[Bug ipa/93429] Missed IPA-CP on by-ref argument directly passed through from caller
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93429 fxue at gcc dot gnu.org changed: What|Removed |Added Status|NEW |RESOLVED Resolution|--- |FIXED --- Comment #2 from fxue at gcc dot gnu.org --- Fixed.
[Bug testsuite/97066] [11 Regression] regression caused by r11-3207
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97066 fxue at gcc dot gnu.org changed: What|Removed |Added Component|tree-optimization |testsuite Last reconfirmed||2020-09-16 Assignee|unassigned at gcc dot gnu.org |fxue at os dot amperecomputing.com Status|UNCONFIRMED |NEW Ever confirmed|0 |1 CC||fxue at gcc dot gnu.org
[Bug tree-optimization/89134] A missing optimization opportunity for a simple branch in loop
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89134 --- Comment #14 from fxue at gcc dot gnu.org --- Author: fxue Date: Thu Nov 7 15:43:01 2019 New Revision: 277923 URL: https://gcc.gnu.org/viewcvs?rev=277923&root=gcc&view=rev Log: Loop split on semi-invariant conditional statement 2019-11-07 Feng Xue PR tree-optimization/89134 * doc/invoke.texi (min-loop-cond-split-prob): Document new --params. * params.def: Add min-loop-cond-split-prob. * tree-ssa-loop-split.c (split_loop): Remove niter parameter, move some outside checks on loop into the function. (split_info): New class. (find_vdef_in_loop, get_control_equiv_head_block): New functions. (find_control_dep_blocks, vuse_semi_invariant_p): Likewise. (ssa_semi_invariant_p, loop_iter_phi_semi_invariant_p): Likewise. (control_dep_semi_invariant_p, stmt_semi_invariant_p_1): Likewise. (stmt_semi_invariant_p, branch_removable_p): Likewise. (get_cond_invariant_branch, compute_added_num_insns): Likewise. (get_cond_branch_to_split_loop, do_split_loop_on_cond): Likewise. (split_loop_on_cond): Likewise. (tree_ssa_split_loops): Add loop split on conditional statement. 2019-11-07 Feng Xue PR tree-optimization/89134 * gcc.dg/tree-ssa/loop-cond-split-1.c: New test. * g++.dg/tree-ssa/loop-cond-split-1.C: New test. * gcc.dg/torture/pr55107.c: Add -fno-split-loops. Added: trunk/gcc/testsuite/g++.dg/tree-ssa/loop-cond-split-1.C trunk/gcc/testsuite/gcc.dg/tree-ssa/loop-cond-split-1.c Modified: trunk/gcc/ChangeLog trunk/gcc/doc/invoke.texi trunk/gcc/params.def trunk/gcc/testsuite/ChangeLog trunk/gcc/testsuite/gcc.dg/torture/pr55107.c trunk/gcc/tree-ssa-loop-split.c
[Bug ipa/91682] IPA-cp can not propagate value for by-ref argument in form of *arg = param op constant
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91682 --- Comment #1 from fxue at gcc dot gnu.org --- Author: fxue Date: Thu Nov 14 03:19:15 2019 New Revision: 278193 URL: https://gcc.gnu.org/viewcvs?rev=278193&root=gcc&view=rev Log: Support extended aggregate jump function in ipa-cp 2019-11-14 Feng Xue PR ipa/91682 * ipa-prop.h (jump_func_type): New value IPA_JF_LOAD_AGG. (ipa_load_agg_data, ipa_agg_value, ipa_agg_value_set): New structs. (ipa_agg_jf_item): Add new field jftype and type, redefine field value. (ipa_agg_jump_function): Remove member function equal_to. (ipa_agg_jump_function_p): Remove typedef. (ipa_copy_agg_values, ipa_release_agg_values): New functions. * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump information for aggregate jump function. (get_ssa_def_if_simple_copy): Add new parameter rhs_stmt to record last definition statement. (load_from_unmodified_param_or_agg): New function. (ipa_known_agg_contents_list): Add new field type and value, remove field constant. (build_agg_jump_func_from_list): Rename parameter const_count to value_count, build aggregate jump function from ipa_load_agg_data. (analyze_agg_content_value): New function. (extract_mem_content): Analyze