[Libreoffice-commits] core.git: tsan-suppress.txt

[Noel Grandin (via logerrit)]

 tsan-suppress.txt |   12 ++++++++++++
 1 file changed, 12 insertions(+)

New commits:
commit 8ed2207983c6d96e8866bd81e53097cb798436e8
Author:     Noel Grandin <noel.grandin at collabora.co.uk>
AuthorDate: Mon Aug 2 14:18:56 2021 +0200
Commit:     Noel Grandin <noel.grandin at collabora.co.uk>
CommitDate: Mon Aug 2 20:05:52 2021 +0200

    more suppressions for TSAN
    
    Change-Id: I96072bdacb805135f0c267a878181035c4b0d810
    Reviewed-on: https://gerrit.libreoffice.org/c/core/+/119890
    Tested-by: Jenkins
    Reviewed-by: Noel Grandin <noel.grandin at collabora.co.uk>

diff --git a/tsan-suppress.txt b/tsan-suppress.txt
index d6130b285df8..9d6c73c557ae 100644
--- a/tsan-suppress.txt
+++ b/tsan-suppress.txt
@@ -35,3 +35,15 @@ race:XBufferedThreadedStream::getNextBlock
 
 # not introduced in stuff the embedded JVM does
 race:libjvm.so
+
+# I think this is OK, because at this point we are doing
+#  if (nRefCount > 1)
+# and we know from our callers that the refcount must be at least one
+# so there is no failure mode
+race:ireallocSequence
+
+# TODO There appears to be a race here, initialisng the
+#   ::com::sun::star::uno::Sequence< T >::s_pType
+# field. But no idea at all how to fix it.
+race:cppu::getTypeFavourUnsigned
+