple_code (gsi_stmt (gsi)) == GIMPLE_SWITCH))
|
|| gimple_code (gsi_stmt (gsi)) == GIMPLE_SWITCH))
|
gsi_remove (&gsi, true);
|
gsi_remove (&gsi, true);
|
|
|
for (ei = ei_start (bb->succs); (e = ei_safe_edge (ei)); )
|
for (ei = ei_start (bb->succs); (e = ei_safe_edge (ei)); )
|
{
|
{
|
if (e->dest != dest_bb)
|
if (e->dest != dest_bb)
|
remove_edge (e);
|
remove_edge (e);
|
else
|
else
|
ei_next (&ei);
|
ei_next (&ei);
|
}
|
}
|
}
|
}
|
|
|
/* Create a duplicate of BB which only reaches the destination of the edge
|
/* Create a duplicate of BB which only reaches the destination of the edge
|
stored in RD. Record the duplicate block in RD. */
|
stored in RD. Record the duplicate block in RD. */
|
|
|
static void
|
static void
|
create_block_for_threading (basic_block bb, struct redirection_data *rd)
|
create_block_for_threading (basic_block bb, struct redirection_data *rd)
|
{
|
{
|
/* We can use the generic block duplication code and simply remove
|
/* We can use the generic block duplication code and simply remove
|
the stuff we do not need. */
|
the stuff we do not need. */
|
rd->dup_block = duplicate_block (bb, NULL, NULL);
|
rd->dup_block = duplicate_block (bb, NULL, NULL);
|
|
|
/* Zero out the profile, since the block is unreachable for now. */
|
/* Zero out the profile, since the block is unreachable for now. */
|
rd->dup_block->frequency = 0;
|
rd->dup_block->frequency = 0;
|
rd->dup_block->count = 0;
|
rd->dup_block->count = 0;
|
|
|
/* The call to duplicate_block will copy everything, including the
|
/* The call to duplicate_block will copy everything, including the
|
useless COND_EXPR or SWITCH_EXPR at the end of BB. We just remove
|
useless COND_EXPR or SWITCH_EXPR at the end of BB. We just remove
|
the useless COND_EXPR or SWITCH_EXPR here rather than having a
|
the useless COND_EXPR or SWITCH_EXPR here rather than having a
|
specialized block copier. We also remove all outgoing edges
|
specialized block copier. We also remove all outgoing edges
|
from the duplicate block. The appropriate edge will be created
|
from the duplicate block. The appropriate edge will be created
|
later. */
|
later. */
|
remove_ctrl_stmt_and_useless_edges (rd->dup_block, NULL);
|
remove_ctrl_stmt_and_useless_edges (rd->dup_block, NULL);
|
}
|
}
|
|
|
/* Hashing and equality routines for our hash table. */
|
/* Hashing and equality routines for our hash table. */
|
static hashval_t
|
static hashval_t
|
redirection_data_hash (const void *p)
|
redirection_data_hash (const void *p)
|
{
|
{
|
edge e = ((const struct redirection_data *)p)->outgoing_edge;
|
edge e = ((const struct redirection_data *)p)->outgoing_edge;
|
return e->dest->index;
|
return e->dest->index;
|
}
|
}
|
|
|
static int
|
static int
|
redirection_data_eq (const void *p1, const void *p2)
|
redirection_data_eq (const void *p1, const void *p2)
|
{
|
{
|
edge e1 = ((const struct redirection_data *)p1)->outgoing_edge;
|
edge e1 = ((const struct redirection_data *)p1)->outgoing_edge;
|
edge e2 = ((const struct redirection_data *)p2)->outgoing_edge;
|
edge e2 = ((const struct redirection_data *)p2)->outgoing_edge;
|
|
|
return e1 == e2;
|
return e1 == e2;
|
}
|
}
|
|
|
/* Given an outgoing edge E lookup and return its entry in our hash table.
|
/* Given an outgoing edge E lookup and return its entry in our hash table.
|
|
|
If INSERT is true, then we insert the entry into the hash table if
|
If INSERT is true, then we insert the entry into the hash table if
|
it is not already present. INCOMING_EDGE is added to the list of incoming
|
it is not already present. INCOMING_EDGE is added to the list of incoming
|
edges associated with E in the hash table. */
|
edges associated with E in the hash table. */
|
|
|
static struct redirection_data *
|
static struct redirection_data *
|
lookup_redirection_data (edge e, edge incoming_edge, enum insert_option insert)
|
lookup_redirection_data (edge e, edge incoming_edge, enum insert_option insert)
|
{
|
{
|
void **slot;
|
void **slot;
|
struct redirection_data *elt;
|
struct redirection_data *elt;
|
|
|
/* Build a hash table element so we can see if E is already
|
/* Build a hash table element so we can see if E is already
|
in the table. */
|
in the table. */
|
elt = XNEW (struct redirection_data);
|
elt = XNEW (struct redirection_data);
|
elt->outgoing_edge = e;
|
elt->outgoing_edge = e;
|
elt->dup_block = NULL;
|
elt->dup_block = NULL;
|
elt->do_not_duplicate = false;
|
elt->do_not_duplicate = false;
|
elt->incoming_edges = NULL;
|
elt->incoming_edges = NULL;
|
|
|
slot = htab_find_slot (redirection_data, elt, insert);
|
slot = htab_find_slot (redirection_data, elt, insert);
|
|
|
/* This will only happen if INSERT is false and the entry is not
|
/* This will only happen if INSERT is false and the entry is not
|
in the hash table. */
|
in the hash table. */
|
if (slot == NULL)
|
if (slot == NULL)
|
{
|
{
|
free (elt);
|
free (elt);
|
return NULL;
|
return NULL;
|
}
|
}
|
|
|
/* This will only happen if E was not in the hash table and
|
/* This will only happen if E was not in the hash table and
|
INSERT is true. */
|
INSERT is true. */
|
if (*slot == NULL)
|
if (*slot == NULL)
|
{
|
{
|
*slot = (void *)elt;
|
*slot = (void *)elt;
|
elt->incoming_edges = XNEW (struct el);
|
elt->incoming_edges = XNEW (struct el);
|
elt->incoming_edges->e = incoming_edge;
|
elt->incoming_edges->e = incoming_edge;
|
elt->incoming_edges->next = NULL;
|
elt->incoming_edges->next = NULL;
|
return elt;
|
return elt;
|
}
|
}
|
/* E was in the hash table. */
|
/* E was in the hash table. */
|
else
|
else
|
{
|
{
|
/* Free ELT as we do not need it anymore, we will extract the
|
/* Free ELT as we do not need it anymore, we will extract the
|
relevant entry from the hash table itself. */
|
relevant entry from the hash table itself. */
|
free (elt);
|
free (elt);
|
|
|
/* Get the entry stored in the hash table. */
|
/* Get the entry stored in the hash table. */
|
elt = (struct redirection_data *) *slot;
|
elt = (struct redirection_data *) *slot;
|
|
|
/* If insertion was requested, then we need to add INCOMING_EDGE
|
/* If insertion was requested, then we need to add INCOMING_EDGE
|
to the list of incoming edges associated with E. */
|
to the list of incoming edges associated with E. */
|
if (insert)
|
if (insert)
|
{
|
{
|
struct el *el = XNEW (struct el);
|
struct el *el = XNEW (struct el);
|
el->next = elt->incoming_edges;
|
el->next = elt->incoming_edges;
|
el->e = incoming_edge;
|
el->e = incoming_edge;
|
elt->incoming_edges = el;
|
elt->incoming_edges = el;
|
}
|
}
|
|
|
return elt;
|
return elt;
|
}
|
}
|
}
|
}
|
|
|
/* Given a duplicate block and its single destination (both stored
|
/* Given a duplicate block and its single destination (both stored
|
in RD). Create an edge between the duplicate and its single
|
in RD). Create an edge between the duplicate and its single
|
destination.