memory store assignment to prepare information for aggregate jump function generation. (determine_known_aggregate_parts): Add new parameter fbi, remove parameter aa_walk_budeget_p. (update_jump_functions_after_inlining): Update aggregate jump function. (ipa_find_agg_cst_for_param): Change type of parameter agg. (try_make_edge_direct_simple_call): Add new parameter new_root. (try_make_edge_direct_virtual_call): Add new parameter new_root and new_root_info. (update_indirect_edges_after_inlining): Pass new argument to try_make_edge_direct_simple_call and try_make_edge_direct_virtual_call. (ipa_write_jump_function): Write aggregate jump function to file. (ipa_read_jump_function): Read aggregate jump function from file. (ipa_agg_value::equal_to): Migrate from ipa_agg_jf_item::equal_to. * ipa-cp.c (ipa_get_jf_arith_result): New function. (ipa_agg_value_from_node): Likewise. (ipa_agg_value_set_from_jfunc): Likewise. (propagate_vals_across_arith_jfunc): Likewise. (propagate_aggregate_lattice): Likewise. (ipa_get_jf_pass_through_result): Call ipa_get_jf_arith_result. (propagate_vals_across_pass_through): Call propagate_vals_across_arith_jfunc. (get_clone_agg_value): Move forward. (propagate_aggs_across_jump_function): Handle value propagation for aggregate jump function. (agg_jmp_p_vec_for_t_vec): Remove. (context_independent_aggregate_values): Replace vec with vec. (copy_plats_to_inter, intersect_with_plats): Likewise. (agg_replacements_to_vector, intersect_with_agg_replacements): Likewise. (intersect_aggregate_with_edge): Likewise. (find_aggregate_values_for_callers_subset): Likewise. (cgraph_edge_brings_all_agg_vals_for_node): Likewise. (estimate_local_effects): Replace vec and vec with vec. (gather_context_independent_values): Likewise. (perform_estimation_of_a_value, decide_whether_version_node): Likewise. * ipa-fnsummary.c (evaluate_conditions_for_known_args): Replace vec with vec. (evaluate_properties_for_edge): Likewise. (estimate_edge_devirt_benefit): Likewise. (estimate_edge_size_and_time): Likewise. (estimate_calls_size_and_time): Likewise. (ipa_call_context::ipa_call_context): Likewise. (estimate_ipcp_clone_size_and_time): Likewise. * ipa-fnsummary.h (ipa_call_context): Replace vec with vec. * ipa-inline-analysis.c (do_estimate_edge_time): Replace vec with vec. (do_estimate_edge_size): Likewise. (do_estimate_edge_hints): Likewise. 2019-11-14 Feng Xue PR ipa/91682 * gcc.dg/ipa/ipcp-agg-10.c: Change dg-scan string. * gcc.dg/ipa/ipcp-agg-11.c: New test. Added: trunk/gcc/testsuite/gcc.dg/ipa/ipcp-agg-11.c Modified: trunk/gcc/ChangeLog trunk/gcc/ipa-cp.c trunk/gcc/ipa-fnsummary.c trunk/gcc/ipa-fnsummary.h trunk/gcc/ipa-inline-analysis.c trunk/gcc/ipa-prop.c trunk/gcc/ipa-prop.h trunk/gcc/testsuite/ChangeLog trunk/gcc/testsuite/gcc.dg/ipa/ipcp-agg-10.c
[Bug ipa/92528] [10 Regression] ICE in ipa_get_parm_lattices since r278219
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92528 --- Comment #7 from fxue at gcc dot gnu.org --- Author: fxue Date: Fri Nov 15 15:03:24 2019 New Revision: 278300 URL: https://gcc.gnu.org/viewcvs?rev=278300&root=gcc&view=rev Log: 2019-11-15 Feng Xue PR ipa/92528 * ipa-prop.c (update_jump_functions_after_inlining): Invalidate aggregate jump function when inlined-to caller has no edge summary. Modified: trunk/gcc/ChangeLog trunk/gcc/ipa-prop.c
[Bug ipa/92133] Support multi versioning on self recursive function
