[Bug jit/63854] New: Fix memory leaks seen in JIT

[Bug jit/63854] New: Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

            Bug ID: 63854
           Summary: Fix memory leaks seen in JIT
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: jit
          Assignee: dmalcolm at gcc dot gnu.org
          Reporter: dmalcolm at gcc dot gnu.org

Created attachment 33959
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=33959&action=edit
Log from valgrind

I'm attaching a log from valgrind on test-factorial.c, running 5 in-process
iterations.

PATH=.:$PATH \
LD_LIBRARY_PATH=. \
LIBRARY_PATH=. \
  valgrind \
    --leak-check=full \
    testsuite/jit/test-factorial.exe 2>&1 | tee test-factorial.valgrind.log

Summary is:
==57820== LEAK SUMMARY:
==57820==    definitely lost: 84,140 bytes in 578 blocks
==57820==    indirectly lost: 257,864 bytes in 3,185 blocks
==57820==      possibly lost: 0 bytes in 0 blocks
==57820==    still reachable: 572,159 bytes in 3,112 blocks
==57820==         suppressed: 0 bytes in 0 blocks

Various things show up as leaking 4 or 5 times; presumably those are
per-invocation leaks that we ought to fix to avoid them affecting jit users.

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #1 from dmalcolm at gcc dot gnu.org ---
Note to self: this was with r217427.

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

dmalcolm at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2014-11-17
     Ever confirmed|0                           |1

--- Comment #2 from dmalcolm at gcc dot gnu.org ---
FWIW am actively working on this; in my current working copy I have the leaks
down to:
   definitely lost: 1,916 bytes in 80 blocks
   indirectly lost: 352 bytes in 4 blocks

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #3 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 19:08:29 2014
New Revision: 217791

URL: https://gcc.gnu.org/viewcvs?rev=217791&root=gcc&view=rev
Log:
PR jit/63854: Fix memory leak within gcc_options

gcc/ChangeLog:
    PR jit/63854
    * opts.c (finalize_options_struct): New.
    * opts.h (finalize_options_struct): New.
    * toplev.c (toplev::finalize): Call finalize_options_struct
    on global_options and global_options_set.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/opts.c
    trunk/gcc/opts.h
    trunk/gcc/toplev.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #4 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 19:11:37 2014
New Revision: 217792

URL: https://gcc.gnu.org/viewcvs?rev=217792&root=gcc&view=rev
Log:
PR jit/63854: Fix memory leak of reginfo.c: valid_mode_changes_obstack

gcc/ChangeLog:
    PR jit/63854
    * reginfo.c (finish_subregs_of_mode): Replace obstack_finish with
    obstack_free when cleaning up valid_mode_changes_obstack.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/reginfo.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #5 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 19:23:06 2014
New Revision: 217793

URL: https://gcc.gnu.org/viewcvs?rev=217793&root=gcc&view=rev
Log:
PR jit/63854: Fix memory leaks within context/pass_manager/dump_manager

gcc/ChangeLog:
    PR jit/63854
    * config/alpha/alpha.c (alpha_option_override): Remove static from
    "handle_trap_shadows_info" and "align_insns_info".
    * config/i386/i386.c (ix86_option_override): Likewise for
    "insert_vzeroupper_info".
    * config/rl78/rl78.c (rl78_asm_file_start): Likewise for
    "rl78_devirt_info" and "rl78_move_elim_info".
    * config/rs6000/rs6000.c (rs6000_option_override): Likewise for
    "analyze_swaps_info".
    * context.c (gcc::context::~context): New.
    * context.h (gcc::context::~context): New.
    * dumpfile.c (dump_files): Add "false" initializers for new field
    "owns_strings".
    (gcc::dump_manager::~dump_manager): New.
    (gcc::dump_manager::dump_register): Add param "take_ownership".
    * dumpfile.h (struct dump_file_info): Add field "owns_strings".
    (gcc::dump_manager::~dump_manager): New.
    (gcc::dump_manager::dump_register): Add param "take_ownership".
    * pass_manager.h (gcc::pass_manager::operator delete): New.
    (gcc::pass_manager::~pass_manager): New.
    * passes.c (pass_manager::register_one_dump_file): Pass "true" to
    new "owns_strings" argument to dump_register.
    (pass_manager::operator delete): New.
    (delete_pass_tree): New function.
    (pass_manager::~pass_manager): New.
    * statistics.c (statistics_early_init): Pass "false" to
    new "owns_strings" argument to dump_register.
    * toplev.c (toplev::finalize): Clean up the context and thus the
    things it owns.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/alpha/alpha.c
    trunk/gcc/config/i386/i386.c
    trunk/gcc/config/rl78/rl78.c
    trunk/gcc/config/rs6000/rs6000.c
    trunk/gcc/context.c
    trunk/gcc/context.h
    trunk/gcc/dumpfile.c
    trunk/gcc/dumpfile.h
    trunk/gcc/pass_manager.h
    trunk/gcc/passes.c
    trunk/gcc/statistics.c
    trunk/gcc/toplev.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #7 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 19:33:49 2014
New Revision: 217795

