Re: [PATCH v4] panic: keep blinking in spite of long spin timer mode

[Andrew Morton <akpm@linux-foundation.org>]

On Mon, 21 Jun 2010 07:52:24 +0900
TAMUKI Shoichi <tamuki@linet.gr.jp> wrote:

> To keep panic_timeout accuracy when running under a hypervisor, the
> current implementation only spins on long time (1 second) calls to
> mdelay.  That brings a good effect, but the problem is the keyboard
> LEDs don't blink at all on that situation.
> 
> This patch changes to call to panic_blink_enter() between every mdelay
> and keeps blinking in spite of long spin timer mode.
> 
> The time to call to mdelay is now 100ms.  Even this change will keep
> panic_timeout accuracy enough when running under a hypervisor.
> 

Thaks for sticking with it - I think the code's a lot better now.  Are
you happy with it?

> ---
>  Changes since v3:
>  - get rid of CONFIG_PANIC_LONGSPIN_TIMER kernel config option
> 
>  Documentation/kernel-parameters.txt |    3 -
>  arch/arm/mach-s3c2440/mach-gta02.c  |   17 ++-----
>  drivers/input/serio/i8042.c         |   25 ++---------
>  include/linux/kernel.h              |    2 
>  kernel/panic.c                      |   58 +++++++++++---------------
>  5 files changed, 37 insertions(+), 68 deletions(-)

And isn't that a nice sight.

> --- a/drivers/input/serio/i8042.c
> +++ b/drivers/input/serio/i8042.c

Dmitry, I think drivers/input/serio/i8042.c is buggy.  It does

	panic_blink = NULL;

as the very last thing on module exit.  However it surely should do
this at least at the _start_ of i8042_exit() and I suspect it really
should be doing this at the start of i8042_shutdown()?

As things stand, a well-timed panic will end up trying to flash LEDs
via hardware which has been "shut down".