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92133 --- Comment #7 from fxue at gcc dot gnu.org --- Author: fxue Date: Mon Dec 2 06:37:30 2019 New Revision: 278893 URL: https://gcc.gnu.org/viewcvs?rev=278893&root=gcc&view=rev Log: Enable recursive function versioning 2019-12-02 Feng Xue PR ipa/92133 * doc/invoke.texi (ipa-cp-max-recursive-depth): Document new option. (ipa-cp-min-recursive-probability): Likewise. * params.opt (ipa-cp-max-recursive-depth): New. (ipa-cp-min-recursive-probability): Likewise. * ipa-cp.c (ipcp_lattice::add_value): Add two new parameters val_p and unlimited. (self_recursively_generated_p): New function. (get_val_across_arith_op): Likewise. (propagate_vals_across_arith_jfunc): Add constant propagation for self-recursive function. (incorporate_penalties): Do not penalize pure self-recursive function. (good_cloning_opportunity_p): Dump node_is_self_scc flag. (propagate_constants_topo): Set node_is_self_scc flag for cgraph node. (get_info_about_necessary_edges): Relax hotness check for edge to self-recursive function. * ipa-prop.h (ipa_node_params): Add new field node_is_self_scc. 2019-12-02 Feng Xue PR ipa/92133 * gcc.dg/ipa/ipa-clone-2.c: New test. Added: trunk/gcc/testsuite/gcc.dg/ipa/ipa-clone-2.c Modified: trunk/gcc/ChangeLog trunk/gcc/doc/invoke.texi trunk/gcc/ipa-cp.c trunk/gcc/ipa-prop.h trunk/gcc/params.opt trunk/gcc/testsuite/ChangeLog
[Bug ipa/92794] [10 Regression] ICE in decide_about_value, at ipa-cp.c:5186
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92794 --- Comment #3 from fxue at gcc dot gnu.org --- What's configure option for 32 be powerpc?
[Bug ipa/92971] New: Suspicious code in cgraph_edge_brings_all_agg_vals_for_node(), ipa-cp.c
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92971
Bug ID: 92971
Summary: Suspicious code in
cgraph_edge_brings_all_agg_vals_for_node(), ipa-cp.c
Product: gcc
Version: unknown
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: ipa
Assignee: unassigned at gcc dot gnu.org
Reporter: fxue at gcc dot gnu.org
CC: marxin at gcc dot gnu.org
Target Milestone: ---
The variable "values" is defined as static, which makes a questionable side
effect. History calls will impact result of current call!
for (i = 0; i < count; i++)
{
static vec values = vNULL;
class ipcp_param_lattices *plats;
[Bug ipa/92794] [10 Regression] ICE in decide_about_value, at ipa-cp.c:5186
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92794 --- Comment #5 from fxue at gcc dot gnu.org --- Author: fxue Date: Thu Dec 19 02:54:40 2019 New Revision: 279561 URL: https://gcc.gnu.org/viewcvs?rev=279561&root=gcc&view=rev Log: Handle aggregate pass-through for self-recursive call (PR ipa/92794) 2019-12-19 Feng Xue PR ipa/92794 * ipa-cp.c (self_recursive_agg_pass_through_p): New function. (intersect_with_plats): Use error_mark_node as place holder when aggregate jump function is simple pass-through for self-recursive call. (intersect_with_agg_replacements): Likewise. (intersect_aggregates_with_edge): Likewise. (find_aggregate_values_for_callers_subset): Likewise. 2019-12-19 Feng Xue PR ipa/92794 * gcc.dg/ipa/92794.c: New test. Added: trunk/gcc/testsuite/gcc.dg/ipa/pr92794.c Modified: trunk/gcc/ChangeLog trunk/gcc/ipa-cp.c trunk/gcc/testsuite/ChangeLog
[Bug ipa/92794] [10 Regression] ICE in decide_about_value, at ipa-cp.c:5186
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92794 --- Comment #7 from fxue at gcc dot gnu.org --- I tested it locally, passed. You can have a try to confirm that.
[Bug middle-end/93041] GCC 10 removes an infinite loop and causes a null pointer to dereferenced
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93041 fxue at gcc dot gnu.org changed: What|Removed |Added CC||fxue at gcc dot gnu.org --- Comment #3 from fxue at gcc dot gnu.org --- For gcc10, by default, a loop containing an exit, although it might not be a real exit, is considered as finite, which is controlled by a newly introduced option -finite-loops. To be compatible with old behavior, you can explicitly add -fno-finite-loops.
