On Power systems with shared configurations of CPUs and memory, there
are some issues with association of additional CPUs and memory to nodes
when hot-adding resources. These patches address some of those problems.
powerpc/numa: Correct the currently broken capability to set the
topology for shared CPUs in LPARs. At boot time for shared CPU
lpars, the topology for each shared CPU is set to node zero, however,
this is now updated correctly using the Virtual Processor Home Node
(VPHN) capabilities information provided by the pHyp. The VPHN handling
in Linux is disabled, if PRRN handling is present.
powerpc/nodes: On systems like PowerPC which allow 'hot-add' of CPU
or memory resources, it may occur that the new resources are to be
inserted into nodes that were not used for these resources at bootup.
In the kernel, any node that is used must be defined and initialized
at boot.
This patch extracts the value of the lowest domain level (number of
allocable resources) from the "rtas" device tree property
"ibm,max-associativity-domains" to use as the maximum number of nodes
to setup as possibly available in the system. This new setting will
override the instruction,
nodes_and(node_possible_map, node_possible_map, node_online_map);
presently seen in the function arch/powerpc/mm/numa.c:initmem_init().
If the property is not present at boot, no operation will be performed
to define or enable additional nodes.
Signed-off-by: Michael Bringmann <[email protected]>
Michael Bringmann (2):
powerpc/numa: Update CPU topology when VPHN enabled
powerpc/nodes: Ensure enough nodes avail for operations
---
Changes in V9:
-- Calculate number of nodes via property "ibm,max-associativity-domains"
powerpc/numa: Correct the currently broken capability to set the
topology for shared CPUs in LPARs. At boot time for shared CPU
lpars, the topology for each shared CPU is set to node zero, however,
this is now updated correctly using the Virtual Processor Home Node
(VPHN) capabilities information provided by the pHyp.
Also, update initialization checks for device-tree attributes to
independently recognize PRRN or VPHN usage.
Signed-off-by: Michael Bringmann <[email protected]>
---
arch/powerpc/include/asm/topology.h | 14 ++++++
arch/powerpc/mm/numa.c | 64 +++++++++++++++++++++++---
arch/powerpc/platforms/pseries/dlpar.c | 2 +
arch/powerpc/platforms/pseries/hotplug-cpu.c | 2 +
4 files changed, 75 insertions(+), 7 deletions(-)
diff --git a/arch/powerpc/include/asm/topology.h b/arch/powerpc/include/asm/topology.h
index dc4e159..85d6428 100644
--- a/arch/powerpc/include/asm/topology.h
+++ b/arch/powerpc/include/asm/topology.h
@@ -98,6 +98,20 @@ static inline int prrn_is_enabled(void)
}
#endif /* CONFIG_NUMA && CONFIG_PPC_SPLPAR */
+#if defined(CONFIG_HOTPLUG_CPU) || defined(CONFIG_NEED_MULTIPLE_NODES)
+#if defined(CONFIG_PPC_SPLPAR)
+extern int timed_topology_update(int nsecs);
+#else
+#define timed_topology_update(nsecs) 0
+#endif /* CONFIG_PPC_SPLPAR */
+#endif /* CONFIG_HOTPLUG_CPU || CONFIG_NEED_MULTIPLE_NODES */
+
+#if defined(CONFIG_PPC_SPLPAR)
+extern void shared_topology_update(void);
+#else
+#define shared_topology_update() 0
+#endif /* CONFIG_PPC_SPLPAR */
+
#include <asm-generic/topology.h>
#ifdef CONFIG_SMP
diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c
index b95c584..3fd4536 100644
--- a/arch/powerpc/mm/numa.c
+++ b/arch/powerpc/mm/numa.c
@@ -29,6 +29,7 @@
#include <linux/seq_file.h>
#include <linux/uaccess.h>
#include <linux/slab.h>
+#include <linux/sched.h>
#include <asm/cputhreads.h>
#include <asm/sparsemem.h>
#include <asm/prom.h>
@@ -906,7 +907,7 @@ void __init initmem_init(void)
/*
* Reduce the possible NUMA nodes to the online NUMA nodes,
- * since we do not support node hotplug. This ensures that we
+ * since we do not support node hotplug. This ensures that we
* lower the maximum NUMA node ID to what is actually present.
