summaryrefslogtreecommitdiffhomepage
path: root/test/syscalls
diff options
context:
space:
mode:
authorAyush Ranjan <ayushranjan@google.com>2021-07-02 18:45:05 -0700
committergVisor bot <gvisor-bot@google.com>2021-07-02 18:47:48 -0700
commitadd8bca5ba53b37096bc653900cb278e11681461 (patch)
tree67c5a45a81fe8a2ded26b66f2ab6681d93da61a6 /test/syscalls
parent3d32a05a35bde4cfed861c274c32bfc55acc19c9 (diff)
[op] Make TCPNonBlockingConnectClose more reasonable.
This test single handedly causes the syscalls:socket_inet_loopback_test test variants to take more than an hour to run on some of our testing environments. Reduce how aggressively this test tries to replicate a fixed flake. This is a regression test. PiperOrigin-RevId: 382849039
Diffstat (limited to 'test/syscalls')
-rw-r--r--test/syscalls/linux/socket_inet_loopback.cc2
1 files changed, 1 insertions, 1 deletions
diff --git a/test/syscalls/linux/socket_inet_loopback.cc b/test/syscalls/linux/socket_inet_loopback.cc
index badc42ec5..9ae0cc59d 100644
--- a/test/syscalls/linux/socket_inet_loopback.cc
+++ b/test/syscalls/linux/socket_inet_loopback.cc
@@ -707,7 +707,7 @@ TEST_P(SocketInetLoopbackTest, TCPNonBlockingConnectClose) {
// Try many iterations to catch a race with socket close and handshake
// completion.
- for (int i = 0; i < 1000; ++i) {
+ for (int i = 0; i < 100; ++i) {
FileDescriptor client = ASSERT_NO_ERRNO_AND_VALUE(
Socket(connector.family(), SOCK_STREAM | SOCK_NONBLOCK, IPPROTO_TCP));
ASSERT_THAT(