Re: [PATCH 01/11] cgroup: move rstat pointers into struct of their own

[JP Kobryn <inwardvessel@gmail.com>]

On 2/20/25 8:53 AM, Yosry Ahmed wrote:
> On Mon, Feb 17, 2025 at 07:14:38PM -0800, JP Kobryn wrote:
>> The rstat infrastructure makes use of pointers for list management.
>> These pointers only exist as fields in the cgroup struct, so moving them
>> into their own struct will allow them to be used elsewhere. The base
>> stat entities are included with them for now.
>>
>> Signed-off-by: JP Kobryn <inwardvessel@gmail.com>
>> ---
>>   include/linux/cgroup-defs.h                   | 90 +-----------------
>>   include/linux/cgroup_rstat.h                  | 92 +++++++++++++++++++
>>   kernel/cgroup/cgroup.c                        |  3 +-
>>   kernel/cgroup/rstat.c                         | 27 +++---
>>   .../selftests/bpf/progs/btf_type_tag_percpu.c |  4 +-
>>   5 files changed, 112 insertions(+), 104 deletions(-)
>>   create mode 100644 include/linux/cgroup_rstat.h
>>
>> diff --git a/include/linux/cgroup-defs.h b/include/linux/cgroup-defs.h
>> index 1b20d2d8ef7c..6b6cc027fe70 100644
>> --- a/include/linux/cgroup-defs.h
>> +++ b/include/linux/cgroup-defs.h
>> @@ -17,7 +17,7 @@
>>   #include <linux/refcount.h>
>>   #include <linux/percpu-refcount.h>
>>   #include <linux/percpu-rwsem.h>
>> -#include <linux/u64_stats_sync.h>
>> +#include <linux/cgroup_rstat.h>
>>   #include <linux/workqueue.h>
>>   #include <linux/bpf-cgroup-defs.h>
>>   #include <linux/psi_types.h>
>> @@ -321,78 +321,6 @@ struct css_set {
>>   	struct rcu_head rcu_head;
>>   };
>>   
>> -struct cgroup_base_stat {
>> -	struct task_cputime cputime;
>> -
>> -#ifdef CONFIG_SCHED_CORE
>> -	u64 forceidle_sum;
>> -#endif
>> -	u64 ntime;
>> -};
>> -
>> -/*
>> - * rstat - cgroup scalable recursive statistics.  Accounting is done
>> - * per-cpu in cgroup_rstat_cpu which is then lazily propagated up the
>> - * hierarchy on reads.
>> - *
>> - * When a stat gets updated, the cgroup_rstat_cpu and its ancestors are
>> - * linked into the updated tree.  On the following read, propagation only
>> - * considers and consumes the updated tree.  This makes reading O(the
>> - * number of descendants which have been active since last read) instead of
>> - * O(the total number of descendants).
>> - *
>> - * This is important because there can be a lot of (draining) cgroups which
>> - * aren't active and stat may be read frequently.  The combination can
>> - * become very expensive.  By propagating selectively, increasing reading
>> - * frequency decreases the cost of each read.
>> - *
>> - * This struct hosts both the fields which implement the above -
>> - * updated_children and updated_next - and the fields which track basic
>> - * resource statistics on top of it - bsync, bstat and last_bstat.
>> - */
>> -struct cgroup_rstat_cpu {
>> -	/*
>> -	 * ->bsync protects ->bstat.  These are the only fields which get
>> -	 * updated in the hot path.
>> -	 */
>> -	struct u64_stats_sync bsync;
>> -	struct cgroup_base_stat bstat;
>> -
>> -	/*
>> -	 * Snapshots at the last reading.  These are used to calculate the
>> -	 * deltas to propagate to the global counters.
>> -	 */
>> -	struct cgroup_base_stat last_bstat;
>> -
>> -	/*
>> -	 * This field is used to record the cumulative per-cpu time of
>> -	 * the cgroup and its descendants. Currently it can be read via
>> -	 * eBPF/drgn etc, and we are still trying to determine how to
>> -	 * expose it in the cgroupfs interface.
>> -	 */
>> -	struct cgroup_base_stat subtree_bstat;
>> -
>> -	/*
>> -	 * Snapshots at the last reading. These are used to calculate the
>> -	 * deltas to propagate to the per-cpu subtree_bstat.
>> -	 */
>> -	struct cgroup_base_stat last_subtree_bstat;
>> -
>> -	/*
>> -	 * Child cgroups with stat updates on this cpu since the last read
>> -	 * are linked on the parent's ->updated_children through
>> -	 * ->updated_next.
>> -	 *
>> -	 * In addition to being more compact, singly-linked list pointing
>> -	 * to the cgroup makes it unnecessary for each per-cpu struct to
>> -	 * point back to the associated cgroup.
>> -	 *
>> -	 * Protected by per-cpu cgroup_rstat_cpu_lock.
>> -	 */
>> -	struct cgroup *updated_children;	/* terminated by self cgroup */
>> -	struct cgroup *updated_next;		/* NULL iff not on the list */
>> -};
>> -
>>   struct cgroup_freezer_state {
>>   	/* Should the cgroup and its descendants be frozen. */
>>   	bool freeze;
>> @@ -517,23 +445,9 @@ struct cgroup {
>>   	struct cgroup *old_dom_cgrp;		/* used while enabling threaded */
>>   
>>   	/* per-cpu recursive resource statistics */
>> -	struct cgroup_rstat_cpu __percpu *rstat_cpu;
>> +	struct cgroup_rstat rstat;
>>   	struct list_head rstat_css_list;
>>   
>> -	/*
>> -	 * Add padding to separate the read mostly rstat_cpu and
>> -	 * rstat_css_list into a different cacheline from the following
>> -	 * rstat_flush_next and *bstat fields which can have frequent updates.
>> -	 */
>> -	CACHELINE_PADDING(_pad_);
>> -
>> -	/*
>> -	 * A singly-linked list of cgroup structures to be rstat flushed.
>> -	 * This is a scratch field to be used exclusively by
>> -	 * cgroup_rstat_flush_locked() and protected by cgroup_rstat_lock.
>> -	 */
>> -	struct cgroup	*rstat_flush_next;
>> -
>>   	/* cgroup basic resource statistics */
>>   	struct cgroup_base_stat last_bstat;
>>   	struct cgroup_base_stat bstat;
>> diff --git a/include/linux/cgroup_rstat.h b/include/linux/cgroup_rstat.h
>> new file mode 100644
>> index 000000000000..f95474d6f8ab
>> --- /dev/null
>> +++ b/include/linux/cgroup_rstat.h
>> @@ -0,0 +1,92 @@
>> +/* SPDX-License-Identifier: GPL-2.0 */
>> +#ifndef _LINUX_RSTAT_H
>> +#define _LINUX_RSTAT_H
>> +
>> +#include <linux/u64_stats_sync.h>
>> +
>> +struct cgroup_rstat_cpu;
> 
> Why do we need the forward declaration instead of just defining struct
> cgroup_rstat_cpu first? Also, why do we need a new header for these
> definitions rather than just adding struct cgroup_rstat to
> cgroup-defs.h?

