[RFC PATCH 2/3] sched/topology: Define and assign sched_domain flag metadata

[Valentin Schneider]

There are some expectations regarding how sched domain flags should be laid
out, but none of them are checked or asserted in
sched_domain_debug_one(). After staring at said flags for a while, I've
come to realize they all (except *one*) fall in either of two categories:

- Shared with children: those flags are set from the base CPU domain
  upwards. Any domain that has it set will have it set in its children. It
  hints at "some property holds true / some behaviour is enabled until this
  level".

- Shared with parents: those flags are set from the topmost domain
  downwards. Any domain that has it set will have it set in its parents. It
  hints at "some property isn't visible / some behaviour is disabled until
  this level".

The odd one out is SD_PREFER_SIBLING, which is cleared below levels with
SD_ASYM_CPUCAPACITY. The change was introduced by commit

  9c63e84db29b ("sched/core: Disable SD_PREFER_SIBLING on asymmetric CPU capacity domains")

as it could break misfit migration on some systems. In light of this, we
might want to change it back to make it fit one of the two categories and
fix the issue another way.

Tweak the sched_domain flag declaration to assign each flag an expected
layout, and include the rationale for each flag "meta type" assignment as a
comment. Consolidate the flag metadata into an array; the index of a flag's
metadata can easily be found with log2(flag), IOW __ffs(flag).

Signed-off-by: Valentin Schneider <valentin.schneider@xxxxxxx>
---
 include/linux/sched/sd_flags.h | 162 +++++++++++++++++++++++++++------
 include/linux/sched/topology.h |  14 ++-
 2 files changed, 147 insertions(+), 29 deletions(-)

diff --git a/include/linux/sched/sd_flags.h b/include/linux/sched/sd_flags.h
index 685bbe736945..1082e4796ffe 100644
--- a/include/linux/sched/sd_flags.h
+++ b/include/linux/sched/sd_flags.h
@@ -3,31 +3,137 @@
  * sched-domains (multiprocessor balancing) flag declarations.
  */
 
