- avoid-cycling-on-vertain-subreg-reloads.patch: Add patch source comment - change-gcc-BASE-VER.patch: Likewise - dont-generate-IF_THEN_ELSE.patch: Likewise - fix-ICE-in-compute_live_loop_exits.patch: Likewise - fix-ICE-in-eliminate_stmt.patch: Likewise - fix-ICE-in-vect_create_epilog_for_reduction.patch: Likewise - fix-ICE-in-vect_stmt_to_vectorize.patch: Likewise - fix-ICE-in-verify_ssa.patch: Likewise - fix-ICE-when-vectorizing-nested-cycles.patch: Likewise - fix-cost-of-plus.patch: Likewise - ipa-const-prop-self-recursion-bugfix.patch: Likewise - simplify-removing-subregs.patch: Likewise - medium-code-mode.patch: Bugfix - fix-when-peeling-for-alignment.patch: Move to ... - fix-PR-92351-When-peeling-for-alignment.patch: ... this - AArch64-Fix-constraints-for-CPY-M.patch: New file - Apply-maximum-nunits-for-BB-SLP.patch: New file - Fix-EXTRACT_LAST_REDUCTION-segfault.patch: New file - Fix-up-push_partial_def-little-endian-bitfield.patch: New file - Fix-zero-masking-for-vcvtps2ph.patch: New file - IRA-Handle-fully-tied-destinations.patch: New file - SLP-VECT-Add-check-to-fix-96837.patch: New file - aarch64-Fix-ash-lr-lshr-mode-3-expanders.patch: New file - aarch64-Fix-bf16-and-matrix-g++-gfortran.patch: New file - aarch64-Fix-mismatched-SVE-predicate-modes.patch: New file - aarch64-fix-sve-acle-error.patch: New file - adjust-vector-cost-and-move-EXTRACT_LAST_REDUCTION-costing.patch: New file - bf16-and-matrix-characteristic.patch: New file - fix-ICE-IPA-compare-VRP-types.patch: New file - fix-ICE-in-affine-combination.patch: New file - fix-ICE-in-pass-vect.patch: New file - fix-ICE-in-vect_update_misalignment_for_peel.patch: New file - fix-addlosymdi-ICE-in-pass-reload.patch: New file - fix-an-ICE-in-vect_recog_mask_conversion_pattern.patch: New file - fix-avx512vl-vcvttpd2dq-2-fail.patch: New file - fix-issue499-add-nop-convert.patch: New file - fix-issue604-ldist-dependency-fixup.patch: New file - modulo-sched-Carefully-process-loop-counter-initiali.patch: New file - re-PR-target-91124-gcc.target-i386-avx512vl-vpshldvd.patch: New file - reduction-paths-with-unhandled-live-stmt.patch: New file - redundant-loop-elimination.patch: New file - sccvn-Improve-handling-of-load-masked-with-integer.patch: New file - speed-up-DDG-analysis-and-fix-bootstrap-compare-debug.patch: New file - store-merging-Consider-also-overlapping-stores-earlier.patch: New file - tree-optimization-96920-another-ICE-when-vectorizing.patch: New file - tree-optimization-97812-fix-range-query-in-VRP-asser.patch: New file - vectorizable-comparison-Swap-operands-only-once.patch: New file - x86-Fix-bf16-and-matrix.patch: New file
719 lines
22 KiB
Diff
719 lines
22 KiB
Diff
This backport contains 2 patchs from gcc main stream tree.
|
||
The commit id of these patchs list as following in the order of time.
|
||
|
||
728c2e5eeaa91cf708f2b1b1f996653a7eebae59
|
||
0001-modulo-sched-speed-up-DDG-analysis-PR90001.patch
|
||
|
||
06d5d63d9944691bb4286e5f6b2422cc97148336
|
||
0001-modulo-sched-fix-bootstrap-compare-debug-issue.patch
|
||
|
||
diff -Nurp a/gcc/ddg.c b/gcc/ddg.c
|
||
--- a/gcc/ddg.c 2020-11-28 18:40:12.371633230 +0800
|
||
+++ b/gcc/ddg.c 2020-11-28 18:38:33.835633230 +0800
|
||
@@ -32,9 +32,6 @@ along with GCC; see the file COPYING3.
