On (02/27/16 11:19), Sergey Senozhatsky wrote:
[..]
> > I think about a compromise. We should try to get the messages
> > out only when kdump is not enabled.
>
> can we zap_locks() if we are on
> nmi_panic()->panic()->console_flush_on_panic() path?
> console_flush_on_panic() is happening after we send out smp_send_stop().
can something like this do the trick?
====8<====
>From 4186873bb4574b4bbb227e7448d56599849de0bd Mon Sep 17 00:00:00 2001
From: Sergey Senozhatsky <sergey.senozhat...@gmail.com>
Subject: [PATCH] printk/nmi: restore printk_func in nmi_panic
When watchdog detects a hardlockup and calls nmi_panic() `printk_func'
must be restored via printk_nmi_exit() call, so panic() will be able
to flush nmi buf and show backtrace and panic message. We also better
explicitly ask nmi to printk_nmi_flush() in console_flush_on_panic(),
because it may be too late to rely on irq work.
Factor out __zap_locks(), and call it from console_flush_on_panic().
This is normally not needed, because logbuf_lock always comes with
IRQ disable/enable magic, but we can panic() from nmi.
Signed-off-by: Sergey Senozhatsky <sergey.senozhat...@gmail.com>
---
include/linux/kernel.h | 6 ++++--
kernel/printk/printk.c | 27 ++++++++++++++++-----------
2 files changed, 20 insertions(+), 13 deletions(-)
diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index f4fa2b2..3ee33d5 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -469,10 +469,12 @@ do {
\
cpu = raw_smp_processor_id(); \
old_cpu = atomic_cmpxchg(&panic_cpu, PANIC_CPU_INVALID, cpu); \
\
- if (old_cpu == PANIC_CPU_INVALID) \
+ if (old_cpu == PANIC_CPU_INVALID) { \
+ printk_nmi_exit(); \
panic(fmt, ##__VA_ARGS__); \
- else if (old_cpu != cpu) \
+ } else if (old_cpu != cpu) { \
nmi_panic_self_stop(regs); \
+ } \
} while (0)
/*
diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
index 9917f69..0a318ed 100644
--- a/kernel/printk/printk.c
+++ b/kernel/printk/printk.c
@@ -1462,6 +1462,15 @@ static void call_console_drivers(int level,
}
}
+static void __zap_locks(void)
+{
+ debug_locks_off();
+ /* If a crash is occurring, make sure we can't deadlock */
+ raw_spin_lock_init(&logbuf_lock);
+ /* And make sure that we print immediately */
+ sema_init(&console_sem, 1);
+}
+
/*
* Zap console related locks when oopsing.
* To leave time for slow consoles to print a full oops,
@@ -1477,11 +1486,7 @@ static void zap_locks(void)
oops_timestamp = jiffies;
- debug_locks_off();
- /* If a crash is occurring, make sure we can't deadlock */
- raw_spin_lock_init(&logbuf_lock);
- /* And make sure that we print immediately */
- sema_init(&console_sem, 1);
+ __zap_locks();
}
int printk_delay_msec __read_mostly;
@@ -2386,15 +2391,15 @@ void console_unblank(void)
*/
void console_flush_on_panic(void)
{
- /*
- * If someone else is holding the console lock, trylock will fail
- * and may_schedule may be set. Ignore and proceed to unlock so
- * that messages are flushed out. As this can be called from any
- * context and we don't want to get preempted while flushing,
- * ensure may_schedule is cleared.
+ __zap_locks();
+
+ /* As this can be called from any context and we don't want
+ * to get preempted while flushing, ensure may_schedule is
+ * cleared.
*/
console_trylock();
console_may_schedule = 0;
+ printk_nmi_flush();
console_unlock();
}
--
2.7.1
