linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Peter Zijlstra <peterz@infradead.org>
To: "Jan H. Schönherr" <jschoenh@amazon.de>
Cc: Ingo Molnar <mingo@redhat.com>,
	linux-kernel@vger.kernel.org, Paul Turner <pjt@google.com>,
	Vincent Guittot <vincent.guittot@linaro.org>,
	Morten Rasmussen <morten.rasmussen@arm.com>,
	Tim Chen <tim.c.chen@linux.intel.com>
Subject: Re: [RFC 00/60] Coscheduling for Linux
Date: Mon, 17 Sep 2018 13:33:15 +0200	[thread overview]
Message-ID: <20180917113315.GS24124@hirez.programming.kicks-ass.net> (raw)
In-Reply-To: <1d86f497-9fef-0b19-50d6-d46ef1c0bffa@amazon.de>

On Fri, Sep 14, 2018 at 06:25:44PM +0200, Jan H. Schönherr wrote:
> On 09/14/2018 01:12 PM, Peter Zijlstra wrote:
> > On Fri, Sep 07, 2018 at 11:39:47PM +0200, Jan H. Schönherr wrote:

> >> B) Why would I want this?
> > 
> >>    In the L1TF context, it prevents other applications from loading
> >>    additional data into the L1 cache, while one application tries to leak
> >>    data.
> > 
> > That is the whole and only reason you did this;
> It really isn't. But as your mind seems made up, I'm not going to bother
> to argue.

> >> D) What can I *not* do with this?
> >> ---------------------------------
> >>
> >> Besides the missing load-balancing within coscheduled task-groups, this
> >> implementation has the following properties, which might be considered
> >> short-comings.
> >>
> >> This particular implementation focuses on SCHED_OTHER tasks managed by CFS
> >> and allows coscheduling them. Interrupts as well as tasks in higher
> >> scheduling classes are currently out-of-scope: they are assumed to be
> >> negligible interruptions as far as coscheduling is concerned and they do
> >> *not* cause a preemption of a whole group. This implementation could be
> >> extended to cover higher scheduling classes. Interrupts, however, are an
> >> orthogonal issue.
> >>
> >> The collective context switch from one coscheduled set of tasks to another
> >> -- while fast -- is not atomic. If a use-case needs the absolute guarantee
> >> that all tasks of the previous set have stopped executing before any task
> >> of the next set starts executing, an additional hand-shake/barrier needs to
> >> be added.
> > 
> > IOW it's completely friggin useless for L1TF.
> 
> Do you believe me now, that L1TF is not "the whole and only reason" I did this? :D

You did mention this work first to me in the context of L1TF, so I might
have jumped to conclusions here.

Also, I have, of course, been looking at (SMT) co-scheduling,
specifically in the context of L1TF, myself. I came up with a vastly
different approach. Tim - where are we on getting some of that posted?

Note; that even though I wrote much of that code, I don't particularly
like it either :-)

  parent reply	other threads:[~2018-09-17 11:33 UTC|newest]

