Re: [PATCH] powerpc/pseries: hibernation/migration should honor topology update policy

2015-10-02 Thread Tyrel Datwyler
Ping?

Don't thing I ever saw anything happen with this patch. On another note
I notice in retrospect the patch subject isn't really stated in
imperative mood. Maybe it would be better as:

powerpc/pseries: make hibernation/migration honor topology update policy

-Tyrel

On 05/05/2015 10:53 AM, Tyrel Datwyler wrote:
> From: Tyrel Datwyler 
> 
> The suspend call paths for hibernation and migration operations call
> stop_topology_update() and start_topology_update() respectively prior to
> suspending the LPAR and upon resume. Topology updating can be
> enabled/disabled from userspace and no check is currently done to determine
> the current policy. This results in topology updates being started upon
> resume from hibernation/migration even in the case where topology updates
> were disabled initially.
> 
> This fixes the issue by storing the current policy and only calling
> start_topology_update() in the case where either PRRN/VPHN were enabled to
> start with.
> 
> Fixes: e04fa61214a3 (powerpc/pseries: Add /proc interface to control topology 
> updates)
> 
> Signed-off-by: Tyrel Datwyler 
> Cc: Nathan Fontenot 
> Cc: Nishanth Aravamudan 
> Cc: sta...@vger.kernel.org
> ---
>  arch/powerpc/include/asm/topology.h  | 5 +
>  arch/powerpc/kernel/rtas.c   | 4 +++-
>  arch/powerpc/mm/numa.c   | 5 +
>  arch/powerpc/platforms/pseries/suspend.c | 5 -
>  4 files changed, 17 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/powerpc/include/asm/topology.h 
> b/arch/powerpc/include/asm/topology.h
> index 5f1048e..44f6519 100644
> --- a/arch/powerpc/include/asm/topology.h
> +++ b/arch/powerpc/include/asm/topology.h
> @@ -63,6 +63,7 @@ static inline void sysfs_remove_device_from_node(struct 
> device *dev,
>  extern int start_topology_update(void);
>  extern int stop_topology_update(void);
>  extern int prrn_is_enabled(void);
> +extern int vphn_is_enabled(void);
>  #else
>  static inline int start_topology_update(void)
>  {
> @@ -76,6 +77,10 @@ static inline int prrn_is_enabled(void)
>  {
>   return 0;
>  }
> +static inline int vphn_is_enabled(void)
> +{
> + return 0;
> +}
>  #endif /* CONFIG_NUMA && CONFIG_PPC_SPLPAR */
> 
>  #include 
> diff --git a/arch/powerpc/kernel/rtas.c b/arch/powerpc/kernel/rtas.c
> index 7a488c1..7ae9992 100644
> --- a/arch/powerpc/kernel/rtas.c
> +++ b/arch/powerpc/kernel/rtas.c
> @@ -906,6 +906,7 @@ int rtas_ibm_suspend_me(u64 handle)
>   DECLARE_COMPLETION_ONSTACK(done);
>   cpumask_var_t offline_mask;
>   int cpuret;
> + int restart_topology_updates = (prrn_is_enabled() || vphn_is_enabled());
> 
>   if (!rtas_service_present("ibm,suspend-me"))
>   return -ENOSYS;
> @@ -957,7 +958,8 @@ int rtas_ibm_suspend_me(u64 handle)
>   if (atomic_read(&data.error) != 0)
>   printk(KERN_ERR "Error doing global join\n");
> 
> - start_topology_update();
> + if (restart_topology_updates)
> + start_topology_update();
> 
>   /* Take down CPUs not online prior to suspend */
>   cpuret = rtas_offline_cpus_mask(offline_mask);
> diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c
> index 5e80621..885d950 100644
> --- a/arch/powerpc/mm/numa.c
> +++ b/arch/powerpc/mm/numa.c
> @@ -1593,6 +1593,11 @@ int prrn_is_enabled(void)
>   return prrn_enabled;
>  }
> 
> +int vphn_is_enabled(void)
> +{
> + return vphn_enabled;
> +}
> +
>  static int topology_read(struct seq_file *file, void *v)
>  {
>   if (vphn_enabled || prrn_enabled)
> diff --git a/arch/powerpc/platforms/pseries/suspend.c 
> b/arch/powerpc/platforms/pseries/suspend.c
> index e76aefa..b5f92e2 100644
> --- a/arch/powerpc/platforms/pseries/suspend.c
> +++ b/arch/powerpc/platforms/pseries/suspend.c
> @@ -147,6 +147,7 @@ static ssize_t store_hibernate(struct device *dev,
>  {
>   cpumask_var_t offline_mask;
>   int rc;
> + int restart_topology_updates = (prrn_is_enabled() || vphn_is_enabled);
> 
>   if (!capable(CAP_SYS_ADMIN))
>   return -EPERM;
> @@ -175,7 +176,9 @@ static ssize_t store_hibernate(struct device *dev,
> 
>   stop_topology_update();
>   rc = pm_suspend(PM_SUSPEND_MEM);
> - start_topology_update();
> +
> + if (restart_topology_updates)
> + start_topology_update();
> 
>   /* Take down CPUs not online prior to suspend */
>   if (!rtas_offline_cpus_mask(offline_mask))
> 

___
Linuxppc-dev mailing list
Linuxppc-dev@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/linuxppc-dev

Re: [PATCH] powerpc/pseries: hibernation/migration should honor topology update policy

2015-05-05 Thread Nathan Fontenot
On 05/05/2015 12:53 PM, Tyrel Datwyler wrote:
> From: Tyrel Datwyler 
> 
> The suspend call paths for hibernation and migration operations call
> stop_topology_update() and start_topology_update() respectively prior to
> suspending the LPAR and upon resume. Topology updating can be
> enabled/disabled from userspace and no check is currently done to determine
> the current policy. This results in topology updates being started upon
> resume from hibernation/migration even in the case where topology updates
> were disabled initially.
> 
> This fixes the issue by storing the current policy and only calling
> start_topology_update() in the case where either PRRN/VPHN were enabled to
> start with.
> 
> Fixes: e04fa61214a3 (powerpc/pseries: Add /proc interface to control topology 
> updates)
> 
> Signed-off-by: Tyrel Datwyler 
> Cc: Nathan Fontenot 
> Cc: Nishanth Aravamudan 
> Cc: sta...@vger.kernel.org

Reviewed-by: Nathan Fontenot 

___
Linuxppc-dev mailing list
Linuxppc-dev@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/linuxppc-dev