|
||
|
||
#ifdef INSN_SCHEDULING
|
||
|
||
-/* A flag indicating that a ddg edge belongs to an SCC or not. */
|
||
-enum edge_flag {NOT_IN_SCC = 0, IN_SCC};
|
||
-
|
||
/* Forward declarations. */
|
||
static void add_backarc_to_ddg (ddg_ptr, ddg_edge_ptr);
|
||
static void add_backarc_to_scc (ddg_scc_ptr, ddg_edge_ptr);
|
||
@@ -188,9 +185,6 @@ create_ddg_dep_from_intra_loop_link (ddg
|
||
else if (DEP_TYPE (link) == REG_DEP_OUTPUT)
|
||
t = OUTPUT_DEP;
|
||
|
||
- gcc_assert (!DEBUG_INSN_P (dest_node->insn) || t == ANTI_DEP);
|
||
- gcc_assert (!DEBUG_INSN_P (src_node->insn) || t == ANTI_DEP);
|
||
-
|
||
/* We currently choose not to create certain anti-deps edges and
|
||
compensate for that by generating reg-moves based on the life-range
|
||
analysis. The anti-deps that will be deleted are the ones which
|
||
@@ -225,9 +219,9 @@ create_ddg_dep_from_intra_loop_link (ddg
|
||
}
|
||
}
|
||
|
||
- latency = dep_cost (link);
|
||
- e = create_ddg_edge (src_node, dest_node, t, dt, latency, distance);
|
||
- add_edge_to_ddg (g, e);
|
||
+ latency = dep_cost (link);
|
||
+ e = create_ddg_edge (src_node, dest_node, t, dt, latency, distance);
|
||
+ add_edge_to_ddg (g, e);
|
||
}
|
||
|
||
/* The same as the above function, but it doesn't require a link parameter. */
|
||
@@ -240,9 +234,6 @@ create_ddg_dep_no_link (ddg_ptr g, ddg_n
|
||
enum reg_note dep_kind;
|
||
struct _dep _dep, *dep = &_dep;
|
||
|
||
- gcc_assert (!DEBUG_INSN_P (to->insn) || d_t == ANTI_DEP);
|
||
- gcc_assert (!DEBUG_INSN_P (from->insn) || d_t == ANTI_DEP);
|
||
-
|
||
if (d_t == ANTI_DEP)
|
||
dep_kind = REG_DEP_ANTI;
|
||
else if (d_t == OUTPUT_DEP)
|
||
@@ -275,16 +266,15 @@ create_ddg_dep_no_link (ddg_ptr g, ddg_n
|
||
static void
|
||
add_cross_iteration_register_deps (ddg_ptr g, df_ref last_def)
|
||
{
|
||
- int regno = DF_REF_REGNO (last_def);
|
||
struct df_link *r_use;
|
||
int has_use_in_bb_p = false;
|
||
- rtx_insn *def_insn = DF_REF_INSN (last_def);
|
||
- ddg_node_ptr last_def_node = get_node_of_insn (g, def_insn);
|
||
- ddg_node_ptr use_node;
|
||
+ int regno = DF_REF_REGNO (last_def);
|
||
+ ddg_node_ptr last_def_node = get_node_of_insn (g, DF_REF_INSN (last_def));
|
||
df_ref first_def = df_bb_regno_first_def_find (g->bb, regno);
|
||
+ ddg_node_ptr first_def_node = get_node_of_insn (g, DF_REF_INSN (first_def));
|
||
+ ddg_node_ptr use_node;
|
||
|
||
- gcc_assert (last_def_node);
|
||
- gcc_assert (first_def);
|
||
+ gcc_assert (last_def_node && first_def && first_def_node);
|
||
|
||
if (flag_checking && DF_REF_ID (last_def) != DF_REF_ID (first_def))
|
||
{
|
||
@@ -303,6 +293,9 @@ add_cross_iteration_register_deps (ddg_p
|
||
|
||
rtx_insn *use_insn = DF_REF_INSN (r_use->ref);
|
||
|
||
+ if (DEBUG_INSN_P (use_insn))
|
||
+ continue;
|
||
+
|
||
/* ??? Do not handle uses with DF_REF_IN_NOTE notes. */
|
||
use_node = get_node_of_insn (g, use_insn);
|
||
gcc_assert (use_node);
|
||
@@ -313,35 +306,28 @@ add_cross_iteration_register_deps (ddg_p
|
||
iteration. Any such upwards exposed use appears before
|
||
the last_def def. */
|
||
create_ddg_dep_no_link (g, last_def_node, use_node,
|
||
- DEBUG_INSN_P (use_insn) ? ANTI_DEP : TRUE_DEP,
|
||
- REG_DEP, 1);
|
||
+ TRUE_DEP, REG_DEP, 1);
|
||
}
|
||
- else if (!DEBUG_INSN_P (use_insn))
|
||
+ else
|
||
{
|
||
/* Add anti deps from last_def's uses in the current iteration
|
||
to the first def in the next iteration. We do not add ANTI
|
||
dep when there is an intra-loop TRUE dep in the opposite
|
||
direction, but use regmoves to fix such disregarded ANTI
|
||
deps when broken. If the first_def reaches the USE then
|
||
- there is such a dep. */
|
||
- ddg_node_ptr first_def_node = get_node_of_insn (g,
|
||
- DF_REF_INSN (first_def));
|
||
-
|
||
- gcc_assert (first_def_node);
|
||
-
|
||
- /* Always create the edge if the use node is a branch in
|
||
- order to prevent the creation of reg-moves.
|
||
- If the address that is being auto-inc or auto-dec in LAST_DEF
|
||
- is used in USE_INSN then do not remove the edge to make sure
|
||
- reg-moves will not be created for that address. */
|
||
- if (DF_REF_ID (last_def) != DF_REF_ID (first_def)
|
||
- || !flag_modulo_sched_allow_regmoves
|
||
+ there is such a dep.
|
||
+ Always create the edge if the use node is a branch in
|
||
+ order to prevent the creation of reg-moves.
|
||
+ If the address that is being auto-inc or auto-dec in LAST_DEF
|
||
+ is used in USE_INSN then do not remove the edge to make sure
|
||
+ reg-moves will not be created for that address. */
|
||
+ if (DF_REF_ID (last_def) != DF_REF_ID (first_def)
|
||
+ || !flag_modulo_sched_allow_regmoves
|
||
|| JUMP_P (use_node->insn)
|
||
- || autoinc_var_is_used_p (DF_REF_INSN (last_def), use_insn)
|
||
+ || autoinc_var_is_used_p (DF_REF_INSN (last_def), use_insn)
|
||
|| def_has_ccmode_p (DF_REF_INSN (last_def)))
|
||
- create_ddg_dep_no_link (g, use_node, first_def_node, ANTI_DEP,
|
||
- REG_DEP, 1);
|
||
-
|
||
+ create_ddg_dep_no_link (g, use_node, first_def_node, ANTI_DEP,
|
||
+ REG_DEP, 1);
|
||
}
|
||
}
|
||
/* Create an inter-loop output dependence between LAST_DEF (which is the
|
||
@@ -351,19 +337,11 @@ add_cross_iteration_register_deps (ddg_p
|
||
defs starting with a true dependence to a use which can be in the
|
||
next iteration; followed by an anti dependence of that use to the
|
||
first def (i.e. if there is a use between the two defs.) */
|
||
- if (!has_use_in_bb_p)
|
||
- {
|
||
- ddg_node_ptr dest_node;
|
||
-
|
||
- if (DF_REF_ID (last_def) == DF_REF_ID (first_def))
|
||
- return;
|
||
-
|
||
- dest_node = get_node_of_insn (g, DF_REF_INSN (first_def));
|
||
- gcc_assert (dest_node);
|
||
- create_ddg_dep_no_link (g, last_def_node, dest_node,
|
||
- OUTPUT_DEP, REG_DEP, 1);
|
||
- }
|
||
+ if (!has_use_in_bb_p && DF_REF_ID (last_def) != DF_REF_ID (first_def))
|
||
+ create_ddg_dep_no_link (g, last_def_node, first_def_node,
|
||
+ OUTPUT_DEP, REG_DEP, 1);
|
||
}
|
||
+
|
||
/* Build inter-loop dependencies, by looking at DF analysis backwards. */
|
||
static void
|
||
build_inter_loop_deps (ddg_ptr g)
|
||
@@ -420,13 +398,9 @@ add_intra_loop_mem_dep (ddg_ptr g, ddg_n
|
||
if (mem_write_insn_p (from->insn))
|
||
{
|
||
if (mem_read_insn_p (to->insn))
|
||
- create_ddg_dep_no_link (g, from, to,
|
||
- DEBUG_INSN_P (to->insn)
|
||
- ? ANTI_DEP : TRUE_DEP, MEM_DEP, 0);
|
||
+ create_ddg_dep_no_link (g, from, to, TRUE_DEP, MEM_DEP, 0);
|
||
else
|
||
- create_ddg_dep_no_link (g, from, to,
|
||
- DEBUG_INSN_P (to->insn)
|
||
- ? ANTI_DEP : OUTPUT_DEP, MEM_DEP, 0);
|
||
+ create_ddg_dep_no_link (g, from, to, OUTPUT_DEP, MEM_DEP, 0);
|
||
}
|
||
else if (!mem_read_insn_p (to->insn))
|
||
create_ddg_dep_no_link (g, from, to, ANTI_DEP, MEM_DEP, 0);
|
||
@@ -444,13 +418,9 @@ add_inter_loop_mem_dep (ddg_ptr g, ddg_n
|
||
if (mem_write_insn_p (from->insn))
|
||
{
|
||
if (mem_read_insn_p (to->insn))
|
||
- create_ddg_dep_no_link (g, from, to,
|
||
- DEBUG_INSN_P (to->insn)
|
||
- ? ANTI_DEP : TRUE_DEP, MEM_DEP, 1);
|
||
+ create_ddg_dep_no_link (g, from, to, TRUE_DEP, MEM_DEP, 1);
|
||
else if (from->cuid != to->cuid)
|
||