|
destination.
|
|
|
Add an additional argument to any PHI nodes at the single
|
Add an additional argument to any PHI nodes at the single
|
destination. */
|
destination. */
|
|
|
static void
|
static void
|
create_edge_and_update_destination_phis (struct redirection_data *rd)
|
create_edge_and_update_destination_phis (struct redirection_data *rd)
|
{
|
{
|
edge e = make_edge (rd->dup_block, rd->outgoing_edge->dest, EDGE_FALLTHRU);
|
edge e = make_edge (rd->dup_block, rd->outgoing_edge->dest, EDGE_FALLTHRU);
|
gimple_stmt_iterator gsi;
|
gimple_stmt_iterator gsi;
|
|
|
rescan_loop_exit (e, true, false);
|
rescan_loop_exit (e, true, false);
|
e->probability = REG_BR_PROB_BASE;
|
e->probability = REG_BR_PROB_BASE;
|
e->count = rd->dup_block->count;
|
e->count = rd->dup_block->count;
|
e->aux = rd->outgoing_edge->aux;
|
e->aux = rd->outgoing_edge->aux;
|
|
|
/* If there are any PHI nodes at the destination of the outgoing edge
|
/* If there are any PHI nodes at the destination of the outgoing edge
|
from the duplicate block, then we will need to add a new argument
|
from the duplicate block, then we will need to add a new argument
|
to them. The argument should have the same value as the argument
|
to them. The argument should have the same value as the argument
|
associated with the outgoing edge stored in RD. */
|
associated with the outgoing edge stored in RD. */
|
for (gsi = gsi_start_phis (e->dest); !gsi_end_p (gsi); gsi_next (&gsi))
|
for (gsi = gsi_start_phis (e->dest); !gsi_end_p (gsi); gsi_next (&gsi))
|
{
|
{
|
gimple phi = gsi_stmt (gsi);
|
gimple phi = gsi_stmt (gsi);
|
source_location locus;
|
source_location locus;
|
int indx = rd->outgoing_edge->dest_idx;
|
int indx = rd->outgoing_edge->dest_idx;
|
|
|
locus = gimple_phi_arg_location (phi, indx);
|
locus = gimple_phi_arg_location (phi, indx);
|
add_phi_arg (phi, gimple_phi_arg_def (phi, indx), e, locus);
|
add_phi_arg (phi, gimple_phi_arg_def (phi, indx), e, locus);
|
}
|
}
|
}
|
}
|
|
|
/* Hash table traversal callback routine to create duplicate blocks. */
|
/* Hash table traversal callback routine to create duplicate blocks. */
|
|
|
static int
|
static int
|
create_duplicates (void **slot, void *data)
|
create_duplicates (void **slot, void *data)
|
{
|
{
|
struct redirection_data *rd = (struct redirection_data *) *slot;
|
struct redirection_data *rd = (struct redirection_data *) *slot;
|
struct local_info *local_info = (struct local_info *)data;
|
struct local_info *local_info = (struct local_info *)data;
|
|
|
/* If this entry should not have a duplicate created, then there's
|
/* If this entry should not have a duplicate created, then there's
|
nothing to do. */
|
nothing to do. */
|
if (rd->do_not_duplicate)
|
if (rd->do_not_duplicate)
|
return 1;
|
return 1;
|
|
|
/* Create a template block if we have not done so already. Otherwise
|
/* Create a template block if we have not done so already. Otherwise
|
use the template to create a new block. */
|
use the template to create a new block. */
|
if (local_info->template_block == NULL)
|
if (local_info->template_block == NULL)
|
{
|
{
|
create_block_for_threading (local_info->bb, rd);
|
create_block_for_threading (local_info->bb, rd);
|
local_info->template_block = rd->dup_block;
|
local_info->template_block = rd->dup_block;
|
|
|
/* We do not create any outgoing edges for the template. We will
|
/* We do not create any outgoing edges for the template. We will
|
take care of that in a later traversal. That way we do not
|
take care of that in a later traversal. That way we do not
|
create edges that are going to just be deleted. */
|
create edges that are going to just be deleted. */
|
}
|
}
|
else
|
else
|
{
|
{
|
create_block_for_threading (local_info->template_block, rd);
|
create_block_for_threading (local_info->template_block, rd);
|
|
|
/* Go ahead and wire up outgoing edges and update PHIs for the duplicate
|
/* Go ahead and wire up outgoing edges and update PHIs for the duplicate
|
block. */
|
block. */
|
create_edge_and_update_destination_phis (rd);
|
create_edge_and_update_destination_phis (rd);
|
}
|
}
|
|
|
/* Keep walking the hash table. */
|
/* Keep walking the hash table. */
|
return 1;
|
return 1;
|
}
|
}
|
|
|
/* We did not create any outgoing edges for the template block during
|
/* We did not create any outgoing edges for the template block during
|
block creation. This hash table traversal callback creates the
|
block creation. This hash table traversal callback creates the
|
outgoing edge for the template block. */
|
outgoing edge for the template block. */
|
|
|
static int
|
static int
|
fixup_template_block (void **slot, void *data)
|
fixup_template_block (void **slot, void *data)
|
{
|
{
|
struct redirection_data *rd = (struct redirection_data *) *slot;
|
struct redirection_data *rd = (struct redirection_data *) *slot;
|
struct local_info *local_info = (struct local_info *)data;
|
struct local_info *local_info = (struct local_info *)data;
|
|
|
/* If this is the template block, then create its outgoing edges
|
/* If this is the template block, then create its outgoing edges
|
and halt the hash table traversal. */
|
and halt the hash table traversal. */
|
if (rd->dup_block && rd->dup_block == local_info->template_block)
|
if (rd->dup_block && rd->dup_block == local_info->template_block)
|
{
|
{
|
create_edge_and_update_destination_phis (rd);
|
create_edge_and_update_destination_phis (rd);
|
return 0;
|
return 0;
|
}
|
}
|
|
|
return 1;
|
return 1;
|
}
|
}
|
|
|
/* Hash table traversal callback to redirect each incoming edge
|
/* Hash table traversal callback to redirect each incoming edge
|
associated with this hash table element to its new destination. */
|
associated with this hash table element to its new destination. */
|
|
|
static int
|
static int
|
redirect_edges (void **slot, void *data)
|
redirect_edges (void **slot, void *data)
|
{
|
{
|
struct redirection_data *rd = (struct redirection_data *) *slot;
|
struct redirection_data *rd = (struct redirection_data *) *slot;
|
struct local_info *local_info = (struct local_info *)data;
|
struct local_info *local_info = (struct local_info *)data;
|
struct el *next, *el;
|
struct el *next, *el;
|
|
|
/* Walk over all the incoming edges associated associated with this
|
/* Walk over all the incoming edges associated associated with this
|
hash table entry. */
|
hash table entry. */
|
for (el = rd->incoming_edges; el; el = next)
|
for (el = rd->incoming_edges; el; el = next)
|
{
|
{
|
edge e = el->e;
|
edge e = el->e;
|
|
|
/* Go ahead and free this element from the list. Doing this now
|
/* Go ahead and free this element from the list. Doing this now
|
avoids the need for another list walk when we destroy the hash
|
avoids the need for another list walk when we destroy the hash
