[rfc 3/4] core/fiber: Put static watermark into stack to track its usage

[Cyrill Gorcunov <gorcunov@gmail.com>]

[-- Attachment #1: fiber-stack-wmark --]
[-- Type: text/plain, Size: 4778 bytes --]

We want to detect a situation where task in fiber is too eager for
stack and close to its exhausting. For this sake upon stack creation
we put 8 marks on last stack page with step of 128 bytes. Such params
allows us to fill ~1/4 of a page, which does seem reasonable but
we might change this params with time.

Since the watermark position is permanent and some task is close
to stack limit we print about the situation once to not spam
a user much and stop putting the mark on recycling.

Closes #3418
---
 src/lib/core/fiber.c |  105 +++++++++++++++++++++++++++++++++++++++++++++++++++
 src/lib/core/fiber.h |    2 
 2 files changed, 107 insertions(+)

Index: tarantool.git/src/lib/core/fiber.c
===================================================================
--- tarantool.git.orig/src/lib/core/fiber.c
+++ tarantool.git/src/lib/core/fiber.c
@@ -104,6 +104,28 @@ static const struct fiber_attr fiber_att
        .flags = FIBER_DEFAULT_FLAGS
 };
 
+/*
+ * Random values generated with uuid.
+ * Try to fit a cache line.
+ */
+static const uint64_t poison_pool[] = {
+	0x74f31d37285c4c37, 0xb10269a05bf10c29,
+	0x0994d845bd284e0f, 0x9ffd4f7129c184df,
+	0x357151e6711c4415, 0x8c5e5f41aafe6f28,
+	0x6917dd79e78049d5, 0xba61957c65ca2465,
+};
+
+/*
+ * We poison by 8 bytes as it natural for stack
+ * step on x86-64. Also 128 byte gap between
+ * poison values should cover a common cases.
+ */
+#define POISON_SIZE	(sizeof(poison_pool) / sizeof(poison_pool[0]))
+#define POISON_GAP	(128 + sizeof(poison_pool[0]))
+#define POISON_OFF	(POISON_GAP / sizeof(poison_pool[0]))
+
+static void fiber_wmark_recycle(struct fiber *fiber);
+
 void
 fiber_attr_create(struct fiber_attr *fiber_attr)
 {
@@ -624,6 +646,7 @@ fiber_recycle(struct fiber *fiber)
 	/* no pending wakeup */
 	assert(rlist_empty(&fiber->state));
 	bool has_custom_stack = fiber->flags & FIBER_CUSTOM_STACK;
+	fiber_wmark_recycle(fiber);
 	fiber_reset(fiber);
 	fiber->name[0] = '\0';
 	fiber->f = NULL;
@@ -710,6 +733,85 @@ page_align_up(void *ptr)
 	return page_align_down(ptr + page_size - 1);
 }
 
+static bool
+stack_has_wmark(void *addr)
+{
+	const uint64_t *src = poison_pool;
+	const uint64_t *dst = addr;
+	size_t i;
+
+	for (i = 0; i < POISON_SIZE; i++) {
+		if (*dst != src[i])
+			return false;
+		dst += POISON_OFF;
+	}
+
+	return true;
+}
+
+static void
+stack_put_wmark(void *addr)
+{
+	const uint64_t *src = poison_pool;
+	uint64_t *dst = addr;
+	size_t i;
+
+	for (i = 0; i < POISON_SIZE; i++) {
+		*dst = src[i];
+		dst += POISON_OFF;
+	}
+}
+
+static void
+fiber_wmark_recycle(struct fiber *fiber)
+{
+	static bool overflow_warned = false;
+
+	if (fiber->stack == NULL || fiber->flags & FIBER_CUSTOM_STACK)
+		return;
+
+	/*
+	 * We are watching for stack overflow in one shot way:
+	 * firstly to not spam a user with messages and secondly
+	 * to relax system from additional operations while our
+	 * watermark is not dynamic.
+	 */
+	if (overflow_warned)
+		return;
+
+	if (!stack_has_wmark(fiber->stack_wmark_ofl)) {
+		if (!overflow_warned) {
+			say_warn("fiber %d is close to a stack limit",
+				 fiber->name, fiber->fid);
+			overflow_warned = true;
+		}
+	}
+}
+
+static void
+fiber_wmark_init(struct fiber *fiber)
+{
+	/*
+	 * No tracking on custom stacks
+	 * in a sake of simplicity.
+	 */
+	if (fiber->flags & FIBER_CUSTOM_STACK) {
+		fiber->stack_wmark_ofl = NULL;
+		return;
+	}
+
+	/*
+	 * Initially we arm last page of a stack
+	 * to catch if we're getting close to
+	 * stack exhausting.
+	 */
+	if (stack_direction < 0)
+		fiber->stack_wmark_ofl = fiber->stack;
+	else
+		fiber->stack_wmark_ofl = fiber->stack + fiber->stack_size - page_size;
+	stack_put_wmark(fiber->stack_wmark_ofl);
+}
+
 static int
 fiber_stack_create(struct fiber *fiber, size_t stack_size)
 {
@@ -757,6 +859,7 @@ fiber_stack_create(struct fiber *fiber,
 	madvise(fiber->stack, fiber->stack_size, MADV_DONTNEED);
 #endif
 
+	fiber_wmark_init(fiber);
 	mprotect(guard, page_size, PROT_NONE);
 	return 0;
 }
@@ -926,8 +1029,10 @@ cord_create(struct cord *cord, const cha
 	/* Record stack extents */
 	tt_pthread_attr_getstack(cord->id, &cord->sched.stack,
 				 &cord->sched.stack_size);
+	cord->sched.stack_wmark_ofl = cord->sched.stack;
 #else
 	cord->sched.stack = NULL;
+	cord->sched.stack_wmark_ofl = NULL;
 	cord->sched.stack_size = 0;
 #endif
 }
Index: tarantool.git/src/lib/core/fiber.h
===================================================================
--- tarantool.git.orig/src/lib/core/fiber.h
+++ tarantool.git/src/lib/core/fiber.h
@@ -348,6 +348,8 @@ struct fiber {
 	struct slab *stack_slab;
 	/** Coro stack addr. */
 	void *stack;
+	/** Stack watermark addr for overflow detection. */
+	void *stack_wmark_ofl;
 	/** Coro stack size. */
 	size_t stack_size;
 	/** Valgrind stack id. */