[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date: Mon, 4 Apr 2022 17:31:02 +0100
From: Lukasz Luba <lukasz.luba@....com>
To: Ionela Voinescu <ionela.voinescu@....com>
Cc: linux-kernel@...r.kernel.org, dietmar.eggemann@....com,
Pierre.Gondois@....com, viresh.kumar@...aro.org, rafael@...nel.org,
daniel.lezcano@...aro.org, linux-pm@...r.kernel.org,
mka@...omium.org, nm@...com, sboyd@...nel.org,
linux-arm-kernel@...ts.infradead.org,
linux-mediatek@...ts.infradead.org, cristian.marussi@....com,
sudeep.holla@....com, matthias.bgg@...il.com
Subject: Re: [RESEND][PATCH 5/8] PM: EM: Remove old debugfs files and print
all 'flags'
Hi Ionela,
Thank you for reviewing these patches.
On 4/4/22 17:02, Ionela Voinescu wrote:
> Hi Lukasz,
>
> IMO the debugfs files were fine as they were:
> - They offered information on units and inefficiencies without having
> to dig into the code to see which bit is for which flag.
> - I believe the artificial EM power values fit under bogoWatts as unit,
> so that part would still be correct.
>
> On the other hand, your new file offers more information: we'd be able
> to see in debugfs whether we're dealing with an artificial EM, despite
> needing a bit more looking over the code to understand the output.
I have consolidated them, so we would support more features in the flag
automatically when there will be a need.
In Android kernel we don't have unfortunately the debugfs compiled-in,
so this information is still only for kernel testing with some Linux
distro.
I have been thinking to switch into sysfs interface, so we could
have it in Android as well. This patch change adding a generic 'flags'
which would be more 'stable' (which is a requirement for sysfs contract)
is also a closer step into that direction. But that is for longer
discussion and not for this $subject.
>
> I don't have a strong opinion and the code looks fine, so:
>
> Reviewed-by: Ionela Voinescu <ionela.voinescu@....com>
>
Powered by blists - more mailing lists