FreeChainXenon/gcc
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

re PR driver/46760 (LTO bootstrap doesn't work with FDO)

Browse source 
PR tree-optimization/46760
	* cgraph.c (cgraph_create_node): Initialize count_materialization_scale.
	* cgraph.h (struct cgraph_node): Add count_materialization_scale.
	* lto-cgraph.c (lto_output_edge): Fix assert.
	(lto_output_node): Output count_materialization_scale.
	(output_profile_summary): Output only runs and sum_max.
	(input_node): Input count_materialization_scale.
	(input_profile_summary): Read data into file specific gcov summary.
	(merge_profile_summaries): New function.
	(input_cgraph): Update call of input_profile_summary;
	call merge_profile_summaries.
	* lto-streamer-in.c (input_cfg): Add count_materialization_scale arg;
	rescale counts at read in.
	(intput_bb): Likewise.
	(input_function): Update call of input_bb.
	(lto_read_body): Update call of input_cfg.
	* lto-streamer.h: Inlclude gcov-io.h
	(lto_file_decl_data): Add gcov_ctr_summary.

From-SVN: r167458
This commit is contained in:
Jan Hubicka 2010-12-04 20:35:23 +01:00 • committed by Jan Hubicka
parent f3007348c3
commit db0bf14f6b
6 changed files with 145 additions and 30 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

21
gcc/ChangeLog
View file

@ -1,3 +1,24 @@
2010-12-03 Jan Hubicka <jh@suse.cz>
PR tree-optimization/46760
* cgraph.c (cgraph_create_node): Initialize count_materialization_scale.
* cgraph.h (struct cgraph_node): Add count_materialization_scale.
* lto-cgraph.c (lto_output_edge): Fix assert.
(lto_output_node): Output count_materialization_scale.
(output_profile_summary): Output only runs and sum_max.
(input_node): Input count_materialization_scale.
(input_profile_summary): Read data into file specific gcov summary.
(merge_profile_summaries): New function.
(input_cgraph): Update call of input_profile_summary;
call merge_profile_summaries.
* lto-streamer-in.c (input_cfg): Add count_materialization_scale arg;
rescale counts at read in.
(intput_bb): Likewise.
(input_function): Update call of input_bb.
(lto_read_body): Update call of input_cfg.
* lto-streamer.h: Inlclude gcov-io.h
(lto_file_decl_data): Add gcov_ctr_summary.
2010-12-03 Dave Korn <dave.korn.cygwin@gmail.com>
* doc/tm.texi.in (Cond. Exec. Macros): Rename node from this ...

1
gcc/cgraph.c
View file

@ -478,6 +478,7 @@ cgraph_create_node (void)
node->previous = NULL;
node->global.estimated_growth = INT_MIN;
node->frequency = NODE_FREQUENCY_NORMAL;
node->count_materialization_scale = REG_BR_PROB_BASE;
ipa_empty_ref_list (&node->ref_list);
cgraph_nodes = node;
cgraph_n_nodes++;

3
gcc/cgraph.h
View file

@ -233,6 +233,9 @@ struct GTY((chain_next ("%h.next"), chain_prev ("%h.previous"))) cgraph_node {
/* Expected number of executions: calculated in profile.c. */
gcov_type count;
/* How to scale counts at materialization time; used to merge
LTO units with different number of profile runs. */
int count_materialization_scale;
/* Unique id of the node. */
int uid;
/* Ordering of all cgraph nodes. */