|
table. */
|
table. */
|
next = el->next;
|
next = el->next;
|
free (el);
|
free (el);
|
|
|
/* Go ahead and clear E->aux. It's not needed anymore and failure
|
/* Go ahead and clear E->aux. It's not needed anymore and failure
|
to clear it will cause all kinds of unpleasant problems later. */
|
to clear it will cause all kinds of unpleasant problems later. */
|
e->aux = NULL;
|
e->aux = NULL;
|
|
|
thread_stats.num_threaded_edges++;
|
thread_stats.num_threaded_edges++;
|
|
|
if (rd->dup_block)
|
if (rd->dup_block)
|
{
|
{
|
edge e2;
|
edge e2;
|
|
|
if (dump_file && (dump_flags & TDF_DETAILS))
|
if (dump_file && (dump_flags & TDF_DETAILS))
|
fprintf (dump_file, " Threaded jump %d --> %d to %d\n",
|
fprintf (dump_file, " Threaded jump %d --> %d to %d\n",
|
e->src->index, e->dest->index, rd->dup_block->index);
|
e->src->index, e->dest->index, rd->dup_block->index);
|
|
|
rd->dup_block->count += e->count;
|
rd->dup_block->count += e->count;
|
rd->dup_block->frequency += EDGE_FREQUENCY (e);
|
rd->dup_block->frequency += EDGE_FREQUENCY (e);
|
EDGE_SUCC (rd->dup_block, 0)->count += e->count;
|
EDGE_SUCC (rd->dup_block, 0)->count += e->count;
|
/* Redirect the incoming edge to the appropriate duplicate
|
/* Redirect the incoming edge to the appropriate duplicate
|
block. */
|
block. */
|
e2 = redirect_edge_and_branch (e, rd->dup_block);
|
e2 = redirect_edge_and_branch (e, rd->dup_block);
|
gcc_assert (e == e2);
|
gcc_assert (e == e2);
|
flush_pending_stmts (e2);
|
flush_pending_stmts (e2);
|
}
|
}
|
else
|
else
|
{
|
{
|
if (dump_file && (dump_flags & TDF_DETAILS))
|
if (dump_file && (dump_flags & TDF_DETAILS))
|
fprintf (dump_file, " Threaded jump %d --> %d to %d\n",
|
fprintf (dump_file, " Threaded jump %d --> %d to %d\n",
|
e->src->index, e->dest->index, local_info->bb->index);
|
e->src->index, e->dest->index, local_info->bb->index);
|
|
|
/* We are using BB as the duplicate. Remove the unnecessary
|
/* We are using BB as the duplicate. Remove the unnecessary
|
outgoing edges and statements from BB. */
|
outgoing edges and statements from BB. */
|
remove_ctrl_stmt_and_useless_edges (local_info->bb,
|
remove_ctrl_stmt_and_useless_edges (local_info->bb,
|
rd->outgoing_edge->dest);
|
rd->outgoing_edge->dest);
|
|
|
/* Fixup the flags on the single remaining edge. */
|
/* Fixup the flags on the single remaining edge. */
|
single_succ_edge (local_info->bb)->flags
|
single_succ_edge (local_info->bb)->flags
|
&= ~(EDGE_TRUE_VALUE | EDGE_FALSE_VALUE | EDGE_ABNORMAL);
|
&= ~(EDGE_TRUE_VALUE | EDGE_FALSE_VALUE | EDGE_ABNORMAL);
|
single_succ_edge (local_info->bb)->flags |= EDGE_FALLTHRU;
|
single_succ_edge (local_info->bb)->flags |= EDGE_FALLTHRU;
|
|
|
/* And adjust count and frequency on BB. */
|
/* And adjust count and frequency on BB. */
|
local_info->bb->count = e->count;
|
local_info->bb->count = e->count;
|
local_info->bb->frequency = EDGE_FREQUENCY (e);
|
local_info->bb->frequency = EDGE_FREQUENCY (e);
|
}
|
}
|
}
|
}
|
|
|
/* Indicate that we actually threaded one or more jumps. */
|
/* Indicate that we actually threaded one or more jumps. */
|
if (rd->incoming_edges)
|
if (rd->incoming_edges)
|
local_info->jumps_threaded = true;
|
local_info->jumps_threaded = true;
|
|
|
return 1;
|
return 1;
|
}
|
}
|
|
|
/* Return true if this block has no executable statements other than
|
/* Return true if this block has no executable statements other than
|
a simple ctrl flow instruction. When the number of outgoing edges
|
a simple ctrl flow instruction. When the number of outgoing edges
|
is one, this is equivalent to a "forwarder" block. */
|
is one, this is equivalent to a "forwarder" block. */
|
|
|
static bool
|
static bool
|
redirection_block_p (basic_block bb)
|
redirection_block_p (basic_block bb)
|
{
|
{
|
gimple_stmt_iterator gsi;
|
gimple_stmt_iterator gsi;
|
|
|
/* Advance to the first executable statement. */
|
/* Advance to the first executable statement. */
|
gsi = gsi_start_bb (bb);
|
gsi = gsi_start_bb (bb);
|
while (!gsi_end_p (gsi)
|
while (!gsi_end_p (gsi)
|
&& (gimple_code (gsi_stmt (gsi)) == GIMPLE_LABEL
|
&& (gimple_code (gsi_stmt (gsi)) == GIMPLE_LABEL
|
|| is_gimple_debug (gsi_stmt (gsi))
|
|| is_gimple_debug (gsi_stmt (gsi))
|
|| gimple_nop_p (gsi_stmt (gsi))))
|
|| gimple_nop_p (gsi_stmt (gsi))))
|
gsi_next (&gsi);
|
gsi_next (&gsi);
|
|
|
/* Check if this is an empty block. */
|
/* Check if this is an empty block. */
|
if (gsi_end_p (gsi))
|
if (gsi_end_p (gsi))
|
return true;
|
return true;
|
|
|
/* Test that we've reached the terminating control statement. */
|
/* Test that we've reached the terminating control statement. */
|
return gsi_stmt (gsi)
|
return gsi_stmt (gsi)
|
&& (gimple_code (gsi_stmt (gsi)) == GIMPLE_COND
|
&& (gimple_code (gsi_stmt (gsi)) == GIMPLE_COND
|
|| gimple_code (gsi_stmt (gsi)) == GIMPLE_GOTO
|
|| gimple_code (gsi_stmt (gsi)) == GIMPLE_GOTO
|
|| gimple_code (gsi_stmt (gsi)) == GIMPLE_SWITCH);
|
|| gimple_code (gsi_stmt (gsi)) == GIMPLE_SWITCH);
|
}
|
}
|
|
|
/* BB is a block which ends with a COND_EXPR or SWITCH_EXPR and when BB
|
/* BB is a block which ends with a COND_EXPR or SWITCH_EXPR and when BB
|
is reached via one or more specific incoming edges, we know which
|
is reached via one or more specific incoming edges, we know which
|
outgoing edge from BB will be traversed.
|
outgoing edge from BB will be traversed.
|
|
|
We want to redirect those incoming edges to the target of the
|
We want to redirect those incoming edges to the target of the
|
appropriate outgoing edge. Doing so avoids a conditional branch
|
appropriate outgoing edge. Doing so avoids a conditional branch
|
and may expose new optimization opportunities. Note that we have
|
and may expose new optimization opportunities. Note that we have
|
to update dominator tree and SSA graph after such changes.
|
to update dominator tree and SSA graph after such changes.
|
|
|
The key to keeping the SSA graph update manageable is to duplicate
|
The key to keeping the SSA graph update manageable is to duplicate
|
the side effects occurring in BB so that those side effects still
|
the side effects occurring in BB so that those side effects still
|
occur on the paths which bypass BB after redirecting edges.
|
occur on the paths which bypass BB after redirecting edges.
|
|
|
We accomplish this by creating duplicates of BB and arranging for
|
We accomplish this by creating duplicates of BB and arranging for
|
the duplicates to unconditionally pass control to one specific
|
the duplicates to unconditionally pass control to one specific
|
successor of BB. We then revector the incoming edges into BB to
|
successor of BB. We then revector the incoming edges into BB to
|
the appropriate duplicate of BB.
|
the appropriate duplicate of BB.