[Bug tree-optimization/93084] [10 regression] Infinite loop in ipa-cp when building clang with LTO+PGO
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93084 fxue at gcc dot gnu.org changed: What|Removed |Added CC||fxue at gcc dot gnu.org --- Comment #5 from fxue at gcc dot gnu.org --- I guess some situation is missed in self_recursively_generated_p(), this causes propagation along scc explode.
[Bug tree-optimization/93084] [10 regression] Infinite loop in ipa-cp when building clang with LTO+PGO
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93084 --- Comment #6 from fxue at gcc dot gnu.org --- Could you share how you build clang with PGO, and train workload?
[Bug tree-optimization/93084] [10 regression] Infinite loop in ipa-cp when building clang with LTO+PGO
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93084 --- Comment #8 from fxue at gcc dot gnu.org --- I checked llvm source code, the recursive call lies in Constant *llvm::ConstantFoldInsertValueInstruction(), but in dump, I noticed that call edge is _ZN4llvm34ConstantFoldInsertValueInstructionEPNS_8ConstantES1_NS_8ArrayRefIjEE/1977266 -> _ZN4llvm34ConstantFoldInsertValueInstructionEPNS_8ConstantES1_NS_8ArrayRefIjEE.localalias/13443902, which means caller and callee are different graph nodes? If so, self_recursively_generated_p () will return false, and cause infinite propagation, please refer to the code: if (!src->val || CS->CALLER != CS->CALLEE->FUNCTION_SYMBOL() () || src->val == val) return false; Another question is that this call edge is in scc, but becomes non-self recursive due to some transformations. And in propagate_vals_across_arith_jfunc(), we check a self-recursive call by comparing source ipcp_lattice with target ipcp_lattice, with the assumption that caller and callee of a self recursive call should have identical ipcp_lattices. if (SRC_LAT != DEST_LAT || param_ipa_cp_max_recursive_depth < 1) return dest_lat->set_contains_variable (); Then is there a case that a ipcp_lattice be shared by different cgraph nodes?
[Bug tree-optimization/93084] [10 regression] Infinite loop in ipa-cp when building clang with LTO+PGO
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93084 --- Comment #9 from fxue at gcc dot gnu.org --- xxx.localalias is gcc-generated as a noninterposable alias to xxx. But I guess target node returned by xxx.localalias->function_symbol() is not xxx. A simple thing we can do is to write a simple case to force generation of .localalias and test that.
[Bug tree-optimization/93084] [10 regression] Infinite loop in ipa-cp when building clang with LTO+PGO
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93084 --- Comment #12 from fxue at gcc dot gnu.org --- Identified the cause, it's my bug, will give a fix soon.
[Bug tree-optimization/93084] [10 regression] Infinite loop in ipa-cp when building clang with LTO+PGO
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93084 --- Comment #13 from fxue at gcc dot gnu.org --- Author: fxue Date: Wed Jan 8 02:55:00 2020 New Revision: 279987 URL: https://gcc.gnu.org/viewcvs?rev=279987&root=gcc&view=rev Log: Find matched aggregate lattice for self-recursive CP (PR ipa/93084) 2020-01-08 Feng Xue PR ipa/93084 * ipa-cp.c (self_recursively_generated_p): Find matched aggregate lattice for a value to check. (propagate_vals_across_arith_jfunc): Add an assertion to ensure finite propagation in self-recursive scc. 2020-01-08 Feng Xue PR ipa/93084 * gcc.dg/ipa/ipa-clone-3.c: New test. Added: trunk/gcc/testsuite/gcc.dg/ipa/ipa-clone-3.c Modified: trunk/gcc/ChangeLog trunk/gcc/ipa-cp.c trunk/gcc/testsuite/ChangeLog
[Bug lto/93166] [10 Regression] ICE in get_info_about_necessary_edges, at ipa-cp.c:4137 since r278893
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93166 --- Comment #3 from fxue at gcc dot gnu.org --- (In reply to Martin Jambor from comment #2) > I have analyzed this ICE and came to the conclusion that the assert is > wrong for polymorphic context lattices - e.g. in the reported case we > always pass the same class to first parameter, which in the recursive > call then serves as a basis to ancestor JF for the second parameter. > > When propagating polymorphic contexts within SCCs we allow creating > new values because for any sane input their number will be limited by > class hierarchy (and potential insane input limited by > param_ipa_cp_value_list_size). I think that is OK too. > > So we can either simply remove the assert or make it active only for > tree lattices. At the moment I'd prefer the former, but Feng, if this > assert proved valuable in development the recursive function > versioning patch, I'll be happy to keep it. What do you think? Sorry for late response due to trip. I'll check it.
