[PATCH] RFC: ARM: Ensure backtraces console printing in sleep mode

[PATCH] RFC: ARM: Ensure backtraces console printing in sleep mode

[Linus Walleij]

From: Aurelien Gerault <aurelien.gerault-nonst@stericsson.com>

Ensuring that oops & panic messages/backtraces will be printed
in console, even if in sleep mode (i.e. when the console is
locked).

So, unlock the console before printing.

Signed-off-by: Aurelien Gerault <aurelien.gerault-nonst@stericsson.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
---
 arch/arm/kernel/traps.c |    4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arch/arm/kernel/traps.c b/arch/arm/kernel/traps.c
index 4928d89..42a5057 100644
--- a/arch/arm/kernel/traps.c
+++ b/arch/arm/kernel/traps.c
@@ -25,6 +25,7 @@
 #include <linux/delay.h>
 #include <linux/init.h>
 #include <linux/sched.h>
+#include <linux/console.h>
 
 #include <linux/atomic.h>
 #include <asm/cacheflush.h>
@@ -273,6 +274,9 @@ void die(const char *str, struct pt_regs *regs, int err)
 	int ret;
 	enum bug_trap_type bug_type = BUG_TRAP_TYPE_NONE;
 
+	/* ensure that backtrace will be printed in console */
+	console_unlock();
+
 	oops_enter();
 
 	raw_spin_lock_irq(&die_lock);
-- 
1.7.9.2

[PATCH] RFC: ARM: Ensure backtraces console printing in sleep mode

[Russell King - ARM Linux]

On Tue, Jun 05, 2012 at 09:28:41AM +0200, Linus Walleij wrote:
> From: Aurelien Gerault <aurelien.gerault-nonst@stericsson.com>
> 
> Ensuring that oops & panic messages/backtraces will be printed
> in console, even if in sleep mode (i.e. when the console is
> locked).
> 
> So, unlock the console before printing.

I don't think this is a good idea - it totally changes the locking, and
probably causes the console lock to become unbalanced, which could be a
problem if the oops was non-fatal to the kernel.

x86 also doesn't play this game either, and we really shouldn't have
different semantics here.