URL: https://gcc.gnu.org/viewcvs?rev=217795&root=gcc&view=rev
Log:
PR jit/63854: Fix memory leak of save_decoded_options

gcc/ChangeLog:
    PR jit/63854
    * toplev.c (toplev::finalize): Clean up save_decoded_options.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/toplev.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #8 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 19:43:20 2014
New Revision: 217796

URL: https://gcc.gnu.org/viewcvs?rev=217796&root=gcc&view=rev
Log:
PR jit/63854: Fix leak of opts_obstack

gcc/ChangeLog:
    PR jit/63854
    * toplev.c (toplev::finalize): Free opts_obstack.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/toplev.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #9 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 19:45:39 2014
New Revision: 217797

URL: https://gcc.gnu.org/viewcvs?rev=217797&root=gcc&view=rev
Log:
PR jit/63854: Fix leak of optimization_summary_obstack

gcc/ChangeLog:
    PR jit/63854
    * ipa-reference.c (ipa_reference_c_finalize): Release
    optimization_summary_obstack.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ipa-reference.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #10 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 19:49:28 2014
New Revision: 217800

URL: https://gcc.gnu.org/viewcvs?rev=217800&root=gcc&view=rev
Log:
PR jit/63854: Add ira_costs_c_finalize

gcc/ChangeLog:
    PR jit/63854
    * ira-costs.c (ira_costs_c_finalize): New function.
    * ira.h (ira_costs_c_finalize): New prototype.
    * toplev.c (toplev::finalize): Call ira_costs_c_finalize.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ira-costs.c
    trunk/gcc/ira.h
    trunk/gcc/toplev.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #11 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 19:57:09 2014
New Revision: 217801

URL: https://gcc.gnu.org/viewcvs?rev=217801&root=gcc&view=rev
Log:
PR jit/63854: Don't leak producer_string in dwarf2out.c

gcc/ChangeLog:
    PR jit/63854
    * dwarf2out.c (dwarf2out_c_finalize): Free producer_string.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/dwarf2out.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #12 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 20:10:00 2014
New Revision: 217802

URL: https://gcc.gnu.org/viewcvs?rev=217802&root=gcc&view=rev
Log:
PR jit/63854: Fix leak of worklist within jit-recording.c

gcc/jit/ChangeLog:
    PR jit/63854
    * jit-recording.c (recording::function::validate): Convert
    "worklist" from vec<> to autovec<> to fix a leak.


Modified:
    trunk/gcc/jit/ChangeLog
    trunk/gcc/jit/jit-recording.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #13 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 20:14:18 2014
New Revision: 217803

URL: https://gcc.gnu.org/viewcvs?rev=217803&root=gcc&view=rev
Log:
PR jit/63854: Fix leak of "avail" within tree-ssa-pre.c

gcc/ChangeLog:
    PR jit/63854
    * tree-ssa-pre.c (do_regular_insertion): Convert "avail" from
    vec<> to auto_vec<> to fix a leak.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/tree-ssa-pre.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #14 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 20:17:44 2014
New Revision: 217804

URL: https://gcc.gnu.org/viewcvs?rev=217804&root=gcc&view=rev
Log:
PR jit/63854: Fix leak of paths within jump threading

gcc/ChangeLog:
    PR jit/63854
    * tree-ssa-threadedge.c (thread_across_edge): Don't just release
    "path", delete it.
    * tree-ssa-threadupdate.c (delete_jump_thread_path): Likewise.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/tree-ssa-threadedge.c
    trunk/gcc/tree-ssa-threadupdate.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #15 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 20:21:19 2014
New Revision: 217806

