linux-kernel - [PATCH] kernel/watchdog: flush all printk nmi buffers when hardlockup detected

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [thread-next>] [day] [month] [year] [list]

Message-ID: <158132813726.1980.17382047082627699898.stgit@buzz>
Date:   Mon, 10 Feb 2020 12:48:57 +0300
From:   Konstantin Khlebnikov <khlebnikov@...dex-team.ru>
To:     Petr Mladek <pmladek@...e.com>,
        Peter Zijlstra <peterz@...radead.org>,
        linux-kernel@...r.kernel.org, Steven Rostedt <rostedt@...dmis.org>,
        Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
        Andrew Morton <akpm@...ux-foundation.org>
Cc:     Dmitry Monakhov <dmtrmonakhov@...dex-team.ru>
Subject: [PATCH] kernel/watchdog: flush all printk nmi buffers when
 hardlockup detected

In NMI context printk() could save messages into per-cpu buffers and
schedule flush by irq_work when IRQ are unblocked. This means message
about hardlockup appears in kernel log only when/if lockup is gone.

Comment in irq_work_queue_on() states that remote IPI aren't NMI safe
thus printk() cannot schedule flush work to another cpu.

This patch adds simple atomic counter of detected hardlockups and
flushes all per-cpu printk buffers in context softlockup watchdog
at any other cpu when it sees changes of this counter.

Signed-off-by: Konstantin Khlebnikov <khlebnikov@...dex-team.ru>
---
 include/linux/nmi.h   |    1 +
 kernel/watchdog.c     |   22 ++++++++++++++++++++++
 kernel/watchdog_hld.c |    1 +
 3 files changed, 24 insertions(+)

diff --git a/include/linux/nmi.h b/include/linux/nmi.h
index 9003e29cde46..8406df72ae5a 100644
--- a/include/linux/nmi.h
+++ b/include/linux/nmi.h
@@ -84,6 +84,7 @@ static inline void reset_hung_task_detector(void) { }
 #if defined(CONFIG_HARDLOCKUP_DETECTOR)
 extern void hardlockup_detector_disable(void);
 extern unsigned int hardlockup_panic;
+extern atomic_t hardlockup_detected;
 #else
 static inline void hardlockup_detector_disable(void) {}
 #endif
diff --git a/kernel/watchdog.c b/kernel/watchdog.c
index b6b1f54a7837..9f5c68fababe 100644
--- a/kernel/watchdog.c
+++ b/kernel/watchdog.c
@@ -92,6 +92,26 @@ static int __init hardlockup_all_cpu_backtrace_setup(char *str)
 }
 __setup("hardlockup_all_cpu_backtrace=", hardlockup_all_cpu_backtrace_setup);
 # endif /* CONFIG_SMP */
+
+atomic_t hardlockup_detected = ATOMIC_INIT(0);
+
+static inline void flush_hardlockup_messages(void)
+{
+	static atomic_t flushed = ATOMIC_INIT(0);
+
+	/* flush messages from hard lockup detector */
+	if (atomic_read(&hardlockup_detected) != atomic_read(&flushed)) {
+		atomic_set(&flushed, atomic_read(&hardlockup_detected));
+		printk_safe_flush();
+	}
+}
+
+#else /* CONFIG_HARDLOCKUP_DETECTOR */
+
+static inline void flush_hardlockup_messages(void)
+{
+}
+
 #endif /* CONFIG_HARDLOCKUP_DETECTOR */
 
 /*
@@ -370,6 +390,8 @@ static enum hrtimer_restart watchdog_timer_fn(struct hrtimer *hrtimer)
 	/* kick the hardlockup detector */
 	watchdog_interrupt_count();
 
+	flush_hardlockup_messages();
+
 	/* kick the softlockup detector */
 	if (completion_done(this_cpu_ptr(&softlockup_completion))) {
 		reinit_completion(this_cpu_ptr(&softlockup_completion));
diff --git a/kernel/watchdog_hld.c b/kernel/watchdog_hld.c
index 247bf0b1582c..a546bc54f6ff 100644
--- a/kernel/watchdog_hld.c
+++ b/kernel/watchdog_hld.c
@@ -154,6 +154,7 @@ static void watchdog_overflow_callback(struct perf_event *event,
 
 		if (hardlockup_panic)
 			nmi_panic(regs, "Hard LOCKUP");
+		atomic_inc(&hardlockup_detected);
 
 		__this_cpu_write(hard_watchdog_warn, true);
 		return;