Thread overview: 114+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-09-07 21:39 [RFC 00/60] Coscheduling for Linux Jan H. Schönherr
2018-09-07 21:39 ` [RFC 01/60] sched: Store task_group->se[] pointers as part of cfs_rq Jan H. Schönherr
2018-09-07 21:39 ` [RFC 02/60] sched: Introduce set_entity_cfs() to place a SE into a certain CFS runqueue Jan H. Schönherr
2018-09-07 21:39 ` [RFC 03/60] sched: Setup sched_domain_shared for all sched_domains Jan H. Schönherr
2018-09-07 21:39 ` [RFC 04/60] sched: Replace sd_numa_mask() hack with something sane Jan H. Schönherr
2018-09-07 21:39 ` [RFC 05/60] sched: Allow to retrieve the sched_domain_topology Jan H. Schönherr
2018-09-07 21:39 ` [RFC 06/60] sched: Add a lock-free variant of resched_cpu() Jan H. Schönherr
2018-09-07 21:39 ` [RFC 07/60] sched: Reduce dependencies of init_tg_cfs_entry() Jan H. Schönherr
2018-09-07 21:39 ` [RFC 08/60] sched: Move init_entity_runnable_average() into init_tg_cfs_entry() Jan H. Schönherr
2018-09-07 21:39 ` [RFC 09/60] sched: Do not require a CFS in init_tg_cfs_entry() Jan H. Schönherr
2018-09-07 21:39 ` [RFC 10/60] sched: Use parent_entity() in more places Jan H. Schönherr
2018-09-07 21:39 ` [RFC 11/60] locking/lockdep: Increase number of supported lockdep subclasses Jan H. Schönherr
2018-09-07 21:39 ` [RFC 12/60] locking/lockdep: Make cookie generator accessible Jan H. Schönherr
2018-09-07 21:40 ` [RFC 13/60] sched: Remove useless checks for root task-group Jan H. Schönherr
2018-09-07 21:40 ` [RFC 14/60] sched: Refactor sync_throttle() to accept a CFS runqueue as argument Jan H. Schönherr
2018-09-07 21:40 ` [RFC 15/60] sched: Introduce parent_cfs_rq() and use it Jan H. Schönherr
2018-09-07 21:40 ` [RFC 16/60] sched: Preparatory code movement Jan H. Schönherr
2018-09-07 21:40 ` [RFC 17/60] sched: Introduce and use generic task group CFS traversal functions Jan H. Schönherr
2018-09-07 21:40 ` [RFC 18/60] sched: Fix return value of SCHED_WARN_ON() Jan H. Schönherr
2018-09-07 21:40 ` [RFC 19/60] sched: Add entity variants of enqueue_task_fair() and dequeue_task_fair() Jan H. Schönherr
2018-09-07 21:40 ` [RFC 20/60] sched: Let {en,de}queue_entity_fair() work with a varying amount of tasks Jan H. Schönherr
2018-09-07 21:40 ` [RFC 21/60] sched: Add entity variants of put_prev_task_fair() and set_curr_task_fair() Jan H. Schönherr
2018-09-07 21:40 ` [RFC 22/60] cosched: Add config option for coscheduling support Jan H. Schönherr
2018-09-07 21:40 ` [RFC 23/60] cosched: Add core data structures for coscheduling Jan H. Schönherr
2018-09-07 21:40 ` [RFC 24/60] cosched: Do minimal pre-SMP coscheduler initialization Jan H. Schönherr
2018-09-07 21:40 ` [RFC 25/60] cosched: Prepare scheduling domain topology for coscheduling Jan H. Schönherr
2018-09-07 21:40 ` [RFC 26/60] cosched: Construct runqueue hierarchy Jan H. Schönherr
2018-09-07 21:40 ` [RFC 27/60] cosched: Add some small helper functions for later use Jan H. Schönherr
2018-09-07 21:40 ` [RFC 28/60] cosched: Add is_sd_se() to distinguish SD-SEs from TG-SEs Jan H. Schönherr
2018-09-07 21:40 ` [RFC 29/60] cosched: Adjust code reflecting on the total number of CFS tasks on a CPU Jan H. Schönherr
2018-09-07 21:40 ` [RFC 30/60] cosched: Disallow share modification on task groups for now Jan H. Schönherr
2018-09-07 21:40 ` [RFC 31/60] cosched: Don't disable idle tick " Jan H. Schönherr
2018-09-07 21:40 ` [RFC 32/60] cosched: Specialize parent_cfs_rq() for hierarchical runqueues Jan H. Schönherr
2018-09-07 21:40 ` [RFC 33/60] cosched: Allow resched_curr() to be called " Jan H. Schönherr
2018-09-07 21:40 ` [RFC 34/60] cosched: Add rq_of() variants for different use cases Jan H. Schönherr
2018-09-07 21:40 ` [RFC 35/60] cosched: Adjust rq_lock() functions to work with hierarchical runqueues Jan H. Schönherr
2018-09-07 21:40 ` [RFC 36/60] cosched: Use hrq_of() for rq_clock() and rq_clock_task() Jan H. Schönherr
2018-09-07 21:40 ` [RFC 37/60] cosched: Use hrq_of() for (indirect calls to) ___update_load_sum() Jan H. Schönherr
2018-09-07 21:40 ` [RFC 38/60] cosched: Skip updates on non-CPU runqueues in cfs_rq_util_change() Jan H. Schönherr
2018-09-07 21:40 ` [RFC 39/60] cosched: Adjust task group management for hierarchical runqueues Jan H. Schönherr
2018-09-07 21:40 ` [RFC 40/60] cosched: Keep track of task group hierarchy within each SD-RQ Jan H. Schönherr
2018-09-07 21:40 ` [RFC 41/60] cosched: Introduce locking for leader activities Jan H. Schönherr
2018-09-07 21:40 ` [RFC 42/60] cosched: Introduce locking for (mostly) enqueuing and dequeuing Jan H. Schönherr
2018-09-07 21:40 ` [RFC 43/60] cosched: Add for_each_sched_entity() variant for owned entities Jan H. Schönherr
