Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.31 as permitted sender) client-ip=23.128.96.31;
Date:   Tue, 05 Dec 2023 12:15:48 +0000
From:   David Woodhouse <dwmw2@infradead.org>
To:     "Rafael J. Wysocki" <rafael@kernel.org>
CC:     pdurrant@amazon.co.uk, bp@alien8.de, dave.hansen@intel.com,
        dave.hansen@linux.intel.com, hdegoede@redhat.com, hpa@zytor.com,
        jalliste@amazon.co.uk, juew@amazon.com, len.brown@intel.com,
        linux-kernel@vger.kernel.org, mingo@redhat.com,
        peterz@infradead.org, rafael.j.wysocki@intel.com,
        rafael@kernel.org, tglx@linutronix.de, usama.arif@bytedance.com,
        x86@kernel.org
Subject: =?US-ASCII?Q?Re=3A_=5BPATCH=5D_x86=3A_intel=5Fepb=3A_Add_earlyp?= =?US-ASCII?Q?aram_option_to_keep_bias_at_performance?=
User-Agent: K-9 Mail for Android
In-Reply-To: <CAJZ5v0iVvMLK_VcPRJ4sW1eOh0EtfcFvKjH5j1y1GbA0Y6q--Q@mail.gmail.com>
References: <53bce76c5c60463eba1372df426a64b9@amazon.co.uk> <1D71D218-5EB6-47DE-A01B-3A66F9F4C74E@infradead.org> <CAJZ5v0iVvMLK_VcPRJ4sW1eOh0EtfcFvKjH5j1y1GbA0Y6q--Q@mail.gmail.com>
Message-ID: <5322CF05-344D-4ADE-B38C-7DCE7F076E0C@infradead.org>
MIME-Version: 1.0
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On 5 December 2023 12:12:09 GMT, "Rafael J=2E Wysocki" <rafael@kernel=2Eorg=
> wrote:
>On Tue, Dec 5, 2023 at 1:00=E2=80=AFPM David Woodhouse <dwmw2@infradead=
=2Eorg> wrote:
>>
>>
>> Paul writes:
>> > The problem is that this will take effect even on a kexec and hence i=
t is throttling
>> > a system that set ENERGY_PERF_BIAS_PERFORMANCE prior to the kexec=2E =
 We use kexec to
>> > live update the host kernel of our systems whilst leaving virtual mac=
hines running=2E
>> > This resetting of the perf bias is having a very detrimental effect o=
n the downtime
>> > of our systems across the live update - about a 7 fold increase=2E
>>
>> It isn't just about kexec, is it? Even in a clean boot why wouldn't we =
want to stay in performance mode until the kernel has *finished* booting?
>
>Because it may overheat during that period=2E
>
>> It's literally adding seconds to the startup time in some cases=2E
>>
>> And yes, we *particularly* care in the kexec case because guests experi=
ence it as excessive steal time=2E But it ain't great in the general case e=
ither, surely?
>
>So IMV it would be perfectly fine to add a command line arg to provide
>the initial value of energy_perf_bias for the ones who know what they
>are doing=2E

We don't even care about setting it to an "initial value" during boot=2E W=
e just want to leave it how it was already set up=2E