| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20230306181723.GA3153@ranerica-svr.sc.intel.com>
Date: Mon, 6 Mar 2023 10:17:23 -0800
From: Ricardo Neri <ricardo.neri-calderon@...ux.intel.com>
To: Ionela Voinescu <ionela.voinescu@....com>
Cc: "Peter Zijlstra (Intel)" <peterz@...radead.org>,
Juri Lelli <juri.lelli@...hat.com>,
Vincent Guittot <vincent.guittot@...aro.org>,
Ricardo Neri <ricardo.neri@...el.com>,
"Ravi V. Shankar" <ravi.v.shankar@...el.com>,
Ben Segall <bsegall@...gle.com>,
Daniel Bristot de Oliveira <bristot@...hat.com>,
Dietmar Eggemann <dietmar.eggemann@....com>,
Len Brown <len.brown@...el.com>, Mel Gorman <mgorman@...e.de>,
"Rafael J. Wysocki" <rafael.j.wysocki@...el.com>,
Srinivas Pandruvada <srinivas.pandruvada@...ux.intel.com>,
Steven Rostedt <rostedt@...dmis.org>,
Tim Chen <tim.c.chen@...ux.intel.com>,
Valentin Schneider <vschneid@...hat.com>, x86@...nel.org,
linux-kernel@...r.kernel.org, "Tim C . Chen" <tim.c.chen@...el.com>
Subject: Re: [PATCH v3 08/10] sched/topology: Remove SHARED_CHILD from
ASYM_PACKING
On Mon, Mar 06, 2023 at 01:10:37PM +0000, Ionela Voinescu wrote:
> Hey,
>
> On Sunday 05 Mar 2023 at 11:08:11 (-0800), Ricardo Neri wrote:
> > On Fri, Mar 03, 2023 at 11:29:52AM +0000, Ionela Voinescu wrote:
> > > Hi Ricardo,
> >
> > Hi Ionela!
> >
> > >
> > > On Monday 06 Feb 2023 at 20:58:36 (-0800), Ricardo Neri wrote:
> > > > Only x86 and Power7 use ASYM_PACKING. They use it differently.
> > > >
> > > > Power7 has cores of equal priority, but the SMT siblings of a core have
> > > > different priorities. Parent scheduling domains do not need (nor have) the
> > > > ASYM_PACKING flag. SHARED_CHILD is not needed. Using SHARED_PARENT would
> > > > cause the topology debug code to complain.
> > > >
> > > > X86 has cores of different priority, but all the SMT siblings of the core
> > > > have equal priority. It needs ASYM_PACKING at the MC level, but not at the
> > > > SMT level (it also needs it at upper levels if they have scheduling groups
> > > > of different priority). Removing ASYM_PACKING from the SMT domain causes
> > > > the topology debug code to complain.
> > > >
> > > > Remove SHARED_CHILD for now. We still need a topology check that satisfies
> > > > both architectures.
> > > >
> > > > Cc: Ben Segall <bsegall@...gle.com>
> > > > Cc: Daniel Bristot de Oliveira <bristot@...hat.com>
> > > > Cc: Dietmar Eggemann <dietmar.eggemann@....com>
> > > > Cc: Len Brown <len.brown@...el.com>
> > > > Cc: Mel Gorman <mgorman@...e.de>
> > > > Cc: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
> > > > Cc: Srinivas Pandruvada <srinivas.pandruvada@...ux.intel.com>
> > > > Cc: Steven Rostedt <rostedt@...dmis.org>
> > > > Cc: Tim C. Chen <tim.c.chen@...el.com>
> > > > Cc: Valentin Schneider <vschneid@...hat.com>
> > > > Cc: x86@...nel.org
> > > > Cc: linux-kernel@...r.kernel.org
> > > > Suggested-by: Valentin Schneider <vschneid@...hat.com>
> > > > Signed-off-by: Ricardo Neri <ricardo.neri-calderon@...ux.intel.com>
> > > > ---
> > > > Changes since v2:
> > > > * Introduced this patch.