-/* Balance when about to become idle */
-SD_FLAG(SD_BALANCE_NEWIDLE,     0)
-/* Balance on exec */
-SD_FLAG(SD_BALANCE_EXEC,        1)
-/* Balance on fork, clone */
-SD_FLAG(SD_BALANCE_FORK,        2)
-/* Balance on wakeup */
-SD_FLAG(SD_BALANCE_WAKE,        3)
-/* Wake task to waking CPU */
-SD_FLAG(SD_WAKE_AFFINE,         4)
-/* Domain members have different CPU capacities */
-SD_FLAG(SD_ASYM_CPUCAPACITY,    5)
-/* Domain members share CPU capacity */
-SD_FLAG(SD_SHARE_CPUCAPACITY,   6)
-/* Domain members share power domain */
-SD_FLAG(SD_SHARE_POWERDOMAIN,   7)
-/* Domain members share CPU pkg resources */
-SD_FLAG(SD_SHARE_PKG_RESOURCES, 8)
-/* Only a single load balancing instance */
-SD_FLAG(SD_SERIALIZE,           9)
-/* Place busy groups earlier in the domain */
-SD_FLAG(SD_ASYM_PACKING,        10)
-/* Prefer to place tasks in a sibling domain */
-SD_FLAG(SD_PREFER_SIBLING,      11)
-/* sched_domains of this level overlap */
-SD_FLAG(SD_OVERLAP,             12)
-/* cross-node balancing */
-SD_FLAG(SD_NUMA,                13)
+#ifndef SD_FLAG
+#define SD_FLAG(x, y, z)
+#endif
+
+/*
+ * Expected flag uses
+ *
+ * SHARED_CHILD: These flags are meant to be set from the base domain upwards.
+ * If a domain has this flag set, all of its children should have it set. This
+ * is usually because the flag describes some shared resource, or because they
+ * are tied to a scheduling behaviour that we want to disable at some point in
+ * the hierarchy for scalability reasons.
+ *
+ * In those cases it doesn't make sense to not have the flag set for a domain
+ * but not have it in (some of) its children: (non-overlapping) sched domains
+ * ALWAYS span their child domains, so operations done with those parent domains
+ * would cover the CPUs in the lower domains that did not have this flag anyway.
+ *
+ *
+ * SHARED_PARENT: These flags are meant to be set from the highest domain
+ * downwards. If a domain has this flag set, all of its parents should have it
+ * set. This is usually for topology properties that start to appear above a
+ * certain level (e.g. domain starts spanning CPUs outside of the base CPU's
+ * socket), which translates to scheduler behaviours that should start taking
+ * effect from a given level.
+ */
+#define SDF_SHARED_CHILD 0x1
+#define SDF_SHARED_PARENT 0x2
+
+/*
+ * Balance when about to become idle
+ *
+ * SHARED_CHILD: Can be set from cpuset.sched_relax_domain_level downwards.
+ */
+SD_FLAG(SD_BALANCE_NEWIDLE,     0, SDF_SHARED_CHILD)
+
+/*
+ * Balance on exec
+ *
+ * SHARED_CHILD: Can be set from NUMA reclaim level downwards.
+ */
+SD_FLAG(SD_BALANCE_EXEC,        1, SDF_SHARED_CHILD)
+
+/*
+ * Balance on fork, clone
+ *
+ * SHARED_CHILD: Can be set from NUMA reclaim level downwards.
+ */
+SD_FLAG(SD_BALANCE_FORK,        2, SDF_SHARED_CHILD)
+
+/*
+ *  Balance on wakeup
+ *
+ * SHARED_CHILD: Can be set from cpuset.sched_relax_domain_level downwards.
+ */
+SD_FLAG(SD_BALANCE_WAKE,        3, SDF_SHARED_CHILD)
+
+/*
+ * Wake task to waking CPU
+ *
+ * SHARED_CHILD: Only for levels below NUMA reclaim distance
+ */
+SD_FLAG(SD_WAKE_AFFINE,         4, SDF_SHARED_CHILD)
+
+/*
+ * Domain members have different CPU capacities
+ *
+ * SHARED_PARENT: CPU capacity asymmetry is visible from a given level
+ * upwards.
+ */
+SD_FLAG(SD_ASYM_CPUCAPACITY,    5, SDF_SHARED_PARENT)
+
+/*
+ * Domain members share CPU capacity
+ *
+ * SHARED_CHILD: CPU capacity is shared from the base domain up to a given level.
+ */
+SD_FLAG(SD_SHARE_CPUCAPACITY,   6, SDF_SHARED_CHILD)
+
+/*
+ * Domain members share power domain
+ *
+ * SHARED_CHILD: Power domains are shared from the base domain up to a given
+ * level.
+ */
+SD_FLAG(SD_SHARE_POWERDOMAIN,   7, SDF_SHARED_CHILD)
+
+/*
+ * Domain members share CPU package resources (i.e. caches)
+ *
+ * SHARED_CHILD: Caches are shared from the base domain up to a given level.
+ */
+SD_FLAG(SD_SHARE_PKG_RESOURCES, 8, SDF_SHARED_CHILD)
+
+/*
+ * Only a single load balancing instance
+ *
+ * SHARED_PARENT: Set for all NUMA levels (incl. NODE), which sit atop
+ * "regular" levels (those built from sched_domain_topology). Could be set
+ * differently, but this is still a property that needs to be propagated to
+ * parents: it doesn't make sense to have it e.g. in the first few levels and
+ * not higher up.
+ */
+SD_FLAG(SD_SERIALIZE,           9, SDF_SHARED_PARENT)
+
+/*
+ * Place busy groups earlier in the domain
+ *
+ * SHARED_CHILD: Set for SMT. Technically could be set further up, but currently
+ * assumed to be set from the base domain upwards (see update_top_cache_domain()).
+ */
+SD_FLAG(SD_ASYM_PACKING,        10, SDF_SHARED_CHILD)
+
+/*
+ * Prefer to place tasks in a sibling domain
+ *
+ * NONE: Set up until domains start spanning NUMA. Close to being a
+ * SHARED_CHILD flag, but cleared below domains with SD_ASYM_CPUCAPACITY.
+ */
+SD_FLAG(SD_PREFER_SIBLING,      11, 0)
+
+/*
+ * sched_domains of this level overlap
+ *
+ * SHARED_PARENT: Set for all NUMA levels above NODE.
+ */
+SD_FLAG(SD_OVERLAP,             12, SDF_SHARED_PARENT)
+
+/*
+ * cross-node balancing
+ *
+ * SHARED_PARENT: Set for all NUMA levels above NODE.
+ */
+SD_FLAG(SD_NUMA,                13, SDF_SHARED_PARENT)
diff --git a/include/linux/sched/topology.h b/include/linux/sched/topology.h
index db7d24c0174b..cdb71ee6f53a 100644
--- a/include/linux/sched/topology.h
+++ b/include/linux/sched/topology.h
@@ -11,10 +11,22 @@
  */
 #ifdef CONFIG_SMP
 
-#define SD_FLAG(name, idx) static const unsigned int name = BIT(idx);
+#define SD_FLAG(name, idx, type) static const unsigned int name = BIT(idx);
 #include <linux/sched/sd_flags.h>
 #undef SD_FLAG
 
+#ifdef CONFIG_SCHED_DEBUG
+#define SD_FLAG(_name, idx, _meta_flags) [idx] = {.meta_flags = _meta_flags, .name = #_name},
+
+static const struct {
+	unsigned int meta_flags;
+	char *name;
+} sd_flag_debug[] = {
+#include <linux/sched/sd_flags.h>
+};
+#undef SD_FLAG
+#endif
+
 #ifdef CONFIG_SCHED_SMT
 static inline int cpu_smt_flags(void)
 {
-- 
2.24.0