ring-buffer: Bring back context level recursive checks

Commit 1a149d7 ("ring-buffer: Rewrite trace_recursive_(un)lock() to be
simpler") replaced the context level recursion checks with a simple counter.
This would prevent the ring buffer code from recursively calling itself more
than the max number of contexts that exist (Normal, softirq, irq, nmi). But
this change caused a lockup in a specific case, which was during suspend and
resume using a global clock. Adding a stack dump to see where this occurred,
the issue was in the trace global clock itself:

  trace_buffer_lock_reserve+0x1c/0x50
  __trace_graph_entry+0x2d/0x90
  trace_graph_entry+0xe8/0x200
  prepare_ftrace_return+0x69/0xc0
  ftrace_graph_caller+0x78/0xa8
  queued_spin_lock_slowpath+0x5/0x1d0
  trace_clock_global+0xb0/0xc0
  ring_buffer_lock_reserve+0xf9/0x390

The function graph tracer traced queued_spin_lock_slowpath that was called
by trace_clock_global. This pointed out that the trace_clock_global() is not
reentrant, as it takes a spin lock. It depended on the ring buffer recursive
lock from letting that happen.

By removing the context detection and adding just a max number of allowable
recursions, it allowed the trace_clock_global() to be entered again and try
to retake the spinlock it already held, causing a deadlock.

Fixes: 1a149d7 ("ring-buffer: Rewrite trace_recursive_(un)lock() to be simpler")
Reported-by: David Weinehall <[email protected]>
Signed-off-by: Steven Rostedt (VMware) <[email protected]>

Author: rostedt

kernel/trace/ring_buffer.c

@@ -2534,39 +2534,67 @@ rb_wakeups(struct ring_buffer *buffer, struct ring_buffer_per_cpu *cpu_buffer)
  * The lock and unlock are done within a preempt disable section.
  * The current_context per_cpu variable can only be modified
  * by the current task between lock and unlock. But it can
- * be modified more than once via an interrupt. There are four
- * different contexts that we need to consider.
+ * be modified more than once via an interrupt. To pass this
+ * information from the lock to the unlock without having to
+ * access the 'in_interrupt()' functions again (which do show
+ * a bit of overhead in something as critical as function tracing,
+ * we use a bitmask trick.
  *
- *  Normal context.
- *  SoftIRQ context
- *  IRQ context
- *  NMI context
+ *  bit 0 =  NMI context
+ *  bit 1 =  IRQ context
+ *  bit 2 =  SoftIRQ context
+ *  bit 3 =  normal context.
  *
- * If for some reason the ring buffer starts to recurse, we
- * only allow that to happen at most 4 times (one for each
- * context). If it happens 5 times, then we consider this a
- * recusive loop and do not let it go further.
+ * This works because this is the order of contexts that can
+ * preempt other contexts. A SoftIRQ never preempts an IRQ
+ * context.
+ *
+ * When the context is determined, the corresponding bit is
+ * checked and set (if it was set, then a recursion of that context
+ * happened).
+ *
+ * On unlock, we need to clear this bit. To do so, just subtract
+ * 1 from the current_context and AND it to itself.
+ *
+ * (binary)
+ *  101 - 1 = 100
+ *  101 & 100 = 100 (clearing bit zero)
+ *
+ *  1010 - 1 = 1001
+ *  1010 & 1001 = 1000 (clearing bit 1)
+ *
+ * The least significant bit can be cleared this way, and it
+ * just so happens that it is the same bit corresponding to
+ * the current context.
  */
 
 static __always_inline int
 trace_recursive_lock(struct ring_buffer_per_cpu *cpu_buffer)
 {
-	if (cpu_buffer->current_context >= 4)
+	unsigned int val = cpu_buffer->current_context;
+	unsigned long pc = preempt_count();
+	int bit;
+
+	if (!(pc & (NMI_MASK | HARDIRQ_MASK | SOFTIRQ_OFFSET)))
+		bit = RB_CTX_NORMAL;
+	else
+		bit = pc & NMI_MASK ? RB_CTX_NMI :
+			pc & HARDIRQ_MASK ? RB_CTX_IRQ :
+			pc & SOFTIRQ_OFFSET ? 2 : RB_CTX_SOFTIRQ;
+
+	if (unlikely(val & (1 << bit)))
 		return 1;
 
-	cpu_buffer->current_context++;
-	/* Interrupts must see this update */
-	barrier();
+	val |= (1 << bit);
+	cpu_buffer->current_context = val;
 
 	return 0;
 }
 
 static __always_inline void
 trace_recursive_unlock(struct ring_buffer_per_cpu *cpu_buffer)
 {
-	/* Don't let the dec leak out */
-	barrier();
-	cpu_buffer->current_context--;
+	cpu_buffer->current_context &= cpu_buffer->current_context - 1;
 }
 
 /**