|
|
|
If NOLOOP_ONLY is true, we only perform the threading as long as it
|
If NOLOOP_ONLY is true, we only perform the threading as long as it
|
does not affect the structure of the loops in a nontrivial way. */
|
does not affect the structure of the loops in a nontrivial way. */
|
|
|
static bool
|
static bool
|
thread_block (basic_block bb, bool noloop_only)
|
thread_block (basic_block bb, bool noloop_only)
|
{
|
{
|
/* E is an incoming edge into BB that we may or may not want to
|
/* E is an incoming edge into BB that we may or may not want to
|
redirect to a duplicate of BB. */
|
redirect to a duplicate of BB. */
|
edge e, e2;
|
edge e, e2;
|
edge_iterator ei;
|
edge_iterator ei;
|
struct local_info local_info;
|
struct local_info local_info;
|
struct loop *loop = bb->loop_father;
|
struct loop *loop = bb->loop_father;
|
|
|
/* ALL indicates whether or not all incoming edges into BB should
|
/* ALL indicates whether or not all incoming edges into BB should
|
be threaded to a duplicate of BB. */
|
be threaded to a duplicate of BB. */
|
bool all = true;
|
bool all = true;
|
|
|
/* To avoid scanning a linear array for the element we need we instead
|
/* To avoid scanning a linear array for the element we need we instead
|
use a hash table. For normal code there should be no noticeable
|
use a hash table. For normal code there should be no noticeable
|
difference. However, if we have a block with a large number of
|
difference. However, if we have a block with a large number of
|
incoming and outgoing edges such linear searches can get expensive. */
|
incoming and outgoing edges such linear searches can get expensive. */
|
redirection_data = htab_create (EDGE_COUNT (bb->succs),
|
redirection_data = htab_create (EDGE_COUNT (bb->succs),
|
redirection_data_hash,
|
redirection_data_hash,
|
redirection_data_eq,
|
redirection_data_eq,
|
free);
|
free);
|
|
|
/* If we thread the latch of the loop to its exit, the loop ceases to
|
/* If we thread the latch of the loop to its exit, the loop ceases to
|
exist. Make sure we do not restrict ourselves in order to preserve
|
exist. Make sure we do not restrict ourselves in order to preserve
|
this loop. */
|
this loop. */
|
if (loop->header == bb)
|
if (loop->header == bb)
|
{
|
{
|
e = loop_latch_edge (loop);
|
e = loop_latch_edge (loop);
|
e2 = (edge) e->aux;
|
e2 = (edge) e->aux;
|
|
|
if (e2 && loop_exit_edge_p (loop, e2))
|
if (e2 && loop_exit_edge_p (loop, e2))
|
{
|
{
|
loop->header = NULL;
|
loop->header = NULL;
|
loop->latch = NULL;
|
loop->latch = NULL;
|
}
|
}
|
}
|
}
|
|
|
/* Record each unique threaded destination into a hash table for
|
/* Record each unique threaded destination into a hash table for
|
efficient lookups. */
|
efficient lookups. */
|
FOR_EACH_EDGE (e, ei, bb->preds)
|
FOR_EACH_EDGE (e, ei, bb->preds)
|
{
|
{
|
e2 = (edge) e->aux;
|
e2 = (edge) e->aux;
|
|
|
if (!e2
|
if (!e2
|
/* If NOLOOP_ONLY is true, we only allow threading through the
|
/* If NOLOOP_ONLY is true, we only allow threading through the
|
header of a loop to exit edges. */
|
header of a loop to exit edges. */
|
|| (noloop_only
|
|| (noloop_only
|
&& bb == bb->loop_father->header
|
&& bb == bb->loop_father->header
|
&& !loop_exit_edge_p (bb->loop_father, e2)))
|
&& !loop_exit_edge_p (bb->loop_father, e2)))
|
{
|
{
|
all = false;
|
all = false;
|
continue;
|
continue;
|
}
|
}
|
|
|
update_bb_profile_for_threading (e->dest, EDGE_FREQUENCY (e),
|
update_bb_profile_for_threading (e->dest, EDGE_FREQUENCY (e),
|
e->count, (edge) e->aux);
|
e->count, (edge) e->aux);
|
|
|
/* Insert the outgoing edge into the hash table if it is not
|
/* Insert the outgoing edge into the hash table if it is not
|
already in the hash table. */
|
already in the hash table. */
|
lookup_redirection_data (e2, e, INSERT);
|
lookup_redirection_data (e2, e, INSERT);
|
}
|
}
|
|
|
/* If we are going to thread all incoming edges to an outgoing edge, then
|
/* If we are going to thread all incoming edges to an outgoing edge, then
|
BB will become unreachable. Rather than just throwing it away, use
|
BB will become unreachable. Rather than just throwing it away, use
|
it for one of the duplicates. Mark the first incoming edge with the
|
it for one of the duplicates. Mark the first incoming edge with the
|
DO_NOT_DUPLICATE attribute. */
|
DO_NOT_DUPLICATE attribute. */
|
if (all)
|
if (all)
|
{
|
{
|
edge e = (edge) EDGE_PRED (bb, 0)->aux;
|
edge e = (edge) EDGE_PRED (bb, 0)->aux;
|
lookup_redirection_data (e, NULL, NO_INSERT)->do_not_duplicate = true;
|
lookup_redirection_data (e, NULL, NO_INSERT)->do_not_duplicate = true;
|
}
|
}
|
|
|
/* We do not update dominance info. */
|
/* We do not update dominance info. */
|
free_dominance_info (CDI_DOMINATORS);
|
free_dominance_info (CDI_DOMINATORS);
|
|
|
/* Now create duplicates of BB.
|
/* Now create duplicates of BB.
|
|
|
Note that for a block with a high outgoing degree we can waste
|
Note that for a block with a high outgoing degree we can waste
|
a lot of time and memory creating and destroying useless edges.
|
a lot of time and memory creating and destroying useless edges.
|
|
|
So we first duplicate BB and remove the control structure at the
|
So we first duplicate BB and remove the control structure at the
|
tail of the duplicate as well as all outgoing edges from the
|
tail of the duplicate as well as all outgoing edges from the
|
duplicate. We then use that duplicate block as a template for
|
duplicate. We then use that duplicate block as a template for
|
the rest of the duplicates. */
|
the rest of the duplicates. */
|
local_info.template_block = NULL;
|
local_info.template_block = NULL;
|
local_info.bb = bb;
|
local_info.bb = bb;
|
local_info.jumps_threaded = false;
|
local_info.jumps_threaded = false;
|
htab_traverse (redirection_data, create_duplicates, &local_info);
|
htab_traverse (redirection_data, create_duplicates, &local_info);
|
|
|
/* The template does not have an outgoing edge. Create that outgoing
|
/* The template does not have an outgoing edge. Create that outgoing
|
edge and update PHI nodes as the edge's target as necessary.
|
edge and update PHI nodes as the edge's target as necessary.