*/
nodes_and(node_possible_map, node_possible_map, node_online_map);
@@ -1148,11 +1149,32 @@ struct topology_update_data {
int new_nid;
};
+#define TOPOLOGY_DEF_TIMER_SECS 60
+
static u8 vphn_cpu_change_counts[NR_CPUS][MAX_DISTANCE_REF_POINTS];
static cpumask_t cpu_associativity_changes_mask;
static int vphn_enabled;
static int prrn_enabled;
static void reset_topology_timer(void);
+static int topology_timer_secs = TOPOLOGY_DEF_TIMER_SECS;
+static int topology_inited;
+static int topology_update_needed;
+
+/*
+ * Change polling interval for associativity changes.
+ */
+int timed_topology_update(int nsecs)
+{
+ if (nsecs > 0)
+ topology_timer_secs = nsecs;
+ else
+ topology_timer_secs = TOPOLOGY_DEF_TIMER_SECS;
+
+ if (vphn_enabled)
+ reset_topology_timer();
+
+ return 0;
+}
/*
* Store the current values of the associativity change counters in the
@@ -1246,6 +1268,12 @@ static long vphn_get_associativity(unsigned long cpu,
"hcall_vphn() experienced a hardware fault "
"preventing VPHN. Disabling polling...\n");
stop_topology_update();
+ break;
+ case H_SUCCESS:
+ printk(KERN_INFO
+ "VPHN hcall succeeded. Reset polling...\n");
+ timed_topology_update(0);
+ break;
}
return rc;
@@ -1323,8 +1351,11 @@ int numa_update_cpu_topology(bool cpus_locked)
struct device *dev;
int weight, new_nid, i = 0;
- if (!prrn_enabled && !vphn_enabled)
+ if (!prrn_enabled && !vphn_enabled) {
+ if (!topology_inited)
+ topology_update_needed = 1;
return 0;
+ }
weight = cpumask_weight(&cpu_associativity_changes_mask);
if (!weight)
@@ -1363,6 +1394,8 @@ int numa_update_cpu_topology(bool cpus_locked)
cpumask_andnot(&cpu_associativity_changes_mask,
&cpu_associativity_changes_mask,
cpu_sibling_mask(cpu));
+ pr_info("Assoc chg gives same node %d for cpu%d\n",
+ new_nid, cpu);
cpu = cpu_last_thread_sibling(cpu);
continue;
}
@@ -1379,6 +1412,9 @@ int numa_update_cpu_topology(bool cpus_locked)
cpu = cpu_last_thread_sibling(cpu);
}
+ if (i)
+ updates[i-1].next = NULL;
+
pr_debug("Topology update for the following CPUs:\n");
if (cpumask_weight(&updated_cpus)) {
for (ud = &updates[0]; ud; ud = ud->next) {
@@ -1433,6 +1469,7 @@ int numa_update_cpu_topology(bool cpus_locked)
out:
kfree(updates);
+ topology_update_needed = 0;
return changed;
}
@@ -1453,6 +1490,14 @@ static void topology_schedule_update(void)
schedule_work(&topology_work);
}
+void shared_topology_update(void)
+{
+ if (firmware_has_feature(FW_FEATURE_VPHN) &&
+ lppaca_shared_proc(get_lppaca()))
+ topology_schedule_update();
+}
+EXPORT_SYMBOL(shared_topology_update);
+
static void topology_timer_fn(unsigned long ignored)
{
if (prrn_enabled && cpumask_weight(&cpu_associativity_changes_mask))
@@ -1469,7 +1514,7 @@ static void topology_timer_fn(unsigned long ignored)
static void reset_topology_timer(void)
{
topology_timer.data = 0;
- topology_timer.expires = jiffies + 60 * HZ;
+ topology_timer.expires = jiffies + topology_timer_secs * HZ;
mod_timer(&topology_timer, topology_timer.expires);
}
@@ -1519,15 +1564,14 @@ int start_topology_update(void)
if (firmware_has_feature(FW_FEATURE_PRRN)) {
if (!prrn_enabled) {
prrn_enabled = 1;
- vphn_enabled = 0;
#ifdef CONFIG_SMP
rc = of_reconfig_notifier_register(&dt_update_nb);
#endif
}
- } else if (firmware_has_feature(FW_FEATURE_VPHN) &&
+ }
+ if (firmware_has_feature(FW_FEATURE_VPHN) &&
lppaca_shared_proc(get_lppaca())) {
if (!vphn_enabled) {
- prrn_enabled = 0;
vphn_enabled = 1;
setup_cpu_associativity_change_counters();
init_timer_deferrable(&topology_timer);
@@ -1550,7 +1594,8 @@ int stop_topology_update(void)
#ifdef CONFIG_SMP
rc = of_reconfig_notifier_unregister(&dt_update_nb);
#endif
- } else if (vphn_enabled) {
+ }
+ if (vphn_enabled) {
vphn_enabled = 0;
rc = del_timer_sync(&topology_timer);
}
@@ -1616,6 +1661,11 @@ static int topology_update_init(void)
if (!proc_create("powerpc/topology_updates", 0644, NULL, &topology_ops))
return -ENOMEM;
+ topology_inited = 1;
+ if (topology_update_needed)
+ bitmap_fill(cpumask_bits(&cpu_associativity_changes_mask),
+ nr_cpumask_bits);
+
return 0;
}
device_initcall(topology_update_init);
diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c
index 3918769..ba9a4a0 100644
--- a/arch/powerpc/platforms/pseries/dlpar.c
+++ b/arch/powerpc/platforms/pseries/dlpar.c
@@ -592,6 +592,8 @@ static ssize_t dlpar_show(struct class *class, struct class_attribute *attr,
static int __init pseries_dlpar_init(void)
{
+ shared_topology_update();
+
pseries_hp_wq = alloc_workqueue("pseries hotplug workqueue",
WQ_UNBOUND, 1);
return sysfs_create_file(kernel_kobj, &class_attr_dlpar.attr);
diff --git a/arch/powerpc/platforms/pseries/hotplug-cpu.c b/arch/powerpc/platforms/pseries/hotplug-cpu.c
index 6afd1ef..5a7fb1e 100644
--- a/arch/powerpc/platforms/pseries/hotplug-cpu.c
+++ b/arch/powerpc/platforms/pseries/hotplug-cpu.c
@@ -356,6 +356,7 @@ static int dlpar_online_cpu(struct device_node *dn)
BUG_ON(get_cpu_current_state(cpu)
!= CPU_STATE_OFFLINE);
cpu_maps_update_done();
+ timed_topology_update(1);
rc = device_online(get_cpu_device(cpu));
if (rc)
goto out;
@@ -522,6 +523,7 @@ static int dlpar_offline_cpu(struct device_node *dn)
set_preferred_offline_state(cpu,
CPU_STATE_OFFLINE);
cpu_maps_update_done();
+ timed_topology_update(1);
rc = device_offline(get_cpu_device(cpu));
if (rc)
goto out;
Michael Bringmann <[email protected]> writes:
> powerpc/numa: Correct the currently broken capability to set the
> topology for shared CPUs in LPARs. At boot time for shared CPU
> lpars, the topology for each shared CPU is set to node zero, however,
> this is now updated correctly using the Virtual Processor Home Node
> (VPHN) capabilities information provided by the pHyp.
>
> Also, update initialization checks for device-tree attributes to
> independently recognize PRRN or VPHN usage.
Did you ever do anything to address Nathan's comments on v4 ?
http://patchwork.ozlabs.org/patch/767587/
Also your change log doesn't describe anything about what the patch does
and why it is the correct fix for the problem.
