From: Steven Rostedt <rostedt@goodmis.org>
To: linux-kernel@vger.kernel.org
Cc: Ingo Molnar <mingo@kernel.org>,
Andrew Morton <akpm@linux-foundation.org>,
Jiri Kosina <jkosina@suse.cz>, Petr Mladek <pmladek@suse.cz>,
"Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Subject: [PATCH 26/26 v5] x86/nmi: Perform a safe NMI stack trace on all CPUs
Date: Fri, 14 Nov 2014 23:59:13 -0500 [thread overview]
Message-ID: <20141115050605.055232587@goodmis.org> (raw)
In-Reply-To: 20141115045847.712848224@goodmis.org
[-- Attachment #1: 0026-x86-nmi-Perform-a-safe-NMI-stack-trace-on-all-CPUs.patch --]
[-- Type: text/plain, Size: 5340 bytes --]
From: "Steven Rostedt (Red Hat)" <rostedt@goodmis.org>
When trigger_all_cpu_backtrace() is called on x86, it will trigger an
NMI on each CPU and call show_regs(). But this can lead to a hard lock
up if the NMI comes in on another printk().
In order to avoid this, when the NMI triggers, it switches the printk
routine for that CPU to call a NMI safe printk function that records the
printk in a per_cpu seq_buf descriptor. After all NMIs have finished
recording its data, the seq_bufs are printed in a safe context.
Link: http://lkml.kernel.org/p/20140619213952.360076309@goodmis.org
Tested-by: Jiri Kosina <jkosina@suse.cz>
Acked-by: Jiri Kosina <jkosina@suse.cz>
Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
---
arch/x86/kernel/apic/hw_nmi.c | 91 ++++++++++++++++++++++++++++++++++++++++---
1 file 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..c95c3e9ce196 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_var_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;
}
--
2.1.1
next prev parent reply other threads:[~2014-11-15 5:07 UTC|newest]
Thread overview: 70+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-11-15 4:58 [PATCH 00/26 v5] trace-seq/seq-buf/x86/printk: Print all stacks from NMI safely Steven Rostedt
2014-11-15 4:58 ` [PATCH 01/26 v5] tracing: Fix trace_seq_bitmask() to start at current position Steven Rostedt
2014-11-15 4:58 ` [PATCH 02/26 v5] tracing: Add trace_seq_has_overflowed() and trace_handle_return() Steven Rostedt
2014-11-15 4:58 ` [PATCH 03/26 v5] blktrace/tracing: Use trace_seq_has_overflowed() helper function Steven Rostedt
2014-11-15 4:58 ` [PATCH 04/26 v5] ring-buffer: Remove check of trace_seq_{puts,printf}() return values Steven Rostedt
2014-11-15 4:58 ` [PATCH 05/26 v5] tracing: Have branch tracer use trace_handle_return() helper function Steven Rostedt
2014-11-15 4:58 ` [PATCH 06/26 v5] tracing: Have function_graph use trace_seq_has_overflowed() Steven Rostedt
2014-11-15 4:58 ` [PATCH 07/26 v5] kprobes/tracing: Use trace_seq_has_overflowed() for overflow checks Steven Rostedt
2014-11-18 14:02 ` Petr Mladek
2014-11-15 4:58 ` [PATCH 08/26 v5] tracing: Do not check return values of trace_seq_p*() for mmio tracer Steven Rostedt
2014-11-18 14:06 ` Petr Mladek
2014-11-15 4:58 ` [PATCH 09/26 v5] tracing/probes: Do not use return value of trace_seq_printf() Steven Rostedt
2014-11-18 14:09 ` Petr Mladek
2014-11-15 4:58 ` [PATCH 10/26 v5] tracing/uprobes: Do not use return values " Steven Rostedt
2014-11-17 5:28 ` Masami Hiramatsu
2014-11-17 5:58 ` Srikar Dronamraju
2014-11-18 14:13 ` Petr Mladek
2014-11-15 4:58 ` [PATCH 11/26 v5] tracing: Do not use return values of trace_seq_printf() in syscall tracing Steven Rostedt