|
|
|
We do this after creating all the duplicates to avoid creating
|
We do this after creating all the duplicates to avoid creating
|
unnecessary edges. */
|
unnecessary edges. */
|
htab_traverse (redirection_data, fixup_template_block, &local_info);
|
htab_traverse (redirection_data, fixup_template_block, &local_info);
|
|
|
/* The hash table traversals above created the duplicate blocks (and the
|
/* The hash table traversals above created the duplicate blocks (and the
|
statements within the duplicate blocks). This loop creates PHI nodes for
|
statements within the duplicate blocks). This loop creates PHI nodes for
|
the duplicated blocks and redirects the incoming edges into BB to reach
|
the duplicated blocks and redirects the incoming edges into BB to reach
|
the duplicates of BB. */
|
the duplicates of BB. */
|
htab_traverse (redirection_data, redirect_edges, &local_info);
|
htab_traverse (redirection_data, redirect_edges, &local_info);
|
|
|
/* Done with this block. Clear REDIRECTION_DATA. */
|
/* Done with this block. Clear REDIRECTION_DATA. */
|
htab_delete (redirection_data);
|
htab_delete (redirection_data);
|
redirection_data = NULL;
|
redirection_data = NULL;
|
|
|
/* Indicate to our caller whether or not any jumps were threaded. */
|
/* Indicate to our caller whether or not any jumps were threaded. */
|
return local_info.jumps_threaded;
|
return local_info.jumps_threaded;
|
}
|
}
|
|
|
/* Threads edge E through E->dest to the edge E->aux. Returns the copy
|
/* Threads edge E through E->dest to the edge E->aux. Returns the copy
|
of E->dest created during threading, or E->dest if it was not necessary
|
of E->dest created during threading, or E->dest if it was not necessary
|
to copy it (E is its single predecessor). */
|
to copy it (E is its single predecessor). */
|
|
|
static basic_block
|
static basic_block
|
thread_single_edge (edge e)
|
thread_single_edge (edge e)
|
{
|
{
|
basic_block bb = e->dest;
|
basic_block bb = e->dest;
|
edge eto = (edge) e->aux;
|
edge eto = (edge) e->aux;
|
struct redirection_data rd;
|
struct redirection_data rd;
|
|
|
e->aux = NULL;
|
e->aux = NULL;
|
|
|
thread_stats.num_threaded_edges++;
|
thread_stats.num_threaded_edges++;
|
|
|
if (single_pred_p (bb))
|
if (single_pred_p (bb))
|
{
|
{
|
/* If BB has just a single predecessor, we should only remove the
|
/* If BB has just a single predecessor, we should only remove the
|
control statements at its end, and successors except for ETO. */
|
control statements at its end, and successors except for ETO. */
|
remove_ctrl_stmt_and_useless_edges (bb, eto->dest);
|
remove_ctrl_stmt_and_useless_edges (bb, eto->dest);
|
|
|
/* And fixup the flags on the single remaining edge. */
|
/* And fixup the flags on the single remaining edge. */
|
eto->flags &= ~(EDGE_TRUE_VALUE | EDGE_FALSE_VALUE | EDGE_ABNORMAL);
|
eto->flags &= ~(EDGE_TRUE_VALUE | EDGE_FALSE_VALUE | EDGE_ABNORMAL);
|
eto->flags |= EDGE_FALLTHRU;
|
eto->flags |= EDGE_FALLTHRU;
|
|
|
return bb;
|
return bb;
|
}
|
}
|
|
|
/* Otherwise, we need to create a copy. */
|
/* Otherwise, we need to create a copy. */
|
update_bb_profile_for_threading (bb, EDGE_FREQUENCY (e), e->count, eto);
|
update_bb_profile_for_threading (bb, EDGE_FREQUENCY (e), e->count, eto);
|
|
|
rd.outgoing_edge = eto;
|
rd.outgoing_edge = eto;
|
|
|
create_block_for_threading (bb, &rd);
|
create_block_for_threading (bb, &rd);
|
create_edge_and_update_destination_phis (&rd);
|
create_edge_and_update_destination_phis (&rd);
|
|
|
if (dump_file && (dump_flags & TDF_DETAILS))
|
if (dump_file && (dump_flags & TDF_DETAILS))
|
fprintf (dump_file, " Threaded jump %d --> %d to %d\n",
|
fprintf (dump_file, " Threaded jump %d --> %d to %d\n",
|
e->src->index, e->dest->index, rd.dup_block->index);
|
e->src->index, e->dest->index, rd.dup_block->index);
|
|
|
rd.dup_block->count = e->count;
|
rd.dup_block->count = e->count;
|
rd.dup_block->frequency = EDGE_FREQUENCY (e);
|
rd.dup_block->frequency = EDGE_FREQUENCY (e);
|
single_succ_edge (rd.dup_block)->count = e->count;
|
single_succ_edge (rd.dup_block)->count = e->count;
|
redirect_edge_and_branch (e, rd.dup_block);
|
redirect_edge_and_branch (e, rd.dup_block);
|
flush_pending_stmts (e);
|
flush_pending_stmts (e);
|
|
|
return rd.dup_block;
|
return rd.dup_block;
|
}
|
}
|
|
|
/* Callback for dfs_enumerate_from. Returns true if BB is different
|
/* Callback for dfs_enumerate_from. Returns true if BB is different
|
from STOP and DBDS_CE_STOP. */
|
from STOP and DBDS_CE_STOP. */
|
|
|
static basic_block dbds_ce_stop;
|
static basic_block dbds_ce_stop;
|
static bool
|
static bool
|
dbds_continue_enumeration_p (const_basic_block bb, const void *stop)
|
dbds_continue_enumeration_p (const_basic_block bb, const void *stop)
|
{
|
{
|
return (bb != (const_basic_block) stop
|
return (bb != (const_basic_block) stop
|
&& bb != dbds_ce_stop);
|
&& bb != dbds_ce_stop);
|
}
|
}
|
|
|
/* Evaluates the dominance relationship of latch of the LOOP and BB, and
|
/* Evaluates the dominance relationship of latch of the LOOP and BB, and
|
returns the state. */
|
returns the state. */
|
|
|
enum bb_dom_status
|
enum bb_dom_status
|
{
|
{
|
/* BB does not dominate latch of the LOOP. */
|
/* BB does not dominate latch of the LOOP. */
|
DOMST_NONDOMINATING,
|
DOMST_NONDOMINATING,
|
/* The LOOP is broken (there is no path from the header to its latch. */
|
/* The LOOP is broken (there is no path from the header to its latch. */
|
DOMST_LOOP_BROKEN,
|
DOMST_LOOP_BROKEN,
|
/* BB dominates the latch of the LOOP. */
|
/* BB dominates the latch of the LOOP. */
|
DOMST_DOMINATING
|
DOMST_DOMINATING
|
};
|
};
|
|
|
static enum bb_dom_status
|
static enum bb_dom_status
|
determine_bb_domination_status (struct loop *loop, basic_block bb)
|
determine_bb_domination_status (struct loop *loop, basic_block bb)
|
{
|
{
|
basic_block *bblocks;
|
basic_block *bblocks;
|
unsigned nblocks, i;
|
unsigned nblocks, i;
|
bool bb_reachable = false;
|
bool bb_reachable = false;
|
edge_iterator ei;
|
edge_iterator ei;
|
edge e;
|
edge e;
|
|
|
#ifdef ENABLE_CHECKING
|
#ifdef ENABLE_CHECKING
|
/* This function assumes BB is a successor of LOOP->header. */
|
/* This function assumes BB is a successor of LOOP->header. */
|
{
|
{
|
bool ok = false;
|
bool ok = false;
|
|
|
FOR_EACH_EDGE (e, ei, bb->preds)
|
FOR_EACH_EDGE (e, ei, bb->preds)
|
{
|
{
|
if (e->src == loop->header)
|
if (e->src == loop->header)
|
{
|
{
|
ok = true;
|
ok = true;
|
break;
|
break;
|
}
|
}
|
}
|
}
|
|
|
gcc_assert (ok);
|
gcc_assert (ok);
|
}
|
}
|
#endif
|
#endif
|
|
|
if (bb == loop->latch)
|
if (bb == loop->latch)
|
return DOMST_DOMINATING;
|
return DOMST_DOMINATING;
|
|
|
/* Check that BB dominates LOOP->latch, and that it is back-reachable
|
/* Check that BB dominates LOOP->latch, and that it is back-reachable
|
from it. */
|
from it. */
|
|
|
bblocks = XCNEWVEC (basic_block, loop->num_nodes);
|
bblocks = XCNEWVEC (basic_block, loop->num_nodes);
|
dbds_ce_stop = loop->header;
|
dbds_ce_stop = loop->header;
|
nblocks = dfs_enumerate_from (loop->latch, 1, dbds_continue_enumeration_p,
|
nblocks = dfs_enumerate_from (loop->latch, 1, dbds_continue_enumeration_p,
|
bblocks, loop->num_nodes, bb);
|
bblocks, loop->num_nodes, bb);
|
for (i = 0; i < nblocks; i++)
|
for (i = 0; i < nblocks; i++)
|
FOR_EACH_EDGE (e, ei, bblocks[i]->preds)
|
FOR_EACH_EDGE (e, ei, bblocks[i]->preds)
|
{
|
{
|
if (e->src == loop->header)
|
if (e->src == loop->header)
|
{
|
{
|
free (bblocks);
|
free (bblocks);
|
return DOMST_NONDOMINATING;
|
return DOMST_NONDOMINATING;
|
}
|
}
|
if (e->src == bb)
|
if (e->src == bb)
|
bb_reachable = true;
|
bb_reachable = true;
|
}
|
}
|
|
|
free (bblocks);
|
free (bblocks);
|
return (bb_reachable ? DOMST_DOMINATING : DOMST_LOOP_BROKEN);
|
return (bb_reachable ? DOMST_DOMINATING : DOMST_LOOP_BROKEN);
|
}
|
}
|
|
|
/* Thread jumps through the header of LOOP. Returns true if cfg changes.
