From: Dario Faggioli <dfaggioli@suse.com>
To: xen-devel@lists.xenproject.org
Cc: George Dunlap <george.dunlap@eu.citrix.com>
Subject: [RFC PATCH v1 11/16] xen: Credit1: support sched_smt_cosched in _csched_cpu_pick()
Date: Sat, 25 Aug 2018 01:36:29 +0200 [thread overview]
Message-ID: <153515378994.8598.11295185673647033946.stgit@Istar.fritz.box> (raw)
In-Reply-To: <153515305655.8598.6054293649487840735.stgit@Istar.fritz.box>
If sched_smt_cosched is enabled, take it into account when choosing on
what pcpu to run a vcpu (e.g., for doing a migration).
Basically, we can only run vcpus of domain A on pcpus of cores where
other vcpus of domain A are running already (and, vice versa, we
absolutely don't want to run a vcpu of domain A on pcpus of cores
where vcpus of another domains are running!).
Signed-off-by: Dario Faggioli <dfaggioli@suse.com>
---
Cc: George Dunlap <george.dunlap@eu.citrix.com>
---
xen/common/sched_credit.c | 25 ++++++++++++++++---------
1 file changed, 16 insertions(+), 9 deletions(-)
diff --git a/xen/common/sched_credit.c b/xen/common/sched_credit.c
index 22327b61fb..81a2c8b384 100644
--- a/xen/common/sched_credit.c
+++ b/xen/common/sched_credit.c
@@ -822,7 +822,7 @@ _csched_cpu_pick(const struct scheduler *ops, struct vcpu *vc, bool_t commit)
cpumask_t *cpus = cpumask_scratch_cpu(vc->processor);
cpumask_t idlers;
cpumask_t *online = cpupool_domain_cpumask(vc->domain);
- struct csched_pcpu *spc = NULL;
+ struct csched_pcpu *spc = NULL, *nspc = NULL;
int cpu = vc->processor;
int balance_step;
@@ -900,6 +900,7 @@ _csched_cpu_pick(const struct scheduler *ops, struct vcpu *vc, bool_t commit)
int migrate_factor;
nxt = cpumask_cycle(cpu, cpus);
+ nspc = CSCHED_PCPU(nxt);
if ( cpumask_test_cpu(cpu, per_cpu(cpu_core_mask, nxt)) )
{
@@ -929,15 +930,21 @@ _csched_cpu_pick(const struct scheduler *ops, struct vcpu *vc, bool_t commit)
weight_cpu > weight_nxt :
weight_cpu * migrate_factor < weight_nxt )
{
- cpumask_and(&nxt_idlers, &nxt_idlers, cpus);
- spc = CSCHED_PCPU(nxt);
- cpu = cpumask_cycle(spc->idle_bias, &nxt_idlers);
- cpumask_andnot(cpus, cpus, per_cpu(cpu_sibling_mask, cpu));
- }
- else
- {
- cpumask_andnot(cpus, cpus, &nxt_idlers);
+ spin_lock(&nspc->core->lock);
+ if ( !sched_smt_cosched ||
+ nspc->core->sdom == NULL || nspc->core->sdom->dom == vc->domain )
+ {
+ cpumask_and(&nxt_idlers, &nxt_idlers, cpus);
+ spc = CSCHED_PCPU(nxt);
+ cpu = cpumask_cycle(spc->idle_bias, &nxt_idlers);
+ cpumask_andnot(cpus, cpus, per_cpu(cpu_sibling_mask, cpu));
+ spin_unlock(&nspc->core->lock);
+ continue;
+ }
+ spin_unlock(&nspc->core->lock);
}
+
+ cpumask_andnot(cpus, cpus, &nxt_idlers);
}
/* Stop if cpu is idle */
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
next prev parent reply other threads:[~2018-08-24 23:36 UTC|newest]
Thread overview: 29+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-08-24 23:35 [RFC PATCH v1 00/16] xen: sched: implement core-scheduling Dario Faggioli
2018-08-24 23:35 ` [RFC PATCH v1 01/16] xen: Credit1: count runnable vcpus, not running ones Dario Faggioli
2018-08-24 23:35 ` [RFC PATCH v1 02/16] xen: Credit1: always steal from pcpus with runnable but not running vcpus Dario Faggioli
2018-08-24 23:35 ` [RFC PATCH v1 03/16] xen: Credit1: do not always tickle an idle pcpu Dario Faggioli
2018-08-24 23:35 ` [RFC PATCH v1 04/16] xen: sched: make the logic for tracking idle core generic Dario Faggioli
2018-08-24 23:35 ` [RFC PATCH v1 05/16] xen: Credit1: track fully idle cores Dario Faggioli
2018-08-24 23:35 ` [RFC PATCH v1 06/16] xen: Credit1: check for fully idle cores when tickling Dario Faggioli
2018-08-24 23:36 ` [RFC PATCH v1 07/16] xen: Credit1: reorg __runq_tickle() code a bit Dario Faggioli
2018-08-24 23:36 ` [RFC PATCH v1 08/16] xen: Credit1: reorg csched_schedule() " Dario Faggioli
2018-08-24 23:36 ` [RFC PATCH v1 09/16] xen: Credit1: SMT-aware domain co-scheduling parameter and data structs Dario Faggioli
2018-08-24 23:36 ` [RFC PATCH v1 10/16] xen: Credit1: support sched_smt_cosched in csched_schedule() Dario Faggioli
2018-08-24 23:36 ` Dario Faggioli [this message]
2018-08-24 23:36 ` [RFC PATCH v1 12/16] xen: Credit1: support sched_smt_cosched in csched_runq_steal() Dario Faggioli
2018-08-24 23:36 ` [RFC PATCH v1 13/16] xen: Credit1: sched_smt_cosched support in __csched_vcpu_is_migrateable() Dario Faggioli
2018-08-24 23:36 ` [RFC PATCH v1 14/16] xen: Credit1: sched_smt_cosched support in __runq_tickle() for pinned vcpus Dario Faggioli
2018-08-24 23:36 ` [RFC PATCH v1 15/16] xen: Credit1: sched_smt_cosched support in __runq_tickle() Dario Faggioli
2018-08-24 23:37 ` [RFC PATCH v1 16/16] xen/tools: tracing of Credit1 SMT domain co-scheduling support Dario Faggioli
2018-09-07 16:00 ` [RFC PATCH v1 00/16] xen: sched: implement core-scheduling Juergen Gross
2018-10-11 17:37 ` Dario Faggioli
2018-10-12 5:15 ` Juergen Gross
2018-10-12 7:49 ` Dario Faggioli
2018-10-12 8:35 ` Juergen Gross
2018-10-12 9:15 ` Dario Faggioli
2018-10-12 9:23 ` Juergen Gross
2018-10-18 10:40 ` Dario Faggioli
2018-10-17 21:36 ` Tamas K Lengyel
2018-10-18 8:16 ` Dario Faggioli
2018-10-18 12:55 ` Tamas K Lengyel
2018-10-18 13:48 ` Dario Faggioli
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=153515378994.8598.11295185673647033946.stgit@Istar.fritz.box \
--to=dfaggioli@suse.com \
--cc=george.dunlap@eu.citrix.com \
--cc=xen-devel@lists.xenproject.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).