[RFC,v3,0/9] pids controller events rework and migration charging

Message ID	20240405170548.15234-1-mkoutny@suse.com (mailing list archive)
Headers	show Received: from smtp-out1.suse.de (smtp-out1.suse.de [195.135.223.130]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 8AB96171667; Fri, 5 Apr 2024 17:05:54 +0000 (UTC) From: =?utf-8?q?Michal_Koutn=C3=BD?= <mkoutny@suse.com> To: cgroups@vger.kernel.org, linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org Cc: Tejun Heo <tj@kernel.org>, Zefan Li <lizefan.x@bytedance.com>, Johannes Weiner <hannes@cmpxchg.org>, Jonathan Corbet <corbet@lwn.net>, Shuah Khan <shuah@kernel.org> Subject: [RFC PATCH v3 0/9] pids controller events rework and migration charging Date: Fri, 5 Apr 2024 19:05:39 +0200 Message-ID: <20240405170548.15234-1-mkoutny@suse.com> Precedence: bulk MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit default: False [-3.51 / 50.00]; BAYES_HAM(-3.00)[100.00%]; MID_CONTAINS_FROM(1.00)[]; NEURAL_HAM_LONG(-1.00)[-1.000]; R_DKIM_ALLOW(-0.20)[suse.com:s=susede1]; NEURAL_HAM_SHORT(-0.20)[-1.000]; MIME_GOOD(-0.10)[text/plain]; MX_GOOD(-0.01)[]; RCVD_VIA_SMTP_AUTH(0.00)[]; ARC_NA(0.00)[]; TO_DN_SOME(0.00)[]; MIME_TRACE(0.00)[0:+]; SPAMHAUS_XBL(0.00)[2a07:de40:b281:104:10:150:64:98:from]; FUZZY_BLOCKED(0.00)[rspamd.com]; RCVD_TLS_ALL(0.00)[]; R_RATELIMIT(0.00)[to_ip_from(RLfcexjtczxtbnj559fj95g6y7)]; FROM_EQ_ENVFROM(0.00)[]; FROM_HAS_DN(0.00)[]; RCPT_COUNT_SEVEN(0.00)[9]; RCVD_COUNT_TWO(0.00)[2]; TO_MATCH_ENVRCPT_ALL(0.00)[]; DBL_BLOCKED_OPENRESOLVER(0.00)[pids.events:url,suse.com:dkim,imap2.dmz-prg2.suse.org:helo,imap2.dmz-prg2.suse.org:rdns]; DKIM_SIGNED(0.00)[suse.com:s=susede1]; DKIM_TRACE(0.00)[suse.com:+]
Series	pids controller events rework and migration charging \| expand [RFC,v3,0/9] pids controller events rework and migration charging [RFC,v3,1/9] cgroup/pids: Remove superfluous zeroing [RFC,v3,2/9] cgroup/pids: Separate semantics of pids.events related to pids.max [RFC,v3,3/9] cgroup/pids: Make event counters hierarchical [RFC,v3,4/9] cgroup/pids: Add pids.events.local [RFC,v3,5/9] selftests: cgroup: Lexicographic order in Makefile [RFC,v3,6/9] selftests: cgroup: Add basic tests for pids controller [RFC,v3,7/9] cgroup/pids: Replace uncharge/charge pair with a single function [RFC,v3,8/9] cgroup/pids: Enforce pids.max on task migrations [RFC,v3,9/9] selftests: cgroup: Add tests pids controller

Message ID

20240405170548.15234-1-mkoutny@suse.com (mailing list archive)

Headers

From: =?utf-8?q?Michal_Koutn=C3=BD?= <mkoutny@suse.com>
To: cgroups@vger.kernel.org,
	linux-doc@vger.kernel.org,
	linux-kernel@vger.kernel.org,
	linux-kselftest@vger.kernel.org
Cc: Tejun Heo <tj@kernel.org>,
	Zefan Li <lizefan.x@bytedance.com>,
	Johannes Weiner <hannes@cmpxchg.org>,
	Jonathan Corbet <corbet@lwn.net>,
	Shuah Khan <shuah@kernel.org>
Subject: [RFC PATCH v3 0/9] pids controller events rework and migration
 charging
Date: Fri,  5 Apr 2024 19:05:39 +0200
Message-ID: <20240405170548.15234-1-mkoutny@suse.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Series

pids controller events rework and migration charging | expand

Message

Michal Koutný April 5, 2024, 5:05 p.m. UTC

The series consists of two parts:
- pids.events rework (originally v2, patches 1-6,
- migration charging, patches 7-9.

The changes are independent in principle, I stacked them for (my)
convenience and because they both deserve RFC:

1) Changed semantics of v2 pids.events
  - similar change was proposed for memory.swap.events:max [1]
2) Migration charging is obsolete concept

How are the new events supposed to be useful?

- pids.events.local:max
  - tells that cgroup's limit is hit (too tight?)
- pids.events.local:max.imposed
  - tells that cgroup's workload was restricted (generalization of
    'cgroup: fork rejected by pids controller in %s' message)
- pids.events:*
  - "only" directs top-down search to cgroups of interest

The migration charging is motivated by apparenty surprising
	pids.current > pids.max
because supervised processes are forked in supervisor's cgroup (more
details in commit cgroup/pids: Enforce pids.max on task migrations too)

Changes from v2 (https://lore.kernel.org/r/20200205134426.10570-1-mkoutny@suse.com)
- implemented pids.events.local (Tejun)
- added migration charging

[1] https://lore.kernel.org/r/20230202155626.1829121-1-hannes@cmpxchg.org/

Michal Koutný (9):
  cgroup/pids: Remove superfluous zeroing
  cgroup/pids: Separate semantics of pids.events related to pids.max
  cgroup/pids: Make event counters hierarchical
  cgroup/pids: Add pids.events.local
  selftests: cgroup: Lexicographic order in Makefile
  selftests: cgroup: Add basic tests for pids controller
  cgroup/pids: Replace uncharge/charge pair with a single function
  cgroup/pids: Enforce pids.max on task migrations
  selftests: cgroup: Add tests pids controller

 Documentation/admin-guide/cgroup-v1/pids.rst |   3 +-
 Documentation/admin-guide/cgroup-v2.rst      |  22 +-
 include/linux/cgroup-defs.h                  |   7 +-
 kernel/cgroup/cgroup.c                       |  16 +-
 kernel/cgroup/pids.c                         | 206 +++++++++----
 tools/testing/selftests/cgroup/Makefile      |  25 +-
 tools/testing/selftests/cgroup/test_pids.c   | 302 +++++++++++++++++++
 7 files changed, 514 insertions(+), 67 deletions(-)
 create mode 100644 tools/testing/selftests/cgroup/test_pids.c


base-commit: 026e680b0a08a62b1d948e5a8ca78700bfac0e6e