|
/* Thread jumps through the header of LOOP. Returns true if cfg changes.
|
If MAY_PEEL_LOOP_HEADERS is false, we avoid threading from entry edges
|
If MAY_PEEL_LOOP_HEADERS is false, we avoid threading from entry edges
|
to the inside of the loop. */
|
to the inside of the loop. */
|
|
|
static bool
|
static bool
|
thread_through_loop_header (struct loop *loop, bool may_peel_loop_headers)
|
thread_through_loop_header (struct loop *loop, bool may_peel_loop_headers)
|
{
|
{
|
basic_block header = loop->header;
|
basic_block header = loop->header;
|
edge e, tgt_edge, latch = loop_latch_edge (loop);
|
edge e, tgt_edge, latch = loop_latch_edge (loop);
|
edge_iterator ei;
|
edge_iterator ei;
|
basic_block tgt_bb, atgt_bb;
|
basic_block tgt_bb, atgt_bb;
|
enum bb_dom_status domst;
|
enum bb_dom_status domst;
|
|
|
/* We have already threaded through headers to exits, so all the threading
|
/* We have already threaded through headers to exits, so all the threading
|
requests now are to the inside of the loop. We need to avoid creating
|
requests now are to the inside of the loop. We need to avoid creating
|
irreducible regions (i.e., loops with more than one entry block), and
|
irreducible regions (i.e., loops with more than one entry block), and
|
also loop with several latch edges, or new subloops of the loop (although
|
also loop with several latch edges, or new subloops of the loop (although
|
there are cases where it might be appropriate, it is difficult to decide,
|
there are cases where it might be appropriate, it is difficult to decide,
|
and doing it wrongly may confuse other optimizers).
|
and doing it wrongly may confuse other optimizers).
|
|
|
We could handle more general cases here. However, the intention is to
|
We could handle more general cases here. However, the intention is to
|
preserve some information about the loop, which is impossible if its
|
preserve some information about the loop, which is impossible if its
|
structure changes significantly, in a way that is not well understood.
|
structure changes significantly, in a way that is not well understood.
|
Thus we only handle few important special cases, in which also updating
|
Thus we only handle few important special cases, in which also updating
|
of the loop-carried information should be feasible:
|
of the loop-carried information should be feasible:
|
|
|
1) Propagation of latch edge to a block that dominates the latch block
|
1) Propagation of latch edge to a block that dominates the latch block
|
of a loop. This aims to handle the following idiom:
|
of a loop. This aims to handle the following idiom:
|
|
|
first = 1;
|
first = 1;
|
while (1)
|
while (1)
|
{
|
{
|
if (first)
|
if (first)
|
initialize;
|
initialize;
|
first = 0;
|
first = 0;
|
body;
|
body;
|
}
|
}
|
|
|
After threading the latch edge, this becomes
|
After threading the latch edge, this becomes
|
|
|
first = 1;
|
first = 1;
|
if (first)
|
if (first)
|
initialize;
|
initialize;
|
while (1)
|
while (1)
|
{
|
{
|
first = 0;
|
first = 0;
|
body;
|
body;
|
}
|
}
|
|
|
The original header of the loop is moved out of it, and we may thread
|
The original header of the loop is moved out of it, and we may thread
|
the remaining edges through it without further constraints.
|
the remaining edges through it without further constraints.
|
|
|
2) All entry edges are propagated to a single basic block that dominates
|
2) All entry edges are propagated to a single basic block that dominates
|
the latch block of the loop. This aims to handle the following idiom
|
the latch block of the loop. This aims to handle the following idiom
|
(normally created for "for" loops):
|
(normally created for "for" loops):
|
|
|
i = 0;
|
i = 0;
|
while (1)
|
while (1)
|
{
|
{
|
if (i >= 100)
|
if (i >= 100)
|
break;
|
break;
|
body;
|
body;
|
i++;
|
i++;
|
}
|
}
|
|
|
This becomes
|
This becomes
|
|
|
i = 0;
|
i = 0;
|
while (1)
|
while (1)
|
{
|
{
|
body;
|
body;
|
i++;
|
i++;
|
if (i >= 100)
|
if (i >= 100)
|
break;
|
break;
|
}
|
}
|
*/
|
*/
|
|
|
/* Threading through the header won't improve the code if the header has just
|
/* Threading through the header won't improve the code if the header has just
|
one successor. */
|
one successor. */
|
if (single_succ_p (header))
|
if (single_succ_p (header))
|
goto fail;
|
goto fail;
|
|
|
if (latch->aux)
|
if (latch->aux)
|
{
|
{
|
tgt_edge = (edge) latch->aux;
|
tgt_edge = (edge) latch->aux;
|
tgt_bb = tgt_edge->dest;
|
tgt_bb = tgt_edge->dest;
|
}
|
}
|
else if (!may_peel_loop_headers
|
else if (!may_peel_loop_headers
|
&& !redirection_block_p (loop->header))
|
&& !redirection_block_p (loop->header))
|
goto fail;
|
goto fail;
|
else
|
else
|
{
|
{
|
tgt_bb = NULL;
|
tgt_bb = NULL;
|
tgt_edge = NULL;
|
tgt_edge = NULL;
|
FOR_EACH_EDGE (e, ei, header->preds)
|
FOR_EACH_EDGE (e, ei, header->preds)
|
{
|
{
|
if (!e->aux)
|
if (!e->aux)
|
{
|
{
|
if (e == latch)
|
if (e == latch)
|
continue;
|
continue;
|
|
|
/* If latch is not threaded, and there is a header
|
/* If latch is not threaded, and there is a header
|
edge that is not threaded, we would create loop
|
edge that is not threaded, we would create loop
|
with multiple entries. */
|
with multiple entries. */
|
goto fail;
|
goto fail;
|
}
|
}
|
|
|
tgt_edge = (edge) e->aux;
|
tgt_edge = (edge) e->aux;
|
atgt_bb = tgt_edge->dest;
|
atgt_bb = tgt_edge->dest;
|
if (!tgt_bb)
|
if (!tgt_bb)
|
tgt_bb = atgt_bb;
|
tgt_bb = atgt_bb;
|
/* Two targets of threading would make us create loop
|
/* Two targets of threading would make us create loop
|
with multiple entries. */
|
with multiple entries. */
|
else if (tgt_bb != atgt_bb)