URL: https://gcc.gnu.org/viewcvs?rev=217806&root=gcc&view=rev
Log:
PR jit/63854: lra.c: Fix leak of point_freq_vec's buffer when calling
lra_inheritance

gcc/ChangeLog:
    PR jit/63854
    * lra.c (lra): After creating live ranges in preparation for call
    to lra_inheritance, set live_p to true.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/lra.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #16 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 20:26:57 2014
New Revision: 217807

URL: https://gcc.gnu.org/viewcvs?rev=217807&root=gcc&view=rev
Log:
PR jit/63854: Add all_late_ipa_passes to GCC_PASS_LISTS

gcc/ChangeLog:
    PR jit/63854
    * pass_manager.h (GCC_PASS_LISTS): Add all_late_ipa_passes.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/pass_manager.h

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #17 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 20:31:52 2014
New Revision: 217808

URL: https://gcc.gnu.org/viewcvs?rev=217808&root=gcc&view=rev
Log:
PR jit/63854: Fix leaking vec in jit

gcc/jit/ChangeLog:
    PR jit/63854
    * jit-playback.c (gcc::jit::playback::compound_type::set_fields):
    Convert param from const vec<playback::field *> & to
    const auto_vec<playback::field *> *.
    (gcc::jit::playback::context::new_function_type): Convert param
    "param_types" from vec<type *> * to const auto_vec<type *> *.
    (gcc::jit::playback::context::new_function): Convert param
    "params" from vec<param *> * to const auto_vec<param *> *.
    (gcc::jit::playback::context::build_call): Convert param "args"
    from vec<rvalue *> to const auto_vec<rvalue *> *.
    (gcc::jit::playback::context::new_call): Likewise.
    (gcc::jit::playback::context::new_call_through_ptr): Likewise.
    (wrapper_finalizer): New function.
    (gcc::jit::playback::wrapper::operator new): Call the finalizer
    variant of ggc_internal_cleared_alloc, supplying
    wrapper_finalizer.
    (gcc::jit::playback::function::finalizer): New.
    (gcc::jit::playback::block::finalizer): New.
    (gcc::jit::playback::source_file::finalizer): New.
    (gcc::jit::playback::source_line::finalizer): New.

    * jit-playback.h
    (gcc::jit::playback::context::new_function_type): Convert param
    "param_types" from vec<type *> * to const auto_vec<type *> *.
    (gcc::jit::playback::context::new_function): Convert param
    "params" from vec<param *> * to const auto_vec<param *> *.
    (gcc::jit::playback::context::new_call): Convert param
    "args" from vec<rvalue *> to const auto_vec<rvalue *> *.
    (gcc::jit::playback::context::new_call_through_ptr): Likewise.
    (gcc::jit::playback::context::build_call): Likewise.
    (gcc::jit::playback::context): Convert fields "m_functions",
    "m_source_files", "m_cached_locations" from vec to auto_vec.
    (gcc::jit::playback::wrapper::finalizer): New virtual function.
    (gcc::jit::playback::compound_type::set_fields): Convert param fro
    const vec<playback::field *> & to
    const auto_vec<playback::field *> *.
    (gcc::jit::playback::function::finalizer): New.
    (gcc::jit::playback::block::finalizer): New.
    (gcc::jit::playback::source_file::finalizer): New.
    (gcc::jit::playback::source_line::finalizer): New.

    * jit-recording.c
    (gcc::jit::recording::function_type::replay_into): Convert local
    from a vec into an auto_vec.
    (gcc::jit::recording::fields::replay_into): Likewise.
    (gcc::jit::recording::function::replay_into): Likewise.
    (gcc::jit::recording::call::replay_into): Likewise.
    (gcc::jit::recording::call_through_ptr::replay_into): Likewise.

