[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAJZ5v0j8Dqro_=wcA5RD3+g-0pp9np3HyVQP9VU0De_JJAf6Pg@mail.gmail.com>
Date: Tue, 13 Aug 2024 13:07:47 +0200
From: "Rafael J. Wysocki" <rafael@...nel.org>
To: Chen Yu <yu.c.chen@...el.com>
Cc: "Rafael J. Wysocki" <rjw@...ysocki.net>, x86 Maintainers <x86@...nel.org>,
LKML <linux-kernel@...r.kernel.org>, Linux PM <linux-pm@...r.kernel.org>,
Thomas Gleixner <tglx@...utronix.de>, Peter Zijlstra <peterz@...radead.org>,
Srinivas Pandruvada <srinivas.pandruvada@...ux.intel.com>,
"Rafael J. Wysocki" <rafael@...nel.org>, Dietmar Eggemann <dietmar.eggemann@....com>,
Ricardo Neri <ricardo.neri@...el.com>, Tim Chen <tim.c.chen@...el.com>
Subject: Re: [PATCH v2 2/3] x86/sched: Add basic support for CPU capacity scaling
Hi,
On Tue, Aug 13, 2024 at 3:27 AM Chen Yu <yu.c.chen@...el.com> wrote:
>
> Hi Rafael,
>
> On 2024-08-12 at 14:42:26 +0200, Rafael J. Wysocki wrote:
> > +void arch_set_cpu_capacity(int cpu, unsigned long cap, unsigned long base_cap,
> > + unsigned long max_freq, unsigned long base_freq)
> > +{
> > + if (static_branch_likely(&arch_hybrid_cap_scale_key)) {
> > + WRITE_ONCE(per_cpu_ptr(arch_cpu_scale, cpu)->capacity,
> > + div_u64(cap << SCHED_CAPACITY_SHIFT, base_cap));
> > + WRITE_ONCE(per_cpu_ptr(arch_cpu_scale, cpu)->freq_ratio,
> > + div_u64(max_freq << SCHED_CAPACITY_SHIFT, base_freq));
> >
>
> Would the capacity update be frequently invoked?
I hope not.
> Just wonder if we could
> first READ_ONCE() to check if the value is already the value we want to
> change to, to avoid one write and less cache snoop overhead (in case other
> CPU reads this CPU's capacity)
Well, I'd rather not complicate the code beyond what is necessary
unless this is demonstrated to make a measurable difference.
Besides, AFAICS the only case in the caller in which the same values
can be passed to arch_set_cpu_capacity() is the CPU offline one and
that should not happen too often.
Powered by blists - more mailing lists