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] [thread-next>] [day] [month] [year] [list]
Message-ID: <f193e4a5-b315-a1bd-7488-7ea0d5e267f8@linaro.org>
Date:   Wed, 27 Sep 2023 17:46:19 +0200
From:   Daniel Lezcano <daniel.lezcano@...aro.org>
To:     "Rafael J. Wysocki" <rafael@...nel.org>
Cc:     "Rafael J. Wysocki" <rjw@...ysocki.net>,
        Linux PM <linux-pm@...r.kernel.org>,
        LKML <linux-kernel@...r.kernel.org>,
        Linux ACPI <linux-acpi@...r.kernel.org>,
        Srinivas Pandruvada <srinivas.pandruvada@...ux.intel.com>,
        Zhang Rui <rui.zhang@...el.com>,
        Lukasz Luba <lukasz.luba@....com>
Subject: Re: [PATCH v1 07/13] thermal: gov_power_allocator: Use trip pointers
 instead of trip indices

On 27/09/2023 17:27, Rafael J. Wysocki wrote:
> On Wed, Sep 27, 2023 at 5:10 PM Daniel Lezcano
> <daniel.lezcano@...aro.org> wrote:
>>
>> On 21/09/2023 19:55, Rafael J. Wysocki wrote:
>>> From: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
>>>
>>> Eliminate the __thermal_zone_get_trip() usage that adds unnecessary
>>> overhead (due to pointless bounds checking and copying of trip point
>>> data) from the power allocator thermal governor and generally make it
>>> use trip pointers instead of trip indices where applicable.
>>
>> Actually the __thermal_zone_get_trip() change was done on purpose to
>> replace the 'throttle' callback index parameter by the trip pointer and
>> removing those call to __thermal_zone_get_trip() while the code was
>> using the trip pointer.
>>
>> IMO, the changes should focus on changing the trip_index parameter by
>> the trip pointer directly in the throttle ops.
> 
> So you would like .throttle() to take a trip pointer argument instead
> of an index?
> 
> The difficulty here is that the user space governor needs to expose
> the index to user space anyway, so it would need to find it if it gets
> a trip pointer instead.
> 
> Not a big deal I suppose, but a bit of extra overhead.
> 
> Also it is easier to switch the governors over to using trip pointers
> internally and then change the .throttle() argument on top of that.
> 
>> The pointer can be
>> retrieved in the handle_thermal_trip() function and passed around for
>> the rest of the actions on this trip point
> 
> Right, except for the user space governor which needs a trip index.
> And the indices are used for tracing too.

Given the userspace governor is going obsolete and the notifications are 
for the userspace, which is slow, we can retrieve the index from the 
throttling ops

-- 
<http://www.linaro.org/> Linaro.org │ Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