Hi,

On 03/06/2018 10:07 AM, Morten Rasmussen wrote:
On Tue, Feb 27, 2018 at 02:18:47PM -0600, Jeremy Linton wrote:
Hi,


First, thanks for taking a look at this.

On 03/01/2018 09:52 AM, Morten Rasmussen wrote:
Hi Jeremy,

On Wed, Feb 28, 2018 at 04:06:19PM -0600, Jeremy Linton wrote:
Now that we have an accurate view of the physical topology
we need to represent it correctly to the scheduler. In the
case of NUMA in socket, we need to assure that the sched domain
we build for the MC layer isn't larger than the DIE above it.

MC shouldn't be larger than any of the NUMA domains either.

Right, that is one of the things this patch is assuring..


To do this correctly, we should really base that on the cache
topology immediately below the NUMA node (for NUMA in socket) >> or below the 
physical package for normal NUMA configurations.

That means we wouldn't support multi-die NUMA nodes?

You mean a bottom level NUMA domain that crosses multiple sockets/dies? That
should work. This patch is picking the widest cache layer below the smallest
of the package or numa grouping. What actually happens depends on the
topology. Given a case where there are multiple dies in a socket, and the
numa domain is at the socket level the MC is going to reflect the caching
topology immediately below the socket. In the case of multiple dies, with a
cache that crosses them in socket, then the MC is basically going to be the
socket, otherwise if the widest cache is per die, or some narrower grouping
(cluster?) then that is what ends up in the MC. (this is easier with some
pictures)

That is more or less what I meant. I think I got confused with the role
of "DIE" level, i.e. that top non-NUMA level, in this. The DIE level
cpumask spans exactly the NUMA node, so IIUC we have three scenarios:

1. Multi-die/socket/physical package NUMA node
    Top non-NUMA level (DIE) spans multiple packages. Bottom NUMA level
    spans multiple multi-package nodes. The MC mask reflects the last-level
    cache within the NUMA node which is most likely per-die or per-cluster
    (inside each die).

2. physical package == NUMA node
    The top non-NUMA (DIE) mask is the same as the core sibling mask.
    If there is cache spanning the entire node, the scheduler topology
    will eliminate a layer (DIE?), so bottom NUMA level would be right on
    top of MC spanning multiple physical packages. If there is no
    node-wide last level cache, DIE is preserved and MC matches the span
    of the last level cache.

3. numa-in-package
    Top non-NUMA (DIE) mask is not reflecting the actual die, but is
    reflecting the NUMA node. MC has a span equal to the largest share
    cache span smaller than or equal to the the NUMA node. If it is
    equal, DIE level is eliminated, otherwise DIE is preserved, but
    doesn't really represent die. Bottom non-NUMA level spans multiple
    in-package NUMA nodes.

As you said, multi-die nodes should work. However, I'm not sure if
shrinking MC to match a cache could cause us trouble, or if it should
just be shrunk to be the smaller of the node mask and core siblings.

Shrinking to the smaller of the numa or package is fairly trivial change, I'm good with that change too.. I discounted it because there might be an advantage in case 2 if the internal hardware is actually a case 3 (or just multiple rings/whatever each with a L3). In those cases the firmware vendor could play around with whatever representation serves them the best.

Unless you have a node-wide last level cache DIE level won't be
eliminated in scenario 2 and 3, and could cause trouble. For
numa-in-package, you can end up with a DIE level inside the node where
the default flags don't favour aggressive spreading of tasks. The same
could be the case for per-package nodes (scenario 2).

Don't we end up redefining physical package to be last level cache
instead of using the PPTT flag for scenario 2 and 3?

I'm not sure I understand, core_siblings isn't changing (its still per package). Only the MC mapping which normally is just core_siblings. For all intents right now this patch is the same as v6, except for the numa-in-package where the MC domain is being shrunk to the node siblings. I'm just trying to setup the code for potential future cases where the LLC isn't equal to the node or package.


I think DIE level should be eliminated for scenario 2 and 3 like it is
for x86.

Ok, that is based on the assumption that MC will always be equal to either the package or node? If that assumption isn't true, then would you keep it, or maybe it doesn't matter?


[...]

This patch creates a set of early cache_siblings masks, then
when the scheduler requests the coregroup mask we pick the
smaller of the physical package siblings, or the numa siblings
and locate the largest cache which is an entire subset of
those siblings. If we are unable to find a proper subset of
cores then we retain the original behavior and return the
core_sibling list.

IIUC, for numa-in-package it is a strict requirement that there is a
cache that span the entire NUMA node? For example, having a NUMA node
consisting of two clusters with per-cluster caches only wouldn't be
supported?

Everything is supported, the MC is reflecting the cache topology. We just
use the physical/numa topology to help us pick which layer of cache topology
lands in the MC. (unless of course we fail to find a PPTT/cache topology, in
which case we fallback to the old behavior of the core_siblings which can
reflect the MPIDR/etc).

I see. For this example we would end up with a "DIE" level and two MC
domains inside each node whether we have the PPTT table and cache
topology or not. I'm just wondering if everyone would be happy with
basing MC on last level cache instead of the smaller of physical package
and NUMA node.

I can't judge that, my idea was simply to provide some flexibility to the firmware. I guess in theory someone who still wanted that split could push a numa domain down to whatever level they wanted to group.