The new header was added so the cgroup_rstat type can be used in bpf
cgroup-defs.h. As for the forward declaration, this was done so that
updated_next and updated children fields of the cgroup_rstat_cpu can
change type from from cgroup to cgroup_rstat.

Regardless, based on the direction we are moving with bpf sharing the
"self" tree, this new header will NOT be needed in v2.

> 
>> +
>> +/*
>> + * rstat - cgroup scalable recursive statistics.  Accounting is done
>> + * per-cpu in cgroup_rstat_cpu which is then lazily propagated up the
>> + * hierarchy on reads.
>> + *
>> + * When a stat gets updated, the cgroup_rstat_cpu and its ancestors are
>> + * linked into the updated tree.  On the following read, propagation only
>> + * considers and consumes the updated tree.  This makes reading O(the
>> + * number of descendants which have been active since last read) instead of
>> + * O(the total number of descendants).
>> + *
>> + * This is important because there can be a lot of (draining) cgroups which
>> + * aren't active and stat may be read frequently.  The combination can
>> + * become very expensive.  By propagating selectively, increasing reading
>> + * frequency decreases the cost of each read.
>> + *
>> + * This struct hosts both the fields which implement the above -
>> + * updated_children and updated_next - and the fields which track basic
>> + * resource statistics on top of it - bsync, bstat and last_bstat.
>> + */
>> +struct cgroup_rstat {
>> +	struct cgroup_rstat_cpu __percpu *rstat_cpu;
>> +
>> +	/*
>> +	 * Add padding to separate the read mostly rstat_cpu and
>> +	 * rstat_css_list into a different cacheline from the following
>> +	 * rstat_flush_next and containing struct fields which can have
>> +	 * frequent updates.
>> +	 */
>> +	CACHELINE_PADDING(_pad_);
>> +	struct cgroup *rstat_flush_next;
>> +};
>> +
>> +struct cgroup_base_stat {
>> +	struct task_cputime cputime;
>> +
>> +#ifdef CONFIG_SCHED_CORE
>> +	u64 forceidle_sum;
>> +#endif
>> +	u64 ntime;
>> +};
>> +
>> +struct cgroup_rstat_cpu {
>> +	/*
>> +	 * Child cgroups with stat updates on this cpu since the last read
>> +	 * are linked on the parent's ->updated_children through
>> +	 * ->updated_next.
>> +	 *
>> +	 * In addition to being more compact, singly-linked list pointing
>> +	 * to the cgroup makes it unnecessary for each per-cpu struct to
>> +	 * point back to the associated cgroup.
>> +	 */
>> +	struct cgroup *updated_children;	/* terminated by self */
>> +	struct cgroup *updated_next;		/* NULL if not on the list */
>> +
>> +	/*
>> +	 * ->bsync protects ->bstat.  These are the only fields which get
>> +	 * updated in the hot path.
>> +	 */
>> +	struct u64_stats_sync bsync;
>> +	struct cgroup_base_stat bstat;
>> +
>> +	/*
>> +	 * Snapshots at the last reading.  These are used to calculate the
>> +	 * deltas to propagate to the global counters.
>> +	 */
>> +	struct cgroup_base_stat last_bstat;
>> +
>> +	/*
>> +	 * This field is used to record the cumulative per-cpu time of
>> +	 * the cgroup and its descendants. Currently it can be read via
>> +	 * eBPF/drgn etc, and we are still trying to determine how to
>> +	 * expose it in the cgroupfs interface.
>> +	 */
>> +	struct cgroup_base_stat subtree_bstat;
>> +
>> +	/*
>> +	 * Snapshots at the last reading. These are used to calculate the
>> +	 * deltas to propagate to the per-cpu subtree_bstat.
>> +	 */
>> +	struct cgroup_base_stat last_subtree_bstat;
>> +};
>> +
>> +#endif	/* _LINUX_RSTAT_H */