[patch 1/2] dump_stack: serialize the output from dump_stack()

[athorlton@sgi.com]

These patches fix up issues with interspersed output from multiple
simultaneous calls to warn or dump_stack on multi-cpu systems.
References: <20130508210102.898396979@asylum.americas.sgi.com>
Content-Disposition: inline; filename=dump-stack-serialize.patch

This patch adds functionality to serialize the output from dump_stack() to 
avoid mangling of the output when dump_stack is called simultaneously from
multiple cpus.

Cc: Vineet Gupta <vgupta@synopsys.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: David S. Miller <davem@davemloft.net>
Cc: Richard Kuo <rkuo@codeaurora.org>
Cc: Jesper Nilsson <jesper.nilsson@axis.com>
Reported-by: Russ Anderson <rja@sgi.com>
Reviewed-by: Robin Holt <holt@sgi.com>
Signed-off-by: Alex Thorlton <athorlton@sgi.com>
---
The original discussion regarding this patch can be found in this thread:
[PATCH] x86: Avoid intermixing cpu dump_stack output on multi-processor systems

 lib/dump_stack.c |   25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)

--- linux.orig/lib/dump_stack.c
+++ linux/lib/dump_stack.c
@@ -7,6 +7,8 @@
 #include <linux/export.h>
 #include <linux/sched.h>
 
+static atomic_t dump_lock = ATOMIC_INIT(-1);
+
 /**
  * dump_stack - dump the current task information and its stack trace
  *
@@ -14,7 +16,30 @@
  */
 void dump_stack(void)
 {
+	int was_locked;
+	int old;
+	int cpu;
+
+	preempt_disable();
+
+retry:
+	cpu = smp_processor_id();
+	old = atomic_cmpxchg(&dump_lock, -1, cpu);
+	if (old == -1) {
+		was_locked = 0;
+	} else if (old == cpu) {
+		was_locked = 1;
+	} else {
+		cpu_relax();
+		goto retry;
+	}
+
 	dump_stack_print_info(KERN_DEFAULT);
 	show_stack(NULL, NULL);
+
+	if (!was_locked)
+		atomic_set(&dump_lock, -1);
+
+	preempt_enable();
 }
 EXPORT_SYMBOL(dump_stack);