    * jit-recording.h (gcc::jit::recording::context): Convert fields
    "m_mementos", "m_compound_types", "m_functions" from vec<> to
    auto_vec <>.
    (gcc::jit::recording::function_type::get_param_types): Convert
    return type from vec<type *> to const vec<type *> &.
    (gcc::jit::recording::function_type): Convert field
    "m_param_types" from a vec<> to an auto_vec<>.
    (gcc::jit::recording::fields): Likewise for field "m_fields".
    (gcc::jit::recording::function::get_params): Convert return type
    from vec <param *> to const vec<param *> &.
    (gcc::jit::recording::function): Convert fields "m_params",
    "m_locals", "m_blocks" from vec<> to auto_vec<>.
    (gcc::jit::recording::block): Likewise for field "m_statements".
    vec<> to auto_vec<>.
    (gcc::jit::recording::call): Likewise for field "m_args".
    (gcc::jit::recording::call_through_ptr): Likewise.


Modified:
    trunk/gcc/jit/ChangeLog
    trunk/gcc/jit/jit-playback.c
    trunk/gcc/jit/jit-playback.h
    trunk/gcc/jit/jit-recording.c
    trunk/gcc/jit/jit-recording.h

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #18 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 20:42:29 2014
New Revision: 217811

URL: https://gcc.gnu.org/viewcvs?rev=217811&root=gcc&view=rev
Log:
PR jit/63854: Fix leak in ipa-icf.c

gcc/ChangeLog:
    PR jit/63854
    * ipa-icf.c (sem_item_optimizer::~sem_item_optimizer): Free each
    congruence_class_group *.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ipa-icf.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #19 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Wed Nov 19 20:45:03 2014
New Revision: 217812

URL: https://gcc.gnu.org/viewcvs?rev=217812&root=gcc&view=rev
Log:
PR jit/63854: Fix leaks in test-fuzzer.c

gcc/testsuite/ChangeLog:
    PR jit/63854
    * jit.dg/test-fuzzer.c (fuzzer_init): Free malloced buffers.
    (make_random_function): Free ff->locals.


Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/jit.dg/test-fuzzer.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #20 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Mon Dec  1 16:13:29 2014
New Revision: 218227

URL: https://gcc.gnu.org/viewcvs?rev=218227&root=gcc&view=rev
Log:
PR jit/63854: Add support for running "make check-jit" under valgrind

gcc/testsuite/ChangeLog:
    PR jit/63854
    * jit.dg/jit.exp (report_leak): New.
    (parse_valgrind_logfile): New.
    (fixed_host_execute): Detect if RUN_UNDER_VALGRIND is present
    in the environment, and if so, run the executable under
    valgrind, capturing valgrind's output to a logfile.  Parse the
    log file, generating PASSes and XFAILs for the summary of leaks.
    Use "wait" before "close": valgrind might not have finished
    writing the log out before we parse it, so we need to wait for
    the spawnee to finish.


Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/jit.dg/jit.exp

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #21 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Mon Dec  1 17:16:04 2014
New Revision: 218230

URL: https://gcc.gnu.org/viewcvs?rev=218230&root=gcc&view=rev
Log:
PR jit/63854: Fix leak in tree-ssa-math-opts.c

gcc/ChangeLog:
    PR jit/63854
    * tree-ssa-math-opts.c (execute_cse_sincos_1): Fix a missing
    release of stmts by converting it to an auto_vec.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/tree-ssa-math-opts.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #23 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Mon Dec  1 17:28:28 2014
New Revision: 218233

URL: https://gcc.gnu.org/viewcvs?rev=218233&root=gcc&view=rev
Log:
PR jit/63854: Fix leak in real.c for i386:init_ext_80387_constants

gcc/ChangeLog:
    PR jit/63854
    * real.c (real_from_string): Add missing mpfr_clear.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/real.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #24 from dmalcolm at gcc dot gnu.org ---
Author: dmalcolm
Date: Mon Dec  1 17:40:03 2014
New Revision: 218234

URL: https://gcc.gnu.org/viewcvs?rev=218234&root=gcc&view=rev
Log:
PR jit/63854: Fix leaks in toyvm.c

gcc/jit/ChangeLog:
    PR jit/63854
    * docs/examples/tut04-toyvm/toyvm.c
    (toyvm_compiled_function): New typedef.
    (toyvm_compiled_func) Rename to...
    (toyvm_compiled_code) ...this.
    (struct toyvm_compiled_function): New struct.
    (toyvm_function_compile): Return a toyvm_compiled_function *
    rather than a toyvm_compiled_func, so that the caller can fully
    clean things up.  Free "funcname".
    (test_script): Update for change to toyvm_function_compile.
    Clean up the toyvm_compiled_function.
    (main): Likewise.
    (docs/intro/tutorial04.rst): Update to reflect the above changes,
    and to better spell out the lifetime of the compiled code.