When a DLPAR happens you modify the VPHN timer to run in 1 nsec, but you
don't wait for it. Why would we not just run the logic synchronously?
It also seems to make VPHN and PRRN no longer exclusive, which looking
at PAPR seems like it might be correct, but is also a major change so
please justify it in detail.
Comments below.
> diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c
> index b95c584..3fd4536 100644
> --- a/arch/powerpc/mm/numa.c
> +++ b/arch/powerpc/mm/numa.c
> @@ -906,7 +907,7 @@ void __init initmem_init(void)
>
> /*
> * Reduce the possible NUMA nodes to the online NUMA nodes,
> - * since we do not support node hotplug. This ensures that we
> + * since we do not support node hotplug. This ensures that we
Please do whitespace/spelling changes in a separate patch.
> * lower the maximum NUMA node ID to what is actually present.
> */
> nodes_and(node_possible_map, node_possible_map, node_online_map);
> @@ -1148,11 +1149,32 @@ struct topology_update_data {
> int new_nid;
> };
>
> +#define TOPOLOGY_DEF_TIMER_SECS 60
> +
> static u8 vphn_cpu_change_counts[NR_CPUS][MAX_DISTANCE_REF_POINTS];
> static cpumask_t cpu_associativity_changes_mask;
> static int vphn_enabled;
> static int prrn_enabled;
> static void reset_topology_timer(void);
> +static int topology_timer_secs = TOPOLOGY_DEF_TIMER_SECS;
> +static int topology_inited;
> +static int topology_update_needed;
None of this code should be in numa.c. Which is not your fault but I'm
inclined to move it before we make it worse.
> +
> +/*
> + * Change polling interval for associativity changes.
> + */
> +int timed_topology_update(int nsecs)
> +{
> + if (nsecs > 0)
> + topology_timer_secs = nsecs;
> + else
> + topology_timer_secs = TOPOLOGY_DEF_TIMER_SECS;
> +
> + if (vphn_enabled)
> + reset_topology_timer();
> +
> + return 0;
> +}
>
> /*
> * Store the current values of the associativity change counters in the
> @@ -1246,6 +1268,12 @@ static long vphn_get_associativity(unsigned long cpu,
> "hcall_vphn() experienced a hardware fault "
> "preventing VPHN. Disabling polling...\n");
> stop_topology_update();
> + break;
> + case H_SUCCESS:
> + printk(KERN_INFO
> + "VPHN hcall succeeded. Reset polling...\n");
We don't need that to hit everyone's console once a minute. Remove it or
pr_debug() if you like.
> @@ -1363,6 +1394,8 @@ int numa_update_cpu_topology(bool cpus_locked)
> cpumask_andnot(&cpu_associativity_changes_mask,
> &cpu_associativity_changes_mask,
> cpu_sibling_mask(cpu));
> + pr_info("Assoc chg gives same node %d for cpu%d\n",
> + new_nid, cpu);
No thanks.
> @@ -1379,6 +1412,9 @@ int numa_update_cpu_topology(bool cpus_locked)
> cpu = cpu_last_thread_sibling(cpu);
> }
>
> + if (i)
> + updates[i-1].next = NULL;
???
> @@ -1453,6 +1490,14 @@ static void topology_schedule_update(void)
> schedule_work(&topology_work);
> }
>
> +void shared_topology_update(void)
> +{
> + if (firmware_has_feature(FW_FEATURE_VPHN) &&
> + lppaca_shared_proc(get_lppaca()))
> + topology_schedule_update();
> +}
> +EXPORT_SYMBOL(shared_topology_update);
There's no reason for that to be exported AFAICS.
> diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c
> index 3918769..ba9a4a0 100644
> --- a/arch/powerpc/platforms/pseries/dlpar.c
> +++ b/arch/powerpc/platforms/pseries/dlpar.c
> @@ -592,6 +592,8 @@ static ssize_t dlpar_show(struct class *class, struct class_attribute *attr,
>
> static int __init pseries_dlpar_init(void)
> {
> + shared_topology_update();
> +
I don't see any reason to call that from here.