- create_ddg_dep_no_link (g, from, to,
|
||
- DEBUG_INSN_P (to->insn)
|
||
- ? ANTI_DEP : OUTPUT_DEP, MEM_DEP, 1);
|
||
+ create_ddg_dep_no_link (g, from, to, OUTPUT_DEP, MEM_DEP, 1);
|
||
}
|
||
else
|
||
{
|
||
@@ -459,13 +429,9 @@ add_inter_loop_mem_dep (ddg_ptr g, ddg_n
|
||
else if (from->cuid != to->cuid)
|
||
{
|
||
create_ddg_dep_no_link (g, from, to, ANTI_DEP, MEM_DEP, 1);
|
||
- if (DEBUG_INSN_P (from->insn) || DEBUG_INSN_P (to->insn))
|
||
- create_ddg_dep_no_link (g, to, from, ANTI_DEP, MEM_DEP, 1);
|
||
- else
|
||
- create_ddg_dep_no_link (g, to, from, TRUE_DEP, MEM_DEP, 1);
|
||
+ create_ddg_dep_no_link (g, to, from, TRUE_DEP, MEM_DEP, 1);
|
||
}
|
||
}
|
||
-
|
||
}
|
||
|
||
/* Perform intra-block Data Dependency analysis and connect the nodes in
|
||
@@ -494,20 +460,10 @@ build_intra_loop_deps (ddg_ptr g)
|
||
sd_iterator_def sd_it;
|
||
dep_t dep;
|
||
|
||
- if (! INSN_P (dest_node->insn))
|
||
- continue;
|
||
-
|
||
FOR_EACH_DEP (dest_node->insn, SD_LIST_BACK, sd_it, dep)
|
||
{
|
||
rtx_insn *src_insn = DEP_PRO (dep);
|
||
- ddg_node_ptr src_node;
|
||
-
|
||
- /* Don't add dependencies on debug insns to non-debug insns
|
||
- to avoid codegen differences between -g and -g0. */
|
||
- if (DEBUG_INSN_P (src_insn) && !DEBUG_INSN_P (dest_node->insn))
|
||
- continue;
|
||
-
|
||
- src_node = get_node_of_insn (g, src_insn);
|
||
+ ddg_node_ptr src_node = get_node_of_insn (g, src_insn);
|
||
|
||
if (!src_node)
|
||
continue;
|
||
@@ -524,8 +480,7 @@ build_intra_loop_deps (ddg_ptr g)
|
||
for (j = 0; j <= i; j++)
|
||
{
|
||
ddg_node_ptr j_node = &g->nodes[j];
|
||
- if (DEBUG_INSN_P (j_node->insn))
|
||
- continue;
|
||
+
|
||
if (mem_access_insn_p (j_node->insn))
|
||
{
|
||
/* Don't bother calculating inter-loop dep if an intra-loop dep
|
||
@@ -564,7 +519,7 @@ create_ddg (basic_block bb, int closing_
|
||
{
|
||
ddg_ptr g;
|
||
rtx_insn *insn, *first_note;
|
||
- int i;
|
||
+ int i, j;
|
||
int num_nodes = 0;
|
||
|
||
g = (ddg_ptr) xcalloc (1, sizeof (struct ddg));
|
||
@@ -576,23 +531,21 @@ create_ddg (basic_block bb, int closing_
|
||
for (insn = BB_HEAD (bb); insn != NEXT_INSN (BB_END (bb));
|
||
insn = NEXT_INSN (insn))
|
||
{
|
||
- if (! INSN_P (insn) || GET_CODE (PATTERN (insn)) == USE)
|
||
+ if (!INSN_P (insn) || GET_CODE (PATTERN (insn)) == USE)
|
||
continue;
|
||
|
||
- if (DEBUG_INSN_P (insn))
|
||
- g->num_debug++;
|
||
- else
|
||
+ if (NONDEBUG_INSN_P (insn))
|
||
{
|
||
if (mem_read_insn_p (insn))
|
||
g->num_loads++;
|
||
if (mem_write_insn_p (insn))
|
||
g->num_stores++;
|
||
+ num_nodes++;
|
||
}
|
||
- num_nodes++;
|
||
}
|
||
|
||
/* There is nothing to do for this BB. */
|
||
- if ((num_nodes - g->num_debug) <= 1)
|
||
+ if (num_nodes <= 1)
|
||
{
|
||
free (g);
|
||
return NULL;
|
||
@@ -607,32 +560,39 @@ create_ddg (basic_block bb, int closing_
|
||
for (insn = BB_HEAD (bb); insn != NEXT_INSN (BB_END (bb));
|
||
insn = NEXT_INSN (insn))
|
||
{
|
||
- if (! INSN_P (insn))
|
||
- {
|
||
- if (! first_note && NOTE_P (insn)
|
||
- && NOTE_KIND (insn) != NOTE_INSN_BASIC_BLOCK)
|
||
- first_note = insn;
|
||
- continue;
|
||
- }
|
||
+ if (LABEL_P (insn) || NOTE_INSN_BASIC_BLOCK_P (insn))
|
||
+ continue;
|
||
+
|
||