Modified:
    trunk/gcc/jit/ChangeLog
    trunk/gcc/jit/docs/examples/tut04-toyvm/toyvm.c
    trunk/gcc/jit/docs/intro/tutorial04.rst

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #25 from David Malcolm <dmalcolm at gcc dot gnu.org> ---
Author: dmalcolm
Date: Thu Dec  4 21:11:04 2014
New Revision: 218403

URL: https://gcc.gnu.org/viewcvs?rev=218403&root=gcc&view=rev
Log:
PR jit/63854: Fix leak of ipa hooks

gcc/ChangeLog:
    PR jit/63854
    * ipa-prop.c (ipa_register_cgraph_hooks): Guard insertion of
    ipa_add_new_function on function_insertion_hook_holder being
    non-NULL.
    * ipa-reference.c (ipa_reference_c_finalize): Remove
    node_removal_hook_holder and node_duplication_hook_holder if
    they've been added to symtab.
    * toplev.c (toplev::finalize): Call ipa_reference_c_finalize
    before cgraph_c_finalize so that the former can access "symtab".


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ipa-prop.c
    trunk/gcc/ipa-reference.c
    trunk/gcc/toplev.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #26 from David Malcolm <dmalcolm at gcc dot gnu.org> ---
Author: dmalcolm
Date: Thu Dec  4 21:23:57 2014
New Revision: 218404

URL: https://gcc.gnu.org/viewcvs?rev=218404&root=gcc&view=rev
Log:
PR jit/63854: Fix double-initialization within tree-pretty-print.c

gcc/ChangeLog:
    PR jit/63854
    * tree-pretty-print.c: Eliminate include of <new>.
    (buffer): Convert this variable from a pretty_printer to a
    pretty_printer *.
    (initialized): Eliminate this variable in favor of the NULL-ness
    of "buffer".
    (print_generic_decl): Update for "buffer" becoming a pointer.
    (print_generic_stmt): Likewise.
    (print_generic_stmt_indented): Likewise.
    (print_generic_expr): Likewise.
    (maybe_init_pretty_print): Likewise, allocating "buffer" on the
    heap and using its non-NULL-ness to ensure idempotency.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/tree-pretty-print.c

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #28 from David Malcolm <dmalcolm at gcc dot gnu.org> ---
Author: dmalcolm
Date: Tue Dec  9 20:46:33 2014
New Revision: 218538

URL: https://gcc.gnu.org/viewcvs?rev=218538&root=gcc&view=rev
Log:
PR jit/63854: Document how to run the jit testsuite under valgrind

gcc/jit/ChangeLog:
    PR jit/63854
    * docs/internals/index.rst (Running under valgrind): New
    subsection.
    (docs/_build/texinfo/libgccjit.texi): Regenerate.

Modified:
    trunk/gcc/jit/ChangeLog
    trunk/gcc/jit/docs/_build/texinfo/libgccjit.texi
    trunk/gcc/jit/docs/internals/index.rst

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

David Malcolm <dmalcolm at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |major

[Bug jit/63854] Fix memory leaks seen in JIT

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

David Binderman <dcb314 at hotmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |dcb314 at hotmail dot com

--- Comment #29 from David Binderman <dcb314 at hotmail dot com> ---
Similar thing in the same area, caught using static analyser cppcheck:

[gcc/jit/jit-playback.c:1791]: (error) Resource leak: f_in

Source code is

  if (!feof (f_in))
    {
      add_error (NULL, "error reading from %s", path);
      free (result);
      return NULL;
    }

  fclose (f_in);

It looks to me like a call to fclose needs to happen just before
the return.

[Bug jit/63854] Fix memory leaks seen in JIT

[dmalcolm at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #30 from David Malcolm <dmalcolm at gcc dot gnu.org> ---
(In reply to David Binderman from comment #29)
> Similar thing in the same area, caught using static analyser cppcheck:
> 
> [gcc/jit/jit-playback.c:1791]: (error) Resource leak: f_in
> 
> Source code is
> 
>   if (!feof (f_in))
>     {
>       add_error (NULL, "error reading from %s", path);
>       free (result);
>       return NULL;
>     }
> 
>   fclose (f_in);
> 
> It looks to me like a call to fclose needs to happen just before
> the return.