+{
+       /* first determine if we are a NUMA in package */
+       const cpumask_t *node_mask = cpumask_of_node(cpu_to_node(cpu));
+       int indx;
+
+       if (!cpumask_subset(node_mask, &cpu_topology[cpu].core_sibling)) {
+               /* not numa in package, lets use the package siblings */
+               node_mask = &cpu_topology[cpu].core_sibling;
+       }
+
+       /*
+        * node_mask should represent the smallest package/numa grouping
+        * lets search for the largest cache smaller than the node_mask.
+        */
+       for (indx = 0; indx < MAX_CACHE_CHECKS; indx++) {
+               cpumask_t *cache_sibs = &cpu_topology[cpu].cache_siblings[indx];
+
+               if (cpu_topology[cpu].cache_id[indx] < 0)
+                       continue;
+
+               if (cpumask_subset(cache_sibs, node_mask))
+                       cpu_topology[cpu].cache_level = indx;

I don't this guarantees that the cache level we found matches exactly
the NUMA node. Taking the two cluster NUMA node example from above, we
would set cache_level to point at the per-cluster cache as it is a
subset of the NUMA node but it would only span half of the node. Or am I
missing something?

I think you got it. If the system is a traditional ARM system with shared
L2's at the cluster level and it doesn't have any L3's/etc and the NUMA node
crosses multiple clusters then you get the cluster L2 grouping in the MC.

I think this is what we want. Particularly, since the newer/larger machines
do have L3+'s contained within their sockets or numa domains, so you end up
with that as the MC.

Okay, thanks for confirming.




+       }
+}
+
  const struct cpumask *cpu_coregroup_mask(int cpu)
  {
+       int *llc = &cpu_topology[cpu].cache_level;
+
+       if (*llc == -1)
+               find_llc_topology_for_cpu(cpu);
+
+       if (*llc != -1)
+               return &cpu_topology[cpu].cache_siblings[*llc];
+
        return &cpu_topology[cpu].core_sibling;

If we don't have any of the cache_sibling masks set up, i.e. we don't
have the cache topology, we would keep looking for it every time
cpu_coregroup_mask() is called. I'm not sure how extensively it is used,
but it could have a performance impact?

Its only called when cores come online/offline (AFAIK).



  }
@@ -221,6 +255,7 @@ static void update_siblings_masks(unsigned int cpuid)
  {
        struct cpu_topology *cpu_topo, *cpuid_topo = &cpu_topology[cpuid];
        int cpu;
+       int idx;
        /* update core and thread sibling masks */
        for_each_possible_cpu(cpu) {
@@ -229,6 +264,16 @@ static void update_siblings_masks(unsigned int cpuid)
                if (cpuid_topo->package_id != cpu_topo->package_id)
                        continue;
+               for (idx = 0; idx < MAX_CACHE_CHECKS; idx++) {
+                       cpumask_t *lsib;
+                       int cput_id = cpuid_topo->cache_id[idx];
+
+                       if (cput_id == cpu_topo->cache_id[idx]) {
+                               lsib = &cpuid_topo->cache_siblings[idx];
+                               cpumask_set_cpu(cpu, lsib);
+                       }

Shouldn't the cache_id validity be checked here? I don't think it breaks
anything though.

It could be, but since its explicitly looking for unified caches its likely
that some of the levels are invalid. Invalid levels get ignored later on so
we don't really care if they are valid here.


Overall, I think this is more or less in line with the MC domain
shrinking I just mentioned in the v6 discussion. It is mostly the corner
cases and assumption about the system topology I'm not sure about.

I think its the corner cases i'm taking care of. The simple fix in v6 is to
take the smaller of core_siblings or node_siblings, but that ignores cases
with split L3s (or the L2 only example above). The idea here is to assure
that MC is following a cache topology. In my mind, it is more a question of
how that is picked. The other way I see to do this, is with a PX domain flag
in the PPTT. We could then pick the core grouping one below that flag. Doing
it that way affords the firmware vendors a lever they can pull to optimize a
given machine for the linux scheduler behavior.

Okay. I think these assumptions/choices should be spelled out somewhere,
either as comments or in the commit message. As said above, I'm not sure
if the simple approach is better or not.

Using the cache span to define the MC level with a numa-in-cluster
switch like some Intel platform seems to have, you could two core being
MC siblings with numa-in-package disabled and them not being siblings
with numa-in-package enabled unless you reconfigure the span of the
caches too and remember to update the ACPI cache topology.

Regarding firmware levers, we don't want vendors to optimize for Linux
scheduler behaviour, but a mechanism to detect how closely related cores
are could make selecting the right mask for MC level easier. As I see
it, we basically have to choose between MC being cache boundary based or
physical package based. This patch implements the former, the simple
solution (core_siblings mask or node_siblings mask) implements the
latter.

Basically, right now (AFAIK) the result is the same because the few machines I have access to have cache layers immediately below those boundaries which are the same size as the package/die.

I'm ok with tossing this patch in favor of something like:

const struct cpumask *cpu_coregroup_mask(int cpu)
{
   const cpumask_t *node_mask = cpumask_of_node(cpu_to_node(cpu));
   if (!cpumask_subset(node_mask, &cpu_topology[cpu].core_sibling))         
   {
      /* not numa in package, lets use the package siblings */
      return &cpu_topology[cpu].core_sibling;
   }
   return node_mask;
}


Mostly, because I want to merge the PPTT parts, and I only added this to clear the NUMA in package borken....

Reply via email to