[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20200911144032.GC12835@gaia>
Date: Fri, 11 Sep 2020 15:40:32 +0100
From: Catalin Marinas <catalin.marinas@....com>
To: Ionela Voinescu <ionela.voinescu@....com>
Cc: will@...nel.org, sudeep.holla@....com, morten.rasmussen@....com,
valentin.schneider@....com, souvik.chakravarty@....com,
viresh.kumar@...aro.org, dietmar.eggemann@....com,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 4/4] arm64: implement CPPC FFH support using AMUs
On Wed, Aug 26, 2020 at 02:03:09PM +0100, Ionela Voinescu wrote:
> +/*
> + * Refer to drivers/acpi/cppc_acpi.c for the description of the functions
> + * below.
> + */
> +bool cpc_ffh_supported(void)
> +{
> + const struct cpumask *cnt_cpu_mask = cpus_with_amu_counters();
> + int cpu = nr_cpu_ids;
> +
> + if (cnt_cpu_mask)
> + cpu = cpumask_any_and(cnt_cpu_mask, cpu_present_mask);
> +
> + if ((cpu >= nr_cpu_ids) || !freq_counters_valid(cpu))
> + return false;
> +
> + return true;
> +}
IIUC, the only need for the cpumask is this function, the others would
have worked just fine with the existing cpu_has_amu_feat(). So you have
a lot more !cnt_cpu_mask checks now.
I wonder whether instead you could add a new function near
cpu_has_amu_feat(), something like get_cpu_with_amu_feat() and do the
cpumask_any_and() in there.
--
Catalin
Powered by blists - more mailing lists