+ if (!first_note && (INSN_P (insn) || NOTE_P (insn)))
|
||
+ first_note = insn;
|
||
+
|
||
+ if (!INSN_P (insn) || GET_CODE (PATTERN (insn)) == USE)
|
||
+ continue;
|
||
+
|
||
if (JUMP_P (insn))
|
||
{
|
||
gcc_assert (!g->closing_branch);
|
||
g->closing_branch = &g->nodes[i];
|
||
}
|
||
- else if (GET_CODE (PATTERN (insn)) == USE)
|
||
+
|
||
+ if (NONDEBUG_INSN_P (insn))
|
||
{
|
||
- if (! first_note)
|
||
- first_note = insn;
|
||
- continue;
|
||
- }
|
||
+ g->nodes[i].cuid = i;
|
||
+ g->nodes[i].successors = sbitmap_alloc (num_nodes);
|
||
+ bitmap_clear (g->nodes[i].successors);
|
||
+ g->nodes[i].predecessors = sbitmap_alloc (num_nodes);
|
||
+ bitmap_clear (g->nodes[i].predecessors);
|
||
+
|
||
+ gcc_checking_assert (first_note);
|
||
+ g->nodes[i].first_note = first_note;
|
||
+
|
||
+ g->nodes[i].aux.count = -1;
|
||
+ g->nodes[i].max_dist = XCNEWVEC (int, num_nodes);
|
||
+ for (j = 0; j < num_nodes; j++)
|
||
+ g->nodes[i].max_dist[j] = -1;
|
||
|
||
- g->nodes[i].cuid = i;
|
||
- g->nodes[i].successors = sbitmap_alloc (num_nodes);
|
||
- bitmap_clear (g->nodes[i].successors);
|
||
- g->nodes[i].predecessors = sbitmap_alloc (num_nodes);
|
||
- bitmap_clear (g->nodes[i].predecessors);
|
||
- g->nodes[i].first_note = (first_note ? first_note : insn);
|
||
- g->nodes[i++].insn = insn;
|
||
+ g->nodes[i++].insn = insn;
|
||
+ }
|
||
first_note = NULL;
|
||
}
|
||
|
||
@@ -668,6 +628,7 @@ free_ddg (ddg_ptr g)
|
||
}
|
||
sbitmap_free (g->nodes[i].successors);
|
||
sbitmap_free (g->nodes[i].predecessors);
|
||
+ free (g->nodes[i].max_dist);
|
||
}
|
||
if (g->num_backarcs > 0)
|
||
free (g->backarcs);
|
||
@@ -792,7 +753,7 @@ create_ddg_edge (ddg_node_ptr src, ddg_n
|
||
e->latency = l;
|
||
e->distance = d;
|
||
e->next_in = e->next_out = NULL;
|
||
- e->aux.info = 0;
|
||
+ e->in_scc = false;
|
||
return e;
|
||
}
|
||
|
||
@@ -820,7 +781,7 @@ add_edge_to_ddg (ddg_ptr g ATTRIBUTE_UNU
|
||
for now that cycles in the data dependence graph contain a single backarc.
|
||
This simplifies the algorithm, and can be generalized later. */
|
||
static void
|
||
-set_recurrence_length (ddg_scc_ptr scc, ddg_ptr g)
|
||
+set_recurrence_length (ddg_scc_ptr scc)
|
||
{
|
||
int j;
|
||
int result = -1;
|
||
@@ -828,17 +789,14 @@ set_recurrence_length (ddg_scc_ptr scc,
|
||
for (j = 0; j < scc->num_backarcs; j++)
|
||
{
|
||
ddg_edge_ptr backarc = scc->backarcs[j];
|
||
- int length;
|
||
int distance = backarc->distance;
|
||
ddg_node_ptr src = backarc->dest;
|
||
ddg_node_ptr dest = backarc->src;
|
||
+ int length = src->max_dist[dest->cuid];
|
||
+
|
||
+ if (length < 0)
|
||
+ continue;
|
||
|
||
- length = longest_simple_path (g, src->cuid, dest->cuid, scc->nodes);
|
||
- if (length < 0 )
|
||
- {
|
||
- /* fprintf (stderr, "Backarc not on simple cycle in SCC.\n"); */
|
||
- continue;
|
||
- }
|
||
length += backarc->latency;
|
||
result = MAX (result, (length / distance));
|
||
}
|
||
@@ -846,9 +804,9 @@ set_recurrence_length (ddg_scc_ptr scc,
|
||
}
|
||
|
||
/* Create a new SCC given the set of its nodes. Compute its recurrence_length
|
||
- and mark edges that belong to this scc as IN_SCC. */
|
||
+ and mark edges that belong to this scc. */
|
||
static ddg_scc_ptr