127
gcc/lto-cgraph.c
View file

@ -302,6 +302,7 @@ lto_output_edge (struct lto_simple_output_block *ob, struct cgraph_edge *edge,
gcc_assert (!(flags & (ECF_LOOPING_CONST_OR_PURE
| ECF_MAY_BE_ALLOCA
| ECF_SIBCALL
| ECF_LEAF
| ECF_NOVOPS)));
}
lto_output_bitpack (&bp);
@ -462,6 +463,7 @@ lto_output_node (struct lto_simple_output_block *ob, struct cgraph_node *node,
lto_output_fn_decl_index (ob->decl_state, ob->main_stream, node->decl);
lto_output_sleb128_stream (ob->main_stream, node->count);
lto_output_sleb128_stream (ob->main_stream, node->count_materialization_scale);
if (tag == LTO_cgraph_analyzed_node)
{
@ -661,12 +663,12 @@ output_profile_summary (struct lto_simple_output_block *ob)
{
if (profile_info)
{
/* We do not output num, it is not terribly useful. */
/* We do not output num, sum_all and run_max, they are not used by
GCC profile feedback and they are difficult to merge from multiple
units. */
gcc_assert (profile_info->runs);
lto_output_uleb128_stream (ob->main_stream, profile_info->runs);
lto_output_sleb128_stream (ob->main_stream, profile_info->sum_all);
lto_output_sleb128_stream (ob->main_stream, profile_info->run_max);
lto_output_sleb128_stream (ob->main_stream, profile_info->sum_max);
lto_output_uleb128_stream (ob->main_stream, profile_info->sum_max);
}
else
lto_output_uleb128_stream (ob->main_stream, 0);
@ -1045,6 +1047,7 @@ input_node (struct lto_file_decl_data *file_data,
node = cgraph_node (fn_decl);
node->count = lto_input_sleb128 (ib);
node->count_materialization_scale = lto_input_sleb128 (ib);
if (tag == LTO_cgraph_analyzed_node)
{
@ -1424,32 +1427,108 @@ static struct gcov_ctr_summary lto_gcov_summary;
/* Input profile_info from IB. */
static void
input_profile_summary (struct lto_input_block *ib)
input_profile_summary (struct lto_input_block *ib,
struct lto_file_decl_data *file_data)
{
unsigned int runs = lto_input_uleb128 (ib);
if (runs)
{
if (!profile_info)
{
profile_info = &lto_gcov_summary;
lto_gcov_summary.runs = runs;
lto_gcov_summary.sum_all = lto_input_sleb128 (ib);
lto_gcov_summary.run_max = lto_input_sleb128 (ib);
lto_gcov_summary.sum_max = lto_input_sleb128 (ib);
}
/* We can support this by scaling all counts to nearest common multiple
of all different runs, but it is perhaps not worth the effort. */
else if (profile_info->runs != runs
|| profile_info->sum_all != lto_input_sleb128 (ib)
|| profile_info->run_max != lto_input_sleb128 (ib)
|| profile_info->sum_max != lto_input_sleb128 (ib))
sorry ("combining units with different profiles is not supported");
/* We allow some units to have profile and other to not have one. This will
just make unprofiled units to be size optimized that is sane. */
file_data->profile_info.runs = runs;
file_data->profile_info.sum_max = lto_input_uleb128 (ib);
if (runs > file_data->profile_info.sum_max)
fatal_error ("Corrupted profile info in %s: sum_max is smaller than runs",
file_data->file_name);
}
}
/* Rescale profile summaries to the same number of runs in the whole unit. */
static void
merge_profile_summaries (struct lto_file_decl_data **file_data_vec)
{
struct lto_file_decl_data *file_data;
unsigned int j;
gcov_unsigned_t max_runs = 0;
struct cgraph_node *node;
struct cgraph_edge *edge;
/* Find unit with maximal number of runs. If we ever get serious about
roundoff errors, we might also consider computing smallest common
multiply. */
for (j = 0; (file_data = file_data_vec[j]) != NULL; j++)
if (max_runs < file_data->profile_info.runs)
max_runs = file_data->profile_info.runs;
if (!max_runs)
return;
/* Simple overflow check. We probably don't need to support that many train
runs. Such a large value probably imply data corruption anyway. */
if (max_runs > INT_MAX / REG_BR_PROB_BASE)
{
sorry ("At most %i profile runs is supported. Perhaps corrupted profile?",
INT_MAX / REG_BR_PROB_BASE);
return;
}
profile_info = &lto_gcov_summary;
lto_gcov_summary.runs = max_runs;
lto_gcov_summary.sum_max = 0;
/* Rescale all units to the maximal number of runs.
sum_max can not be easily merged, as we have no idea what files come from
the same run. We do not use the info anyway, so leave it 0. */
for (j = 0; (file_data = file_data_vec[j]) != NULL; j++)
if (file_data->profile_info.runs)
{
int scale = ((REG_BR_PROB_BASE * max_runs
+ file_data->profile_info.runs / 2)
/ file_data->profile_info.runs);
lto_gcov_summary.sum_max = MAX (lto_gcov_summary.sum_max,
(file_data->profile_info.sum_max
* scale
+ REG_BR_PROB_BASE / 2)
/ REG_BR_PROB_BASE);
}
/* Watch roundoff errors. */
if (lto_gcov_summary.sum_max < max_runs)
lto_gcov_summary.sum_max = max_runs;
/* If merging already happent at WPA time, we are done. */
if (flag_ltrans)
return;
/* Now compute count_materialization_scale of each node.
During LTRANS we already have values of count_materialization_scale
computed, so just update them. */
for (node = cgraph_nodes; node; node = node->next)
if (node->local.lto_file_data->profile_info.run_max)
{
int scale;
if (node->local.lto_file_data->profile_info.runs)
scale =
((node->count_materialization_scale * max_runs
+ node->local.lto_file_data->profile_info.run_max / 2)
/ node->local.lto_file_data->profile_info.run_max);
else
scale = node->count_materialization_scale;
node->count_materialization_scale = scale;
if (scale < 0)
fatal_error ("Profile information in %s corrupted",
file_data->file_name);
if (scale == REG_BR_PROB_BASE)
continue;
for (edge = node->callees; edge; edge = edge->next_callee)
edge->count = ((edge->count * scale + REG_BR_PROB_BASE / 2)
/ REG_BR_PROB_BASE);
node->count = ((node->count * scale + REG_BR_PROB_BASE / 2)
/ REG_BR_PROB_BASE);
}
}
/* Input and merge the cgraph from each of the .o files passed to
lto1. */
@ -1473,7 +1552,7 @@ input_cgraph (void)
&data, &len);
if (!ib)
fatal_error ("cannot find LTO cgraph in %s", file_data->file_name);
input_profile_summary (ib);
input_profile_summary (ib, file_data);
file_data->cgraph_node_encoder = lto_cgraph_encoder_new ();
nodes = input_cgraph_1 (file_data, ib);
lto_destroy_simple_input_block (file_data, LTO_section_cgraph,
@ -1499,6 +1578,8 @@ input_cgraph (void)
VEC_free (cgraph_node_ptr, heap, nodes);
VEC_free (varpool_node_ptr, heap, varpool);
}
merge_profile_summaries (file_data_vec);
/* Clear out the aux field that was used to store enough state to
tell which nodes should be overwritten. */