[Bug lto/93166] [10 Regression] ICE in get_info_about_necessary_edges, at ipa-cp.c:4137 since r278893
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93166 --- Comment #4 from fxue at gcc dot gnu.org --- Created attachment 47676 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=47676&action=edit all-in-one test case
[Bug lto/93166] [10 Regression] ICE in get_info_about_necessary_edges, at ipa-cp.c:4137 since r278893
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93166 --- Comment #5 from fxue at gcc dot gnu.org --- To add a specialized class for tree lattices is a way, but will complicate the code. We can simple remove the assertion. And a one-file test case was made from original files.
[Bug lto/93166] [10 Regression] ICE in get_info_about_necessary_edges, at ipa-cp.c:4137 since r278893
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93166 --- Comment #7 from fxue at gcc dot gnu.org --- (In reply to Jeffrey A. Law from comment #6) > I've got state, so I'll take this as well. Here is a patch at https://gcc.gnu.org/ml/gcc-patches/2020-01/msg00068.html, waiting for review from someone.
[Bug ipa/93429] New: Missed IPA-CP on by-ref argument directly passed through from caller
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93429
Bug ID: 93429
Summary: Missed IPA-CP on by-ref argument directly passed
through from caller
Product: gcc
Version: unknown
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: ipa
Assignee: unassigned at gcc dot gnu.org
Reporter: fxue at gcc dot gnu.org
CC: marxin at gcc dot gnu.org
Target Milestone: ---
Here is a case,
f1 (int *p1)
{
... = *p1;
...
}
f2 (int *p2)
{
*p2 = 2;
f1 (p2);
}
It is easy to that in f1(), *p1 should be 2 after a simple propagation from
f2() to f1(). But current IPA never do that if argument to f1() is a simple
pass-through of a parameter of f2(), as p2 -> p1.
[Bug ipa/93707] New: ICE in perlbench from SPEC2017
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93707 Bug ID: 93707 Summary: ICE in perlbench from SPEC2017 Product: gcc Version: unknown Status: UNCONFIRMED Severity: normal Priority: P3 Component: ipa Assignee: unassigned at gcc dot gnu.org Reporter: fxue at gcc dot gnu.org CC: marxin at gcc dot gnu.org Target Milestone: --- Compiling options "-flto --param ipa-cp-eval-threshold=1" will trigger an ICE in the GCC and perlbench benchmark in Spec2017. during IPA pass: cp lto1: internal compiler error: in find_more_scalar_values_for_callers_subset, at ipa-cp.c:4709 0x1698187 find_more_scalar_values_for_callers_subset ../.././gcc/ipa-cp.c:4709 0x169f7d3 decide_about_value ../.././gcc/ipa-cp.c:5490 0x169fdc3 decide_whether_version_node ../.././gcc/ipa-cp.c:5537 0x169fdc3 ipcp_decision_stage ../.././gcc/ipa-cp.c:5718 0x169fdc3 ipcp_driver ../.././gcc/ipa-cp.c:5901
[Bug ipa/93707] ICE in perlbench from SPEC2017
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93707 fxue at gcc dot gnu.org changed: What|Removed |Added Status|UNCONFIRMED |NEW Last reconfirmed||2020-02-12 Assignee|unassigned at gcc dot gnu.org |fxue at gcc dot gnu.org Ever confirmed|0 |1
[Bug ipa/93707] ICE in perlbench from SPEC2017
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93707 --- Comment #1 from fxue at gcc dot gnu.org --- Created attachment 47825 --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=47825&action=edit A simplified case
[Bug ipa/93707] ICE in perlbench from SPEC2017
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93707 --- Comment #2 from fxue at gcc dot gnu.org --- For a self-recursive function, a for-all-contexts clone could generate an edge whose callee is not the function. Thus, to check whether an edge stands for a recursive call during cloning, we should not use ordinary way of comparing caller and callee of the edge.
