janus-valgrind.supp

# annoying glib2 object type registration stuff
{
	glib2-type-register-fundamental
	Memcheck:Leak
	fun:calloc
	fun:g_malloc0
	...
	fun:g_type_register_fundamental
	...
}
{
	glib2-type-register-fundamental-realloc
	Memcheck:Leak
	...
	fun:g_realloc
	...
	fun:g_type_register_fundamental
	...
}
{
	glib2-type-register-static
	Memcheck:Leak
	fun:malloc
	fun:g_malloc
	...
	fun:type_node_any_new_W
	fun:g_type_register_static
	...
}
{
	glib2-type-register-static-realloc
	Memcheck:Leak
	...
	fun:g_realloc
	...
	fun:type_node_any_new_W
	fun:g_type_register_static
	...
}
{
	glib2-type-add-interface-static
	Memcheck:Leak
	fun:malloc
	fun:g_malloc
	...
	fun:type_add_interface_Wm
	fun:g_type_add_interface_static
	...
}
{
	glib2-dl-init-gobject-type-data
	Memcheck:Leak
	fun:calloc
	fun:g_malloc0
	fun:type_data_make_W
	fun:gobject_init_ctor
	...
	fun:_dl_init
	...
}
{
	glib2-dl-init-gobject-type-node
	Memcheck:Leak
	fun:calloc
	fun:g_malloc0
	...
	fun:type_node_fundamental_new_W
	fun:gobject_init_ctor
	...
	fun:_dl_init
	...
}
{
	glib2-signal-new-class
	Memcheck:Leak
	fun:calloc
	fun:g_malloc0
	...
	fun:g_signal_type_cclosure_new
	fun:g_signal_new
	...
	fun:g_object_do_class_init
	...
}
{
	glib2-signal-new-class-intern
	Memcheck:Leak
	fun:calloc
	fun:g_malloc0
	...
	fun:g_signal_type_cclosure_new
	fun:g_signal_new
	...
	fun:g_cancellable_class_intern_init
	...
}
{
	nice-socket-new-vtable-base
	Memcheck:Leak
	fun:malloc
	fun:g_malloc
	...
	fun:type_iface_vtable_base_init_Wm
	...
	fun:nice_udp_bsd_socket_new
	...
}
{
	nice-socket-new-class
	Memcheck:Leak
	fun:calloc
	fun:g_malloc0
	...
	fun:g_socket_class_intern_init
	...
	fun:nice_udp_bsd_socket_new
	...
}

{
	nice-agent-class-init
	Memcheck:Leak
	fun:calloc
	fun:g_malloc0
	...
	fun:g_type_create_instance
	...
	fun:nice_agent_class_init
	...
}

# probably a bug in sofia-sip
{
	sofia-sip-su-home
	Memcheck:Leak
	fun:calloc
	...
	fun:su_home_new
	...
	fun:main
}

# GLib Valgrind suppressions file
#
# This provides a list of suppressions for all of GLib (including GIO), for all
# Valgrind tools (memcheck, drd, helgrind, etc.) for the false positives and
# deliberate one-time leaks which GLib causes to be reported when running under
# Valgrind.
#
# When running an application which links to GLib under Valgrind, you can pass
# this suppression file to Valgrind using --suppressions=/path/to/glib-2.0.supp.
#
# http://valgrind.org/docs/manual/manual-core.html#manual-core.suppress
#
# Note that there is currently no way for Valgrind to load this automatically
# (https://bugs.kde.org/show_bug.cgi?id=160905), so the best GLib can currently
# do is to install this file as part of its development package.
#
# This file should be updated if GLib introduces a new deliberate one-time leak,
# or another false race positive in Valgrind: please file bugs at:
#
# https://bugzilla.gnome.org/enter_bug.cgi?product=glib

{
	gnutls-init-calloc
	Memcheck:Leak
	fun:calloc
	...
	fun:gtls_gnutls_init
}

{
	gnutls-init-realloc
	Memcheck:Leak
	fun:realloc
	...
	fun:gtls_gnutls_init
}

{
	g-tls-backend-gnutls-init
	Memcheck:Leak
	fun:g_once_impl
	fun:g_tls_backend_gnutls_init
}

