Re: [PATCH v2] reboot: Backup orderly_poweroff
From: Ingo Molnar
Date: Thu Jan 14 2016 - 04:05:38 EST
* Keerthy <j-keerthy@xxxxxx> wrote:
> orderly_poweroff is triggered when a graceful shutdown
> of system is desired. This may be used in many critical states of the
> kernel such as when subsystems detects conditions such as critical
> temperature conditions. However, in certain conditions in system
> boot up sequences like those in the middle of driver probes being
> initiated, userspace will be unable to power off the system in a clean
> manner and leaves the system in a critical state. In cases like these,
> the /sbin/poweroff will return success (having forked off to attempt
> powering off the system. However, the system overall will fail to
> completely poweroff (since other modules will be probed) and the system
> is still functional with no userspace (since that would have shut itself
> However, there is no clean way of detecting such failure of userspace
> powering off the system. In such scenarios, it is necessary for a backup
> workqueue to be able to force a shutdown of the system when orderly
> shutdown is not successful after a configurable time period.
> Signed-off-by: Keerthy <j-keerthy@xxxxxx>
> Suggested-by: Eduardo Valentin <edubezval@xxxxxxxxx>
> Reported-by: Nishanth Menon <nm@xxxxxx>
> Links to previous discussion can be found here:
> Boot tested on DRA7.
> changes in v2:
> * Changed #ifdef to #if CONFIG_SHUTDOWN_BACKUP_DELAY_MS
> arch/Kconfig | 7 +++++++
> kernel/reboot.c | 23 ++++++++++++++++++-----
> 2 files changed, 25 insertions(+), 5 deletions(-)
> Index: linux/arch/Kconfig
> --- linux.orig/arch/Kconfig 2016-01-11 15:26:07.732173131 +0530
> +++ linux/arch/Kconfig 2016-01-11 15:26:07.728173205 +0530
> @@ -37,6 +37,18 @@
> def_bool y
> depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI && !PPC64
> +config SHUTDOWN_BACKUP_DELAY_MS
> + int "Backup shutdown delay in milli-seconds"
> + default 0
> + help
> + The number of milliseconds to delay before backup workqueue
> + executes attempting to poweroff the system after the
> + orderly_poweroff function has failed to complete.
> + If set to 0, the backup workqueue is not active. The value
> + should be conservatively configured based on userspace latencies
> + expected for a given system.
I don't really understand this. In what circumstances can a reboot fail?
I think that is what should be fixed: a reboot should never fail, instead of
introducing some sort of fragile timeout based method.