Merge tag 'trace-seq-buf-3.19' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace

Pull nmi-safe seq_buf printk update from Steven Rostedt:
 "This code is a fork from the trace-3.19 pull as it needed the
  trace_seq clean ups from that branch.

  This code solves the issue of performing stack dumps from NMI context.
  The issue is that printk() is not safe from NMI context as if the NMI
  were to trigger when a printk() was being performed, the NMI could
  deadlock from the printk() internal locks.  This has been seen in
  practice.

  With lots of review from Petr Mladek, this code went through several
  iterations, and we feel that it is now at a point of quality to be
  accepted into mainline.

  Here's what is contained in this patch set:

   - Creates a "seq_buf" generic buffer utility that allows a descriptor
     to be passed around where functions can write their own "printk()"
     formatted strings into it.  The generic version was pulled out of
     the trace_seq() code that was made specifically for tracing.

   - The seq_buf code was change to model the seq_file code.  I have a
     patch (not included for 3.19) that converts the seq_file.c code
     over to use seq_buf.c like the trace_seq.c code does.  This was
     done to make sure that seq_buf.c is compatible with seq_file.c.  I
     may try to get that patch in for 3.20.

   - The seq_buf.c file was moved to lib/ to remove it from being
     dependent on CONFIG_TRACING.

   - The printk() was updated to allow for a per_cpu "override" of the
     internal calls.  That is, instead of writing to the console, a call
     to printk() may do something else.  This made it easier to allow
     the NMI to change what printk() does in order to call dump_stack()
     without needing to update that code as well.

   - Finally, the dump_stack from all CPUs via NMI code was converted to
     use the seq_buf code.  The caller to trigger the NMI code would
     wait till all the NMIs finished, and then it would print the
     seq_buf data to the console safely from a non NMI context

  One added bonus is that this code also makes the NMI dump stack work
  on PREEMPT_RT kernels.  As printk() includes sleeping locks on
  PREEMPT_RT, printk() only writes to console if the console does not
  use any rt_mutex converted spin locks.  Which a lot do"

* tag 'trace-seq-buf-3.19' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
  x86/nmi: Fix use of unallocated cpumask_var_t
  printk/percpu: Define printk_func when printk is not defined
  x86/nmi: Perform a safe NMI stack trace on all CPUs
  printk: Add per_cpu printk func to allow printk to be diverted
  seq_buf: Move the seq_buf code to lib/
  seq-buf: Make seq_buf_bprintf() conditional on CONFIG_BINARY_PRINTF
  tracing: Add seq_buf_get_buf() and seq_buf_commit() helper functions
  tracing: Have seq_buf use full buffer
  seq_buf: Add seq_buf_can_fit() helper function
  tracing: Add paranoid size check in trace_printk_seq()
  tracing: Use trace_seq_used() and seq_buf_used() instead of len
  tracing: Clean up tracing_fill_pipe_page()
  seq_buf: Create seq_buf_used() to find out how much was written
  tracing: Add a seq_buf_clear() helper and clear len and readpos in init
  tracing: Convert seq_buf fields to be like seq_file fields
  tracing: Convert seq_buf_path() to be like seq_path()
  tracing: Create seq_buf layer in trace_seq

1 files changed, 86 insertions, 5 deletions

diff --git a/arch/x86/kernel/apic/hw_nmi.c b/arch/x86/kernel/apic/hw_nmi.c
index 6a1e71bde323..6873ab925d00 100644
--- a/arch/x86/kernel/apic/hw_nmi.c
+++ b/arch/x86/kernel/apic/hw_nmi.c
@@ -18,6 +18,7 @@
 #include <linux/nmi.h>
 #include <linux/module.h>
 #include <linux/delay.h>
+#include <linux/seq_buf.h>
 
 #ifdef CONFIG_HARDLOCKUP_DETECTOR
 u64 hw_nmi_get_sample_period(int watchdog_thresh)
@@ -29,14 +30,35 @@ u64 hw_nmi_get_sample_period(int watchdog_thresh)
 #ifdef arch_trigger_all_cpu_backtrace
 /* For reliability, we're prepared to waste bits here. */
 static DECLARE_BITMAP(backtrace_mask, NR_CPUS) __read_mostly;
+static cpumask_t printtrace_mask;
+
+#define NMI_BUF_SIZE		4096
+
+struct nmi_seq_buf {
+	unsigned char		buffer[NMI_BUF_SIZE];
+	struct seq_buf		seq;
+};
+
+/* Safe printing in NMI context */
+static DEFINE_PER_CPU(struct nmi_seq_buf, nmi_print_seq);
 
 /* "in progress" flag of arch_trigger_all_cpu_backtrace */
 static unsigned long backtrace_flag;
 