|
||
-create_scc (ddg_ptr g, sbitmap nodes)
|
||
+create_scc (ddg_ptr g, sbitmap nodes, int id)
|
||
{
|
||
ddg_scc_ptr scc;
|
||
unsigned int u = 0;
|
||
@@ -866,16 +824,18 @@ create_scc (ddg_ptr g, sbitmap nodes)
|
||
ddg_edge_ptr e;
|
||
ddg_node_ptr n = &g->nodes[u];
|
||
|
||
+ gcc_assert (n->aux.count == -1);
|
||
+ n->aux.count = id;
|
||
+
|
||
for (e = n->out; e; e = e->next_out)
|
||
if (bitmap_bit_p (nodes, e->dest->cuid))
|
||
{
|
||
- e->aux.count = IN_SCC;
|
||
+ e->in_scc = true;
|
||
if (e->distance > 0)
|
||
add_backarc_to_scc (scc, e);
|
||
}
|
||
}
|
||
|
||
- set_recurrence_length (scc, g);
|
||
return scc;
|
||
}
|
||
|
||
@@ -1018,7 +978,7 @@ check_sccs (ddg_all_sccs_ptr sccs, int n
|
||
ddg_all_sccs_ptr
|
||
create_ddg_all_sccs (ddg_ptr g)
|
||
{
|
||
- int i;
|
||
+ int i, j, k, scc, way;
|
||
int num_nodes = g->num_nodes;
|
||
auto_sbitmap from (num_nodes);
|
||
auto_sbitmap to (num_nodes);
|
||
@@ -1038,7 +998,7 @@ create_ddg_all_sccs (ddg_ptr g)
|
||
ddg_node_ptr dest = backarc->dest;
|
||
|
||
/* If the backarc already belongs to an SCC, continue. */
|
||
- if (backarc->aux.count == IN_SCC)
|
||
+ if (backarc->in_scc)
|
||
continue;
|
||
|
||
bitmap_clear (scc_nodes);
|
||
@@ -1049,10 +1009,52 @@ create_ddg_all_sccs (ddg_ptr g)
|
||
|
||
if (find_nodes_on_paths (scc_nodes, g, from, to))
|
||
{
|
||
- scc = create_scc (g, scc_nodes);
|
||
+ scc = create_scc (g, scc_nodes, sccs->num_sccs);
|
||
add_scc_to_ddg (sccs, scc);
|
||
}
|
||
}
|
||
+
|
||
+ /* Init max_dist arrays for FloydâWarshall-like
|
||
+ longest patch calculation algorithm. */
|
||
+ for (k = 0; k < num_nodes; k++)
|
||
+ {
|
||
+ ddg_edge_ptr e;
|
||
+ ddg_node_ptr n = &g->nodes[k];
|
||
+
|
||
+ if (n->aux.count == -1)
|
||
+ continue;
|
||
+
|
||
+ n->max_dist[k] = 0;
|
||
+ for (e = n->out; e; e = e->next_out)
|
||
+ if (e->distance == 0 && g->nodes[e->dest->cuid].aux.count == n->aux.count)
|
||
+ n->max_dist[e->dest->cuid] = e->latency;
|
||
+ }
|
||
+
|
||
+ /* Run main Floid-Warshall loop. We use only non-backarc edges
|
||
+ inside each scc. */
|
||
+ for (k = 0; k < num_nodes; k++)
|
||
+ {
|
||
+ scc = g->nodes[k].aux.count;
|
||
+ if (scc != -1)
|
||
+ {
|
||
+ for (i = 0; i < num_nodes; i++)
|
||
+ if (g->nodes[i].aux.count == scc)
|
||
+ for (j = 0; j < num_nodes; j++)
|
||
+ if (g->nodes[j].aux.count == scc
|
||
+ && g->nodes[i].max_dist[k] >= 0
|
||
+ && g->nodes[k].max_dist[j] >= 0)
|
||
+ {
|
||
+ way = g->nodes[i].max_dist[k] + g->nodes[k].max_dist[j];
|
||
+ if (g->nodes[i].max_dist[j] < way)
|
||
+ g->nodes[i].max_dist[j] = way;
|
||
+ }
|
||
+ }
|
||
+ }
|
||
+
|
||
+ /* Calculate recurrence_length using max_dist info. */
|
||
+ for (i = 0; i < sccs->num_sccs; i++)
|
||
+ set_recurrence_length (sccs->sccs[i]);
|
||
+
|
||
order_sccs (sccs);
|
||
|
||
if (flag_checking)
|
||
@@ -1155,72 +1157,4 @@ find_nodes_on_paths (sbitmap result, ddg
|
||
return bitmap_and (result, reachable_from, reach_to);
|
||
}
|
||
|
||
-
|
||
-/* Updates the counts of U_NODE's successors (that belong to NODES) to be
|
||
- at-least as large as the count of U_NODE plus the latency between them.
|
||
- Sets a bit in TMP for each successor whose count was changed (increased).