20
gcc/lto-streamer-in.c
View file

@ -719,7 +719,8 @@ make_new_block (struct function *fn, unsigned int index)
/* Read the CFG for function FN from input block IB. */
static void
input_cfg (struct lto_input_block *ib, struct function *fn)
input_cfg (struct lto_input_block *ib, struct function *fn,
int count_materialization_scale)
{
unsigned int bb_count;
basic_block p_bb;
@ -752,7 +753,8 @@ input_cfg (struct lto_input_block *ib, struct function *fn)
if (bb == NULL)
bb = make_new_block (fn, index);
edge_count = lto_input_uleb128 (ib);
edge_count = (lto_input_uleb128 (ib) * count_materialization_scale
+ REG_BR_PROB_BASE / 2) / REG_BR_PROB_BASE;
/* Connect up the CFG. */
for (i = 0; i < edge_count; i++)
@ -1066,7 +1068,8 @@ input_gimple_stmt (struct lto_input_block *ib, struct data_in *data_in,
static void
input_bb (struct lto_input_block *ib, enum LTO_tags tag,
struct data_in *data_in, struct function *fn)
struct data_in *data_in, struct function *fn,
int count_materialization_scale)
{
unsigned int index;
basic_block bb;
@ -1079,7 +1082,8 @@ input_bb (struct lto_input_block *ib, enum LTO_tags tag,
index = lto_input_uleb128 (ib);
bb = BASIC_BLOCK_FOR_FUNCTION (fn, index);
bb->count = lto_input_sleb128 (ib);
bb->count = (lto_input_sleb128 (ib) * count_materialization_scale
+ REG_BR_PROB_BASE / 2) / REG_BR_PROB_BASE;
bb->loop_depth = lto_input_sleb128 (ib);
bb->frequency = lto_input_sleb128 (ib);
bb->flags = lto_input_sleb128 (ib);
@ -1253,12 +1257,14 @@ input_function (tree fn_decl, struct data_in *data_in,
DECL_INITIAL (fn_decl) = lto_input_tree (ib, data_in);
gcc_assert (DECL_INITIAL (fn_decl));
DECL_SAVED_TREE (fn_decl) = NULL_TREE;
node = cgraph_node (fn_decl);
/* Read all the basic blocks. */
tag = input_record_start (ib);
while (tag)
{
input_bb (ib, tag, data_in, fn);
input_bb (ib, tag, data_in, fn,
node->count_materialization_scale);
tag = input_record_start (ib);
}
@ -1300,7 +1306,6 @@ input_function (tree fn_decl, struct data_in *data_in,
gimple_set_body (fn_decl, bb_seq (ei_edge (ei)->dest));
}
node = cgraph_node (fn_decl);
fixup_call_stmt_edges (node, stmts);
execute_all_ipa_stmt_fixups (node, stmts);
@ -1393,6 +1398,7 @@ lto_read_body (struct lto_file_decl_data *file_data, tree fn_decl,
{
struct function *fn = DECL_STRUCT_FUNCTION (fn_decl);
struct lto_in_decl_state *decl_state;
struct cgraph_node *node = cgraph_node (fn_decl);
push_cfun (fn);
init_tree_ssa (fn);
@ -1402,7 +1408,7 @@ lto_read_body (struct lto_file_decl_data *file_data, tree fn_decl,
gcc_assert (decl_state);
file_data->current_decl_state = decl_state;
input_cfg (&ib_cfg, fn);
input_cfg (&ib_cfg, fn, node->count_materialization_scale);
/* Set up the struct function. */
input_function (fn_decl, data_in, &ib_main);

3
gcc/lto-streamer.h
View file

@ -31,6 +31,7 @@ along with GCC; see the file COPYING3. If not see
#include "vec.h"
#include "vecprim.h"
#include "alloc-pool.h"
#include "gcov-io.h"
/* Define when debugging the LTO streamer. This causes the writer
to output the numeric value for the memory address of the tree node
@ -610,6 +611,8 @@ struct GTY(()) lto_file_decl_data
/* Symbol resolutions for this file */
VEC(ld_plugin_symbol_resolution_t,heap) * GTY((skip)) resolutions;
struct gcov_ctr_summary GTY((skip)) profile_info;
};
typedef struct lto_file_decl_data *lto_file_decl_data_ptr;