{
	p11-tokens-init
	Memcheck:Leak
	fun:calloc
	...
	fun:create_tokens_inlock
	fun:initialize_module_inlock_reentrant
}

{
	gobject-init-malloc
	Memcheck:Leak
	fun:malloc
	...
	fun:gobject_init_ctor
}

{
	gobject-init-realloc
	Memcheck:Leak
	fun:realloc
	...
	fun:gobject_init_ctor
}

{
	gobject-init-calloc
	Memcheck:Leak
	fun:calloc
	...
	fun:gobject_init_ctor
}

{
	g-type-register-dynamic
	Memcheck:Leak
	fun:malloc
	...
	fun:g_type_register_dynamic
}

{
	g-type-register-static
	Memcheck:Leak
	fun:malloc
	...
	fun:g_type_register_static
}

{
	g-type-register-static-realloc
	Memcheck:Leak
	fun:realloc
	...
	fun:g_type_register_static
}

{
	g-type-register-static-calloc
	Memcheck:Leak
	fun:calloc
	...
	fun:g_type_register_static
}

{
	g-type-add-interface-dynamic
	Memcheck:Leak
	fun:malloc
	...
	fun:g_type_add_interface_dynamic
}

{
	g-type-add-interface-static
	Memcheck:Leak
	fun:malloc
	...
	fun:g_type_add_interface_static
}

{
	g-test-rand-init
	Memcheck:Leak
	fun:calloc
	...
	fun:g_rand_new_with_seed_array
	fun:test_run_seed
	...
	fun:g_test_run
}

{
	g-test-rand-init2
	Memcheck:Leak
	fun:calloc
	...
	fun:g_rand_new_with_seed_array
	...
	fun:get_global_random
	...
	fun:g_test_init
}

{
	g-quark-table-new
	Memcheck:Leak
	fun:g_hash_table_new
	...
	fun:quark_new
}

{
	g-quark-table-resize
	Memcheck:Leak
	fun:g_hash_table_resize
	...
	fun:quark_new
}

{
	g-type-interface-init
	Memcheck:Leak
	fun:malloc
	...
	fun:type_iface_vtable_base_init_Wm
}

{
	g-type-class-init
	Memcheck:Leak
	fun:calloc
	...
	fun:type_class_init_Wm
}

{
	g-io-module-default-singleton-malloc
	Memcheck:Leak
	fun:malloc
	...
	fun:g_type_create_instance
	...
	fun:_g_io_module_get_default
}

{
	g-io-module-default-singleton-module
	Memcheck:Leak
	fun:calloc
	...
	fun:g_module_open
	...
	fun:_g_io_module_get_default
}

{
	g-get-language-names
	Memcheck:Leak
	fun:malloc
	...
	fun:g_get_language_names
}

{
	g-static-mutex
	Memcheck:Leak
	fun:malloc
	...
	fun:g_static_mutex_get_mutex_impl
}

{
	g-system-thread-init
	Memcheck:Leak
	fun:calloc
	...
	fun:g_system_thread_new
}

{
	g-io-module-default-proxy-resolver-gnome
	Memcheck:Leak
	fun:calloc
	...
	fun:g_proxy_resolver_gnome_init
	...
	fun:_g_io_module_get_default
}

# memcheck checks that the third argument to ioctl() is a valid pointer, but
# some ioctls use that argument as an integer
{
   ioctl-with-non-pointer-param
   Memcheck:Param
   ioctl(generic)
   fun:ioctl
   fun:btrfs_reflink_with_progress
}

{
	g-private-get
	drd:ConflictingAccess
	fun:g_private_get
}
{
	g-private-get-helgrind
	Helgrind:Race
	fun:g_private_get
}


{
	g-private-set
	drd:ConflictingAccess
	fun:g_private_set
}
{
	g-private-set-helgrind
	Helgrind:Race
	fun:g_private_set
}

{
	g-type-construct-free
	drd:ConflictingAccess
	fun:g_type_free_instance
}
{
	g-type-construct-free-helgrind
	Helgrind:Race
	fun:g_type_free_instance
}

{
	g-variant-unref
	drd:ConflictingAccess
	fun:g_variant_unref
}
{
	g-variant-unref-helgrind
	Helgrind:Race
	fun:g_variant_unref
}