|
||
- Returns nonzero if any count was changed. */
|
||
-static int
|
||
-update_dist_to_successors (ddg_node_ptr u_node, sbitmap nodes, sbitmap tmp)
|
||
-{
|
||
- ddg_edge_ptr e;
|
||
- int result = 0;
|
||
-
|
||
- for (e = u_node->out; e; e = e->next_out)
|
||
- {
|
||
- ddg_node_ptr v_node = e->dest;
|
||
- int v = v_node->cuid;
|
||
-
|
||
- if (bitmap_bit_p (nodes, v)
|
||
- && (e->distance == 0)
|
||
- && (v_node->aux.count < u_node->aux.count + e->latency))
|
||
- {
|
||
- v_node->aux.count = u_node->aux.count + e->latency;
|
||
- bitmap_set_bit (tmp, v);
|
||
- result = 1;
|
||
- }
|
||
- }
|
||
- return result;
|
||
-}
|
||
-
|
||
-
|
||
-/* Find the length of a longest path from SRC to DEST in G,
|
||
- going only through NODES, and disregarding backarcs. */
|
||
-int
|
||
-longest_simple_path (struct ddg * g, int src, int dest, sbitmap nodes)
|
||
-{
|
||
- int i;
|
||
- unsigned int u = 0;
|
||
- int change = 1;
|
||
- int num_nodes = g->num_nodes;
|
||
- auto_sbitmap workset (num_nodes);
|
||
- auto_sbitmap tmp (num_nodes);
|
||
-
|
||
-
|
||
- /* Data will hold the distance of the longest path found so far from
|
||
- src to each node. Initialize to -1 = less than minimum. */
|
||
- for (i = 0; i < g->num_nodes; i++)
|
||
- g->nodes[i].aux.count = -1;
|
||
- g->nodes[src].aux.count = 0;
|
||
-
|
||
- bitmap_clear (tmp);
|
||
- bitmap_set_bit (tmp, src);
|
||
-
|
||
- while (change)
|
||
- {
|
||
- sbitmap_iterator sbi;
|
||
-
|
||
- change = 0;
|
||
- bitmap_copy (workset, tmp);
|
||
- bitmap_clear (tmp);
|
||
- EXECUTE_IF_SET_IN_BITMAP (workset, 0, u, sbi)
|
||
- {
|
||
- ddg_node_ptr u_node = &g->nodes[u];
|
||
-
|
||
- change |= update_dist_to_successors (u_node, nodes, tmp);
|
||
- }
|
||
- }
|
||
- return g->nodes[dest].aux.count;
|
||
-}
|
||
-
|
||
#endif /* INSN_SCHEDULING */
|
||
diff -Nurp a/gcc/ddg.h b/gcc/ddg.h
|
||
--- a/gcc/ddg.h 2020-03-12 19:07:21.000000000 +0800
|
||
+++ b/gcc/ddg.h 2020-11-28 18:38:33.835633230 +0800
|
||
@@ -64,6 +64,10 @@ struct ddg_node
|
||
sbitmap successors;
|
||
sbitmap predecessors;
|
||
|
||
+ /* Temporary array used for Floyd-Warshall algorithm to find
|
||
+ scc recurrence length. */
|
||
+ int *max_dist;
|
||
+
|
||
/* For general use by algorithms manipulating the ddg. */
|
||
union {
|
||
int count;
|
||
@@ -95,11 +99,8 @@ struct ddg_edge
|
||
ddg_edge_ptr next_in;
|
||
ddg_edge_ptr next_out;
|
||
|
||
- /* For general use by algorithms manipulating the ddg. */
|
||
- union {
|
||
- int count;
|
||
- void *info;
|
||
- } aux;
|
||
+ /* Is true when edge is already in scc. */
|
||
+ bool in_scc;
|
||
};
|
||
|
||
/* This structure holds the Data Dependence Graph for a basic block. */
|
||
@@ -115,9 +116,6 @@ struct ddg
|
||
int num_loads;
|
||
int num_stores;
|
||
|
||
- /* Number of debug instructions in the BB. */
|
||
- int num_debug;
|
||
-
|
||
/* This array holds the nodes in the graph; it is indexed by the node
|
||
cuid, which follows the order of the instructions in the BB. */
|
||
ddg_node_ptr nodes;
|
||
@@ -178,7 +176,6 @@ ddg_all_sccs_ptr create_ddg_all_sccs (dd
|
||
void free_ddg_all_sccs (ddg_all_sccs_ptr);
|
||
|
||
int find_nodes_on_paths (sbitmap result, ddg_ptr, sbitmap from, sbitmap to);
|
||
-int longest_simple_path (ddg_ptr, int from, int to, sbitmap via);
|
||
|
||