> > > >
> > > > Changes since v1:
> > > > * N/A
> > > > ---
> > > > include/linux/sched/sd_flags.h | 5 +----
> > > > 1 file changed, 1 insertion(+), 4 deletions(-)
> > > >
> > > > diff --git a/include/linux/sched/sd_flags.h b/include/linux/sched/sd_flags.h
> > > > index 57bde66d95f7..800238854ba5 100644
> > > > --- a/include/linux/sched/sd_flags.h
> > > > +++ b/include/linux/sched/sd_flags.h
> > > > @@ -132,12 +132,9 @@ SD_FLAG(SD_SERIALIZE, SDF_SHARED_PARENT | SDF_NEEDS_GROUPS)
> > > > /*
> > > > * Place busy tasks earlier in the domain
> > > > *
> > > > - * SHARED_CHILD: Usually set on the SMT level. Technically could be set further
> > > > - * up, but currently assumed to be set from the base domain
> > > > - * upwards (see update_top_cache_domain()).
> > > > * NEEDS_GROUPS: Load balancing flag.
> > > > */
> > > > -SD_FLAG(SD_ASYM_PACKING, SDF_SHARED_CHILD | SDF_NEEDS_GROUPS)
> > > > +SD_FLAG(SD_ASYM_PACKING, SDF_NEEDS_GROUPS)
> > >
> > > While this silences the warning one would have gotten when removing
> > > SD_ASYM_PACKING from SMT level, it will still result in sd_asym_packing
> > > being NULL for these systems, which breaks nohz balance. That is because
> > > highest_flag_domain() still stops searching at the first level without
> > > the flag set, in this case SMT, even if levels above have the flag set.
> >
> > You are absolutely right! This how this whole discussion started. It
> > slipped my mind.
> >
> > >
> > > Maybe highest_flag_domain() should be changed to take into account the
> > > metadata flags?
> >
> > What about the patch below? Search will stop if the flag has
> > SDF_SHARED_CHILD as it does today. Otherwise it will search all the
> > domains.
> >
> > --- a/kernel/sched/sched.h
> > +++ b/kernel/sched/sched.h
> > @@ -1773,6 +1773,12 @@ queue_balance_callback(struct rq *rq,
> > for (__sd = rcu_dereference_check_sched_domain(cpu_rq(cpu)->sd); \
> > __sd; __sd = __sd->parent)
> >
> > +#define SD_FLAG(name, mflags) (name * !!((mflags) & SDF_SHARED_CHILD)) |
> > +static const unsigned int SD_SHARED_CHILD_MASK =
> > +#include <linux/sched/sd_flags.h>
> > +0;
> > +#undef SD_FLAG
> > +
> > /**
> > * highest_flag_domain - Return highest sched_domain containing flag.
> > * @cpu: The CPU whose highest level of sched domain is to
> > @@ -1781,15 +1787,19 @@ queue_balance_callback(struct rq *rq,
> > * for the given CPU.
> > *
> > * Returns the highest sched_domain of a CPU which contains the given flag.
> > - */
> > +*/
> ^^^
> likely an unintended change
Yes! I will remove it in the patch I post.
> > static inline struct sched_domain *highest_flag_domain(int cpu, int flag)
> > {
> > struct sched_domain *sd, *hsd = NULL;
> >
> > for_each_domain(cpu, sd) {
> > - if (!(sd->flags & flag))
> > + if (sd->flags & flag) {
> > + hsd = sd;
> > + continue;
> > + }
> > +
>
> There might be room for a comment here:
> /*
> * If the flag is not set and is known to be shared with lower
> * domains, stop the search, as it won't be found further up.
> */
Sure, I can and a comment to this effect.
> > + if (flag & SD_SHARED_CHILD_MASK)
> > break;
> > - hsd = sd;
> > }
> >
> > return hsd;
>
> It looks nice and sane to me - I've not compiled or tested it :).
Thank you very much for your feedback!
BR,
Ricardo
Powered by blists - more mailing lists