[Bug ipa/93799] [10 Regression] ice in propagate_vals_across_arith_jfunc, at ipa-cp.c:2039
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93799 --- Comment #2 from fxue at gcc dot gnu.org --- Duplicate of pr93763
[Bug ipa/93823] [10 Regression] ICE: in find_more_scalar_values_for_callers_subset, at ipa-cp.c:4709 due to -fipa-cp
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93823 --- Comment #2 from fxue at gcc dot gnu.org --- Duplicate of pr93707
[Bug ipa/91089] IPA-cp does not setup proper cost model for switch default case in function versioning
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91089 --- Comment #3 from fxue at gcc dot gnu.org --- Author: fxue Date: Tue Sep 17 12:30:08 2019 New Revision: 275802 URL: https://gcc.gnu.org/viewcvs?rev=275802&root=gcc&view=rev Log: PR ipa/91089 - Setup predicate for switch default case in IPA 2019-09-17 Feng Xue PR ipa/91089 * doc/invoke.texi (ipa-max-switch-predicate-bounds): Document new option. * params.def (PARAM_IPA_MAX_SWITCH_PREDICATE_BOUNDS): New. * ipa-fnsummary.c (set_switch_stmt_execution_predicate): Add predicate for switch default case using range analysis information. 2019-09-17 Feng Xue PR ipa/91089 * gcc.dg/ipa/pr91089.c: New test. Added: trunk/gcc/testsuite/gcc.dg/ipa/pr91089.c Modified: trunk/gcc/ChangeLog trunk/gcc/doc/invoke.texi trunk/gcc/ipa-fnsummary.c trunk/gcc/params.def trunk/gcc/testsuite/ChangeLog
[Bug ipa/91088] IPA-cp cost evaluation is too conservative for "if (f(param) cmp const_val)" condition
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=91088 --- Comment #4 from fxue at gcc dot gnu.org --- Author: fxue Revision: 277054 Modified property: svn:log Modified: svn:log at Wed Oct 16 07:41:24 2019 -- --- svn:log (original) +++ svn:log Wed Oct 16 07:41:24 2019 @@ -1,1 +1,36 @@ -generalized IPA predicate on parameter +Generalized IPA predicate on parameter reference + +2019-10-16 Feng Xue + +PR ipa/91088 +* doc/invoke.texi (ipa-max-param-expr-ops): Document new option. +* params.def (PARAM_IPA_MAX_PARAM_EXPR_OPS): New. +* ipa-predicat.h (struct expr_eval_op): New struct. +(expr_eval_ops): New typedef. +(struct condition): Add type and param_ops fields, remove size field. +(add_condition): Replace size parameter with type parameter, add +param_ops parameter. +* ipa-predicat.c (expr_eval_ops_equal_p): New function. +(predicate::add_clause): Add comparisons on type and param_ops. +(dump_condition): Add debug dump for param_ops. +(remap_after_inlining): Adjust call arguments to add_condition. +(add_condition): Replace size parameter with type parameter, add +param_ops parameter. Unshare constant value used in conditions. +* ipa-fnsummary.c (evaluate_conditions_for_known_args): Fold +parameter expressions using param_ops. +(decompose_param_expr): New function. +(set_cond_stmt_execution_predicate): Use call to decompose_param_expr +to replace call to unmodified_parm_or_parm_agg_item. +(set_switch_stmt_execution_predicate): Likewise. +(will_be_nonconstant_expr_predicate): Likewise. Replace usage of size +with type. +(inline_read_section): Read param_ops from summary stream. +(ipa_fn_summary_write): Write param_ops to summary stream. + +2019-10-16 Feng Xue + +PR ipa/91088 +* gcc.dg/ipa/pr91088.c: New test. +* gcc.dg/ipa/pr91089.c: Add sub-test for range analysis. +* g++.dg/tree-ssa/ivopts-3.C: Force a function to be noinline. +
[Bug testsuite/92125] New test gcc.dg/ipa/pr91088.c introduced in r277054 fails
