test/crimson: disable stall-detector on aarch64 to fix asan failures

Ceph crimson uses seastar on debug mode which would enable ASAN, while stall-detector uses glibc backtrace function which would cause ASAN failures on aarch64.
Reason see scylladb/scylladb#15090 (comment)

Because arm ci servers in lab are "elderly", causing stall happened often, this PR is to disable stall-detector until seastar upstream migrated to libunwind, see scylladb/seastar#1878
Seastar does not provide a function to disable stall-detector, at present, increase stall-detector time to avoid stall-detector timeout on arm platform.

fixes: https://tracker.ceph.com/issues/65635

Signed-off-by: cailianchun <arm7star@qq.com>

1 files changed, 13 insertions, 0 deletions

diff --git a/src/test/crimson/seastar_runner.h b/src/test/crimson/seastar_runner.h
index 63cc50d9f05..590eef13adf 100644
--- a/src/test/crimson/seastar_runner.h
+++ b/src/test/crimson/seastar_runner.h
@@ -71,6 +71,19 @@ struct SeastarRunner {
     auto ret = app.run(argc, argv, [this] {
       on_end.reset(new seastar::readable_eventfd);
       return seastar::now().then([this] {
+// FIXME: The stall detector uses glibc backtrace function to
+// collect backtraces, this causes ASAN failures on ARM.
+// For now we just extend timeout duration to 10000h in order to
+// get the same effect as disabling the stall detector which is not provided by seastar.
+// the ticket about migrating to libunwind: https://github.com/scylladb/seastar/issues/1878
+// Will remove once the ticket fixed.
+// Ceph ticket see: https://tracker.ceph.com/issues/65635
+#ifdef __aarch64__
+	seastar::smp::invoke_on_all([] {
+	  using namespace std::chrono;
+	  seastar::engine().update_blocked_reactor_notify_ms(duration_cast<milliseconds>(10000h));
+	}).get();
+#endif
 	begin_signaled = true;
 	[[maybe_unused]] auto r = ::eventfd_write(begin_fd.get(), APP_RUNNING);
 	assert(r == 0);