All of lore.kernel.org
 help / color / mirror / Atom feed
* [patch V3 0/8] extensible prctl task isolation interface and vmstat sync
@ 2021-08-24 15:24 Marcelo Tosatti
  2021-08-24 15:24 ` [patch V3 1/8] add basic task isolation prctl interface Marcelo Tosatti
                   ` (8 more replies)
  0 siblings, 9 replies; 27+ messages in thread
From: Marcelo Tosatti @ 2021-08-24 15:24 UTC (permalink / raw)
  To: linux-kernel
  Cc: Nitesh Lal, Nicolas Saenz Julienne, Frederic Weisbecker,
	Christoph Lameter, Juri Lelli, Peter Zijlstra, Alex Belits,
	Peter Xu


The logic to disable vmstat worker thread, when entering
nohz full, does not cover all scenarios. For example, it is possible
for the following to happen:

1) enter nohz_full, which calls refresh_cpu_vm_stats, syncing the stats.
2) app runs mlock, which increases counters for mlock'ed pages.
3) start -RT loop

Since refresh_cpu_vm_stats from nohz_full logic can happen _before_
the mlock, vmstat shepherd can restart vmstat worker thread on
the CPU in question.

To fix this, add task isolation prctl interface to quiesce
deferred actions when returning to userspace.

The patchset is based on ideas and code from the 
task isolation patchset from Alex Belits:
https://lwn.net/Articles/816298/

Please refer to Documentation/userspace-api/task_isolation.rst
(patch 2) for details.

Note: the prctl interface is independent of nohz_full=.

---------

v3: 

 - Split in smaller patches		 (Nitesh Lal).
 - Misc cleanups			 (Nitesh Lal).
 - Clarify nohz_full is not a dependency (Nicolas Saenz).
 - Incorrect values for prctl definitions (kernel robot).
 - Save configured state, so applications  
   can activate externally configured
   task isolation parameters.
 - Remove "system default" notion (chisol should
   make it obsolete).
 - Update documentation: add new section with explanation
   about configuration/activation and code example.
 - Update samples.
 - Report configuration/activation state at
   /proc/pid/task_isolation.
 - Condense dirty information of per-CPU vmstats counters 
   in a bool.
 - In-kernel KVM support.
 - Add support to configure inheritance on fork and exec.

v2:

- Finer-grained control of quiescing (Frederic Weisbecker / Nicolas Saenz).

- Avoid potential regressions by allowing applications
  to use ISOL_F_QUIESCE_DEFMASK (whose default value
  is configurable in /sys/).         (Nitesh Lal / Nicolas Saenz).

v2 can be found at:
https://lore.kernel.org/patchwork/project/lkml/list/?series=510225


---

 Documentation/userspace-api/task_isolation.rst |  281 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 arch/x86/kvm/x86.c                             |    3 
 fs/proc/base.c                                 |   68 +++++++++++++++++++
 include/linux/sched.h                          |    5 +
 include/linux/task_isolation.h                 |  131 ++++++++++++++++++++++++++++++++++++++
 include/linux/vmstat.h                         |   17 ++++
 include/uapi/linux/prctl.h                     |   27 +++++++
 init/init_task.c                               |    3 
 kernel/Makefile                                |    2 
 kernel/entry/common.c                          |    2 
 kernel/exit.c                                  |    2 
 kernel/fork.c                                  |   23 ++++++
 kernel/sys.c                                   |   26 +++++++
 kernel/task_isolation.c                        |  315 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 mm/vmstat.c                                    |  167 ++++++++++++++++++++++++++++++++++++------------
 samples/Kconfig                                |    7 ++
 samples/Makefile                               |    1 
 samples/task_isolation/Makefile                |    9 ++
 samples/task_isolation/task_isol.c             |   83 ++++++++++++++++++++++++
 samples/task_isolation/task_isol.h             |    9 ++
 samples/task_isolation/task_isol_userloop.c    |   56 ++++++++++++++++
 21 files changed, 1194 insertions(+), 43 deletions(-)


^ permalink raw reply	[flat|nested] 27+ messages in thread

end of thread, other threads:[~2021-09-10 13:49 UTC | newest]

Thread overview: 27+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-08-24 15:24 [patch V3 0/8] extensible prctl task isolation interface and vmstat sync Marcelo Tosatti
2021-08-24 15:24 ` [patch V3 1/8] add basic task isolation prctl interface Marcelo Tosatti
2021-08-24 15:24 ` [patch V3 2/8] add prctl task isolation prctl docs and samples Marcelo Tosatti
2021-08-26  9:59   ` Frederic Weisbecker
2021-08-26 12:11     ` Marcelo Tosatti
2021-08-26 19:15       ` Christoph Lameter
2021-08-26 20:37         ` Marcelo Tosatti
2021-08-27 13:08       ` Frederic Weisbecker
2021-08-27 14:44         ` Marcelo Tosatti
2021-08-30 11:38           ` Frederic Weisbecker
2021-09-01 13:11   ` Nitesh Lal
2021-09-01 17:34     ` Marcelo Tosatti
2021-09-01 17:49       ` Nitesh Lal
2021-08-24 15:24 ` [patch V3 3/8] task isolation: sync vmstats on return to userspace Marcelo Tosatti
2021-09-10 13:49   ` nsaenzju
2021-08-24 15:24 ` [patch V3 4/8] procfs: add per-pid task isolation state Marcelo Tosatti
2021-08-24 15:24 ` [patch V3 5/8] task isolation: sync vmstats conditional on changes Marcelo Tosatti
2021-08-25  9:46   ` Christoph Lameter
2021-08-24 15:24 ` [patch V3 6/8] KVM: x86: call isolation prepare from VM-entry code path Marcelo Tosatti
2021-08-24 15:24 ` [patch V3 7/8] mm: vmstat: move need_update Marcelo Tosatti
2021-08-24 15:24 ` [patch V3 8/8] mm: vmstat_refresh: avoid queueing work item if cpu stats are clean Marcelo Tosatti
2021-08-25  9:30   ` Christoph Lameter
2021-09-01 13:05   ` Nitesh Lal
2021-09-01 17:32     ` Marcelo Tosatti
2021-09-01 18:33       ` Marcelo Tosatti
2021-09-03 17:38         ` Nitesh Lal
2021-08-25 10:02 ` [patch V3 0/8] extensible prctl task isolation interface and vmstat sync Marcelo Tosatti

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.