linux/drivers/base
Darren Hart db1e59483d topology: make core_mask include at least cluster_siblings
Ampere Altra defines CPU clusters in the ACPI PPTT. They share a Snoop
Control Unit, but have no shared CPU-side last level cache.

cpu_coregroup_mask() will return a cpumask with weight 1, while
cpu_clustergroup_mask() will return a cpumask with weight 2.

As a result, build_sched_domain() will BUG() once per CPU with:

BUG: arch topology borken
the CLS domain not a subset of the MC domain

The MC level cpumask is then extended to that of the CLS child, and is
later removed entirely as redundant. This sched domain topology is an
improvement over previous topologies, or those built without
SCHED_CLUSTER, particularly for certain latency sensitive workloads.
With the current scheduler model and heuristics, this is a desirable
default topology for Ampere Altra and Altra Max system.

Rather than create a custom sched domains topology structure and
introduce new logic in arch/arm64 to detect these systems, update the
core_mask so coregroup is never a subset of clustergroup, extending it
to cluster_siblings if necessary. Only do this if CONFIG_SCHED_CLUSTER
is enabled to avoid also changing the topology (MC) when
CONFIG_SCHED_CLUSTER is disabled.

This has the added benefit over a custom topology of working for both
symmetric and asymmetric topologies. It does not address systems where
the CLUSTER topology is above a populated MC topology, but these are not
considered today and can be addressed separately if and when they
appear.

The final sched domain topology for a 2 socket Ampere Altra system is
unchanged with or without CONFIG_SCHED_CLUSTER, and the BUG is avoided:

For CPU0:

CONFIG_SCHED_CLUSTER=y
CLS  [0-1]
DIE  [0-79]
NUMA [0-159]

CONFIG_SCHED_CLUSTER is not set
DIE  [0-79]
NUMA [0-159]

Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rafael@kernel.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Vincent Guittot <vincent.guittot@linaro.org>
Cc: D. Scott Phillips <scott@os.amperecomputing.com>
Cc: Ilkka Koskinen <ilkka@os.amperecomputing.com>
Cc: <stable@vger.kernel.org> # 5.16.x
Suggested-by: Barry Song <song.bao.hua@hisilicon.com>
Reviewed-by: Barry Song <song.bao.hua@hisilicon.com>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Acked-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Darren Hart <darren@os.amperecomputing.com>
Link: https://lore.kernel.org/r/c8fe9fce7c86ed56b4c455b8c902982dc2303868.1649696956.git.darren@os.amperecomputing.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2022-04-20 18:57:20 +02:00
..
firmware_loader firmware_loader: EXTRA_FIRMWARE does not support compressed files 2022-02-25 12:14:42 +01:00
power Merge branches 'pm-sleep', 'pm-domains' and 'pm-docs' 2022-03-18 18:29:21 +01:00
regmap regmap: Updates for v5.18 2022-03-21 18:22:42 -07:00
test driver core: Simplify async probe test code by using ktime_ms_delta() 2021-12-29 10:57:22 +01:00
arch_numa.c mm: percpu: add generic pcpu_populate_pte() function 2022-01-20 08:52:52 +02:00
arch_topology.c topology: make core_mask include at least cluster_siblings 2022-04-20 18:57:20 +02:00
attribute_container.c driver core: attribute_container: fix W=1 warnings 2021-05-14 13:37:10 +02:00
auxiliary.c Documentation/auxiliary_bus: Move the text into the code 2021-12-03 16:41:50 +01:00
base.h software nodes: Split software_node_notify() 2021-07-16 19:17:05 +02:00
bus.c kobject: remove kset from struct kset_uevent_ops callbacks 2021-12-28 11:26:18 +01:00
cacheinfo.c cacheinfo: clear cache_leaves(cpu) in free_cache_attributes() 2021-07-21 17:29:40 +02:00
class.c block: remove genhd.h 2022-02-02 07:49:59 -07:00
component.c component: Add common helper for compare/release functions 2022-02-25 12:16:12 +01:00
container.c
core.c block: remove genhd.h 2022-02-02 07:49:59 -07:00
cpu.c sched/isolation: Use single feature type while referring to housekeeping cpumask 2022-02-16 15:57:55 +01:00
dd.c drivers/base/dd.c : Remove the initial value of the global variable 2022-03-18 14:31:31 +01:00
devcoredump.c devcoredump: remove contact information 2021-06-04 15:05:44 +02:00
devres.c devres: fix typos in comments 2022-03-18 14:30:12 +01:00
devtmpfs.c Driver core changes for 5.18-rc1 2022-03-28 12:41:28 -07:00
driver.c drivers: base: Convert to printk alias functions 2020-07-10 14:16:44 +02:00
firmware.c
hypervisor.c
init.c drivers/base/node: consolidate node device subsystem initialization in node_dev_init() 2022-03-22 15:57:10 -07:00
isa.c bus: Make remove callback return void 2021-07-21 11:53:42 +02:00
Kconfig devtmpfs: mount with noexec and nosuid 2021-12-30 13:54:42 +01:00
Makefile mm/memory_hotplug: remove CONFIG_MEMORY_HOTPLUG_SPARSE 2021-11-06 13:30:42 -07:00
map.c driver: base: Prefer unsigned int to bare use of unsigned 2021-07-21 17:30:09 +02:00
memory.c drivers/base/memory: clarify adding and removing of memory blocks 2022-03-22 15:57:10 -07:00
module.c
node.c drivers/base/memory: determine and store zone for single-zone memory blocks 2022-03-22 15:57:10 -07:00
pinctrl.c
platform-msi.c platform-msi: Simplify platform device MSI code 2021-12-16 22:22:19 +01:00
platform.c platform: use dev_err_probe() in platform_get_irq_byname() 2022-02-08 12:22:07 +01:00
property.c device property: Don't split fwnode_get_irq*() APIs in the code 2022-02-14 18:57:41 +01:00
soc.c base: soc: Make soc_device_match() simpler and easier to read 2022-03-18 14:28:07 +01:00
swnode.c software node: fix wrong node passed to find nargs_prop 2021-12-22 18:26:18 +01:00
syscore.c syscore: Use pm_pr_dbg() for syscore_{suspend,resume}() 2020-09-08 13:32:06 +02:00
topology.c topology/sysfs: Hide PPIN on systems that do not support it. 2022-04-20 18:56:39 +02:00
trace.c devres: Enable trace events 2021-06-15 17:14:36 +02:00
trace.h devres: Enable trace events 2021-06-15 17:14:36 +02:00
transport_class.c scsi: drivers: base: Propagate errors through the transport component 2020-01-15 22:55:37 -05:00