+static void print_seq_line(struct nmi_seq_buf *s, int start, int end)
+{
+	const char *buf = s->buffer + start;
+
+	printk("%.*s", (end - start) + 1, buf);
+}
+
 void arch_trigger_all_cpu_backtrace(bool include_self)
 {
+	struct nmi_seq_buf *s;
+	int len;
+	int cpu;
 	int i;
-	int cpu = get_cpu();
+	int this_cpu = get_cpu();
 
 	if (test_and_set_bit(0, &backtrace_flag)) {
 		/*
@@ -49,7 +71,17 @@ void arch_trigger_all_cpu_backtrace(bool include_self)
 
 	cpumask_copy(to_cpumask(backtrace_mask), cpu_online_mask);
 	if (!include_self)
-		cpumask_clear_cpu(cpu, to_cpumask(backtrace_mask));
+		cpumask_clear_cpu(this_cpu, to_cpumask(backtrace_mask));
+
+	cpumask_copy(&printtrace_mask, to_cpumask(backtrace_mask));
+	/*
+	 * Set up per_cpu seq_buf buffers that the NMIs running on the other
+	 * CPUs will write to.
+	 */
+	for_each_cpu(cpu, to_cpumask(backtrace_mask)) {
+		s = &per_cpu(nmi_print_seq, cpu);
+		seq_buf_init(&s->seq, s->buffer, NMI_BUF_SIZE);
+	}
 
 	if (!cpumask_empty(to_cpumask(backtrace_mask))) {
 		pr_info("sending NMI to %s CPUs:\n",
@@ -65,11 +97,58 @@ void arch_trigger_all_cpu_backtrace(bool include_self)
 		touch_softlockup_watchdog();
 	}
 
+	/*
+	 * Now that all the NMIs have triggered, we can dump out their
+	 * back traces safely to the console.
+	 */
+	for_each_cpu(cpu, &printtrace_mask) {
+		int last_i = 0;
+
+		s = &per_cpu(nmi_print_seq, cpu);
+		len = seq_buf_used(&s->seq);
+		if (!len)
+			continue;
+
+		/* Print line by line. */
+		for (i = 0; i < len; i++) {
+			if (s->buffer[i] == '\n') {
+				print_seq_line(s, last_i, i);
+				last_i = i + 1;
+			}
+		}
+		/* Check if there was a partial line. */
+		if (last_i < len) {
+			print_seq_line(s, last_i, len - 1);
+			pr_cont("\n");
+		}
+	}
+
 	clear_bit(0, &backtrace_flag);
 	smp_mb__after_atomic();
 	put_cpu();
 }
 
+/*
+ * It is not safe to call printk() directly from NMI handlers.
+ * It may be fine if the NMI detected a lock up and we have no choice
+ * but to do so, but doing a NMI on all other CPUs to get a back trace
+ * can be done with a sysrq-l. We don't want that to lock up, which
+ * can happen if the NMI interrupts a printk in progress.
+ *
+ * Instead, we redirect the vprintk() to this nmi_vprintk() that writes
+ * the content into a per cpu seq_buf buffer. Then when the NMIs are
+ * all done, we can safely dump the contents of the seq_buf to a printk()
+ * from a non NMI context.
+ */
+static int nmi_vprintk(const char *fmt, va_list args)
+{
+	struct nmi_seq_buf *s = this_cpu_ptr(&nmi_print_seq);
+	unsigned int len = seq_buf_used(&s->seq);
+
+	seq_buf_vprintf(&s->seq, fmt, args);
+	return seq_buf_used(&s->seq) - len;
+}
+
 static int
 arch_trigger_all_cpu_backtrace_handler(unsigned int cmd, struct pt_regs *regs)
 {
@@ -78,12 +157,14 @@ arch_trigger_all_cpu_backtrace_handler(unsigned int cmd, struct pt_regs *regs)
 	cpu = smp_processor_id();
 
 	if (cpumask_test_cpu(cpu, to_cpumask(backtrace_mask))) {
-		static arch_spinlock_t lock = __ARCH_SPIN_LOCK_UNLOCKED;
+		printk_func_t printk_func_save = this_cpu_read(printk_func);
 
-		arch_spin_lock(&lock);
+		/* Replace printk to write into the NMI seq */
+		this_cpu_write(printk_func, nmi_vprintk);
 		printk(KERN_WARNING "NMI backtrace for cpu %d\n", cpu);
 		show_regs(regs);
-		arch_spin_unlock(&lock);
+		this_cpu_write(printk_func, printk_func_save);
+
 		cpumask_clear_cpu(cpu, to_cpumask(backtrace_mask));
 		return NMI_HANDLED;
 	}