Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <12013659.O9o76ZdvQC@kreacher> <61311732eeea1f45e85537e911e4bb024c0a30b7.camel@intel.com>
In-Reply-To: <61311732eeea1f45e85537e911e4bb024c0a30b7.camel@intel.com>
From:   "Rafael J. Wysocki" <rafael@kernel.org>
Date:   Sun, 17 Jul 2022 21:39:01 +0200
Message-ID: <CAJZ5v0j+FTX4UF-9Y0BQc2mYXQiphsnkt07CALhF7BPtSdDxgg@mail.gmail.com>
Subject: Re: [PATCH] intel: thermal: PCH: Drop ACPI_FADT_LOW_POWER_S0 check
To:     Zhang Rui <rui.zhang@intel.com>
Cc:     "Rafael J. Wysocki" <rjw@rjwysocki.net>,
        Linux PM <linux-pm@vger.kernel.org>,
        Linux ACPI <linux-acpi@vger.kernel.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Mario Limonciello <mario.limonciello@amd.com>,
        Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Sun, Jul 17, 2022 at 8:14 AM Zhang Rui <rui.zhang@intel.com> wrote:
>
> On Thu, 2022-07-14 at 21:11 +0200, Rafael J. Wysocki wrote:
> > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> >
> > If ACPI_FADT_LOW_POWER_S0 is not set, this doesn't mean that low-
> > power
> > S0 idle is not usable.  It merely means that using S3 on the given
> > system is more beneficial from the energy saving perspective than
> > using
> > low-power S0 idle, as long as S3 is supported.
>
> Agreed.
>
> >
> > Suspend-to-idle is still a valid suspend mode if
> > ACPI_FADT_LOW_POWER_S0
> > is not set and the pm_suspend_via_firmware() check in
> > pch_wpt_suspend()
> > is sufficient to distinguish suspend-to-idle from S3, so drop the
> > confusing ACPI_FADT_LOW_POWER_S0 check.
>
> the cooling delay in the suspend callback is to make sure PCH
> temperature won't block S0ix during s2idle. So if S0ix is not
> supported, it is meaningless to invoke the cooling delay during s2idle.

But there is no way to determine whether or not S0ix is supported.  In
particular, ACPI_FADT_LOW_POWER_S0 is not one.

> so the problem is that we don't have an indicator for S0ix capability.
> And this also applies to drivers/rtc/rtc-cmos.c, where we use ACPI SCI
> for runtime RTC wakeup instead of HPET interrupt on "S0ix capable"
> platforms because the HPET timer may block S0ix.

"S0ix capable" doesn't matter.  What matters is whether or not the
current transition under way is into S0 or into suspend-to-idle.  In
the latter case there is no reason to avoid doing whatever is done in
the expectation that S0ix may be entered going forward.