2014-11-15 4:58 ` [PATCH 12/26 v5] tracing: Remove return values of most trace_seq_*() functions Steven Rostedt
2014-11-18 14:18 ` Petr Mladek
2014-11-15 4:59 ` [PATCH 13/26 v5] tracing: Fix return value of ftrace_raw_output_prep() Steven Rostedt
2014-11-18 14:24 ` Petr Mladek
2014-11-15 4:59 ` [PATCH 14/26 v5] tracing: Create seq_buf layer in trace_seq Steven Rostedt
2014-11-19 14:51 ` Petr Mladek
2014-11-19 15:08 ` Steven Rostedt
2014-11-15 4:59 ` [PATCH 15/26 v5] tracing: Convert seq_buf_path() to be like seq_path() Steven Rostedt
2014-11-15 4:59 ` [PATCH 16/26 v5] tracing: Convert seq_buf fields to be like seq_file fields Steven Rostedt
2014-11-15 4:59 ` [PATCH 17/26 v5] tracing: Add a seq_buf_clear() helper and clear len and readpos in init Steven Rostedt
2014-11-15 4:59 ` [PATCH 18/26 v5] seq_buf: Create seq_buf_used() to find out how much was written Steven Rostedt
2014-11-18 15:02 ` Petr Mladek
2014-11-19 15:49 ` Steven Rostedt
2014-11-19 16:30 ` Petr Mladek
2014-11-15 4:59 ` [PATCH 19/26 v5] tracing: Use trace_seq_used() and seq_buf_used() instead of len Steven Rostedt
2014-11-17 17:32 ` Steven Rostedt
2014-11-17 19:11 ` [PATCH 1/2] tracing: Clean up tracing_fill_pipe_page() Steven Rostedt
2014-11-18 16:15 ` Petr Mladek
2014-11-19 16:17 ` Steven Rostedt
2014-11-19 16:51 ` Petr Mladek
2014-11-19 17:12 ` Steven Rostedt
2014-11-17 19:12 ` [PATCH 2/2] tracing: Use trace_seq_used() and seq_buf_used() instead of len Steven Rostedt
2014-11-18 16:33 ` Petr Mladek
2014-11-18 17:37 ` Steven Rostedt
2014-11-19 11:40 ` Petr Mladek
2014-11-19 13:48 ` Steven Rostedt
2014-11-19 14:40 ` Petr Mladek
2014-11-19 15:01 ` Steven Rostedt
2014-11-19 16:00 ` Steven Rostedt
2014-11-19 16:44 ` Petr Mladek
2014-11-15 4:59 ` [PATCH 20/26 v5] seq_buf: Add seq_buf_can_fit() helper function Steven Rostedt
2014-11-17 17:36 ` Steven Rostedt
2014-11-18 0:07 ` Joe Perches
2014-11-18 0:27 ` Steven Rostedt
2014-11-18 0:35 ` Joe Perches
2014-11-18 0:56 ` Steven Rostedt
2014-11-18 1:07 ` Joe Perches
2014-11-18 1:24 ` Steven Rostedt
2014-11-18 1:48 ` Joe Perches
2014-11-18 2:37 ` Steven Rostedt
2014-11-18 2:50 ` Joe Perches
2014-11-18 3:00 ` Steven Rostedt
2014-11-18 16:40 ` Petr Mladek
2014-11-15 4:59 ` [PATCH 21/26 v5] tracing: Have seq_buf use full buffer Steven Rostedt
2014-11-15 4:59 ` [PATCH 22/26 v5] tracing: Add seq_buf_get_buf() and seq_buf_commit() helper functions Steven Rostedt
2014-11-15 4:59 ` [PATCH 23/26 v5] seq-buf: Make seq_buf_bprintf() conditional on CONFIG_BINARY_PRINTF Steven Rostedt
2014-11-15 4:59 ` [PATCH 24/26 v5] seq_buf: Move the seq_buf code to lib/ Steven Rostedt
2014-11-15 4:59 ` [PATCH 25/26 v5] printk: Add per_cpu printk func to allow printk to be diverted Steven Rostedt
2014-11-15 4:59 ` Steven Rostedt [this message]
2014-11-18 17:02 ` [PATCH 26/26 v5] x86/nmi: Perform a safe NMI stack trace on all CPUs Petr Mladek
2014-11-15 5:08 ` [PATCH 00/26 v5] trace-seq/seq-buf/x86/printk: Print all stacks from NMI safely Steven Rostedt
2014-11-18 3:21 ` Steven Rostedt
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20141115050605.055232587@goodmis.org \
--to=rostedt@goodmis.org \
--cc=akpm@linux-foundation.org \
--cc=jkosina@suse.cz \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@kernel.org \
--cc=paulmck@linux.vnet.ibm.com \
--cc=pmladek@suse.cz \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).