Thanks.  I've fixed this in r219487.

[Bug jit/63854] Fix memory leaks seen in JIT

[ienkovich at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854
Bug 63854 depends on bug 64003, which changed state.

Bug 64003 Summary: valgrind complains about get_attr_length_nobnd in insn-attrtab.c from i386.md
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64003

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |FIXED

[Bug jit/63854] Fix memory leaks seen in JIT

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #33 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Alex Coplan <acoplan@gcc.gnu.org>:

https://gcc.gnu.org/g:b46584d7836bee011facdf946fd7241c748d66f0

commit r11-2799-gb46584d7836bee011facdf946fd7241c748d66f0
Author: Alex Coplan <alex.coplan@arm.com>
Date:   Fri Aug 21 14:26:11 2020 +0100

    driver: Fix several memory leaks [PR63854]

    This patch fixes several memory leaks in the driver, all of which relate
    to the handling of static specs. We introduce functions
    set_static_spec_{shared,owned}() which are used to enforce proper memory
    management when updating the strings in the static_specs table.

    This is achieved by making use of the alloc_p field in the table
    entries. Similarly to set_spec(), each time we update an entry, we check
    whether alloc_p is set, and free the old value if so. We then set
    alloc_p correctly based on whether we "own" this memory or whether we're
    just taking a pointer to a shared string which we shouldn't free.

    The following table shows the number of leaks found by AddressSanitizer
    when running a minimal libgccjit program on AArch64. The test program
    does the whole libgccjit compilation cycle in a loop (including acquiring
    and releasing the context), and the table below shows the number of leaks
    for different iterations of that loop.

    +--------------+-----+-----+------+---------------+
    | # of runs >  | 1   | 2   | 3    | Leaks per run |
    +--------------+-----+-----+------+---------------+
    | Before patch | 463 | 940 | 1417 | 477           |
    +--------------+-----+-----+------+---------------+
    | After patch  | 416 | 846 | 1276 | 430           |
    +--------------+-----+-----+------+---------------+

    gcc/ChangeLog:

            PR jit/63854
            * gcc.c (set_static_spec): New.
            (set_static_spec_owned): New.
            (set_static_spec_shared): New.
            (driver::maybe_putenv_COLLECT_LTO_WRAPPER): Use
            set_static_spec_owned() to take ownership of lto_wrapper_file
            such that it gets freed in driver::finalize.
            (driver::maybe_run_linker): Use set_static_spec_shared() to
            ensure that we don't try and free() the static string "ld",
            also ensuring that any previously-allocated string in
            linker_name_spec is freed. Likewise with argv0.
            (driver::finalize): Use set_static_spec_shared() when resetting
            specs that previously had allocated strings; remove if(0)
            around call to free().

[Bug jit/63854] Fix memory leaks seen in JIT

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=63854

--- Comment #34 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jeff Law <law@gcc.gnu.org>:

https://gcc.gnu.org/g:bc84b61b00792ef38e1c2cd6b7280087e4fd2b57

commit r12-7723-gbc84b61b00792ef38e1c2cd6b7280087e4fd2b57
Author: Marc Nieper-Wißkirchen <marc@nieper-wisskirchen.de>
Date:   Sat Mar 19 13:42:26 2022 -0400

    [PATCH] gcc: pass-manager: Fix memory leak. [PR jit/63854]

    Before the patch, compiling the hello world example of libgccjit with
    the external driver under Valgrind shows a loss of 12,611 (48 direct)
    bytes.  After the patch, no memory leaks are reported anymore.
    (Memory leaks occurring when using the internal driver are mostly in
    the driver code in gcc/gcc.c and have to be fixed separately.)

    The patch has been tested by fully bootstrapping the compiler with the
    frontends C, C++, Fortran, LTO, ObjC, JIT and running the test suite
    under a x86_64-pc-linux-gnu host.

    gcc/ChangeLog:

            PR jit/63854
            * hash-traits.h (struct typed_const_free_remove): New.
            (struct free_string_hash): New.
            * pass_manager.h: Use free_string_hash.
            * passes.cc (pass_manager::register_pass_name): Use
free_string_hash.
            (pass_manager::~pass_manager): Delete allocated m_name_to_pass_map.