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=92125 --- Comment #2 from fxue at gcc dot gnu.org --- Author: fxue Date: Thu Oct 17 09:55:37 2019 New Revision: 277095 URL: https://gcc.gnu.org/viewcvs?rev=277095&root=gcc&view=rev Log: PR testsuite/92125 2019-10-17 Feng Xue PR testsuite/92125 * gcc.dg/ipa/pr91088.c: Change char conversion to bitand. Modified: trunk/gcc/testsuite/ChangeLog trunk/gcc/testsuite/gcc.dg/ipa/pr91088.c
[Bug tree-optimization/89713] Optimize away an empty loop whose finiteness can not be analytically determined
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=89713 --- Comment #5 from fxue at gcc dot gnu.org --- Author: fxue Date: Thu Jun 13 04:17:42 2019 New Revision: 272234 URL: https://gcc.gnu.org/viewcvs?rev=272234&root=gcc&view=rev Log: PR tree-optimization/89713 - Assume loop with an exit is finite gcc/ChangeLog: * doc/invoke.texi (-ffinite-loops): Document new option. * common.opt (-ffinite-loops): New option. * tree-ssa-dce.c (mark_stmt_if_obviously_necessary): Mark IFN_GOACC_LOOP calls as necessary. * tree-ssa-loop-niter.c (finite_loop_p): Assume loop with an exit is finite. * omp-offload.c (oacc_xform_loop): Skip lowering if return value of IFN_GOACC_LOOP call is not used. * opts.c (default_options_table): Enable -ffinite-loops at -O2+. gcc/testsuite/ChangeLog: * g++.dg/tree-ssa/empty-loop.C: New test. * gcc.dg/tree-ssa/dce-2.c: New test. * gcc.dg/const-1.c: Add -fno-finite-loops option. * gcc.dg/graphite/graphite.exp: Likewise. * gcc.dg/loop-unswitch-1.c: Likewise. * gcc.dg/predict-9.c: Likewise. * gcc.dg/pure-2.c: Likewise. * gcc.dg/tree-ssa/20040211-1.c: Likewise. * gcc.dg/tree-ssa/loop-10.c: Likewise. * gcc.dg/tree-ssa/split-path-6.c: Likewise. * gcc.dg/tree-ssa/ssa-thread-12.c: Likewise. libgomp/ChangeLog: * testsuite/libgomp.oacc-c-c++-common/pr84955-1.c: New test. Added: trunk/gcc/testsuite/g++.dg/tree-ssa/empty-loop.C trunk/gcc/testsuite/gcc.dg/tree-ssa/dce-2.c trunk/libgomp/testsuite/libgomp.oacc-c-c++-common/pr84955-1.c Modified: trunk/gcc/ChangeLog trunk/gcc/common.opt trunk/gcc/doc/invoke.texi trunk/gcc/omp-offload.c trunk/gcc/opts.c trunk/gcc/testsuite/ChangeLog trunk/gcc/testsuite/gcc.dg/const-1.c trunk/gcc/testsuite/gcc.dg/graphite/graphite.exp trunk/gcc/testsuite/gcc.dg/loop-unswitch-1.c trunk/gcc/testsuite/gcc.dg/predict-9.c trunk/gcc/testsuite/gcc.dg/pure-2.c trunk/gcc/testsuite/gcc.dg/tree-ssa/20040211-1.c trunk/gcc/testsuite/gcc.dg/tree-ssa/loop-10.c trunk/gcc/testsuite/gcc.dg/tree-ssa/split-path-6.c trunk/gcc/testsuite/gcc.dg/tree-ssa/ssa-thread-12.c trunk/gcc/tree-ssa-dce.c trunk/gcc/tree-ssa-loop-niter.c trunk/libgomp/ChangeLog
[Bug ipa/90401] Missed propagation of by-ref constant argument to callee function
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=90401 --- Comment #3 from fxue at gcc dot gnu.org --- Author: fxue Date: Fri Jun 14 02:34:48 2019 New Revision: 272282 URL: https://gcc.gnu.org/viewcvs?rev=272282&root=gcc&view=rev Log: PR ipa/90401 gcc/ChangeLog: * ipa-prop.c (add_to_agg_contents_list): New function. (clobber_by_agg_contents_list_p): Likewise. (extract_mem_content): Likewise. (get_place_in_agg_contents_list): Delete. (determine_known_aggregate_parts): Renamed from determine_locally_known_aggregate_parts. New parameter aa_walk_budget_p. gcc/testsuite/ChangeLog: * gcc.dg/ipa/ipcp-agg-10.c: New test. Added: trunk/gcc/testsuite/gcc.dg/ipa/ipcp-agg-10.c Modified: trunk/gcc/ChangeLog trunk/gcc/ipa-prop.c trunk/gcc/testsuite/ChangeLog
