From d48d2e6b914582df5fb589113fc4201c64806591 Mon Sep 17 00:00:00 2001 From: Micael Karlberg Date: Mon, 29 Jul 2019 14:55:53 +0200 Subject: [esock|test] Race during socket close Add a sleep to reduce the chance that process exit has not been "processed" by the nif code. Also increased the test case timetrap for the sc_ctrl_proc_exit group of test cases. OTP-15904 --- erts/emulator/test/socket_SUITE.erl | 19 +++++++++++++------ 1 file changed, 13 insertions(+), 6 deletions(-) diff --git a/erts/emulator/test/socket_SUITE.erl b/erts/emulator/test/socket_SUITE.erl index 62e770b339..ff5a4e7e2e 100644 --- a/erts/emulator/test/socket_SUITE.erl +++ b/erts/emulator/test/socket_SUITE.erl @@ -12156,7 +12156,7 @@ sc_cpe_socket_cleanup_tcp4(suite) -> sc_cpe_socket_cleanup_tcp4(doc) -> []; sc_cpe_socket_cleanup_tcp4(_Config) when is_list(_Config) -> - ?TT(?SECS(5)), + ?TT(?SECS(30)), tc_try(sc_cpe_socket_cleanup_tcp4, fun() -> InitState = #{domain => inet, @@ -12176,7 +12176,7 @@ sc_cpe_socket_cleanup_tcp6(suite) -> sc_cpe_socket_cleanup_tcp6(doc) -> []; sc_cpe_socket_cleanup_tcp6(_Config) when is_list(_Config) -> - ?TT(?SECS(5)), + ?TT(?SECS(30)), tc_try(sc_cpe_socket_cleanup_tcp6, fun() -> has_support_ipv6() end, fun() -> @@ -12197,7 +12197,7 @@ sc_cpe_socket_cleanup_tcpL(suite) -> sc_cpe_socket_cleanup_tcpL(doc) -> []; sc_cpe_socket_cleanup_tcpL(_Config) when is_list(_Config) -> - ?TT(?SECS(5)), + ?TT(?SECS(30)), tc_try(sc_cpe_socket_cleanup_tcpL, fun() -> has_support_unix_domain_socket() end, fun() -> @@ -12218,7 +12218,7 @@ sc_cpe_socket_cleanup_udp4(suite) -> sc_cpe_socket_cleanup_udp4(doc) -> []; sc_cpe_socket_cleanup_udp4(_Config) when is_list(_Config) -> - ?TT(?SECS(5)), + ?TT(?SECS(30)), tc_try(sc_cpe_socket_cleanup_udp4, fun() -> InitState = #{domain => inet, @@ -12239,7 +12239,7 @@ sc_cpe_socket_cleanup_udp6(suite) -> sc_cpe_socket_cleanup_udp6(doc) -> []; sc_cpe_socket_cleanup_udp6(_Config) when is_list(_Config) -> - ?TT(?SECS(5)), + ?TT(?SECS(30)), tc_try(sc_cpe_socket_cleanup_udp6, fun() -> has_support_ipv6() end, fun() -> @@ -12260,7 +12260,7 @@ sc_cpe_socket_cleanup_udpL(suite) -> sc_cpe_socket_cleanup_udpL(doc) -> []; sc_cpe_socket_cleanup_udpL(_Config) when is_list(_Config) -> - ?TT(?SECS(5)), + ?TT(?SECS(30)), tc_try(sc_cpe_socket_cleanup_udpL, fun() -> has_support_unix_domain_socket() end, fun() -> @@ -12366,8 +12366,15 @@ sc_cpe_socket_cleanup(InitState) -> ERROR end end}, + + ?SEV_SLEEP(?SECS(5)), + %% The reason we get closed, is that as long as there is a ref to %% the resource (socket), then it will not be garbage collected. + %% Note that its still a race that the nif has processed that the + %% "controlling process" has terminated. There really is no + %% proper timeout for this, but the 5 seconds "should" be enough... + %% We should really have some way to subscribe to socket events... #{desc => "verify no socket (closed)", cmd => fun(#{owner := Pid, sock := Sock} = _State) -> case socket:getopt(Sock, otp, controlling_process) of -- cgit v1.2.3