It could just as easily be a machine init call in the file where it lives.
cheers
On 08/23/2017 06:41 AM, Michael Ellerman wrote:
> Michael Bringmann <[email protected]> writes:
>
>> powerpc/numa: Correct the currently broken capability to set the
>> topology for shared CPUs in LPARs. At boot time for shared CPU
>> lpars, the topology for each shared CPU is set to node zero, however,
>> this is now updated correctly using the Virtual Processor Home Node
>> (VPHN) capabilities information provided by the pHyp.
>>
>> Also, update initialization checks for device-tree attributes to
>> independently recognize PRRN or VPHN usage.
>
> Did you ever do anything to address Nathan's comments on v4 ?
>
> http://patchwork.ozlabs.org/patch/767587/
Looking at this patch I do not see that VPHN is always enabled.
>
>
> Also your change log doesn't describe anything about what the patch does
> and why it is the correct fix for the problem.
>
> When a DLPAR happens you modify the VPHN timer to run in 1 nsec, but you
> don't wait for it. Why would we not just run the logic synchronously?
>
> It also seems to make VPHN and PRRN no longer exclusive, which looking
> at PAPR seems like it might be correct, but is also a major change so
> please justify it in detail.
This is correct, they are not exclusive. When we first added PRRN support
we mistakenly thought they were exclusive which is why the code currently
only starts PRRN, or VPHN if PRRN is not present.
>
> Comments below.
>
>
>> diff --git a/arch/powerpc/mm/numa.c b/arch/powerpc/mm/numa.c
>> index b95c584..3fd4536 100644
>> --- a/arch/powerpc/mm/numa.c
>> +++ b/arch/powerpc/mm/numa.c
>> @@ -906,7 +907,7 @@ void __init initmem_init(void)
>>
>> /*
>> * Reduce the possible NUMA nodes to the online NUMA nodes,
>> - * since we do not support node hotplug. This ensures that we
>> + * since we do not support node hotplug. This ensures that we
>
> Please do whitespace/spelling changes in a separate patch.
>
>> * lower the maximum NUMA node ID to what is actually present.
>> */
>> nodes_and(node_possible_map, node_possible_map, node_online_map);
>> @@ -1148,11 +1149,32 @@ struct topology_update_data {
>> int new_nid;
>> };
>>
>> +#define TOPOLOGY_DEF_TIMER_SECS 60
>> +
>> static u8 vphn_cpu_change_counts[NR_CPUS][MAX_DISTANCE_REF_POINTS];
>> static cpumask_t cpu_associativity_changes_mask;
>> static int vphn_enabled;
>> static int prrn_enabled;
>> static void reset_topology_timer(void);
>> +static int topology_timer_secs = TOPOLOGY_DEF_TIMER_SECS;
>> +static int topology_inited;
>> +static int topology_update_needed;
>
> None of this code should be in numa.c. Which is not your fault but I'm
> inclined to move it before we make it worse.
Agreed. Perhaps this should all be in mm/vphn.c
-Nathan
>
>> +
>> +/*
>> + * Change polling interval for associativity changes.
>> + */
>> +int timed_topology_update(int nsecs)
>> +{
>> + if (nsecs > 0)
>> + topology_timer_secs = nsecs;
>> + else
>> + topology_timer_secs = TOPOLOGY_DEF_TIMER_SECS;
>> +
>> + if (vphn_enabled)
>> + reset_topology_timer();
>> +
>> + return 0;
>> +}
>>
>> /*
>> * Store the current values of the associativity change counters in the
>> @@ -1246,6 +1268,12 @@ static long vphn_get_associativity(unsigned long cpu,
>> "hcall_vphn() experienced a hardware fault "
>> "preventing VPHN. Disabling polling...\n");
>> stop_topology_update();
>> + break;
>> + case H_SUCCESS:
>> + printk(KERN_INFO
>> + "VPHN hcall succeeded. Reset polling...\n");
>
> We don't need that to hit everyone's console once a minute. Remove it or
> pr_debug() if you like.