2018-09-07 21:40 ` [RFC 44/60] cosched: Perform various rq_of() adjustments in scheduler code Jan H. Schönherr
2018-09-07 21:40 ` [RFC 45/60] cosched: Continue to account all load on per-CPU runqueues Jan H. Schönherr
2018-09-07 21:40 ` [RFC 46/60] cosched: Warn on throttling attempts of non-CPU runqueues Jan H. Schönherr
2018-09-07 21:40 ` [RFC 47/60] cosched: Adjust SE traversal and locking for common leader activities Jan H. Schönherr
2018-09-07 21:40 ` [RFC 48/60] cosched: Adjust SE traversal and locking for yielding and buddies Jan H. Schönherr
2018-09-07 21:40 ` [RFC 49/60] cosched: Adjust locking for enqueuing and dequeueing Jan H. Schönherr
2018-09-07 21:40 ` [RFC 50/60] cosched: Propagate load changes across hierarchy levels Jan H. Schönherr
2018-09-07 21:40 ` [RFC 51/60] cosched: Hacky work-around to avoid observing zero weight SD-SE Jan H. Schönherr
2018-09-07 21:40 ` [RFC 52/60] cosched: Support SD-SEs in enqueuing and dequeuing Jan H. Schönherr
2018-09-07 21:40 ` [RFC 53/60] cosched: Prevent balancing related functions from crossing hierarchy levels Jan H. Schönherr
2018-09-07 21:40 ` [RFC 54/60] cosched: Support idling in a coscheduled set Jan H. Schönherr
2018-09-07 21:40 ` [RFC 55/60] cosched: Adjust task selection for coscheduling Jan H. Schönherr
2018-09-07 21:40 ` [RFC 56/60] cosched: Adjust wakeup preemption rules " Jan H. Schönherr
2018-09-07 21:40 ` [RFC 57/60] cosched: Add sysfs interface to configure coscheduling on cgroups Jan H. Schönherr
2018-09-07 21:40 ` [RFC 58/60] cosched: Switch runqueues between regular scheduling and coscheduling Jan H. Schönherr
2018-09-07 21:40 ` [RFC 59/60] cosched: Handle non-atomicity during switches to and from coscheduling Jan H. Schönherr
2018-09-07 21:40 ` [RFC 60/60] cosched: Add command line argument to enable coscheduling Jan H. Schönherr
2018-09-10  2:50   ` Randy Dunlap
2018-09-12  0:24 ` [RFC 00/60] Coscheduling for Linux Nishanth Aravamudan
2018-09-12 19:34   ` Jan H. Schönherr
2018-09-12 23:15     ` Nishanth Aravamudan
2018-09-13 11:31       ` Jan H. Schönherr
2018-09-13 18:16         ` Nishanth Aravamudan
2018-09-12 23:18     ` Jan H. Schönherr
2018-09-13  3:05       ` Nishanth Aravamudan
2018-09-13 19:19 ` [RFC 61/60] cosched: Accumulated fixes and improvements Jan H. Schönherr
2018-09-26 17:25   ` Nishanth Aravamudan
2018-09-26 21:05     ` Nishanth Aravamudan
2018-10-01  9:13       ` Jan H. Schönherr
2018-09-14 11:12 ` [RFC 00/60] Coscheduling for Linux Peter Zijlstra
2018-09-14 16:25   ` Jan H. Schönherr
2018-09-15  8:48     ` Task group cleanups and optimizations (was: Re: [RFC 00/60] Coscheduling for Linux) Jan H. Schönherr
2018-09-17  9:48       ` Peter Zijlstra
2018-09-18 13:22         ` Jan H. Schönherr
2018-09-18 13:38           ` Peter Zijlstra
2018-09-18 13:54             ` Jan H. Schönherr
2018-09-18 13:42           ` Peter Zijlstra
2018-09-18 14:35           ` Rik van Riel
2018-09-19  9:23             ` Jan H. Schönherr
2018-11-23 16:51           ` Frederic Weisbecker
2018-12-04 13:23             ` Jan H. Schönherr
2018-09-17 11:33     ` Peter Zijlstra [this message]
2018-11-02 22:13       ` [RFC 00/60] Coscheduling for Linux Nishanth Aravamudan
2018-09-17 12:25     ` Peter Zijlstra
2018-09-26  9:58       ` Jan H. Schönherr
2018-09-27 18:36         ` Subhra Mazumdar
2018-11-23 16:29           ` Frederic Weisbecker
2018-09-17 13:37     ` Peter Zijlstra
2018-09-26  9:35       ` Jan H. Schönherr
2018-09-18 14:40     ` Rik van Riel
2018-09-24 15:23       ` Jan H. Schönherr
2018-09-24 18:01         ` Rik van Riel
2018-09-18  0:33 ` Subhra Mazumdar
2018-09-18 11:44   ` Jan H. Schönherr
2018-09-19 21:53     ` Subhra Mazumdar
2018-09-24 15:43       ` Jan H. Schönherr
2018-09-27 18:12         ` Subhra Mazumdar
2018-10-04 13:29 ` Jon Masters
2018-10-17  2:09 ` Frederic Weisbecker
2018-10-19 11:40   ` Jan H. Schönherr
2018-10-19 14:52     ` Frederic Weisbecker
2018-10-19 15:16     ` Rik van Riel
2018-10-19 15:33       ` Frederic Weisbecker
2018-10-19 15:45         ` Rik van Riel
2018-10-19 19:07           ` Jan H. Schönherr
2018-10-19  0:26 ` Subhra Mazumdar
2018-10-26 23:44   ` Jan H. Schönherr
2018-10-29 22:52     ` Subhra Mazumdar
2018-10-26 23:05 ` Subhra Mazumdar
2018-10-27  0:07   ` Jan H. Schönherr

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20180917113315.GS24124@hirez.programming.kicks-ass.net \
    --to=peterz@infradead.org \
    --cc=jschoenh@amazon.de \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@redhat.com \
    --cc=morten.rasmussen@arm.com \
    --cc=pjt@google.com \
    --cc=tim.c.chen@linux.intel.com \
    --cc=vincent.guittot@linaro.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body. 
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).