{
	g-unix-signals-main
	drd:ConflictingAccess
	fun:_g_main_create_unix_signal_watch
}
{
	g-unix-signals-dispatch
	drd:ConflictingAccess
	...
	fun:dispatch_unix_signals*
}
{
	g-unix-signals-dispatch-helgrind
	Helgrind:Race
	...
	fun:dispatch_unix_signals*
}
{
	g-unix-signals-other
	drd:ConflictingAccess
	fun:g_unix_signal_watch*
}
{
	g-unix-signals-other-helgrind
	Helgrind:Race
	fun:g_unix_signal_watch*
}
{
	g-unix-signals-handler
	drd:ConflictingAccess
	fun:g_unix_signal_handler*
}
{
	g-unix-signals-handler-helgrind
	Helgrind:Race
	fun:g_unix_signal_handler*
}
{
	g-unix-signals-worker
	drd:ConflictingAccess
	fun:glib_worker_main
}
{
	g-unix-signals-worker-helgrind
	Helgrind:Race
	fun:glib_worker_main
}

{
	g-wakeup-acknowledge
	drd:ConflictingAccess
	fun:read
	fun:g_wakeup_acknowledge
}

{
	g-type-fundamental
	drd:ConflictingAccess
	fun:g_type_fundamental
}
{
	g-type-fundamental-helgrind
	Helgrind:Race
	fun:g_type_fundamental
}
{
	g-type-class-peek-static
	drd:ConflictingAccess
	fun:g_type_class_peek_static
}
{
	g-type-class-peek-static-helgrind
	Helgrind:Race
	fun:g_type_class_peek_static
}
{
	g-type-is-a
	drd:ConflictingAccess
	...
	fun:g_type_is_a
}
{
	g-type-is-a-helgrind
	Helgrind:Race
	...
	fun:g_type_is_a
}

{
	g-inet-address-get-type
	drd:ConflictingAccess
	fun:g_inet_address_get_type
}
{
	g-inet-address-get-type-helgrind
	Helgrind:Race
	fun:g_inet_address_get_type
}

# From: https://github.com/fredericgermain/valgrind/blob/master/glibc-2.X-drd.supp
{
	drd-libc-stdio
	drd:ConflictingAccess
	obj:*/lib*/libc-*
}
{
	drd-libc-recv
	drd:ConflictingAccess
	fun:recv
}
{
	drd-libc-send
	drd:ConflictingAccess
	fun:send
}

# GSources do an opportunistic ref count check
{
	g-source-set-ready-time
	drd:ConflictingAccess
	fun:g_source_set_ready_time
}
{
	g-source-set-ready-time-helgrind
	Helgrind:Race
	fun:g_source_set_ready_time
}

{
	g-source-iter-next
	Helgrind:Race
	fun:g_source_iter_next
	fun:g_main_context_*
	fun:g_main_context_iterate
}

{
	g-object-instance-private
	drd:ConflictingAccess
	fun:*_get_instance_private
}
{
	g-object-instance-private-helgrind
	Helgrind:Race
	fun:*_get_instance_private
}

# GLib legitimately calls pthread_cond_signal without a mutex held
{
	g-task-thread-complete
	drd:CondErr
	...
	fun:g_cond_signal
	fun:g_task_thread_complete
}
{
	g-task-thread-complete
	Helgrind:Misc
	...
	fun:g_cond_signal
	fun:g_task_thread_complete
}

# False positive, but I can't explain how (FIXME)
{
	g-task-cond
	Helgrind:Misc
	...
	fun:g_cond_clear
	fun:g_task_finalize
}

# Real race, but is_cancelled() is an opportunistic function anyway
{
	g-cancellable-is-cancelled
	Helgrind:Race
	fun:g_cancellable_is_cancelled
}

# False positive
{
	g-main-context-cond
	Helgrind:Misc
	...
	fun:g_cond_clear
	fun:g_main_context_unref
}

# False positives
{
	g-source-unlocked
	Helgrind:Race
	fun:g_source_*_unlocked
}
{
	g-source-internal
	Helgrind:Race
	fun:g_source_*_internal
}

# False positive
{
	g_object_real_dispose
	Helgrind:Race
	fun:g_object_real_dispose
}

# False positive
{
	g_object_new_valist
	Helgrind:Race
	...
	fun:g_object_new_valist
}