bool autoinc_var_is_used_p (rtx_insn *, rtx_insn *);
|
||
|
||
diff -Nurp a/gcc/modulo-sched.c b/gcc/modulo-sched.c
|
||
--- a/gcc/modulo-sched.c 2020-03-12 19:07:21.000000000 +0800
|
||
+++ b/gcc/modulo-sched.c 2020-11-28 18:38:33.835633230 +0800
|
||
@@ -370,7 +370,7 @@ doloop_register_get (rtx_insn *head, rtx
|
||
: prev_nondebug_insn (tail));
|
||
|
||
for (insn = head; insn != first_insn_not_to_check; insn = NEXT_INSN (insn))
|
||
- if (!DEBUG_INSN_P (insn) && reg_mentioned_p (reg, insn))
|
||
+ if (NONDEBUG_INSN_P (insn) && reg_mentioned_p (reg, insn))
|
||
{
|
||
if (dump_file)
|
||
{
|
||
@@ -429,7 +429,7 @@ res_MII (ddg_ptr g)
|
||
if (targetm.sched.sms_res_mii)
|
||
return targetm.sched.sms_res_mii (g);
|
||
|
||
- return ((g->num_nodes - g->num_debug) / issue_rate);
|
||
+ return g->num_nodes / issue_rate;
|
||
}
|
||
|
||
|
||
@@ -2156,11 +2156,7 @@ sms_schedule_by_order (ddg_ptr g, int mi
|
||
ddg_node_ptr u_node = &ps->g->nodes[u];
|
||
rtx_insn *insn = u_node->insn;
|
||
|
||
- if (!NONDEBUG_INSN_P (insn))
|
||
- {
|
||
- bitmap_clear_bit (tobe_scheduled, u);
|
||
- continue;
|
||
- }
|
||
+ gcc_checking_assert (NONDEBUG_INSN_P (insn));
|
||
|
||
if (bitmap_bit_p (sched_nodes, u))
|
||
continue;
|
||
@@ -3162,9 +3158,6 @@ ps_has_conflicts (partial_schedule_ptr p
|
||
{
|
||
rtx_insn *insn = ps_rtl_insn (ps, crr_insn->id);
|
||
|
||
- if (!NONDEBUG_INSN_P (insn))
|
||
- continue;
|
||
-
|
||
/* Check if there is room for the current insn. */
|
||
if (!can_issue_more || state_dead_lock_p (curr_state))
|
||
return true;
|
||
diff -Nurp a/gcc/testsuite/gcc.c-torture/execute/pr70127-debug-sms.c b/gcc/testsuite/gcc.c-torture/execute/pr70127-debug-sms.c
|
||
--- a/gcc/testsuite/gcc.c-torture/execute/pr70127-debug-sms.c 1970-01-01 08:00:00.000000000 +0800
|
||
+++ b/gcc/testsuite/gcc.c-torture/execute/pr70127-debug-sms.c 2020-11-28 18:38:33.835633230 +0800
|
||
@@ -0,0 +1,23 @@
|
||
+/* { dg-additional-options "-fcompare-debug -fmodulo-sched" } */
|
||
+
|
||
+struct S { int f; signed int g : 2; } a[1], c = {5, 1}, d;
|
||
+short b;
|
||
+
|
||
+__attribute__((noinline, noclone)) void
|
||
+foo (int x)
|
||
+{
|
||
+ if (x != 1)
|
||
+ __builtin_abort ();
|
||
+}
|
||
+
|
||
+int
|
||
+main ()
|
||
+{
|
||
+ while (b++ <= 0)
|
||
+ {
|
||
+ struct S e = {1, 1};
|
||
+ d = e = a[0] = c;
|
||
+ }
|
||
+ foo (a[0].g);
|
||
+ return 0;
|
||
+}
|
||
diff -Nurp a/gcc/testsuite/gcc.dg/torture/pr87197-debug-sms.c b/gcc/testsuite/gcc.dg/torture/pr87197-debug-sms.c
|
||
--- a/gcc/testsuite/gcc.dg/torture/pr87197-debug-sms.c 1970-01-01 08:00:00.000000000 +0800
|
||
+++ b/gcc/testsuite/gcc.dg/torture/pr87197-debug-sms.c 2020-11-28 18:38:33.835633230 +0800
|
||
@@ -0,0 +1,36 @@
|
||
+/* { dg-do compile } */
|
||
+/* { dg-additional-options "-fcompare-debug -fmodulo-sched --param sms-min-sc=1" } */
|
||
+
|
||
+int a, c, e, f, g;
|
||
+void
|
||
+h (int i)
|
||
+{
|
||
+ a = i;
|
||
+}
|
||
+void
|
||
+j (char *i, long k)
|
||
+{
|
||
+ while (k--)
|
||
+ c = *i++;
|
||
+}
|
||
+void
|
||
+l (unsigned char *i, long k)
|
||
+{
|
||
+ unsigned char *b = i + k;
|
||
+ while (i < b)
|
||
+ {
|
||
+ h (*i);
|
||
+ i++;
|
||
+ }
|
||
+}
|
||
+void
|
||
+m ()
|
||
+{
|
||
+ while (e)
|
||
+ {
|
||
+ float d = g;
|
||
+ l ((char *) &d, sizeof (g));
|
||
+ if (f)
|
||
+ j ((char *) &d, sizeof (g));
|
||
+ }
|
||
+}
|