linux-kernel - Re: [PATCH v5 00/13] SCMI Notifications Core Support

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]

Date:   Mon, 23 Mar 2020 08:54:41 +0000
From:   Cristian Marussi <cristian.marussi@....com>
To:     Lukasz Luba <lukasz.luba@....com>, linux-kernel@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org
Cc:     sudeep.holla@....com, james.quinlan@...adcom.com,
        Jonathan.Cameron@...wei.com
Subject: Re: [PATCH v5 00/13] SCMI Notifications Core Support

Hi Lukasz

On 3/18/20 9:01 AM, Lukasz Luba wrote:
> Hi Cristian,
> 
> On 3/16/20 3:03 PM, Cristian Marussi wrote:
>> Hi all,
>>
>> this series wants to introduce SCMI Notification Support, built on top of
>> the standard Kernel notification chain subsystem.
>>
>> At initialization time each SCMI Protocol takes care to register with the
>> new SCMI notification core the set of its own events which it intends to
>> support.
>>
>> Using the API exposed via scmi_handle.notify_ops a Kernel user can register
>> its own notifier_t callback (via a notifier_block as usual) against any
>> registered event as identified by the tuple:
>>
>>         (proto_id, event_id, src_id)
>>
>> where src_id represents a generic source identifier which is protocol
>> dependent like domain_id, performance_id, sensor_id and so forth.
>> (users can anyway do NOT provide any src_id, and subscribe instead to ALL
>>   the existing (if any) src_id sources for that proto_id/evt_id combination)
>>
>> Each of the above tuple-specified event will be served on its own dedicated
>> blocking notification chain, dynamically allocated on-demand when at least
>> one user has shown interest on that event.
>>
>> Upon a notification delivery all the users' registered notifier_t callbacks
>> will be in turn invoked and fed with the event_id as @action param and a
>> generated custom per-event struct _report as @data param.
>> (as in include/linux/scmi_protocol.h)
>>
>> The final step of notification delivery via users' callback invocation is
>> instead delegated to a pool of deferred workers (Kernel cmwq): each
>> SCMI protocol has its own dedicated worker and dedicated queue to push
>> events from the rx ISR to the worker.
>>
> 
> Could you give an example how the notification would be delivered
> further to the upper layers, like hwmon driver, cpufreq or thermal?

Sure. I tested registering various callbacks against PERF events (since they're
what I have available on my platform implementation); I used probe() in scmi-genpd
and scmi-cpufreq to register generic testing callbacks like:

scmi_cpufrq_probe()

...
	src_id = 0x00;
	handle->notify_ops->register_event_notifier(handle, SCMI_PROTOCOL_PERF,
						    0x1, &src_id, &my_cpufreq_nb);

and removing similarly in remove().

I'll send you my debug patches that includes genpd/cpufreq callbacks registration and
an additional dummy driver that just registers perf callbacks so you can have a better idea
of the intended usage.

> For example, for sensor protocol which delivers event
> SENSOR_TRIP_POINT_EVENT indicating a trip point was crossed.
>

Regarding sensors it could be something like:


	sensor_id = 0x00;

	handle->notify_ops->register_event_notifier(handle, SCMI_PROTOCOL_SENSOR,
						    0x0, &sensor_id, &my_sensor_nb);

	handle->sensor_ops->trip_point_config(handle, sensor_id, 0, 10000);
  
Note that the notification core takes care on its own of enabling the specific events generation
when you register the first callback for a specific event as identified by (proto_id, event_id, src_id)
but in the case of the sensor protocol you'll need to explicitly setup the trip_point too.

As a result when the trip point is crossed you'll receive in your callback a sensor_report
(as in include/linux/scmi_protocol.h) as your data arg.

I just realized that being the notification enable method (trip_notify) also exposed by the ops directly,
I should probably add some sort of alert for a user...since it is not meant to be used directly when
the notification subsystem is being used....I'll think about this pitfall.

> Would it be possible for:
> drivers/hwmon/scmi-hwmon.c
> to get this temperature events like an interrupt?
> 

Sorry I did not get what you mean here.

Regards

Cristian

> I couldn't find it in the implementation of the registered handlers.
> 
> Regards,
> Lukasz