>
>> @@ -1363,6 +1394,8 @@ int numa_update_cpu_topology(bool cpus_locked)
>> cpumask_andnot(&cpu_associativity_changes_mask,
>> &cpu_associativity_changes_mask,
>> cpu_sibling_mask(cpu));
>> + pr_info("Assoc chg gives same node %d for cpu%d\n",
>> + new_nid, cpu);
>
> No thanks.
>
>> @@ -1379,6 +1412,9 @@ int numa_update_cpu_topology(bool cpus_locked)
>> cpu = cpu_last_thread_sibling(cpu);
>> }
>>
>> + if (i)
>> + updates[i-1].next = NULL;
>
> ???
>
>> @@ -1453,6 +1490,14 @@ static void topology_schedule_update(void)
>> schedule_work(&topology_work);
>> }
>>
>> +void shared_topology_update(void)
>> +{
>> + if (firmware_has_feature(FW_FEATURE_VPHN) &&
>> + lppaca_shared_proc(get_lppaca()))
>> + topology_schedule_update();
>> +}
>> +EXPORT_SYMBOL(shared_topology_update);
>
> There's no reason for that to be exported AFAICS.
>
>> diff --git a/arch/powerpc/platforms/pseries/dlpar.c b/arch/powerpc/platforms/pseries/dlpar.c
>> index 3918769..ba9a4a0 100644
>> --- a/arch/powerpc/platforms/pseries/dlpar.c
>> +++ b/arch/powerpc/platforms/pseries/dlpar.c
>> @@ -592,6 +592,8 @@ static ssize_t dlpar_show(struct class *class, struct class_attribute *attr,
>>
>> static int __init pseries_dlpar_init(void)
>> {
>> + shared_topology_update();
>> +
>
> I don't see any reason to call that from here.
>
> It could just as easily be a machine init call in the file where it lives.
>
>
> cheers
>
Nathan Fontenot <[email protected]> writes:
> On 08/23/2017 06:41 AM, Michael Ellerman wrote:
>> Michael Bringmann <[email protected]> writes:
>>
>>> powerpc/numa: Correct the currently broken capability to set the
>>> topology for shared CPUs in LPARs. At boot time for shared CPU
>>> lpars, the topology for each shared CPU is set to node zero, however,
>>> this is now updated correctly using the Virtual Processor Home Node
>>> (VPHN) capabilities information provided by the pHyp.
>>>
>>> Also, update initialization checks for device-tree attributes to
>>> independently recognize PRRN or VPHN usage.
>>
>> Did you ever do anything to address Nathan's comments on v4 ?
>>
>> http://patchwork.ozlabs.org/patch/767587/
>
> Looking at this patch I do not see that VPHN is always enabled.
You mean *this* patch? Or v4?
I think you mean this patch, in which case I agree.
>> Also your change log doesn't describe anything about what the patch does
>> and why it is the correct fix for the problem.
>>
>> When a DLPAR happens you modify the VPHN timer to run in 1 nsec, but you
>> don't wait for it. Why would we not just run the logic synchronously?
>>
>> It also seems to make VPHN and PRRN no longer exclusive, which looking
>> at PAPR seems like it might be correct, but is also a major change so
>> please justify it in detail.
>
> This is correct, they are not exclusive. When we first added PRRN support
> we mistakenly thought they were exclusive which is why the code currently
> only starts PRRN, or VPHN if PRRN is not present.
OK.
So we need a patch that does that and only that, and clearly explains
why we're doing that and why it's the correct thing to do.
Then a 2nd patch can fiddle with the timer, if we must.
...
>>> +static int topology_timer_secs = TOPOLOGY_DEF_TIMER_SECS;
>>> +static int topology_inited;
>>> +static int topology_update_needed;
>>
>> None of this code should be in numa.c. Which is not your fault but I'm
>> inclined to move it before we make it worse.
>
> Agreed. Perhaps this should all be in mm/vphn.c
Actually I was thinking platforms/pseries/vphn.c
cheers