Merge 'Minor abort on OOM changes' from Travis Downs

At Redpanda we are moving to enable abort-on-OOM by default, and as part of this it is useful to be able to _disable_ as well as enable this setting (current behavior is that it starts out disabled and the reactor does a one-time enable based on the command line flag `--abort-on-seastar-bad-alloc`.

This series implements the ability to disable it and also to get its status.

A small test is included.

Closes #1506

* https://github.com/scylladb/seastar:
  Update memory namespace comment.
  Replace deprecated enable abort call
  Add a test for enable/disable abort on OOM
  Allow application to query the abort on OOM state
  Allow abort on OOM to be set both ways

include/seastar/core/memory.hh

@@ -95,9 +95,11 @@ namespace seastar {
 ///
 /// Often, the best way to debug an allocation failure is a coredump. This
 /// feature allows dumping core on allocation failures, containing the stack of
-/// the failed allocation, by means of aborting. To enable set the
-/// `abort_on_seastar_bad_alloc` configuration option or the respective command
-/// line flag.
+/// the failed allocation, by means of aborting. To enable this behavior, set
+/// `abort_on_seastar_bad_alloc` in `reactor_options` or pass the
+/// `--abort-on-seastar-bad-alloc` command line flag. Additionally, applications
+/// may enable or disable this functionality globally at runtime by calling
+/// `set_abort_on_allocation_failure()`.
 ///
 /// ### Dump diagnostics report
 ///
@@ -139,6 +141,8 @@ static constexpr size_t huge_page_size =
 void configure(std::vector<resource::memory> m, bool mbind,
         std::optional<std::string> hugetlbfs_path = {});
 
+// A deprecated alias for set_abort_on_allocation_failure(true).
+[[deprecated("use set_abort_on_allocation_failure(true) instead")]]
 void enable_abort_on_allocation_failure();
 
 class disable_abort_on_alloc_failure_temporarily {
@@ -223,6 +227,30 @@ void set_reclaim_hook(
 
 /// \endcond
 
+/// \brief Set the global state of the abort on allocation failure behavior.
+///
+/// If enabled, an allocation failure (i.e., the requested memory
+/// could not be allocated even after reclaim was attempted), will
+/// generally result in `abort()` being called. If disabled, the
+/// failure is reported to the caller, e.g., by throwing a
+/// `std::bad_alloc` for C++ allocations such as new, or returning
+/// a null pointer from malloc.
+///
+/// Note that even if the global state is set to enabled, the
+/// `disable_abort_on_alloc_failure_temporarily` class may override
+/// the behavior tepmorarily on a given shard. That is, abort only
+/// occurs if abort is globablly enabled on this shard _and_ there
+/// are no `disable_abort_on_alloc_failure_temporarily` objects
+/// currently alive on this shard.
+void set_abort_on_allocation_failure(bool enabled);
+
+/// \brief Determine the abort on allocation failure mode.
+///
+/// Return true if the global abort on allocation failure behavior
+/// is enabled, or false otherwise. Always returns false if the
+/// default (system) allocator is being used.
+bool is_abort_on_allocation_failure();
+
 class statistics;
 
 /// Capture a snapshot of memory allocation statistics for this lcore.

src/core/memory.cc

@@ -100,6 +100,10 @@ disable_abort_on_alloc_failure_temporarily::~disable_abort_on_alloc_failure_temp
     --abort_on_alloc_failure_suppressed;
 }
 
+void enable_abort_on_allocation_failure() {
+    set_abort_on_allocation_failure(true);
+}
+
 static std::pmr::polymorphic_allocator<char> static_malloc_allocator{std::pmr::get_default_resource()};;
 std::pmr::polymorphic_allocator<char>* malloc_allocator{&static_malloc_allocator};
 
@@ -1600,8 +1604,12 @@ class disable_report_on_alloc_failure_temporarily {
 static std::atomic<bool> abort_on_allocation_failure{false};
 static std::atomic<alloc_failure_kind> dump_diagnostics_on_alloc_failure_kind{alloc_failure_kind::critical};
 
-void enable_abort_on_allocation_failure() {
-    abort_on_allocation_failure.store(true, std::memory_order_seq_cst);
+void set_abort_on_allocation_failure(bool enabled) {
+    abort_on_allocation_failure.store(enabled, std::memory_order_seq_cst);
+}
+
+bool is_abort_on_allocation_failure() {
+    return abort_on_allocation_failure;
 }
 
 void set_dump_memory_diagnostics_on_alloc_failure_kind(alloc_failure_kind kind) {
@@ -2268,8 +2276,14 @@ scoped_heap_profiling::scoped_heap_profiling() noexcept {
 scoped_heap_profiling::~scoped_heap_profiling() {
 }
 
-void enable_abort_on_allocation_failure() {
-    seastar_logger.warn("Seastar compiled with default allocator, will not abort on bad_alloc");
+void set_abort_on_allocation_failure(bool enabled) {
+    if (enabled) {
+        seastar_logger.warn("Seastar compiled with default allocator, will not abort on bad_alloc");
+    }
+}
+
+bool is_abort_on_allocation_failure() {
+    return false;
 }
 
 reclaimer::reclaimer(std::function<reclaiming_result ()> reclaim, reclaimer_scope) {

src/core/reactor.cc

@@ -4205,7 +4205,7 @@ void smp::configure(const smp_options& smp_opts, const reactor_options& reactor_
     }
 
     if (reactor_opts.abort_on_seastar_bad_alloc) {
-        memory::enable_abort_on_allocation_failure();
+        memory::set_abort_on_allocation_failure(true);
     }
 
     if (reactor_opts.dump_memory_diagnostics_on_alloc_failure_kind) {

tests/unit/alloc_test.cc

@@ -198,6 +198,20 @@ SEASTAR_TEST_CASE(test_realloc_nullptr) {
     return make_ready_future<>();
 }
 
+SEASTAR_TEST_CASE(test_enable_abort_on_oom) {
+    bool original = seastar::memory::is_abort_on_allocation_failure();
+
+    seastar::memory::set_abort_on_allocation_failure(false);
+    BOOST_CHECK(!seastar::memory::is_abort_on_allocation_failure());
+
+    seastar::memory::set_abort_on_allocation_failure(true);
+    BOOST_CHECK(seastar::memory::is_abort_on_allocation_failure());
+
+    seastar::memory::set_abort_on_allocation_failure(original);
+
+    return make_ready_future<>();
+}
+
 void * volatile sink;
 
 SEASTAR_TEST_CASE(test_bad_alloc_throws) {