2 # This file is part of the LibreOffice project.
4 # This Source Code Form is subject to the terms of the Mozilla Public
5 # License, v. 2.0. If a copy of the MPL was not distributed with this
6 # file, You can obtain one at http://mozilla.org/MPL/2.0/.
9 # Use via environment variable TSAN_OPTIONS=suppressions=.../tsan-suppress.txt
11 # There looks to be a lock ordering problem here, but I can't see how it could
12 # actually be hit in practice.
13 deadlock:cppuhelper::ServiceManager::disposing()
14 deadlock:cppuhelper::ServiceManager::loadImplementation
15 deadlock:AffineBridge::v_callInto_v
17 # Ignore stuff in external DBUS library.
18 # Some kind of dbus lock issue when we call it from psp::CUPSManager.
20 race:g_dbus_connection_signal_unsubscribe
21 race:g_path_get_basename
22 race:g_date_time_format_iso8601
24 # inside an assert so I don't care.
25 race:cppu::OWeakConnectionPoint::acquire
26 race:AffineBridge::v_enter
29 # right now, I'm not interested in deadlocks at all, too many false+
32 # This is checking SAL_STRING_IS_STATIC, which is safe because that is written at compile time.
33 # Depending on the optimisation/debug level, the call stack may have different leaves, hence
34 # specifying this in more than one way.
35 race:rtl::str::acquire<_rtl_uString>
36 race:rtl::str::release<_rtl_uString>
37 race:rtl_uString_acquire
38 race:rtl_uString_release
39 race:rtl_uString_assign
41 # I've convinced myself this is a false+, caused by ping-ponging the buffer between two
42 # threads, but I might be wrong.
43 race:XBufferedThreadedStream::getNextBlock
45 # I am not interested in stuff the embedded JVM does.
48 # I think this is OK, because at this point we are doing
50 # and we know from our callers that the refcount must be at least one
51 # so there is no failure mode.
54 # TODO There appears to be a race here, initialising the
55 # ::com::sun::star::uno::Sequence< T >::s_pType
56 # field. But no idea at all how to fix it.
57 race:cppu::getTypeFavourUnsigned
60 # This is all inside GIO/Glib, no idea what it is doing.
62 race:slab_allocator_alloc_chunk
63 race:g_source_destroy_internal
64 race:g_source_unref_internal
66 race:g_socket_send_message_with_timeout
67 race:g_idle_source_new
74 # Python build, lock order inversion
77 # Not interested in deadlock issues when they involve the shutdown path
78 # - the shutdown path is special, so if we see a lock-ordering here that conflicts with
79 # a "normal" lock-ordering path, that is highly unlikely to result in a real issue.
82 # Sometimes tsan will report data-races in these, who knows why, they should be thread-safe
87 # There is a lock-cycle here, together with the mutexes acquired in
88 # SvtSysLocaleOptions::SvtSysLocaleOptions
89 # SvtSysLocale::SvtSysLocale
90 # but I can't see how it could lead to a problem in practice
91 deadlock:ItemHolder1::impl_addItem
93 # ignore warning about nRefCount in sal_Sequence.
94 # <sberg> noelgrandin, we generally assume that sal_Int32 is small enough, and hardware sane enough,
95 # that reading it will always give a sane value, i.e., if the read happens in parallel with a write,
96 # the read produces either the old, original value, or the newly written one.
97 race:cppu::createEmptySequence
98 race:uno_type_sequence_reference2One