linux-kernel - Re: ACPI binding with SCMI

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

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <20230830091337.4rsik2tza6hqa22r@bogus>
Date:   Wed, 30 Aug 2023 10:13:37 +0100
From:   Sudeep Holla <sudeep.holla@....com>
To:     Pawandeep Oza <poza@....qualcomm.com>
Cc:     "Pawandeep Oza (QUIC)" <quic_poza@...cinc.com>,
        "rafael@...nel.org" <rafael@...nel.org>,
        Sudeep Holla <sudeep.holla@....com>,
        "linux-arm-kernel@...ts.infradead.org" 
        <linux-arm-kernel@...ts.infradead.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-acpi@...r.kernel.org" <linux-acpi@...r.kernel.org>
Subject: Re: ACPI binding with SCMI

On Tue, Aug 29, 2023 at 03:10:34PM +0000, Pawandeep Oza wrote:
> I am looking for scmi binding to ACPI. Basically, SCMI based perf control.
> We have ACPI based system, so I am looking for where I can describe (some
> way to describe doorbell and share memory in ACPI), and then scmi perf can
> bind to it via some sort of transport (perhaps fastchannels ? )

OK, the Section 5.2 ACPI-based Transport in the SCMI spec can be more clearer
IMO. It does state:
" SCMI FastChannels can be represented as ACPI System Memory and used
  in the Continuous Performance Control (CPC) object when using ACPI 
  Collaborative Processor Performance Control (CPPC)."

It doesn't cover the fact that CPPC is compatible only with SCMI fastchannels
without doorbells. Just a plain MMIO register to set the perf.

The normal PCC mailbox works fine if the CPPC Desired Perf Register is
a PCC address space based GAS. However the command is not 1:1 compatible
with SCMI perf. This is one of the reason why the protocols 0x0-0xF was
reserved in SCMI to be compatible with the ACPI CPPC way of using PCC
for perf.

Hope this helps and I didn't make it more complicated for you.

--
Regards,
Sudeep