|
else if (tgt_bb != atgt_bb)
|
goto fail;
|
goto fail;
|
}
|
}
|
|
|
if (!tgt_bb)
|
if (!tgt_bb)
|
{
|
{
|
/* There are no threading requests. */
|
/* There are no threading requests. */
|
return false;
|
return false;
|
}
|
}
|
|
|
/* Redirecting to empty loop latch is useless. */
|
/* Redirecting to empty loop latch is useless. */
|
if (tgt_bb == loop->latch
|
if (tgt_bb == loop->latch
|
&& empty_block_p (loop->latch))
|
&& empty_block_p (loop->latch))
|
goto fail;
|
goto fail;
|
}
|
}
|
|
|
/* The target block must dominate the loop latch, otherwise we would be
|
/* The target block must dominate the loop latch, otherwise we would be
|
creating a subloop. */
|
creating a subloop. */
|
domst = determine_bb_domination_status (loop, tgt_bb);
|
domst = determine_bb_domination_status (loop, tgt_bb);
|
if (domst == DOMST_NONDOMINATING)
|
if (domst == DOMST_NONDOMINATING)
|
goto fail;
|
goto fail;
|
if (domst == DOMST_LOOP_BROKEN)
|
if (domst == DOMST_LOOP_BROKEN)
|
{
|
{
|
/* If the loop ceased to exist, mark it as such, and thread through its
|
/* If the loop ceased to exist, mark it as such, and thread through its
|
original header. */
|
original header. */
|
loop->header = NULL;
|
loop->header = NULL;
|
loop->latch = NULL;
|
loop->latch = NULL;
|
return thread_block (header, false);
|
return thread_block (header, false);
|
}
|
}
|
|
|
if (tgt_bb->loop_father->header == tgt_bb)
|
if (tgt_bb->loop_father->header == tgt_bb)
|
{
|
{
|
/* If the target of the threading is a header of a subloop, we need
|
/* If the target of the threading is a header of a subloop, we need
|
to create a preheader for it, so that the headers of the two loops
|
to create a preheader for it, so that the headers of the two loops
|
do not merge. */
|
do not merge. */
|
if (EDGE_COUNT (tgt_bb->preds) > 2)
|
if (EDGE_COUNT (tgt_bb->preds) > 2)
|
{
|
{
|
tgt_bb = create_preheader (tgt_bb->loop_father, 0);
|
tgt_bb = create_preheader (tgt_bb->loop_father, 0);
|
gcc_assert (tgt_bb != NULL);
|
gcc_assert (tgt_bb != NULL);
|
}
|
}
|
else
|
else
|
tgt_bb = split_edge (tgt_edge);
|
tgt_bb = split_edge (tgt_edge);
|
}
|
}
|
|
|
if (latch->aux)
|
if (latch->aux)
|
{
|
{
|
/* First handle the case latch edge is redirected. */
|
/* First handle the case latch edge is redirected. */
|
loop->latch = thread_single_edge (latch);
|
loop->latch = thread_single_edge (latch);
|
gcc_assert (single_succ (loop->latch) == tgt_bb);
|
gcc_assert (single_succ (loop->latch) == tgt_bb);
|
loop->header = tgt_bb;
|
loop->header = tgt_bb;
|
|
|
/* Thread the remaining edges through the former header. */
|
/* Thread the remaining edges through the former header. */
|
thread_block (header, false);
|
thread_block (header, false);
|
}
|
}
|
else
|
else
|
{
|
{
|
basic_block new_preheader;
|
basic_block new_preheader;
|
|
|
/* Now consider the case entry edges are redirected to the new entry
|
/* Now consider the case entry edges are redirected to the new entry
|
block. Remember one entry edge, so that we can find the new
|
block. Remember one entry edge, so that we can find the new
|
preheader (its destination after threading). */
|
preheader (its destination after threading). */
|
FOR_EACH_EDGE (e, ei, header->preds)
|
FOR_EACH_EDGE (e, ei, header->preds)
|
{
|
{
|
if (e->aux)
|
if (e->aux)
|
break;
|
break;
|
}
|
}
|
|
|
/* The duplicate of the header is the new preheader of the loop. Ensure
|
/* The duplicate of the header is the new preheader of the loop. Ensure
|
that it is placed correctly in the loop hierarchy. */
|
that it is placed correctly in the loop hierarchy. */
|
set_loop_copy (loop, loop_outer (loop));
|
set_loop_copy (loop, loop_outer (loop));
|
|
|
thread_block (header, false);
|
thread_block (header, false);
|
set_loop_copy (loop, NULL);
|
set_loop_copy (loop, NULL);
|
new_preheader = e->dest;
|
new_preheader = e->dest;
|
|
|
/* Create the new latch block. This is always necessary, as the latch
|
/* Create the new latch block. This is always necessary, as the latch
|
must have only a single successor, but the original header had at
|
must have only a single successor, but the original header had at
|
least two successors. */
|
least two successors. */
|
loop->latch = NULL;
|
loop->latch = NULL;
|
mfb_kj_edge = single_succ_edge (new_preheader);
|
mfb_kj_edge = single_succ_edge (new_preheader);
|
loop->header = mfb_kj_edge->dest;
|
loop->header = mfb_kj_edge->dest;
|
latch = make_forwarder_block (tgt_bb, mfb_keep_just, NULL);
|
latch = make_forwarder_block (tgt_bb, mfb_keep_just, NULL);
|
loop->header = latch->dest;
|
loop->header = latch->dest;
|
loop->latch = latch->src;
|
loop->latch = latch->src;
|
}
|
}
|
|
|
return true;
|
return true;
|
|
|
fail:
|
fail:
|
/* We failed to thread anything. Cancel the requests. */
|
/* We failed to thread anything. Cancel the requests. */
|
FOR_EACH_EDGE (e, ei, header->preds)
|
FOR_EACH_EDGE (e, ei, header->preds)
|
{
|
{
|
e->aux = NULL;
|
e->aux = NULL;
|
}
|
}
|
return false;
|
return false;
|
}
|
}
|
|
|
/* Walk through the registered jump threads and convert them into a
|
/* Walk through the registered jump threads and convert them into a
|
form convenient for this pass.
|
form convenient for this pass.
|
|
|
Any block which has incoming edges threaded to outgoing edges
|
Any block which has incoming edges threaded to outgoing edges
|
will have its entry in THREADED_BLOCK set.
|
will have its entry in THREADED_BLOCK set.
|
|
|
Any threaded edge will have its new outgoing edge stored in the
|
Any threaded edge will have its new outgoing edge stored in the
|
original edge's AUX field.
|
original edge's AUX field.
