[RFC PATCH v1 00/16] xen: sched: implement core-scheduling

From: Dario Faggioli <dfaggioli@suse.com>
To: xen-devel@lists.xenproject.org
Cc: Wei Liu <wei.liu2@citrix.com>,
	George Dunlap <George.Dunlap@eu.citrix.com>,
	Andrew Cooper <andrew.cooper3@citrix.com>,
	Ian Jackson <ian.jackson@eu.citrix.com>,
	Bhavesh Davda <bhavesh.davda@oracle.com>,
	Jan Beulich <jbeulich@suse.com>
Subject: [RFC PATCH v1 00/16] xen: sched: implement core-scheduling
Date: Sat, 25 Aug 2018 01:35:11 +0200	[thread overview]
Message-ID: <153515305655.8598.6054293649487840735.stgit@Istar.fritz.box> (raw)

Hello,

As anticipated here:
https://lists.xenproject.org/archives/html/xen-devel/2018-08/msg02052.html

Here's a preliminary version of my work, trying to implement
core-scheduling in Xen.

First of all, this deals with Credit1 only. I have patches for Credit2,
and I've been working on having them ready by today, but I could not
defeat the latest bugs. :-/
I'll post them when back from vacation. Just let me anticipate, that
doing something like this in Credit2, is a lot simpler than what you
see here for Credit1.

Even these patches that I'm posting are not perfect, and In fact there
are some TODOs and XXXs --both in the changelogs and in the code.

They give me a system that boots, where I can do basic stuff (like
playing with dom0, creating guests, etc), and where the constraint of
only scheduling vcpus from one domain at a time on pcpus that are part
of the same core is, as far as I've seen, respected.

There are still cases where the behavior is unideal, e.g., we could
make a better use of some of the cores which are, some of the times,
left idle.

There are git branches here:
 https://gitlab.com/dfaggioli/xen.git rel/sched/core-scheduling-RFCv1
 https://github.com/fdario/xen.git rel/sched/core-scheduling-RFCv1

Any comment is more than welcome.

Regards,
Dario
---
Dario Faggioli (16):
      xen: Credit1: count runnable vcpus, not running ones
      xen: Credit1: always steal from pcpus with runnable but not running vcpus
      xen: Credit1: do not always tickle an idle pcpu
      xen: sched: make the logic for tracking idle core generic.
      xen: Credit1: track fully idle cores.
      xen: Credit1: check for fully idle cores when tickling
      xen: Credit1: reorg __runq_tickle() code a bit.
      xen: Credit1: reorg csched_schedule() code a bit.
      xen: Credit1: SMT-aware domain co-scheduling parameter and data structs
      xen: Credit1: support sched_smt_cosched in csched_schedule()
      xen: Credit1: support sched_smt_cosched in _csched_cpu_pick()
      xen: Credit1: support sched_smt_cosched in csched_runq_steal().
      xen: Credit1: sched_smt_cosched support in __csched_vcpu_is_migrateable().
      xen: Credit1: sched_smt_cosched support in __runq_tickle() for pinned vcpus.
      xen: Credit1: sched_smt_cosched support in __runq_tickle().
      xen/tools: tracing of Credit1 SMT domain co-scheduling support

 docs/misc/xen-command-line.markdown |   11 +
 tools/xentrace/xenalyze.c           |   85 +++++-
 xen/common/sched_credit.c           |  532 +++++++++++++++++++++++++++--------
 xen/common/sched_credit2.c          |  128 +++-----
 xen/common/schedule.c               |    7 
 xen/include/xen/sched.h             |   42 +++
 6 files changed, 603 insertions(+), 202 deletions(-)
--
<<This happens because I choose it to happen!>> (Raistlin Majere)
-----------------------------------------------------------------
Dario Faggioli, Ph.D, http://about.me/dario.faggioli
Software Engineer @ SUSE https://www.suse.com/

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel