[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <9b108208-cb52-68d8-8a16-8a33c3824d74@amd.com>
Date: Thu, 4 Jul 2024 20:17:22 +0530
From: "Gupta, Akshay" <Akshay.Gupta@....com>
To: Greg KH <gregkh@...uxfoundation.org>
Cc: linux-hwmon@...r.kernel.org, linux-kernel@...r.kernel.org,
linux@...ck-us.net, arnd@...db.de, naveenkrishna.chatradhi@....com
Subject: Re: [PATCH 0/6] misc: add amd side-band functionality
On 7/4/2024 5:18 PM, Greg KH wrote:
> Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
>
>
> On Thu, Jul 04, 2024 at 11:16:18AM +0000, Akshay Gupta wrote:
>> AMD's APML interface provides system management functionality accessed by
>> the BMC. Sbrmi driver under hwmon subsystem, which is probed as an i2c
>> driver and reports power using APML specified protocol.
> What is "APML"? What is "BMC"?
My bad, I will provide expansion for the acronyms in the first usage.
>
>> However, APML interface defines few other protocols to support
>> full system management functionality out-of-band.
> What is out-of-band here?
Out-of-band management is term used for Baseboard Management Controller
(BMC) talking to
system management unit(IP in the processor, AMD's documentation called
this link as side band interface.
We are using the terms out-of-band and side band interface as per context.
>
>> This patchset is an attempt to add all APML core functionality in one place
>> and provide hwmon and user interface
>> 1. Move the i2c client probe and sbrmi core functionality from drivers/hwmon
>> to drivers/misc/
>> 2. Add a platform device, which probes the hwmon/sbrmi and continues to
>> report power using the symbol exported by the misc/sbrmi-core.
> Are you sure this is a platform device? It better not be a pci one...
It is not a pci device, will remove the platform device and register
hwmon device in misc.
>
>> 3. Convert i2c to regmap which provides multiple benefits
>> over direct smbus APIs.
> What are those?
We are planning to add i2c/i3c support and 1byte/2 byte register addressing.
>
>> 4. Register a misc device which provides
>> a. An ioctl interface through node /dev/sbrmiX
> Why?
>
>> b. Open-sourced and widely used https://github.com/amd/esmi_oob_library
>> will continue to provide user-space programmable API.
> Will this use the new ioctl api? If it's not present, what will it use
> instead?
This is an attempt to upstream the out of tree kernel module(which
defines this ioctl), open sourced at:
https://github.com/amd/apml_modules/ and continue to support
https://github.com/amd/esmi_oob_library
which depends only on this ioctl.
Thanks,
Akshay
> thanks,
>
> greg k-h
Powered by blists - more mailing lists