|
|
|
This form avoids the need to walk all the edges in the CFG to
|
This form avoids the need to walk all the edges in the CFG to
|
discover blocks which need processing and avoids unnecessary
|
discover blocks which need processing and avoids unnecessary
|
hash table lookups to map from threaded edge to new target. */
|
hash table lookups to map from threaded edge to new target. */
|
|
|
static void
|
static void
|
mark_threaded_blocks (bitmap threaded_blocks)
|
mark_threaded_blocks (bitmap threaded_blocks)
|
{
|
{
|
unsigned int i;
|
unsigned int i;
|
bitmap_iterator bi;
|
bitmap_iterator bi;
|
bitmap tmp = BITMAP_ALLOC (NULL);
|
bitmap tmp = BITMAP_ALLOC (NULL);
|
basic_block bb;
|
basic_block bb;
|
edge e;
|
edge e;
|
edge_iterator ei;
|
edge_iterator ei;
|
|
|
for (i = 0; i < VEC_length (edge, threaded_edges); i += 2)
|
for (i = 0; i < VEC_length (edge, threaded_edges); i += 2)
|
{
|
{
|
edge e = VEC_index (edge, threaded_edges, i);
|
edge e = VEC_index (edge, threaded_edges, i);
|
edge e2 = VEC_index (edge, threaded_edges, i + 1);
|
edge e2 = VEC_index (edge, threaded_edges, i + 1);
|
|
|
e->aux = e2;
|
e->aux = e2;
|
bitmap_set_bit (tmp, e->dest->index);
|
bitmap_set_bit (tmp, e->dest->index);
|
}
|
}
|
|
|
/* If optimizing for size, only thread through block if we don't have
|
/* If optimizing for size, only thread through block if we don't have
|
to duplicate it or it's an otherwise empty redirection block. */
|
to duplicate it or it's an otherwise empty redirection block. */
|
if (optimize_function_for_size_p (cfun))
|
if (optimize_function_for_size_p (cfun))
|
{
|
{
|
EXECUTE_IF_SET_IN_BITMAP (tmp, 0, i, bi)
|
EXECUTE_IF_SET_IN_BITMAP (tmp, 0, i, bi)
|
{
|
{
|
bb = BASIC_BLOCK (i);
|
bb = BASIC_BLOCK (i);
|
if (EDGE_COUNT (bb->preds) > 1
|
if (EDGE_COUNT (bb->preds) > 1
|
&& !redirection_block_p (bb))
|
&& !redirection_block_p (bb))
|
{
|
{
|
FOR_EACH_EDGE (e, ei, bb->preds)
|
FOR_EACH_EDGE (e, ei, bb->preds)
|
e->aux = NULL;
|
e->aux = NULL;
|
}
|
}
|
else
|
else
|
bitmap_set_bit (threaded_blocks, i);
|
bitmap_set_bit (threaded_blocks, i);
|
}
|
}
|
}
|
}
|
else
|
else
|
bitmap_copy (threaded_blocks, tmp);
|
bitmap_copy (threaded_blocks, tmp);
|
|
|
BITMAP_FREE(tmp);
|
BITMAP_FREE(tmp);
|
}
|
}
|
|
|
|
|
/* Walk through all blocks and thread incoming edges to the appropriate
|
/* Walk through all blocks and thread incoming edges to the appropriate
|
outgoing edge for each edge pair recorded in THREADED_EDGES.
|
outgoing edge for each edge pair recorded in THREADED_EDGES.
|
|
|
It is the caller's responsibility to fix the dominance information
|
It is the caller's responsibility to fix the dominance information
|
and rewrite duplicated SSA_NAMEs back into SSA form.
|
and rewrite duplicated SSA_NAMEs back into SSA form.
|
|
|
If MAY_PEEL_LOOP_HEADERS is false, we avoid threading edges through
|
If MAY_PEEL_LOOP_HEADERS is false, we avoid threading edges through
|
loop headers if it does not simplify the loop.
|
loop headers if it does not simplify the loop.
|
|
|
Returns true if one or more edges were threaded, false otherwise. */
|
Returns true if one or more edges were threaded, false otherwise. */
|
|
|
bool
|
bool
|
thread_through_all_blocks (bool may_peel_loop_headers)
|
thread_through_all_blocks (bool may_peel_loop_headers)
|
{
|
{
|
bool retval = false;
|
bool retval = false;
|
unsigned int i;
|
unsigned int i;
|
bitmap_iterator bi;
|
bitmap_iterator bi;
|
bitmap threaded_blocks;
|
bitmap threaded_blocks;
|
struct loop *loop;
|
struct loop *loop;
|
loop_iterator li;
|
loop_iterator li;
|
|
|
/* We must know about loops in order to preserve them. */
|
/* We must know about loops in order to preserve them. */
|
gcc_assert (current_loops != NULL);
|
gcc_assert (current_loops != NULL);
|
|
|
if (threaded_edges == NULL)
|
if (threaded_edges == NULL)
|
return false;
|
return false;
|
|
|
threaded_blocks = BITMAP_ALLOC (NULL);
|
threaded_blocks = BITMAP_ALLOC (NULL);
|
memset (&thread_stats, 0, sizeof (thread_stats));
|
memset (&thread_stats, 0, sizeof (thread_stats));
|
|
|
mark_threaded_blocks (threaded_blocks);
|
mark_threaded_blocks (threaded_blocks);
|
|
|
initialize_original_copy_tables ();
|
initialize_original_copy_tables ();
|
|
|
/* First perform the threading requests that do not affect
|
/* First perform the threading requests that do not affect
|
loop structure. */
|
loop structure. */
|
EXECUTE_IF_SET_IN_BITMAP (threaded_blocks, 0, i, bi)
|
EXECUTE_IF_SET_IN_BITMAP (threaded_blocks, 0, i, bi)
|
{
|
{
|
basic_block bb = BASIC_BLOCK (i);
|
basic_block bb = BASIC_BLOCK (i);
|
|
|
if (EDGE_COUNT (bb->preds) > 0)
|
if (EDGE_COUNT (bb->preds) > 0)
|
retval |= thread_block (bb, true);
|
retval |= thread_block (bb, true);
|
}
|
}
|
|
|
/* Then perform the threading through loop headers. We start with the
|
/* Then perform the threading through loop headers. We start with the
|
innermost loop, so that the changes in cfg we perform won't affect
|
innermost loop, so that the changes in cfg we perform won't affect
|
further threading. */
|
further threading. */
|
FOR_EACH_LOOP (li, loop, LI_FROM_INNERMOST)
|
FOR_EACH_LOOP (li, loop, LI_FROM_INNERMOST)
|
{
|
{
|
if (!loop->header
|
if (!loop->header
|
|| !bitmap_bit_p (threaded_blocks, loop->header->index))
|
|| !bitmap_bit_p (threaded_blocks, loop->header->index))
|
continue;
|
continue;
|
|
|
retval |= thread_through_loop_header (loop, may_peel_loop_headers);
|
retval |= thread_through_loop_header (loop, may_peel_loop_headers);
|
}
|
}
|
|
|
statistics_counter_event (cfun, "Jumps threaded",
|
statistics_counter_event (cfun, "Jumps threaded",
|
thread_stats.num_threaded_edges);
|
thread_stats.num_threaded_edges);
|
|
|
free_original_copy_tables ();
|
free_original_copy_tables ();
|
|
|
BITMAP_FREE (threaded_blocks);
|
BITMAP_FREE (threaded_blocks);
|
threaded_blocks = NULL;
|
threaded_blocks = NULL;
|
VEC_free (edge, heap, threaded_edges);
|
VEC_free (edge, heap, threaded_edges);
|
threaded_edges = NULL;
|
threaded_edges = NULL;
|
|
|
if (retval)
|
if (retval)
|
loops_state_set (LOOPS_NEED_FIXUP);
|
loops_state_set (LOOPS_NEED_FIXUP);
|
|
|
return retval;
|
return retval;
|
}
|
}
|
|
|
/* Register a jump threading opportunity. We queue up all the jump
|
/* Register a jump threading opportunity. We queue up all the jump
|
threading opportunities discovered by a pass and update the CFG
|
threading opportunities discovered by a pass and update the CFG
|
and SSA form all at once.
|
and SSA form all at once.
|
|
|
E is the edge we can thread, E2 is the new target edge, i.e., we
|
E is the edge we can thread, E2 is the new target edge, i.e., we
|
are effectively recording that E->dest can be changed to E2->dest
|
are effectively recording that E->dest can be changed to E2->dest
|
after fixing the SSA graph. */
|
after fixing the SSA graph. */
|
|
|
void
|
void
|
register_jump_thread (edge e, edge e2)
|
register_jump_thread (edge e, edge e2)
|
{
|
{
|
if (threaded_edges == NULL)
|
if (threaded_edges == NULL)
|
threaded_edges = VEC_alloc (edge, heap, 10);
|
threaded_edges = VEC_alloc (edge, heap, 10);
|
|
|
VEC_safe_push (edge, heap, threaded_edges, e);
|
VEC_safe_push (edge, heap, threaded_edges, e);
|
VEC_safe_push (edge, heap, threaded_edges, e2);
|
VEC_safe_push (edge, heap, threaded_edges, e2);
|
}
|
}
|
|
|